7
2016
Original scientific paper
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements as a Result of Changes in Planting Geometry Simon A. Ackerman, Stefan Seifert, Pierre A. Ackerman, Thomas Seifert Abstract Traditionally, the removal of entire rows at regular intervals through thinning compartments has been applied to facilitate access to mechanised timber harvesting operations in South Africa. These row thinnings have essentially involved the removal of every 7th row in a standard 2.7×2.7 m planting regime, resulting in a machine trail width of 5.4 m and a theoretical distance to the furthest tree of 8.1 m. A simulation study, based on alternative planting geometries, investigated the effect on harvesting in terms of harvesting productivity, system costs and impact on stand structure. Compartments of different planting geometries ranging from 2.7×2.7 m to 2.5×2.9 m, 2.4×3 m and 2.3×3.1 m at two thinning reference ages were simulator generated. These compartments were then simulator thinned and harvested in the simulation. Results showed that the boom reach of the harvester is optimised by extending row removal from the 7th to the 9th row. At the same time, machine trail length per hectare was reduced by 20%. This creates more productive area for tree growth, potentially reduced residual stand impacts, and increases the proportion of selectively harvested trees per hectare. The increased distance between row thinning removals enhanced the potential volume harvested trail length (m3/m) and in turn led up to a 8% increase in harvesting productivity, up to a 21% increase in forwarding productivity and a reduction in total costs of up to 7% when changing planting geometry from 2.7×2.7 m to 2.3×3.1 m and 2.4×3.0 m, for first and second thinning. Keywords: harvesting, simulation, thinning, planting geometry, productivity, system costing, optimisation
1. Introduction The advent of more advanced mechanised timber harvesting systems has identified the potential of possibly modifying planting geometries and thinning practices (Bredenkamp 1984). One of the alternatives considered is that of row thinnings where an entire row or rows are removed at predetermined intervals throughout the compartment. However, a balance needs to be achieved between improved harvesting efficiency and potential losses by eliminating a portion of the selective thinning process (Bredenkamp 1984). Croat. j. for. eng. 37(2016)1
It had been found that, if the execution of these two entirely different thinning systems were not well aligned (i.e. selective thinning is carried out first without identifying the trees to be removed in the rows removals), it results in an irregular stand structure along the removed rows (Ackerman et al. 2013). Suboptimal tree volume growth and tree form is a further consequence (Ackerman et al. 2013). The study simulated both felling and subsequent timber extraction operations in virtually constructed stands, where both access rows had been removed and
1
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
selective thinning applied between these rows. The proviso was that the simulation exercise had to maintain regular stand structure to satisfy optimal stand development. The use of an aggregation index (R), as proposed by Clark and Evans (1954), was applied as a measure of irregularity in the stands and, as an indicator of stand occupation efficiency, appears to have not been applied in South Africa forestry before. Similarly the application of computer simulation, now widely used in forest operations research worldwide (Asikainen 1995, 2001, 2010), was used to test different planting geometries on thinning harvesting productivity and cost. Simulation offers effective systems evaluation potential as alternatives (harvesting systems and management regimes) can be tested virtually without actual implementation of the said systems in the field (Talbot et al. 2003, Hogg et al. 2010, Pretzsch et al. 2002a).
2. Objective The objective of the study was to quantify the consequences of alternative planting geometries to the conventional 2.7×2.7 m on mechanised cut-to-length CTL) harvesting. The study questioned whether the modification of planting geometry: Þ reduced machine trail length per hectare still maintaining suitable access for the harvesting machines; Þ maintained compartment tree spacing regularity when simulated thinnings are done; Þ increased harvesting productivity with reduced harvesting system costs.
3. Materials and methods In South African forestry research, information on tree characteristics in compartments (individual tree models for DBH and height based on competition) and time consumption models for harvesting (time study data) is scarce. For this reason and for the sake of generating simulated stands and time consumptions, species growth models and harvesting system time models were sourced from worldwide research. These models were assumed to be representative to the work done for the area of operation in this paper. The procedure followed by the investigation into changing planting geometries and simulation is summarised as a flow chart in Fig. 1. The study was based on simulated compartments that were generated based on real data to mimic a re-
2
Fig. 1 Flow chart of the procedure followed for thinning and harvesting of compartments to maintain stand regularity alistic tree size distribution. Spatial adjustments of virtually generated compartments were done through a computer simulator and were based on existing silvicultural prescriptions for saw-timber production (Table 1). Various alternative initial planting geometries returning the same final stems per hectare (SPHA) as prescribed were tested during the simulation. The simulated planting geometries took into account the physical characteristics and limitations of both the harvester and forwarder that were to be used in the study for the harvesting simulation of both first and second thinnings.
3.1 Determining tree characteristics to develop computer simulated compartments The first step in the process to determine new planting geometries involved using pre-thinning enuTable 1 Standard establishment and thinning prescriptions in South Africa Action
Desired density
Spacing (initial)
2.7×2.7 m
Stems per hectare planted (SPHA)
1371 SPHA
First thinning (age 8)
650 SPHA
Second thinning (age 13)
400 SPHA
Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
meration tree data for compartments at thinning ages 8 and 13 years. This would establish the tree characteristics for each thinning age. The data set contained information for compartments of the same Site Index (SI20) of 20. This data was used to develop DBH and height data representative of trees at the two particular thinning reference ages in a compartment. However, applying this tree data randomly to a grid position does not sufficiently mimic the reaction of trees to growing space, nor to genetic variations. The reason is that compartment structure is not a purely random process. Competition between trees leads to a distinct tree dimension (Seifert 2003), relating to spatial pattern and compartment structure (Pretzsch 1997), where larger trees suppress their smaller neighbours. These spatial structures, resulting from competitive processes, had to be taken into account. The structure generator, developed by Pretzsch et al. (2002b), was used for creating realistic diameter distributions and spatial distributions of trees. The results where validated with data from the existing trial plots. Tree diameters and heights were manually increased in proportion to the mean DBH and height based on pre-thinning enumeration data between first and second thinning. As a standard in South African growth and yield modelling, natural mortality is not taken into account in heavily thinned stands (Kotze et al. 2012), as evident between first and second thinning. This approach was applied to all the various alternative planting geometries investigated.
3.2 Determining optimal tree spacing and planting geometry The second step involved matching machine size (and limitations) to planting geometries and adjusting these to various alternatives, while still maintaining the conventional tree spacing (2.7×2.7 m – 1370 trees/ha). Traditionally, machine trails for this geometry and others have been placed along the seventh row at right angles to tree rows. The removal of the seventh row for mechanised harvesting at this espacement results in a machine trail 4.5 m wide with a distance of 18.9 m between machine trails and an average required reach distance from either side of 9.45 m for the harvester boom.
S. A. Ackerman et al.
By adjusting the distances of trees within and between the rows, the alternative planting geometries in Table 2 were proposed. Distance between machine trials, width of the machine trails and length of machine trail per hectare were used as criteria for selecting the spacing geometry to be used in the study. Table 2 Breakdown of various planting spacings tested Spacing x–y
Rows to be removed
2.7×2.7 m
7th and 8th
2.5×2.9 m
7th, 8th and 9th
2.4×3.1 m
7th, 8th and 9th
2.3×3.0 m
7th, 8th and 9th
3.2.1 Machine limitations used to determine minimum planting spacing A Tigercat harvester and forwarder CTL system was selected for this study (Table 3), since these machines were already in operation on the plantation where the data was collected. A trail width of 1 m wider than the machine was considered a feasible criterion for the different planting geometries to prevent damage to stems and to limit tree root disturbances (Table 3). 3.2.2 Planting geometries used in thinning and harvesting simulations Using the machine limitations (Table 3), a selection system was developed to test the feasibility of various planting geometries from Table 2. The aim of the evaluation was to increase the distance between machine trails as much as possible (> 7th row), thus reducing the machine trail length per hectare and ensuring the distance between machine trail was equal to or less than 20 m so that the harvester boom could reach trees from the machine trail (10 m to the middle of the inter-row). Matching these criteria would limit stand impact and maximise the harvester boom reach. Even row (8th) spacing was excluded from the simulations due to the centre point between two machine
Table 3 Machine limitations based on boom reach and machine track width for Tigercat harvesters and forwarders (Tigercat 2011) Machine
Machine type
Boom reach, max
Boom reach, telescopic
Machine width
Payload
Tigercat H822c
Tracked harvester
8.91 m
11.07 m
3.43 m
–
Tigercat 1075B
Forwarder
7.83 m
N/A
3.30 m (bunk)
14,000 kg
Croat. j. for. eng. 37(2016)1
3
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
trails possibly falling exactly on the same tree row. This would lead to sub-optimal harvesting, as the machine would essentially have to harvest four rows from one machine trail and only three from the other, thus not utilising absolute boom reach on one side.
3.3 Stand simulations Following the process of matching machine specifications to various planting geometries, spatial tree lists containing x-and y-coordinates, DBH and height information of a 1.5 ha compartment were created in Excel. This was done for each of the planting geometries selected for the study. As a standard, the x-value always indicates the planting spacing used where a row of trees are removed for a machine trail. These were used as input into a specially designed simulation programme for thinning and harvesting, which was coded in the statistical language R (R Core Team 2012). A thinning from below was simulated for each stand. In this process trees that were marked as thinned were harvested by a harvesting simulator. 3.3.1 Thinning Thinning from below generally concentrates on the removal of trees that are smaller in relation to the neighbours in the same growing area, thus relieving competition (Murray and von Gadow 1991, Kassier 1993, Pukkala and Miina 1998, Pretzsch 2009). The thinning was simulated with a rule based algorithm without stochastic components, as this would have created an additional source of variance. As a consequence of this deterministic approach, a repeated application of the algorithm to the same stand would have resulted in the removal of the same trees. Input for the programme was the targeted final stem number per hectare as related to the size of the plantation area to be thinned (Ntarget). The programme would evaluate neighbouring trees in relation to a particular tree to determine the growing area and the growth status of the centre tree. Within the programme, a defined local search radius for tree neighbours around a target tree from the Ntarget was calculated by estimating the average growing area per tree (Eq. 1).
Agrow =
10, 000 × m2 N target
(1)
The local search radius for neighbouring trees was determined as 2.5 times the radius of a circle with the same area as Agrow (Eq. 2).
rgrow = ( Agrow / 0)
(2)
Each of the tree neighbours within the search radius were used to calculate the local stem density, a
4
DBH rank of the target (centre) tree to its neighbours, the proportion of the trees thinned to the target tree and a flag to mark if the distance to the nearest neighbour was less than rgrow. The local density was divided by the maximum density found in the stand. In order to make the values rateable, they were linearly transformed to be in a range between 0 and 1. This operation was done sequentially for all the trees in the stand. Lastly, the values calculated were summed up to determine a potential for a tree to be removed in the thinning process. The summed values were then ranked, and the trees with the highest potential to be thinned to the target SPHA were marked »to be removed« (TBR) and the rest were marked »not to be removed« (NTBR) as flags in the output. To limit the effect of stand edges on thinning, a subset of 1 ha subset was taken from the middle of the stand. A measure of aggregation (R) (Clark and Evans 1954) was used to determine the uniformity of the spacing in the stand after thinning. This measure of aggregation provided a test to evaluate the efficacy of the thinning algorithm. The particular data preparation and outputs for first and second thinnings are described below. 3.3.2 Simulated marking for thinning Before the first thinning simulation, the rows that were thinned for the extraction trails (7th or 9th row) were removed from the dataset as this would be done in practice. The full data set, with these row trees removed, was then thinned and trees TBR to the desired stand density (including removed row trees) and NTBR trees were marked. The row trees were then reintroduced as TBR for further analysis. The resulting dataset with the marked trees (row thinned and selectively) was then used as input for the spatial harvesting simulation. The second thinning simulation followed the same procedure, based on the stem numbers resulting from the first thinning operation except for the fact that no further row-thinnings were applied. 3.3.3 Harvesting In the harvesting simulation process, the spatial reach of a harvester moving along a skid trail was simulated. Based on x-and y-coordinates of trees and the flag for TBR or NTBR trees, individual tree harvesting was conducted. Each individual skid trail location (defined by start and end) was used as an input to the simulator. The output identified all the trees around the machine trail that could be reached by a 10 m boom, flagging them as accessible. If trees were attributed as acCroat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
S. A. Ackerman et al.
Fig. 2 a) harvester boom swath area and b) tree reach polygon cessible and marked, TBR would be flagged as harvested for a particular harvesting stop. These stops were determined and calculated using a harvesting simulator. The simulator was used to estimate the influence of spatial stand structure, extraction rows and stem number reduction on harvesting costs, and was designed and implemented using R (R Core Team 2012). This simulator was able to estimate the least number of position changes (harvesting position) of the harvester along a predetermined machine trail, and the number of trees harvested at each position. The simulation was based on pure geometry using only the tree positions and the line on which the harvester moved on the machine trail. The reach of the boom and the tree coordinates were used to identify the optimal point from which most trees could be harvested, (Fig. 2a and b). From a start position, the harvester moved forward on the machine trail to the first optimal point at which most trees could be reached. From this first stop, once all the harvestable trees had been virtually harvested, the next optimal point was selected and the harvester moved forward to that point. It was assumed that all trees in the polygon of Fig. 2a could be reached by the harvester head from the harvester position, the boom swath area. This, in reality may not be the case. The next step was to define the area from which a specific tree could be reached by the harvester, the tree reach polygon (Fig. 2b). The tree reach polygon can be derived by calculating all possible harvester positions from which the harvester boom can reach the targeted Croat. j. for. eng. 37(2016)1
harvestable tree. Geometrically this equals the inversion of the boom swath area in Fig. 2a. By intersecting the tree reach polygon with the machine trail, a new harvester stop line segment was created (Fig. 2). If the harvester was on this line segment, the boom could reach a particular tree. The procedure followed a sequence to find the optimal position to harvest most trees from a position, without reversing, assuming that this would match the strategy of a real harvesting operator. Selection of the nearest trees to harvest and the line selection for each stop are shown in Fig. 3. The intersection of the tree reach polygon (Fig. 2b) with the machine trail line defines the line of the segment where trees will be harvested for that stop. All tree polygons (Fig. 3), which intersect the starting line segment, are added to the list of harvested trees. When no more trees intersect the segment, the maximum number of trees that can be harvested from that line segment has been found and the endpoint of this segment is used as the new harvester position. These steps were repeated until the harvester had reached the end of this machine trail. This process allowed each harvested tree to be assigned to a specific harvester stop position. The total number of harvesting stops and the distance between stops were recorded. The accumulated distance along the machine trail was also calculated. A tree volume, based on the DBH and height values, was assigned to each harvested tree using the Schumacher and Hall function with parameters for P. patula (Bredenkamp 2012). The volume per harvesting
5
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
stop was totalled for each row with the distance between harvesting stops and accumulated distance travelled along the machine trail.
3.4 Harvester and forwarder productivity
Fig. 3 Nearest tree to harvesting stop and tree selection polygons inverted and translated to the tree position
Element
Table 4 Time element calculations used to determine time consumption in simulated operation Time calculation
2 Harvesting
Harvester
1 Driving a) Moving boom to cut b) Felling c) Processing
d) Boom in e) Clearing debris 1 Travel empty
33 m/cmin (Eliasson et al. 1999) 0.1 cmin/tree (Nurminen et al. 2006) t=0.093+0.101x (Nurminen et al. 2006) t=time (cmin/tree); x=volume of the tree t=0.0359+1.1368x (Nurminen et al. 2006) t=time (cmin/tree); x=tree volume 0.049 cmin/tree (Nurminen et al. 2006) 0.017 cmin/tree (Nurminen et al. 2006) 56 m/cmin (Nurminen et al. 2006)
2 Load
Forwarder
0.211
First thinning t=2.022+ x (Nurminen et al. 2006) t=time (cmin/tree); x=volume of the tree Second thinning
3 Travel partially loaded 3 Travel loaded 4 Unloading
6
t=2.777+
0.211 (Nurminen et al. 2006) x
t=time (cmin/tree); x=volume of the tree 26.7 m/cmin (Nurminen et al. 2006) 43.9 m/cmin (Nurminen et al. 2006) *0.569 cmin/m3 (Nurminen et al. 2006) *Based on mixed sawtimber loads
Volumes harvested at each harvesting stop were calculated. In order to determine the productivity of the harvesting system, the time taken to harvest and forward the timber needed to be determined. Time consumption was determined using existing time study functions with the harvesting and forwarding time consumption broken up into time elements. Due to actual time studies not being within the scope of the project in South Africa, element times and machine speeds were taken from studies by Eliasson et al. (1999) and Nurminen et al. (2006), respectively, for Nordic countries (Table 4). Based on the output from the harvesting simulations, a harvested volume for each harvesting stop was allocated for each machine trail that would have been harvested. The forwarder would then load timber from each of these harvesting stops. The simulated work method for each machine is described as follows. A harvester cycle starts at the base of the first machine trail and moves to the first harvesting stop as determined by the harvesting simulation. All the trees for that particular harvesting stop are assumed to be harvested and processed. Once the harvesting is complete, the next cycle starts with the machine moving north to the next harvesting stop (Fig. 4). At the end (highest x-and y-coordinate) of the machine trail, the machine moves to the base of the next machine trail and the simulation starts again. As with the harvester (Fig. 4), the forwarder would move into the stand from the start of machine trail one. It would then travel empty along the trail to the first timber stack, load and travel partially loaded to the next stack and continue loading. This was repeated until the forwarder was fully loaded to its capacity of 20,000 kg or 18.86 m3 (Table 3) for a Tigercat 1075B. This figure is based on a direct conversion of weight to volume of 1.06 tonnes to m3 provided by Bredenkamp (2012). Once the forwarder reaches the end of the machine trail, it is moved to the next one. At the point where the forwarder is full, it stops loading and travels full back down the machine trail to the nearest road where timber is unloaded. The machine then travels unloaded back to the last unfinished stack or a new stack to continue the process. Information gathered from the machine work methods and the time models was used to calculate Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
S. A. Ackerman et al.
Table 5 Costs (South African Rand) and costing assumptions for machines and attachments used in system costings (G. Olsen pers. comm. 2012, J. van Heerden pers. comm. 2013) Item
H822C Harvester
1075B Forwarder
Fixed cost inputs
Fig. 4 Simulation steps for harvester and forwarder for harvesting and loading time allocation the time taken to harvest 1 m3 of timber for each scenario and it was then compared to the standard spacing (2.7×2.7 m). Inputs to fixed and variable costs were based on standard industry data and input from the machine dealers. Operator, licensing, insurance, other miscellaneous costs and delays were not taken into account. Based on this information (Table 5), machine costs were determined for each scenario using a standard machine costing model (Eliasson 2013).
3.5 Statistical analysis A Levene-test for variance homogeneity was used to check for violations of the assumptions of homogenous variance between groups. Analysis of variance (ANOVA) was used to determine whether there were significant differences between the test criteria in planting geometries. In some cases, heteroscedasticity prohibited traditional t-tests and ANOVA. A nonparametric Welch’s t-test was used in these cases; this test is more robust against homoscedasticity violations. Subsequently, to determine further differences between planting geometries, a Bonferroni multiple hypothesis test or a Tamhane T2 test were applied, depending on homoscedastic or heteroscedasticity of variance respectively (Lyman Ott 1990).
4. Results 4.1 Harvesting thinnings from optimised stand structure 4.1.1 Determining the optimal tree geometry The planting geometry selection process found that the following planting geometries 2.5×2.9 m, Croat. j. for. eng. 37(2016)1
Machine cost
R4’056’754.00
R4’728’538.00
Harvesting attachment
R1’319’985.00
No attachment
Machine life
18,000 hrs
18,000 hrs
Harvesting attachment life
18,000 hrs
NA
Salvage cost machine, %
10
10
Salvage cost attachment, %
0
NA
Interest rate, % Insurance, registration, set-up and garaging costs
9
9
R 0.00
R 0.00
Variable cost inputs Fuel costs
R 11.60 (Feb, 2013) R 11.60 (Feb, 2013)
Fuel consumption
28 l/hr
12 l/hr
20%
10%
100
100
100
NA
2
8
Cost per track/tyre
R 155,000.00
R 42,000.00
Life of track/tyres
9000 hrs
8000 hrs
Cutter bar life
61.2 PMH
NA
Cutter bar cost
R 1500.00
NA
Chain life
38.25 PMH
NA
Chain cost
R 500.00
NA
Sprocket life
612 PMH
NA
Sprocket cost
R 1100.00
NA
Operator inputs
–
–
Operators per shift
1
1
Oil cost of fuel cost Maintenance cost machine, % Maintenance cost attachment, % Number of tracks/tyres
No operator costs were taken into account Productivity inputs Working days per year Shifts per day Hours per shift Productivity per hour Machine utilisation
240
240
2
2
9 9 Based in time study Based in time study information information 85% 85%
2.3×3.1 m and 2.4×3.0 m (Table 6), were suitable alternatives for the conventional 2.7×2.7 m geometry; i.e. the control.
7
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
Table 6 Acceptable planting geometries based on rows removed, machine trail length and closest tree distance Planting geometry m×m 2.7×2.7 2.5×2.9
Machine trail width m
Distance to furthest tree m
Row remove machine trail
Spacing between trails m*
Trail length ha–1 m
Number of rows removed ha–1
5.4
9.45
7th
18.9
599.4
6
10.0
th
22.5
500.0
5
th
5.0
9
2.3×3.1
4.6
9.2
9
21.6
504.0
5
2.4×3.0
4.8
9.6
9th
20.7
506.0
5
*Measured from the mid-point of the machine trails
The alternatives reduced the length of machine trail ha–1 by between 99.4 m/ha and 93.4 m/ha. The number of tree rows removed per hectare was reduced by adjusting the width between the skid trails in all cases. In all the proposed planting geometries, the distance to the furthest tree was within the maximum reach of the harvester boom (10 m). In order to test the efficiency of the thinning in maintaining an evenly distributed tree structure, a Clark and Evans aggregation (R) index was carried out on the tree distribution before and after thinning. The results of this analysis appear in Table 7. 4.1.2 Virtual harvesting of sample stands Harvested volume data of the virtually thinned stands are shown in Table 8. The results show the removed and remaining volume after each thinning, mean volume harvested at each harvesting stop and the mean distance between the harvesting stops. The mean differences between the different planting geometries (control vs potential scenarios) and the abovementioned criteria were compared. 4.1.2.1 Volume harvested per stop for each planting geometry ANOVA analysis results for differences between the mean volumes harvested at each harvesting stop on machine trails for each planting geometry are shown in Fig. 5. Analysis of the data indicates that there were significant differences (p<0.05) between mean harvested volume at each harvesting stop for both first and second thinning. A post hoc analysis using a Bonferroni multiple comparison test found that there were significant differences (p<0.05) between volume harvested at each stop for all of the geometries in the first thinning, except for the control and 2.4×3.0 m planting geometry. In the second thinning, there were no significant differences (p>0.05) between volume harvested at each stop for all of the geometries, except for a significant
8
Table 7 Clark and Evans (R) index for stands before and after thinning Thinning
First
Second
Clark and Evan aggregation index, R
Planting geometry m×m
Before thinning
After thinning
2.7×2.7
1.863
1.098
2.5×2.9
1.760
1.132
2.4×3.0
1.701
1.124
2.3×3.1
1.641
1.156
2.7×2.7
1.425
1.126
2.5×2.9
1.398
1.100
2.4×3.0
1.386
1.196
2.3×3.1
1.641
1.156
difference between 2.5×2.9 m and 2.3×3.1 m geometries. 4.1.2.2 Distance between harvesting stops for each planting geometry A Welch t-test showed differences between the mean distances between harvesting stops on machine trails for each of the planting geometries (Fig. 6). The results of this test show that there were significant differences (p<0.05) between the distances between harvesting stops in both first and second thinning. A Tamhane T2 multiple comparison indicates significant differences between all the geometries except for the control and 2.4×3.0 m and the control and 2.3×3.1 m planting geometries in first thinning. In the second thinning, there were no significant differences between any of the combinations except for the control and 2.5×2.9 m planting geometry. 4.1.2.3 Harvesting time per harvesting stop for each planting geometry ANOVA analysis was done on the first thinning data; it is, however, necessary to make a Welch t-test Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
S. A. Ackerman et al.
Table 8 Harvested data before initial thinning and after first or second thinning Thinning
First
Second
Total volume, m3/ha
Means per harvesting stop
Planting geometry m×m
Removed
Remaining
Volume, m
s
Distance, m
s
2.7×2.7
30.37
46.96
0.41
0.08
7.91
0.14
2.5×2.9
27.66
48.13
0.26
0.03
5.19
1.02
2.4×3.0
30.27
46.96
0.42
0.08
7.23
0.72
2.3×3.1
28.56
47.46
0.51
0.05
9.05
0.43
2.7×2.7
35.85
93.89
0.91
0.17
12.85
1.28
2.5×2.9
35.31
90.87
0.76
0.20
10.38
1.39
2.4×3.0
35.98
89.91
0.88
0.12
11.64
2.03
2.3×3.1
39.02
90.57
1.00
0.12
11.86
1.12
on the second thinning data, too (Fig. 7). The results show that there were significant differences between the mean harvesting times at each harvesting stop. Significant differences were also found between all of the planting geometries in first thinning operations except for the control and the 2.4×3.0 m planting geometry (Bonferroni multiple comparison test). The second thinning showed no significant differences between the geometries, except between the 2.5×2.9 m and the 2.3×3.1 m geometries (Tamhane T2 multiple comparison test).
3
4.1.3 Time study and cycle times Harvester cycles, volume and production achieved in the two thinning operations for each planting geometry are shown in Table 9. The number of cycles depended on the number of harvesting stops determined by the harvesting simulator. In the first thinning, production was reduced between the control and the remaining planting geometries, while in the second thinning the opposite was true as an increase was evident. Forwarder cycles (Table 10) were limited by the load capacity of the
Fig. 5 Mean volume harvested for each stop (a) first thinning and (b) second thinning for each planting geometry Croat. j. for. eng. 37(2016)1
9
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
Fig. 6 Mean distance travelled between harvesting stops for (a) first thinning and (b) second thinning for each planting geometry
Fig. 7 Mean time consumption to harvest trees for each harvesting stop for first thinning (a) and second thinning (b) for each planting geometry forwarder, and in most cases only one full load was possible (18.86 m3) followed by a partial load. However, in the second thinning on the 2.3x3.1 m geometry, the additional volume to the machine trail led to two full loads and one partial third load being forwarded. The lowest production was found in 2.5x2.9 m planting geometry; there was, however, a general increase in production from the control to the remaining planting geometries.
10
4.1.4 Machine and systems costing The results of the machine costing and system costing are shown in Table 11. In first and second thinning, the most expensive thinning operation (total costs) was for the 2.5x2.9 m planting geometry (R 306.76·m-3 and R 139.90·m-3). In the first thinning, the cheapest system was that of the 2.3x3.1 m planting geometry (R 236.78·m-3). The second thinning showed a reduction in cost between the control and the remaining planting geometries. Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1â&#x20AC;&#x201C;15)
S. A. Ackerman et al.
Table 9 Harvester total cycles, time taken, volume, productive machine hours (PMH) and volume per PMH for each geometry and thinning Thinning
First
Second
Planting geometry, mxm
Cycles
Time
Volume
PMH
m3/PMH
2.7x2.7
78
259.66
30.75
4.33
7.11
2.5x2.9
119
240.95
28.22
4.02
7.03
2.4x3.0
72
244.74
28.50
4.08
6.99
2.3x3.1
58
251.79
28.84
4.20
6.87
2.7x2.7
47
132.2
35.70
2.20
16.20
2.5x2.9
54
122.88
35.61
2.05
17.39
2.4x3.0
44
124.78
36.20
2.08
17.41
2.3x3.1
43
134.34
39.24
2.24
17.53
5. Discussion 5.1 Planting geometry changes The alternative planting geometries that were compared in this simulation study (Table 6) indicated that a 20% reduction in machine trail length (from 599.4 m/ha to 500 m/ha) is possible when compared to the standard 2.7x2.7 m planting geometry (the control). A reduction in machine trail length has a number of advantages. Large gaps in the canopy, created by the cutting out of rows for machine trails in standard planting geometries, were reduced in size or limited. Furthermore, the likelihood of damage to residual trees during harvesting, purely because there are fewer trails, is also reduced (Hunt and Krueger 1960, Ohman 1970, Kromhout and Bosman 1982, Vasiliauskas 2001). However, in some cases, the distance between machine trails can cause the harvester head at full boom reach to lose control of the harvest tree. The resultant uncontrolled fall of the harvested tree can in some cases lead to residual tree damage (FrĂśding 1992 and SirĂŠn 1992), if not monitored effectively. It could be assumed, based on works of Warkotsch et al. (1994) and Bettinger et al. (1998), that fewer trails also resulted in reducing the potential of soil damage in terms of soil compaction and displacement. Similarly, the reduction in gaps in the canopy and irregular stand structure also reduce the negative effects on branchiness of the planted trees (Seifert 2003, Ackerman et al. 2013). CTL harvesting, as applied in this study, generally shows reduced stand impact over tree-length and fulltree harvesting systems (Wang et al. 2005). This has great advantage over the traditional planting geometries. Croat. j. for. eng. 37(2016)1
5.2 Stand regularity after thinning Alternative planting geometries and a thinning algorithm were developed to provide realistic thinning output while maintaining stand regularity. The aggregation index, (R) (Clark and Evans 1954) showed that the thinning algorithm was effective in terms of maintaining regular stand spacing. The aim of the simulator was to avoid clustering of the trees and to maintain a (R) value higher than 1.0. All the aggregation index results were higher than this threshold (Table 7). This illustrates that the stands were thinned to a random distribution with no clustering.
5.3 Harvesting and forwarding productivity 5.3.1 Harvester As expected, volumes per harvesting stop on machine trails increased with a reduction in machine trail length (Table 9). This was also closely associated with the distance between harvesting/loading stops and the time consumption for harvesting at each stop. In all cases, the 2.5x2.9 m planting geometry consumed less time than the control (2.7x2.7 m) and all other alternatives due to the lower volume per stop and shorter distances between stops. There were, however, many more stops per hectare than for the other geometries. There was an overall increase in time consumed at each harvesting stop in the first as opposed to the second thinning. This was due to higher stem numbers (of lower piece volume) in the younger stand harvested. The individual tree volume in this simulation did not influence time consumption. The harvester boom movement related activities were the main driver of
11
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1â&#x20AC;&#x201C;15)
Table 10 Forwarder cycle times and volumes per cycle for each thinning and geometry and total time and volume per hour Thinning
First
Second
Cycle one
Cycle two
Cycle three
Total
Planting geometry, mxm
Time
Volume
Time
Volume
Time
Volume
Time
Volume
2.7x2.7
144.78
18.86
101.02
11.89
NA
NA
245.80
2.5x2.9
233.07
18.86
116.65
9.36
NA
NA
2.4x3.0
137.5
18.86
88.97
9.64
NA
2.3x3.1
115.84
18.86
64.93
9.98
2.7x2.7
85.11
18.86
107.22
2.5x2.9
107.31
18.86
2.4x3.0
97.94
2.3x3.1
89.09
PMH
m3/PMH
30.75
4.1
7.51
349.72
28.22
5.83
4.84
NA
226.47
28.5
3.77
7.55
NA
NA
180.77
28.84
3.01
9.57
16.84
NA
NA
192.33
35.7
3.21
11.14
112.9
16.75
NA
NA
220.21
35.61
3.67
9.7
18.86
81.67
17.34
NA
NA
179.61
36.2
2.99
12.09
18.86
81.31
18.86
15.09
1.52
185.50
39.24
3.09
12.69
this. In other words, due to the individual tree volume being less in first thinnings, the multiple boom movements did not translate into a potentially higher volume harvested (Eliasson and Lageson 1999, Talbot et al. 2003). This phenomenon will potentially decrease productivity of the system in first thinnings (Belbo 2010). Analysis of the scenario data revealed that the distance a harvester moved between harvesting stops and the volumes harvested at each stop influenced each other. In order to optimise machine working and movement time, a balance between these two factors would greatly increase the productivity. This is supported by results in other studies (Talbot et al. 2003). When deciding on a feasible alternative to the control (2.5x2.9 m, 2.4x3.0 m and 2.3x3.1 m), the productivity results for the harvester were inconclusive in the first thinning mainly due to the great number of small trees. One would assume that the spacing geometries
with the highest volume per harvesting stop, the shortest distance between stops and lowest total harvesting time consumption would appear to be the best alternative. Harvester productivity decreased by between 1 and 3% in the first thinning and increased by between 7 and 8% in the second thinning. This was, however, a net increase in productivity over the two thinning operations. There was a general increase in productivity between geometries 2.4x3.0 m and 2.3x3.1 m when compared with the control. It is evident that these were the best suited alternatives to change planting geometry at this point.
5.3.2 Forwarder Forwarder productivity depended on the distance travelled between loading points and the volume available at each stop in the scenario simulation (Table 11).
Table 11 Results of machine costing for first and second thinning for harvesting and forwarding operations (South African Rand) Thinning
First
Second
12
Planting geometry, mxm
Harvester cost, R/m3
Forwarder cost, R/m3
Total system cost, R/m3
2.7x2.7
153.06
99.86
252.92
2.5x2.9
154.81
154.95
306.76
2.4x3.0
155.69
99.33
255.02
2.3x3.1
158.41
78.37
236.78
2.7x2.7
67.18
67.32
134.50
2.5x2.9
62.58
77.32
139.90
2.4x3.0
62.51
62.03
124.54
2.3x3.1
62.08
59.10
121.18
Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1â&#x20AC;&#x201C;15)
The grapple size influences the number of times the boom had to be deployed. While boom movement influenced time consumed loading the forwarder, as with the harvester, travel time did not have a great effect on the productivity. The main influence of productivity, evident from this study, was the increase in forwarder productivity when volume per harvesting stop increased. Similar travelling distances between harvesting stops were found in the simulation between the control, 2.4x3.0 m and 2.3x3.1 m, showing the importance of the volume per stop as a factor driving productivity increases. Overall productivity increases of between 21% (first thinning) and 12% (second thinning) could be achieved by using alternative planting geometries. Similar to that of the harvester, 2.4x3.0 m and 2.3x3.1 m were the most productive planting geometries for the forwarder.
5.4 Harvesting system cost In general, there was a decrease in cost/m3 between the control and the alternative planting geometries (Table 11). The planting geometries that led to the lowest costs were 2.4x3.0 m and 2.3x3.1 m in both first and second thinning operations. These two systems yielded an overall reduction in cost of 7% (R 16.14 m-3) and 10% (R 13.32 m-3) in first and second thinning, respectively. As discussed above, these two planting geometries did not significantly differ from each other in terms of volume per harvesting stop, distance between harvesting stop and time consumption per harvesting stop. However, a reduction of R 18.24 m-3 and R 3.66 m-3 could be achieved in first and second thinning operations, respectively, when choosing between 2.4x3.0 m and 2.3x3.1 m planting geometries; the latter having the lowest cost. The results show evident financial benefit of adopting alternative planting geometries to the control one. However, by changing the planting geometry the potential cost reduction can make these thinnings more competitive for the current systems.
6. Conclusion When optimising the planting geometries for mechanised thinning operations, it was found that the thinning simulator can effectively maintain stand regularity thus proving the efficacy of the method for the purpose of this study, and the overall system productivity could be increased by up to 8% and 21%, respectively, in harvester and forwarder productivity if the planting geometry was changed. This showed that rectangular geometries were superior to standard Croat. j. for. eng. 37(2016)1
S. A. Ackerman et al.
quadratic planting geometries, resulting in the possibility of achieving a cost reduction of up to 7% in first and 10% in second thinnings. Adding to the understanding of stand characteristics, the development and application of a computer based harvesting simulation model has once again highlighted the power of simulation techniques in providing answers to these complex issues. Financial decisions to implement changes in stand management require the ability to test these scenarios without the associated risks involved by trial and error applications. This work has also attempted to change mindsets by exploring alternatives to standard, square planting geometries by showing that small adjustments can potentially improve overall harvesting productivity and costs and reduce damage to the stands. The benefit of maintaining stand regularity in terms of tree growth characteristics and volume increment is evident. Furthermore, the objective of implementing other planting geometries, while maintaining stand regularity, has also shown to improve harvesting productivity and reduce overall harvesting system cost in a simulation environment. Marrying the thinning and harvesting simulator with stand and tree distance dependent growth, simulators would provide scenario testing for the whole forestry value chain. This would ensure that parts of this unique value chain do not work in isolation, but provide detailed feedback throughout the system. This research has made a start at developing this interaction, where aspects of Operations Research are not seen in isolation but as a combined field for all forestry disciplines. Developing these links and interactions between silviculture, growth and yield and harvesting will benefit the forestry industry and increase its overall competitiveness.
Acknowledgements Many thanks to Merensky Ltd and the THRIP/NRF and Green Landscapes Project in NRFâ&#x20AC;&#x2122;s GCSSR programme for financial means to conduct the study. The authors would like to thank Elizabeth Gleasure for her time and insight in the preparation of this paper.
7. References Ackerman, S.A., Ackerman, P.A., Seifert, T., 2013: Effects of irregular stand structure on tree growth, crown extension and branchiness of plantation grown Pinus patula. Southern Forests 75(4): 247â&#x20AC;&#x201C;256. Asikainen, A., 1995: Discrete-event simulation of mechanized wood-harvesting systems. Dissertation thesis, Research notes 38. University of Joensuu, Faculty of Forestry. 98 p.
13
S. A. Ackerman et al.
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15)
Asikainen, A., 2001: Simulation of logging and barge transport of wood from forests on islands. International Journal of Forest Engineering 12(2): 43–50.
Kromhout, C.P., Bosman, D.L., 1982: The influence of short rotation forestry on wood production for sawnwood and veneer. South African Forestry Journal 120(1): 11–18.
Asikainen, A., 2010: Simulation of stump crushing and truck transport of chips. Scandinavian Journal of Forest Research 25(3): 245–250.
Lyman Ott, R., 1990: Introduction to Statistical Methods and Data Analysis, California: Duxbury Press. 1051 p.
Belbo, H., 2010: Comparison of two working methods for small tree harvesting with a multi tree felling head mounted on farm tractor. Silva Fennica 44(3): 453–464. Bettinger, P., Bettinger, K.A., Boston, K., 1998: Correlation among spatial and non-spatial variables describing a cut-tolength thinning site in the Pacific Northwest, USA. Forest Ecology and Management 104(1–3): 139–149. Bredenkamp, B.V., 1984: Row thinnings do not adversely affect yields or form of the final crop in improved Pinus taeda. South African Forestry Journal 131(1): 28–33. Bredenkamp, B.V., 2012: The volume and mass of logs and standing trees. In B. V. Brendenkamp and S. J. Upfold, eds. South African Forestry Handbook. Menlo Park: Southern African Institute of Forestry, 239–267. Clark, P., Evans, F., 1954: Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology 35(4): 445–453. Eliasson, L., 2013: Machine cost calculation model. Avialable at: http://www.forestenergy.org/pages/costing-model-machine-cost-calculation/ Eliasson, L., Bengtsson, J., Cedergren, J., Lageson, H., 1999: Comparison of single-grip harvester productivity in clearand shelterwood cutting. Journal of Forest Engineering 10(1): 43–48. Eliasson, L., Lageson, H., 1999: Simulation study of a singlegrip harvester in thinning from below and thinning from above. Scandinavian Journal of Forest Research 14(6): 589– 595. Fröding, A., 1992: Thinning damage to coniferous stands in Sweden. Dissertation, Swedish University of Agriculture, Department of Operational Efficiency, 49 p. Hogg, G., Pulkki, R., Ackerman, P.A., 2010: Multi-stem mechanized harvesting operation analysis – application of Arena 9 discrete-event simulation software in Zululand, South Africa. International Journal of Forest Engineering 21(2): 14–22. Hunt, J., Krueger, K.W., 1960: Decay associated with thinning wounds in and Douglas-fir hemlock. Journal of Forestry 60(5): 336–340. Kassier, H.W., 1993: Dynamics of diameter and height distributions in even-aged pine plantations. Unpublished Doctorate Thesis, Stellenbosch: University of Stellenbosch. 167 p. Kotze, H., Kassier, H.W., Fletcher, Y., Morley, T., 2012: Growth modelling and yield tables. In B.V. Bredenkamp and S. Upfold, (eds). South African Forestry Handbook. Menlo Park: Southern African Institute of Forestry, 175–210.
14
Murray, D.M., von Gadow, K., 1991: Relationships between the diameter distributions before and after thinning. Forest Science 37(2): 552–559. Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363. Ohman, J.H., 1970: Value loss from skidding wounds in sugar maple and yellow birch. Journal of Forestry 68(4): 226–230. Pretzsch, H., 1997: Analysis and modeling of spatial stand structures. Methodological considerations based on mixed beech-larch stands in Lower Saxony. Forest Ecology and Management 97(3): 237–253. Pretzsch, H., 2009: Forest Dynamics, Growth and Yield, Berlin, Heidelberg: Springer Berlin Heidelberg, 665 p. Pretzsch, H., Biber, P., Ďurský, J., von Gadow, K., Hasenauer, H., Kändler, G., Kenk, G., Kublin, E., Nagel, J., Pukkala, T., Skovsgaard, J.P., Sodtke, R., Sterba, H., 2002a: Recommendations for standardized documentation and further development of forest growth simulators. Forstwissenschaftliches Centralblatt 121(3): 138–151. Pretzsch, H., Biber, P., Ďurský, J., 2002b: The single tree-based stand simulator SILVA: construction, application and evaluation. Forest Ecology and Management 162(1): 3–21. Pukkala, T., Miina, J., 1998: Tree-selection algorithms for optimizing thinning using a distance-dependent growth model. Canadian Journal of Forest Research 28(5): 693–702. R Core Team, 2012: R – A language and environment for statistical computing. Available at: http://www.r-project.org/ Seifert, T., 2003: Integration of wood quality, bucking and sorting in growth simulators sensitive to silvicultural treatment (In German with english captions). PhD Thesis. Technische Universität München, 314 p. Sirén, M., 1992: Harvennuspuun korjuujälki ja sen merkitys. Summary: The impact of thinning operations on the site. Teho 1: 8–10, 31. Talbot, B., Nordfjell, T., Suadicani, K., 2003: Assessing the utility of two integrated harvester-forwarder machine concepts through stand-level simulation. International Journal of Forest Engineering 14(2): 31–43. Tigercat 2011: Tigercat Pocket Product Guide 10th ed., Brantford, Canada. Vasiliauskas, R., 2001: Damage to trees due to forestry operations and its pathological significance in temperate forests: a literature review. Forestry 74(4): 319–336. Croat. j. for. eng. 37(2016)1
Mechanised Pine Thinning Harvesting Simulation: Productivity and Cost Improvements ... (1–15) Wang, J., Greene, W., 1999: An interactive simulation system for modeling stands, harvests, and machines. International Journal of Forest Engineering 10(1): 81–99.
S. A. Ackerman et al.
Warkotsch, P.W., van Huyssteen, L., Olsen, G.J., 1994: Identification and quantification of soil compaction due to various harvesting methods – a case study. South African Forestry Journal 170(1): 7–15.
Authors’ address:
Received: March 17, 2015 Accepted: May 2, 2015 Croat. j. for. eng. 37(2016)1
Simon A. Ackerman, MSc.* e-mail: simon.ackerman@icfr.ukzn.ac.za Stefan Seifert, PhD. e-mail: s.seifert@zukomplex.de Pierre A. Ackerman, PhD. e-mail: packer@sun.ac.za Thomas Seifert, PhD. e-mail: seifert@sun.ac.za University of Stellenbosch Department of Forest and Wood Science Private Bag X1 7602 Matieland SOUTH AFRICA * Corresponding author
15
Original scientific paper
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia Mohammad Reza Ghaffariyan, Mark Brown, Mauricio Acuna, John McGrath Abstract Mallee plantations have been integrated into wheat farms in Western Australia as a large-scale and multi-purpose woody crop since the 1990s. Mallee describes the growing habit of certain eucalypt species that grow with multiple stems shooting from an underground crown root (lignotuber), usually to a height of up to 10 meters. These types of plantations could be a considerable source of biomass to produce renewable energy. In this project the supply chain of Mallee was modelled using BIOPLAN’s linear programming model to investigate the impact of tree size, extraction distance and transport distance on supply chain costs. The harvesting system included a feller-buncher, front end loader, in-field chipper and truck. The mobile Bruks chipper was found to be more efficient than Peterson Pacific to chip Mallee trees. The results indicated that harvesting larger tree sizes can slightly diminish chipping cost. Extraction cost was very sensitive to the extraction distance in this case study. Long transport distances in larger management area (to meet higher energy demands) will highly increase the transport cost. From optimised supply chain cost and sensitivity analysis, the best practice for efficient Mallee biomass supply chain was suggested as following: harvesting Mallee trees when reaching larger size (about 0.3 m3 for a tree consisting of multiple stems with an average DBH of 5 cm to 10 cm per each stem), planning average extraction distance to be shorter than 1000–1500 m, establishing the Mallee plantations closer to energy plant with transport distance shorter than 100 km (with a radius of 50–75 km providing an effective compromise between cost and distance) or alternatively installing new bioenergy plants no farther than 100 km from existing Mallee plantations. Keywords: harvesting, chipping, productivity, operating cost, supply chain, optimisation
1. Introduction Mallee plantations have formed the basis of a processing industry in Australia for more than 100 years because of their natural abundance of eucalyptus oil. They can also be integrated into wheat belts to reduce soil salinity, give shade and shelter for animals, reduce erosion by acting as windbreaks and store carbon. Once established, the biomass can be harvested every few years. As the tree resprouts, or coppices, from the underground crown root (lignotuber), there is no need to replant (http://biomassproducer.com.au). Mallee describes the growing habit of certain eucalypt species that grow with multiple stems shooting from an underground crown root (lignotuber), usually to a height of up to 10 meters. Mallee eucalypts grow in the semiarid areas of southern Australia in: New South Wales, Croat. j. for. eng. 37(2016)1
north-western Victoria, southern South Australia and southern Western Australia. Biomass yield of 10–20 green metric tonnes (GMt) per hectare per year can be achieved when Mallees are grown in widely spaced two-row belts in alley systems in regions with adequate rainfall and suitable soil types (http://biomassproducer.com.au). Since the early 1990s almost 13,000 ha of Mallees have been established in Western Australia (WA). Eight different species and subspecies have been utilised, seven of which occur naturally in Western Australia (URS Australia, 2008). These large-scale plantations integrated into wheat farms (Wu et al. 2008) have been established as multi-purpose woody crop (Nuberg 1998, Spinelli et al. 2013). These integrated plantations can be a considerable source of woody biomass to produce renewable energy. However, unlike the forestry biomass supply chain
17
M. R. Ghaffariyan et al.
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17–25)
(Ghaffariyan et al. 2013a), the Mallee plantations established in specific rows within agricultural farms require appropriate combination of harvesting equipment and working method in order to make a profitable operation, while reducing the operating cost of biomass collection and minimum damages to the environment including agricultural land and plantations area. Some of the factors that influence operating costs of the biomass supply chain include moisture content (MC) (Acuna et al. 2012, Gautam et al. 2013, Visser et al. 2014), harvesting equipment efficiency and transportation distance (Kühmaier et al. 2007), capacity of the plant, efficiency of the combustion (Röser et al. 2011). In this case study the research objectives included: Þ c omparing the operating costs of different chippers to select the least expensive chipper for chipping Mallee trees; Þo ptimising the biomass supply chain for Mallee plantations using linear programming tool (called BIOPLAN) to minimise supply chain operating cost; Þv erifying the impact of wood extraction distance on the supply chain operating cost; Þ a nalysing the impact of Mallee tree size on operating cost; Þ i dentifying the relationship between transport distance, total harvesting volume per area and supply chain cost; Þ i dentifying maximum allowable transport distance for establishing the farms or building new energy plant.
2. Materials and methods 2.1 Study area Since natural drying rates of Mallee trees were not available in Western Australia, the results from a similar natural drying case study of harvesting residues of a Eucalyptus globulus plantation was applied in this modelling exercise (Ghaffariyan et al. 2013a). The site was located near the town of Rocky Gully in Western Australia (Ghaffariyan et al. 2013a). The site was about 30 km away from the weather station, but due to relatively consistent weather patterns in the area and to respect budget limitations, a dedicated weather station at the study site was not used. Study samples (12 samples per each sampling time, total of 120 samples per study period) were taken from a 103 m long, 4.8 m wide and 2.9 m tall pile of residues. The samples were collected from three cross sections (with the same spacing between each) at the top from the inner parts of the pile wherever possible, centre and bottom of the
18
pile, and their moisture content (MC) was measured on a monthly basis from August 2011 to August 2012. Each wood sample (disk of 1–2 kg) was obtained with the help of a chainsaw and contained normal biomass components (bazrk, leaves, small branches). The wood samples were stored in plastic bags and then oven dried at 105°C for a few days for MC measurements, which in turn were used to develop natural drying curves over time. In addition, total rainfall per month (mm) and average min. and max. temperatures were collected (Ghaffariyan 2013) from the closest weather station located in Rocky Gully (station 009964, Australian Government, Bureau of Meteorology). Based on the long term climate data (rainfall and max. and min. temperature), and taking the MC curve generated from the drying study as the basis, a number of other natural drying curves with different starting date of storage were estimated based on the approach described by Acuna et al. 2012. The case study area for modelling Mallee biomass supply chain was located near Katanning and Collie in Western Australia. It was assumed that from 14,000 ha of farm lands about 130,000 GMt of Mallee will be produced (as basic scenario) based on internal growth and yield modelling done by the department of the Western Australian Government responsible for the development of the Malley management system with wheat growers for the past 20 years. The harvesting system was assumed to be a combination of feller-buncher to fell the trees, a front-end loader to extract the bunches from the machine operating trails to the road side or chipping place (Spinelli et al. 2013) and a mobile chipper to chip the trees directly into trucks.
2.2 Method The first step of the modelling exercise was to determine the most efficient mobile chipper with the lowest operating cost for the same tree size, operation type and chip discharge place. A Peterson Pacific chipper (Spinelli et al. 2013) was compared with a Bruks 805.2 Table 1 Machine specifications of Bruks mobile chipper Model Base Engine, to power the chipper
Bruks 805.2 STC mobile chipper Forwarder-mounted (Ecolog forwarder, 300 HP, 223.8 kW) Scania diesel engine, 450 HP, 335.7 kW
Maximum diameter of logs to chip 50 cm Forwarder load capacity, chipper, bin and chips
19,500 kg
Croat. j. for. eng. 37(2016)1
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17–25)
STC mobile chipper (Ghaffariyan et al. 2012) mounted on a truck (Table 1) using a chipping productivity-cost model developed using Australian and Italian chipping operations (Ghaffariyan et al. 2013c). The second step of the modelling involved a sensitivity analysis using BIOPLAN, a tool specifically designed and implemented to optimise biomass supply chains (Acuna et al. 2012). Based on a linear programming model, the BIOPLAN tool was adopted in this project to investigate the impact of extraction distance, tree size, total harvesting weight and transportation distance on supply chain costs. Using natural drying curves as an explicit parameter, the objective of the model was to minimize total supply costs including harvesting, storage, chipping and transport for the Mallee supply chain. 2.2.1 Description of the optimisation model The BIOPLAN tool was used to determine the optimal supply of Mallee chips that satisfies the demand at the power plant. In the supply chain optimisation model, decisions on the volumes of Mallee to be harvested are made on a monthly basis and storage at the roadside of this material is allowed for a period of up to 24 months. This is a nominal time period and can be modified in BIOPLAN. Thus the optimal drying period will be determined after running the linear programming model, which will not exceed the maximum nominal drying period established in the tool (in this case 24 months). It is assumed that the chips produced from Mallee are consumed during the same month they arrive at the energy plant and, therefore, there are no costs associated with the storage of chips at the energy plant. In addition, the energy content of the chips being supplied from the Mallee plantations must meet the power plant monthly demand (tonnes) in year 2 (production year). It is assumed that a plant demand is a monthly volume of chips during the production year (year 2 for our modelling purposes), but the raw material (Mallee trees) may be harvested and stacked at the roadside for drying as from January Year 1. Thus, the optimal solution specifies when and how much to harvest (e.g. 100 m3 in March Year 1) and for how long to stack the logs before chipping and transport to the power plant (e.g. until January Year 2). The model provides the results in a series of matrices including among others: Þ t onnes and corresponding solid volume of Mallee to be harvested in each month (a decision variable); Þ l oose volume (lv) of chips produced at the roadside in each period; Croat. j. for. eng. 37(2016)1
M. R. Ghaffariyan et al.
Þn umber of truck loads delivered to the energy plant; Þ e nergy content of chips (GJ) arriving at the power plant. Summary tables also provide mean energy content (GJ) per m3 and tonne; Þh arvesting, extraction, chipping, storage and transportation costs. In addition, BIOPLAN estimates the total cost for the whole supply chain and total cost by activity (harvesting, storage, chipping and transportation) as well as total energy of the fuel supplied to the plant in GJ. 2.2.2 Parameters of the model The model parameters are listed in Table 2. Net calorific value was obtained from Perez et al. 2006 for E. Globules, as accurate figures for Mallee trees were not available. The basic wood density was assumed to be about 535 kg/solid m3 based on given information by Western Australia Plantation Resources (WAPRES). Woody biomass loss due to storage and manipulation was assumed to be 2% (Acuna et al. 2012, Laitila 2006). Volume and payload of trucks were gathered in field studies carried out by the authors in Western Australia (Ghaffariyan et al. 2013a). To calculate the cost associated with letting biomass dry, we added this amount to the cost of harvesting, extraction and piling the biomass following the same approach presented by Roise et al. 2013. The following variable definitions are used: »CP« is the cost to pile a GMt, »CH« is the cost to harvest and skid a GMt, »r« is the monthly interest rate (assumed to be 0.50% per month) and »T« is the length of drying in months. Then the drying cost at time of delivery is the future value of all the cost before drying the wood Table 2 Parameters and conversion factors used in the analysis Parameters/conversion factors Energy content of E. globulus at 0% MC, MJ/kg Basic density, kg/solid m3 3
Bulk density, kg/loose m
Solid content, chips from residues 3
3
Ratio loose m to solid m
Value 17.38 535 224.7 0.42 2.38
Truck payload, tonnes
40
Truck volume, loose m3
70
Transport distance, km
50
Material loss rate, %/month
2.0
Interest rate, %/month
0.58
19
M. R. Ghaffariyan et al.
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17–25)
(Eq. 1). Future versions of the tool will include the interest on the stumpage price paid to the farmers as well as the interest on the growing costs during the drying period (time between harvest and chipping of the logs). Cost of the dried material $ / GMt = (CH + CP ) × (1 + i )n
(1)
2.2.3 Mathematical optimising model The supply chain optimisation model was developed using linear programming and was implemented using the What’sBest® solver package for MS-Excel. Once the tables and solver engine were setup, a Visual Basic program was written to execute the model. The data sets, parameters and variables used in the mathematical formulation of the model are presented in Table 3. Objective function (FO) The objective function (Eq. 2) minimizes the total supply chain costs ($) associated with biomass harvesting, storage, chipping and transport.
)
(
FO = ∑SOLIDVOLt ,p × HARVESTC + STORAGEC t ,p + CHIPPINGC + ∑LOOSEVOLt ,p × TRANSPORTC (2)
t ,p
t ,p
Constraints Eq. 2 ensures that the energy content of the chips supplied satisfy the monthly demand at the plant.
∑
t≤p
LOOSEVOLt ,p × ENERGYt ,p ≥ DEMANDp ∀p ∈ P
(3)
Eq. 3 ensures that an even volume of Mallee trees are harvested evenly in each year. This allows for continuous work for the harvesting and haulage contractors.
∑ SOLIDVOL
p
t ,p
= ∑ SOLIDVOLt +1,p ∀t ∈{1…11,13…23}
(4)
p
Table 3 Sets, parameters and variables used in the mathematical formulation of the model Term
Definition Set
t,p = periods
t ∈T = {1…24} , p ∈ P = {13…24}
Parameters a
Conversion factor from m3 solid to m3 loose
DEMANDp
Energy demand in period p at the energy plant
ENERGYCt,p
Energy content of chips produced in period p from material harvested in period t, respectively. Depends on the moisture content of the material that is chipped
HARVESTC
Harvesting and extraction cost, $/m3 solid
STORAGECt,p
Storage cost ($/m3 solid) of whole trees stored at the roadside from period t to p (t ≤ p)
CHIPPINGC
Chipping cost ($/m3 solid) for whole trees chipped at the roadside
TRANPORTC
Transportation cost ($/m3) for tree chips (loose volume) transported to the energy plant
Variables SOLIDVOLt,p
Solid volume of trees harvested in period t, and stored at the roadside until period p for chipping at the roadside
LOOSEVOLt,p
SOLIDVOLt,p × a = Loose volume of chips from trees harvested in period t, and stored at the roadside until period p for chipping
20
Croat. j. for. eng. 37(2016)1
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17â&#x20AC;&#x201C;25)
2.2.4 Sensitivity analysis The sensitivity analysis was carried out to determine and quantify the impact of the mentioned operational factors on the supply chain cost. To run the analysis, one parameter was changed within its operational limits while holding the other parameters constant. Then the costs for different values of each parameter were graphed using a bar chart. For the sensitivity analysis, the average extraction distance was varied from 250 m (maximum distance of 500 m) to 2500 m (maximum distance of 5000 m) to study the impacts upon the costs for short and long distances. Tree size ranged from 0.1 m3 to 0.3 m3 in the analysis to verify its impact upon supply chain cost. The transport distance was varied from 25 km to 150 km for three levels of harvesting weight including 130,000 GMt per year, 200,000 GMt per year and 300,000 GMt per year to see how the operating cost changed for the range of parameters. Minimum and maximum acceptable ranges of moisture content of the delivered chips were assumed to be 10% and 40%, respectively, in this study.
3. Results 3.1 Chipping Chipping productivity and cost depend on tree size, chipper power, chip discharge place and type of operations (Ghaffariyan et al. 2013c). The chipping cost predicting model was run for both chippers considering the following factors; Bruks mobile chipper (based on truck) purchase price: $ 550,000, Peterson Pacific purchase price (truck mounted): $ 1,050,000, fuel consumption of Bruks chipper: 54.6 l/h, fuel consumption of Peterson Pacific: 100 l/h, Tree size: 0.1 m3, chip discharge: directly into trucks and type of operation: biomass operation. Based on the results obtained, the chipping cost for the Bruks chipper and Peterson Pacific chippers were 10.73 $/GMt and 11.00 $/GMt, respectively.
M. R. Ghaffariyan et al.
supply chain cost was found at total minimised costs of $11,750,561 to meet the energy demand of 53,000 MWh per month. Operating cost of the supply chain was 45.1 $/GMt (18.5 $/MWh). Table 4 presents the minimised operating costs of the supply chain.
3.3 Sensivity analysis 3.3.1 Impact of extraction distance on supply chain costs The average extraction distance was varied from 250 m to 2500 m (Spinelli et al. 2013), while the other parameters were held constant as described in the basic senario (Fig. 1). Longer extraction distances increase the extraction cost on a linear fashion (Spinelli et al. 2013). Longer extraction distances also result in longer travelling time for mobile chipper to move along the larger road side piles, which may impact the chipping costs slightly (Ghaffariyan et al. 2012). Harvesting cost in Fig. 1 is the sum of felling and extraction costs. 3.3.2 Impact of tree size on supply chain costs Decreasing chipping cost reduced total operating cost per GMt by only a small but significant amount (Fig. 2). Increasing tree size from 0.1 m3 to 0.3 m3 (while holding other factors constant) decreased proper chipping cost from 15.8 $/GMt (tree size of 0.1 m3) to 13.8 $/GMt (tree size of 0.3 m3) as larger tree size would increase the chipper productivity (Ghaffariyan et al. 2013c). Based on the availabe extraction productivity model in the study area on Mallee tree harvest-
3.2 Optimised supply chain for basic scenario For a tree size of 0.2 m3, average extraction distance of 1500 m, transport distance of 50 km, annual interest rate of 6% and harvesting volume of 130,000 GMt per year (plantation area of about 14,000 ha), the optimised Table 4 Minimised operating cost ($/GMt) of the supply chain Harvesting 19.3
Storage
Chipping
Transport
Total
0.3
14.7
10.8
45.1
Croat. j. for. eng. 37(2016)1
Fig. 1 Impact of different extraction distances on supply chain costs
21
M. R. Ghaffariyan et al.
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17â&#x20AC;&#x201C;25)
3.3.3 Impact of transport distance on supply chain costs As expected, changing transport distance from 25 km to 75 km, considering constant level of the other factors (harvesting volume of 130,000 GMt per year), resulted in a higher operating cost. A longer transport distance increases travelling time per turn and will then increase transportation costs. In this case study, transportation cost increased from 5.4 $/GMt with transport distance of 25 km to 16.2 $/km for transport distance of 75 km (10 km increase in distance results in additional cost of 2.2 $/GMt for transportaton). Total operating cost increased due to higher transportation cost for longer distances (Fig. 3).
Fig. 2 Impact of different tree sizes on supply chain costs
3.3.4 Impact of harvesting volume/transport distance on supply chain costs Total supply chain cost ($/GMt) was calculated for three levels of harvesting volume and supply points located at a close, medium and far distance from energy plant (depending on the size of managemnt area) as described in Table 5. Table 5 Harvesting volumes and transport distances for different scenarios Senario
Harvesting volume GMt/ha
Energy demand GWh/month
Plantation area, ha
Transport distance, km
A
130,000
53
14,000
<50
B
200,000
82
25,000
<100
C
300,000
122
32,000
<150
Transportation cost was increased for longer distances (while the other costs per GMt including harvesting, chipping and storage did not change) resulting in higher supply chain cost. Increasing harvesting volume per area and transport distance increased the total supply chain cost by a linear relationship (Fig. 4).
Fig. 3 Impact of transport distance on supply chain costs ing (Spinelli et al. 2013), the extraction costs would not change highly by varying tree size from 0.1 m3 to 0.3 m3 as Spinelli et al. (2014) emphasized on the impact of extraction distance and load weight. Tree size was not a significant variable in their extraction productivity prediciting model. However, tree size impacts the chipping cost as shown in Fig. 2 and larger tree size results in lower cost.
22
4. Discussion With a significant resource of Mallee planted as row crops in Western Australia, understanding the cost drivers for the supply chain is critical for mobilising and expanding the resource for commercial purposes. Emphasis in the development of potential supply chains for energy has been on harvesting systems that are productive in the very small tree sizes. The results of this optimised modelling show that tree size has a significant impact on the costs of the supply chains. This was due to the impact of tree size on chipping Croat. j. for. eng. 37(2016)1
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17â&#x20AC;&#x201C;25)
Fig. 4 Impact of harvesting volume per area per year on supply chain cost for different transport distances productivity (Fig. 2), where larger tree size resulted in higher productivity of the chipper. This impact reduced the chipping costs. Other chipping studies by Watson et al. (1986), Spinelli and Hartsough (2006) and Ghaffariyan et al. (2013b) found similar relationship between tree size and chipping productivity. Using the available forest harvesting system of a feller buncher, extraction and transportation of the chipped material will have a high impact on the costs. The row planting configuration used for Mallee plantings in WA, if not carefully planned and managed, will directly increase the delivered costs. Going from a 500 m to a 2000 m maximum extraction distance increases the delivered cost by 75%, while reducing the tree size three-fold from 0.3 m3 per tree to 0.1 m3 only increases the delivered costs by less than 5%. Spinelli and Hartsough (2001) compared front-end loader with a grapple skidder for extracting short rotation Eucalypt plantation in California. The loader was 40â&#x20AC;&#x201C;60% more productive than the grapple skidder, depending on extraction distance. According to their study, front-end loaders might be proper extraction machines for short rotation plantations, where tree characteristics, terrain and soil conditions allow their use. In their study, increasing distances (ranging from 37 m to 366 m) decreased the extraction productivity of the loader (and skidder), which is similar to our study results although we assumed longer extraction distances to be planned due to the size and shape of Mallee plantations in Western Australia. In our case study on Mallee plantaCroat. j. for. eng. 37(2016)1
M. R. Ghaffariyan et al.
tions, the front-end loader seemed to be an effective machine. Long-term planning as to where to establish future Mallee plantations and establish industrial users for the Mallee need to take into account the impact of transport costs. As shown in the modelling, if a 200,000 GMt demand has to extend its reach from an average distance of 75 km to 100 km, the total cost of supply is increased over 12%, which can easily be the difference between a commercial facility being viable or not. Other study on utilising the harvesting residues of eucalypt plantations in Western Australia (Ghaffariyan et al. 2013a) considered a variation of 20 km to 120 km for transportation distance, where the transportation cost increased significantly for longer distances due to longer time required for travelling between plantations and mill (KĂźhmaier et al. 2007; Sikanen et al. 2005). In addition to transport distances, other studies in Europe (Sosa et al. 2015a, Sosa et al. 2015b) have also concluded that truck configuration can also have a substantial impact on transport costs. These studies have also shown that the volume of chips to satisfy the demand of power plants could be very sensitive to changes in MC, which may have a significant impact on the spatial distribution of the supply points and the corresponding delivery volumes; so future studies in WA should further investigate the effect of these parameters. In addition, special consideration will have to be given to the scale of the energy plants that are planned to be built in WA. Optimally, matching supply and demand volumes in complex logistics scenarios will demand more sophisticated planning tools and require more efforts in operations planning. Yield of Mallee and other biomass materials per unit area, as well as the spatial location of these feedstocks will also determine the optimum size of the power plants (Cameron et al. 2007). Finally, future studies should investigate the impact of storage time on the quality and losses of the biomass products. Despite the positive impacts on supply costs, extending the storage period in order to further reduce moisture content may cause a loss of drying matter content due to fibre deterioration. To avoid dry matter content losses, it has been suggested that biomass products should remain at the stand to dry for a few days/weeks before been transported to roadside for further drying or chipping (Routa et al. 2015). This practice could also help reduce extraction costs. In addition to drying matter content losses, reduction in moisture content could negatively affect chip quality and chipping costs. Good quality wood chip fuel is produced by machines with sharp knives,
23
M. R. Ghaffariyan et al.
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17–25)
with the ability to vary the size of chips produced to meet end-user specifications (Kofman 2006). Chipping stems with reduced moisture content will increase the maintenance costs and blunt knives more often, which in turn will increase the amount of fines and the amount of overlong particles, producing chips with a less defined shape.
5. Conclusions Modelling has demonstrated that a properly sized chipper for the resource can have an impact on the chipping cost, thus providing a 0.25 ($/GMt) lower chipping cost. Extraction distance has the largest impact on overall supply costs within the expected ranges of operating situations, with the tested range of extraction distance effectively doubling the delivered chip cost. That is due to the fact that, as extraction distance increases, the front-end loader becomes increasingly inefficient. In reality, operators are likely to change technique and deploy a proper forwarder, thus dampening the effect of the increasing extraction distance. Tree size, like most forest harvesting operations, has a significant impact on the harvest costs but its impact on the delivered chip costs is limited to less than 5%, meaning that it is likely to be a suitable system even for small trees. Transport distance plays an important role in delivered chip costs with distances over 100 km tending to exceed economically viable supply costs being sought by industry.
Acknowledgements The authors would like to thank Raffaele Spinelli, PhD. for providing useful information on operating costs. They would also like to thank two reviewers of CJFE for providing valuable suggestions.
6. References Acuna, M., Anttila, P., Sikanen, L., Prinz, R., Asikainen, A., 2012: Predicting and controlling moisture content to optimise forest biomass logistics. Croatian Journal of Forest Engineering 33 (2): 225–238. Cameron, J., Kumar, A., Flynn, P., 2007: The impact of feedstock cost on technology selection and optimum size. Biomass Bioenergy 31(2/3): 137–44. Gautam, S., Pulkki, R., Shahi, Ch., Leitch, M., 2012: Fuel quality changes in full tree logging residue during storage in roadside slash piles in North western Ontario. Biomass Bioenergy 42: 43–50. Ghaffariyan, M.R., Sessions, J., Brown, M., 2012: Evaluating productivity, cost, chip quality and biomass recovery for a
24
mobile chipper in Australian road side chipping operations. Journal of Forest Science 58(2): 530–535. Ghaffariyan, M.R., 2013: The natural drying process of logs and harvesting residues – preliminary results. Hobart, Australia. Australian Forest Operations Research Alliance (AFORA), Industry bulletin No.: 2, 3 p. Ghaffariyan, M.R., Acuna, M., Brown, M., 2013a: Analysing the effect of five operational factors on forest residue supply chain costs: A case study in Western Australia. Biomass and Bioenergy 59: 486–493. Ghaffariyan, M.R., Sessions, J., Brown, M. 2013b: Roadside chipping in a first thinning Operation for Radiata Pine in South Australia. Croatian Journal of Forest Engineering 34(1): 91–101. Ghaffariyan, M.R., Spinelli, R., Brown, M., 2013c: A model to predict productivity of different chipping operations. Southern Forests: a Journal of Forest Science 75(3): 129–136. http://biomassproducer.com.au/producing-biomass/biomass-types/trees/mallee/#.Vcf4B9Yw-cw Kofman, P., 2006: Quality wood chip fuel. COFORD connects. Harvesting/Transportation No.6. Dublin, Ireland, 4 p. Kühmaier, M., Kanzian, C., Holzleitner, F., Stampfer. K., 2007: Wertschöpfungskette Waldhackgut Optimierung von Ernte, Transport und Logistik. Projektstudie im Auftrag von BMLFUW, Land Niederösterreich, Stadt Wien und ÖBf AG. Institut für Forsttechnik, Department für Wald und Bodenwissenschaften, Wien: Universität für Bodenkultur, 283 p. Laitila, J., 2006: Cost and sensitive analysis tools for forest energy procurement chains. Forest Studies/ Metsandulikud Uurimused 45(1): 5–10. Nuberg, I.K., 1998: Effect of shelter on temperate crops: a review to define research for Australian conditions. Agroforestry Systems 41(1): 3–34. Perez, S., Renedo, C.J., Ortiz, A., Manana, M., Silio, D., Peredo, J., 2006: Comparison of energy potential of the Eucalyptus globulus and Eucalyptus nitens. International conference on renewable energy and power quality; April 5–6, Mallorca, Spain; ICREPQ, 5 p. Röser, D., Mola-Yudego, B., Sikanen, L., Prinz, R., Gritten, D., Emer, B., Väätäinen, K., Erkkilä, A., 2011: Natural drying treatments during seasonal storage of wood for bioenergy in different European locations. Biomass and Bioenergy 35(10): 4238–4247. Routaa, J., Kolström, M., Ruotsalainen, J., Sikanen, L., 2015. Precision measurement of forest harvesting residue moisture change and dry matter losses by constant weight monitoring. International Journal of Forest Engineering 26(1): 71–83. Sikanen, L., Asikainen, A., Lehikoinen, M., 2005: Transport control of forest fuels by fleet manager, mobile terminals and GPS. Biomass Bioenergy 28(2):183–191. Sosa, A., Acuna, M., McDonnell, K., Devlin, G., 2015: Controlling moisture content and truck configurations to model Croat. j. for. eng. 37(2016)1
Optimised Harvesting Cost for Mallee Supply Chain in Western Australia (17–25)
M. R. Ghaffariyan et al.
and optimise biomass supply chain logistics in Ireland. Applied Energy 137: 338–351.
URS Australia, 2012: Oil mallee industry development plan for Western Australia. Industry development plan, 106 p.
Sosa, A., Acuna, M., McDonnell, K., Devlin, G., 2015: Managing the moisture content of wood biomass for the optimisation of Ireland’s transport supply strategy to bioenergy markets and competing industries. Energy 86: 354–368.
Visser, R., Berkett, H., Spinelli, R. 2014: Determining the effect of storage conditions on the natural drying of radiata pine logs for energy use. New Zealand Journal of Forestry Science 44(3): 1–8.
Spinelli, R., Hartsough, B.R. 2006: Harvesting SRF poplar for pulpwood; experience in the Pacific Northwest. Biomass and bioenergy 30(5): 439–445.
Watson, W.F, Sabo, R.F., Stokes, B.J. 1986: Productivity of in-woods chippers processing understory biomass. In: A proceedings of the Council on Forest Engineering, Improving productivity through Forest Engineering; September 29 – October 2; Mobile, AL, USA: 69–72.
Spinelli, R., Brown, M., Giles, R., Huxtable, D., Relano, R., Magagnotti, N., 2014: Harvesting alternatives for mallee agroforestry plantations in Western Australia. Agroforestry Systems 88(3): 479–487.
Wu, H., Fu, Q., Giles, R., Bartle, J., 2008: Production of Mallee Biomass in Western Australia: Energy Balance Analysis. Energy & Fuels 22(1): 190–198.
Authors’ address: Mohammad Reza Ghaffariyan, PhD.* e-mail: ghafari901@yahoo.com Prof. Mark Brown, PhD. e-mail: mbrown2@usc.edu.au University of the Sunshine Coast Locked Bag 4 4558 Maroochydore, Queensland AUSTRALIA Mauricio Acuna, PhD. e-mail: macuna@usc.edu.au University of the Sunshine Coast Private bag 12 7001 Hobart AUSTRALIA
Received: March 09, 2015 Accepted: August 25, 2015 Croat. j. for. eng. 37(2016)1
John McGrath, PhD. e-mail: John.mcgrath1206@gmail.com McGrath Consulting 13 Zenith Street, Shelley, Western Australia AUSTRALIA * Corresponding author
25
Original scientific paper
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages, Densities and Thinning Intensities Piotr S. Mederski, Mariusz Bembenek, Zbigniew Karaszewski, Agnieszka Łacka, Anna Szczepańska-Álvarez, Martyna Rosińska Abstract In economic terms, the main limiting factors in harvester application in thinning operations are the stand age and thinning intensity with respect to tree size. Furthermore, harvested mean tree size depends on initial stand density but also on the number of trees cut per hectare. The objective of the research was to estimate the impact of: Þ stand age (class), Þ increasing stand density in each age class (AC), Þ increasing number of trees for harvesting in each AC, Þ thinning intensity, on harvester productivity. 17, 19 and 20 sample plots were established within 3rd (AC3) 4th (AC4) and 5th (AC5) age classes, respectively. In each AC, sample plots were selected that had an increasing number of trees per hectare: 563÷1603, 323÷868 and 476÷836 trees ha–1, in AC3, AC4 and AC5, respectively. Also, in each AC, an increasing number of trees per hectare for harvesting was selected: 130÷853, 80÷315 and 108÷282, in AC3, AC4 and AC5, respectively, with the relevant increasing thinning intensity: 35÷84, 21÷77 and 34÷88 m3 ha–1. In each AC, the stands were divided according to different thinning intensity (THI): a<30, 30≤b≤60 and c>60 m3 ha–1, respectively. A Komatsu 931.1 harvester was used for the thinning operation in each stand. The lowest mean productivity was observed in AC3 (18.57 m3 h–1), which was statistically different to AC4 and AC5 (22.24 and 22.60 m3 h–1, respectively). Within each AC, productivity lowered as the number of trees per hectare increased in the initial stand. The productivity decreased in AC3 and AC5 with the increasing number of trees for harvesting, which was not the case in AC4. In relation to the THIs, the lowest mean productivity was obtained in THIa (16.19 m3 h–1), which was statistically different to THIb and THIc (21.44 and 21.98 m3 h–1, respectively). An increasing THI only influenced productivity positively in AC4 and AC5. It can be concluded that the productivity of the Komatsu 931.1 harvester increased along with: Þ older AC, Þ decreasing number of trees in the initial stand in each AC, Þ lowering number of trees for harvesting in AC3 and AC5, Þ increasing THI in only AC4 and AC5. Finally, in the present model, the larger the mean DBH of the trees for harvesting, the greater the productivity. However, the mean DBH has to be considered in conjunction with the number of trees for harvesting (which depends on AC and THI, as variables in the model) when productivity is analysed. Keywords: thinning operation, productivity curves, Scots pine (Pinus sylvestris L.)
1. Introduction Working on harvester productivity curves for Polish conditions is meaningful at this stage, as a large Croat. j. for. eng. 37(2016)1
number of harvesters are in operation. Since 1987, when the first harvester was introduced in Poland (Moskalik 2002), their numbers have grown considerably. Between 2006 and 2008, the number of harvesters
27
P. S. Mederski et al.
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
grew from 21 to ca. 170 (Kusiak 2008, Sowa 2009). A survey from 2011 and 2012 revealed that there were 351 harvesters in Poland, 16 of which were owned by the State Forests (Żabierek and Wojtkowiak 2012). Currently, it is estimated that there are ca. 450 harvesters operating in Poland, which are able to harvest ca. 30–35% of the total annual volume, estimated to 38 million m3 of timber. For this reason, competition between forest contractors is very high, which leads to a lowering of the price of timber harvesting within an open tender process. Therefore, it is in the interests of entrepreneurs to find out which thinning intensity is acceptable for a low price. Estimating harvester productivity is a basic step towards calculating the costs of forest operations. Special attention has to be paid to harvesters, the purchasing cost (ca. € 400,000) of which has to be balanced with sufficient working hours and an annual cut of 24,000 m3 in thinnings (Więsik 1998). It is essential, therefore, to allocate the appropriate machine for thinning operations in order to achieve satisfactory productivity. Studies completed so far have taken into account various factors, which have a direct impact on harvester productivity. These factors can be divided into four main groups: Þ stand conditions, Þ tree characteristics, Þ terrain conditions, Þ operator skills. Within the first group, »stand conditions«, productivity depends on: stand density and thinning intensity (Eliasson 1999), type of thinning and harvested volume per hectare (Suadicani and Fjeld 2001), frequency of tending operations or lack of them (Gerasimov et al. 2012), standard cuttings or stand damaged by wind (Szewczyk et al. 2014) and the spatial distribution of strip roads (Mederski 2006). Within the second group, »tree characteristics«, productivity depends on: size of selected trees (Iwaoka et al. 1999, Wang and Haarla 2002, Visser and Stampfer 2003, Nurminen et al. 2006), tree species, especially conifers versus broadleaves (Mederski 2006, Spinelli et al. 2010, Danilović et al. 2011, Visser and Spinelli 2012, Mederski 2013, Bembenek et al. 2015), tree shape and its morphological features (Evanson and McConchie 1996, Suchomel et al. 2012), thickness of branches (Glöde 1999) and criteria for tree selection for thinning (Eliasson and Lageson 1999). The third factor influencing productivity »terrain conditions« includes studies on slope gradient, terrain configuration and bearing capacity (Stampfer 1999, Picchio et al. 2012). Finally, harvester productivity also depends on the
28
level of operator skills (Purfürst 2010, Purfürst and Erler 2011). Additionally, it should be noted that harvester productivity has to be referenced to a certain decade or point in time. Nurminen et al. (2006) point out that progress in harvester development in time can positively influence productivity. Nurminen et al. (2006) confirmed that higher productivities were achieved for pine (by 14–35%), spruce (by 12–34%) and birch (by 5–21%) in comparison with data collected in the previous decade by Kuitto et al. (1994, as cited by Nurminen et al. 2006). From the above mentioned factors influencing productivity, tree size is the most common and most often studied. The influence of tree size on productivity is referred to as »piece-size law« – the bigger the piece (tree), the higher the productivity, as described by Visser and Spinelli (2012). Initially, this concept was described by Speidel (1952, as cited by Berg et al. 2014) as the »law of mass per piece«. It is also important to mention that at some point a piece, which is too large, can influence productivity negatively: a tree which is too large for machine capacity (size and power) may not be processed effectively; a point which was explained well by Visser and Spinelli (2012). Taking into account the above mentioned research results, it was hypothesised that higher productivity can be achieved when: Þ the stand is older, Þ there is a smaller number of trees for cutting, although with the same mean thinning intensity (THI), Þ the THI is higher. In fact, all of these three factors include the »piecesize law«. Therefore, the aim of this paper was to find out the differences in productivity in pure pine stands characterised by increasing: Þ age, Þ number of trees per hectare in the initial stand in each age class (AC), Þ number of trees for harvesting in each AC, Þ thinning intensity in each AC, as all of these factors influence mean DBH of harvested trees.
2. Material and methods Pure Scots pine (Pinus sylvestris L.) stands were selected for study in Drawno Forest District, northwest Poland (E 15°50’–16°00’, N 53°10’–53°13’). The research was carried out in pure pine stands grown in the same soil, site and weather conditions. The stand compartments were divided according to age class and the number of trees per hectare. 56 sample Croat. j. for. eng. 37(2016)1
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
P. S. Mederski et al.
Table 1 Stand characteristics AC4
AC5
Trees Mean Trees Mean ha–1 DBH ha–1 DBH
Compartment
Trees Mean Trees Mean ha–1 DBH ha–1 DBH
THI ratio1
THI, m3 ha–1
Harvested trees
THI, m3 ha–1
Whole stand trees
THI ratio1
Harvested trees
Compartment
Whole stand trees
THI, m3 ha–1
Harvested trees
THI ratio1
Compartment
Sample plot
AC3 Whole stand trees
1
107f I
563
25.9
143
24.0
0.9
53
103k
323
33.1
105
29.8
0.9
51
96h
476
28.2
108
25.0
0.9
34
2
107fII
703
23.4
130
21.3
0.9
35
90d
460
27.5
98
23.6
0.9
35
127a
518
27.5
188
24.5
0.9
88
3
93c
863
19.4
273
17.1
0.9
45
80c
483
28.1
95
24.0
0.9
34
134c
534
25.5
124
21.1
0.8
37
4
31l
917
20.1
250
18.4
0.9
57
95d
513
26.4
98
23.6
0.9
34
170a
564
27.9
170
25.1
0.9
72
5
138h
957
19.5
397
17.1
0.9
62
95a
540
25.7
125
24.0
0.9
39
97b
594
25.7
186
24.6
1.0
61
6
32k
1043
19.7
367
17.8
0.9
38
116c
560
25.5
105
22.8
0.9
27
98hI
594
26.5
172
23.2
0.9
56
7
172c 1070
Trees Mean Trees Mean ha–1 DBH ha–1 DBH
19.9
403
17.9
0.9
73
151a
570
24.6
153
21.9
0.9
36
99g
596
27.1
264
24.0
0.9
65
1083
18.6
360
16.8
0.9
57
152b
593
24.7
138
22.1
0.9
38
122a
618
24.6
176
22.1
0.9
49
9
166g 1097
20.3
410
16.9
0.8
53
93g
625
23.7
185
21.2
0.9
44
13a
632
24.1
212
21.0
0.9
55
10
37aI
1123
19.6
413
16.7
0.9
60
119aI
660
21.3
195
18.3
0.9
26
73f
634
24.6
238
22.0
0.9
53
11
12a
1153
18.5
423
18.1
1.0
72
89b
663
28.9
80
21.4
0.7
21
98hII
640
25.0
154
20.5
0.8
46
12
41a
1270
19.2
567
15.4
0.8
71
153f
663
24.7
193
20.7
0.8
44
155b
644
26.4
162
23.2
0.9
57
13
80g
1270
20.4
397
19.1
0.9
84
14c
683
23.7
200
22.3
0.9
55
154a
648
25.9
200
22.3
0.9
67
14
94aI
1297
17.4
513
14.9
0.9
66
172f
695
25.6
195
23.4
0.9
65
72b
682
24.7
282
22.1
0.9
77
15
37aII 1403
18.3
490
16.6
0.9
64
84c
708
24.3
300
22.0
0.9
77
143d
708
21.8
248
19.0
0.9
59
16
40c
1603
15.4
853
16.8
1.1
62
119aII 708
20.5
200
17.4
0.8
27
5g
720
22.3
238
20.0
0.9
47
17
94aII 1603
15.4
630
13.2
0.9
55
83b
748
24.1
163
21.5
0.9
76
78l
756
21.5
252
19.6
0.9
57
8
30i
18
–
–
–
–
–
–
–
171h
768
23.8
255
21.0
0.9
61
100f
758
23.7
172
21.1
0.9
68
19
–
–
–
–
–
–
–
46b
868
22.3
315
20.4
0.9
56
6fI
836
21.8
176
21.5
1.0
53
20
–
–
–
–
–
–
–
–
–
–
–
–
–
–
6fII
836
23.6
156
18.8
0.8
48
Mean
–
1119
19.5
413
17.5
0.9
59
–
622
25.2
168
22.2
0.9
44
–
649
24.9
194
22.0
0.9
57
Median
–
1097
19.5
403
17.1
0.9
60
–
660
24.7
163
22.0
0.9
39
–
637
24.9
181
22.1
0.9
56
1
Min
–
563
15.4
130
13.2
0.8
35
–
323
20.5
80
17.4
0.7
21
–
476
21.5
108
18.8
0.8
34
Max
–
1603
25.9
853
24.0
1.1
84
–
868
33.1
315
29.8
0.9
77
–
836
28.2
282
25.1
1.0
88
Sd
–
280
2.5
174
2.4
0.06
13
–
126
2.9
69
2.6
0.04
16
–
98
2.0
47
2.0
0.04
13
n
–
17
17
17
17
17
17
–
19
19
19
19
19
19
–
20
20
20
20
20
20
thinning intensity ratio – understood us ratio of mean DBH of extracted trees to mean DBH of whole stand trees (Lagesson 1997, Mederski 2006)
plots were selected: 17, 19 and 20 in the AC3 (41÷60 y.o.), AC4 (61÷80 y.o.) and AC5 (81÷100 y.o.) In the compartments, sample plots were marked with an area of 0.3, 0.4 and 0.5 ha in the stands of AC3, AC4 and AC5. Bigger sample plots were selected in older stands characterised by a lower number of trees, though sufficient for the experiment. In each sample plot, the same pattern of strip roads was designed, with a maximum width of up to 4 m and a distance between them of 20 m (from axis to axis of the strip road, Mederski 2006). The selected sample plots had an increasing number of trees within each AC (Table 1). This depended Croat. j. for. eng. 37(2016)1
mostly on the stand health (due to pest and fungi development), as all of them grew in similar soil conditions with the same site index. There were 563 to 1603 trees ha–1 in AC3, 323 to 868 trees ha–1 in AC4, and 476 to 836 trees ha–1 AC5. A higher mean number of trees were selected for thinning in AC3 (413 per ha), than in AC4 and AC5 (168 and 194 per ha, respectively). On each sample plot, the DBHs of all the harvested trees were measured with an electronic calliper with an accuracy of 0.1 cm. The number was marked with white paint on each measured tree. The mean DBH in AC3 was 17.5 cm, which was lower than in AC4 and AC5 (22.2 and 22.0 cm, respec-
29
P. S. Mederski et al.
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
tively). The stands of AC5 presented similar mean values in terms of the number of harvested trees per ha as well as the mean DBH in comparison with AC4 (Table 1). It was found that most of the AC5 stands had different thinning schedules in terms of time, and the thinning was postponed to a later time – from AC4 to AC5. In each AC, division according to thinning intensity (THI) was applied: a<30, 30≤b≤60 and c>60 m3 ha–1, respectively. This division was carried out after statistical analysis, which suggested borders at ca. 30 and 60 m3 ha–1. Finally, in each AC, sample plots were of an increasing number of trees and of increasing thinning intensity groups. However, only in AC4 all thinning intensities were recorded: THIa, THIb and THIc. In AC3 and AC5, there were only THIb and THIc. Silvicultural treatments were prescribed according to current standards: more intensive thinning was proposed in stands with a bigger number of trees, with the idea that only one intervention was expected in one decade. Positive thinning was applied, which means that trees of lower importance were selected to make the best possible conditions for future trees. For the thinning operation, a Komatsu 931.1 harvester was used with a powerful 193 kW (7.4 l stroke volume) engine. The machine was equipped with a CRH 22 boom, with a reach of up to 9.8 m and a lifting torque (gross) 217 kNm. The Komatsu 365 harvester head had three steel feed rollers and five delimbing knives (four moving). Thinning was carried out by two harvester operators aged 39 and 44, both with a 7-year experience. The two operators worked in different shifts on sample plots selected randomly. On all the sample plots, the same types of assortments were harvested: 2.85, 2.50 and 2.45 m long saw logs, pulp wood and industrial wood, respectively. Time studies were carried out with respect to the productive machine hour (PMH) without delays (Mederski 2006). Productivity (P) was calculated as:
P=
V TPMH
(1)
where: V volume of harvested timber, m3; TPMH time of productive machine hour (moving, crane and head positioning, cutting, felling, delimbing and bucking; without delays), h. All the delays were excluded from the study in order to only compare the pure productive time from each sample plot. In fact, there were some delays and repairs during the study, however, they occurred randomly and were not related to particular stand condi-
30
tions (AC or THI); therefore they were excluded from further analysis. Thanks to this, the field studies were more accurate, but the analyses were performed at the level of total PMH per sample plot and total volume of harvested timber per sample plot. The timber volume under bark was taken from the harvester computer. In order to compare the mean productivity with respect to the experimental variables, prior to the variance analysis, the Lilliefors test, for the normal distribution of data (Thode 2002), was done followed by the Bartlett test of homogeneity of variances (Zar 1999) in the analysed model. A multiple analysis of variance was done with respect to the estimated interaction between the analysed factors. Based on the interaction plots, interactions between factors were chosen (Ott 1984). To discover which means were significantly different from each other, the Tukey’s post hoc test was used for factors significant in ANOVA (Ott 1984). Statistical inference was performed at significance level α = 0.05. Pearson’s correlation matrix was determined for the studied characteristics. Based on the results of ANOVA and values of correlation coefficients, the multiple regression was proposed, where the influence of the experimental factors statistically different in the analysis of variance on the mean productivity per PMH was determined. The program package R (3.0.2) was used for the calculations (R Development Core Team 2013).
3. Results In each AC, productivity lowered as the number of trees increased within the considered AC (Fig. 1). As the trend lines show, the highest productivities were mostly in AC5 and the lowest in AC3. Also in AC3, an increasing number of trees in the initial stand had the biggest impact on lowering productivity. In AC4, this trend was the weakest. In some of the stands in AC4 (end of the curve), higher productivities were achieved than in AC5. What is very important here is that, generally, the productivities in AC3 were lower than in AC5, even though the average THIs were similar: 59 and 57 m3 ha–1, respectively (Table 1). Using the Lilliefors test, based on KolomogorovSmirnov statistics, it was shown that the analysed factor – productivity – was normally distributed (D=0.1028, p-value=0.1481). With the application of the Bartlett test, it was decided that there was no evidence to reject the null hypothesis for the test of homogeneity of variance for the analysed factors (K-squared=2.1152, df=2, Croat. j. for. eng. 37(2016)1
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
P. S. Mederski et al.
Table 3 Tukey’s tests for mean productivities obtained within the analysed ACs and THIs, a=0.05; identical superscripts in column »groups« denote no significant difference between mean values (according to Tukey’s HSD test)
AC
THI
Fig. 1 Changes in productivities in numbered sample plots with trend lines (order of sample plots is according to increasing number of trees in the initial stands, as in Table 1) p-value=0.3473). The multiple analysis of variance for productivity (Table 2) showed significant differences in mean effects in relation to AC and THI, and for those factors, Tukey’s post hoc tests were carried out (Table 3). In contrast, shift and operator had no significant impact on productivity (Table 2). Considering the division of stands into age classes, the lowest mean productivity was observed in AC3 (18.57 m3 h–1), which was statistically different to AC4 and AC5 (22.24 and 22.60 m3 h–1, respectively, Table 2 Anova Table (Type III tests) Sum. sq.
D.f.
F value
p-value
AC
210.6
2
8.3873
0.0009 ***
THI
218.4
2
8.6975
0.0007 ***
SHIFT
60.4
2
2.4049
0.1032
OPERATOR
7.7
1
0.6171
0.4367
AC: SHIFT
63.2
4
1.2587
0.3022
AC: OPERATOR
25.4
2
1.0101
0.3733
SHIFT: OPERATOR
18.3
2
0.7272
0.4895
RESIDUALS
502.1
40
–
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Croat. j. for. eng. 37(2016)1
–
Mean
Std. error
r
Min.
Max.
Groups
3
18.57
0.9198
17
12.64 27.60
a
4
22.24
1.0960
19
12.99 32.61
b
5
22.60
0.8834
20
17.84 30.90
b
a
16.19
1.1337
4
12.99 18.07
a
b
21.44
0.7398
32
12.64 32.61
b
c
21.98
1.0793
20
15.56 30.90
b
Table 3) as Tukey’s test has shown (HSD=2.829186; r.harmonic=18.58). In AC3, a high number of trees were cut (413 per hectare) with a low mean DBH of 17.5 cm (Table 1). Low productivity was achieved in AC3 even when mean thinning intensity was the highest (59 m3 ha–1) in comparison with AC4 and AC5, where on average 44 and 57 m3 ha–1 was harvested, respectively. At the same time, the mean productivities achieved in the stands within the lowest THIa amounted to 16.19 m3 h–1 and were statistically different (Tukey test, HSD=4.052411; r.harmonic=9.06) from those from stands of THIb and THIc (21.44 and 21.98 m3 PMH–1, respectively, Table 3). These low productivities occurred only in AC4, where either a small number of trees per hectare were harvested (sample plots 6 and 11) or they were of small mean diameter (sample plots 10 and 16, Table 1). In the case of AC3, the increasing THI did not influence the productivity at all. In fact, it lowered slightly when more intensive thinning was applied (Fig. 2a). Furthermore, in AC3, the productivity decreased considerably, when more trees were harvested (Fig. 2d) with a smaller mean DBH (sample plots 14, 15, 16 and 17, Table 1). This was not the case in the older stands (AC4 and AC5 with thicker trees compared to AC3, Table 1) and especially in AC4 where increased THI had the biggest impact on productivity results (Fig. 2a). In principle, the productivity depended on the mean DBH of the harvested trees: the larger the DBH, the higher the productivity (Fig. 2b). In the case of the impact of the mean DBH on productivity in each AC, the data were more clustered along the curves in comparison with the THI, which suggests that »piece-size law« had the biggest impact on the final result within the considered stand conditions.
31
P. S. Mederski et al.
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
Fig. 2 Productivity as a function of different stand parameters (black curves represent mean values for all data)
Confirmation of this impact was obtained when the correlation matrix was built, where productivity was correlated with the mean DBH of the harvested trees (expressed by the highest factor: 0.78, Table 4). As the AC groups (3, 4 and 5) and THI groups (a, b and c) had a statistically significant impact on productivity, a model of multiple regression was proposed with the DBHmean of the harvested trees:
where: Y mean productivity per PMH; DBHmean mean DBH of harvested trees; δi Kronecker’s delta, 1 stand of i − age class ( of i − thinning intensity ) di = in other case 0
the proposed model, the determination factor Y = –7.8920 + 1.2494 × DBH mean – 0.8587d 4 – 1.3237d 5 + 3.7631d bFor + 5.2550 dc R2 was 0.7168, and the highest significance of estimat DBH mean – 0.8587d 4 – 1.3237d 5 + 3.7631d b + 5.2550d c (2) ed factor: DBHmean (Table 5).
32
Croat. j. for. eng. 37(2016)1
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
P. S. Mederski et al.
Table 4 Correlation matrix Number of whole stand trees
Number of harvested trees
Mean DBH of whole stand trees
Mean DBH of harvested trees
THI
Harvested volume
Mean volume from one tree
Productivity
Number of whole stand trees
1.00
0.81
–0.70
–0.67
0.47
0.11
–0.55
–0.46
Number of harvested trees
0.81
1.00
–0.71
–0.62
0.56
0.04
–0.72
–0.36
Mean DBH of whole stand trees
–0.70
–0.71
1.00
0.94
–0.21
0.46
0.83
0.64
Mean DBH of harvested trees
–0.67
–0.62
0.94
1.00
–0.09
0.47
0.72
0.78
THI
0.47
0.56
–0.21
–0.09
1.00
0.50
–0.31
0.22
Harvested volume
0.11
0.04
0.46
0.47
0.50
1.00
0.41
0.37
Mean volume from one tree
–0.55
–0.72
0.83
0.72
–0.31
0.41
1.00
0.35
Productivity
–0.46
–0.36
0.64
0.78
0.22
0.37
0.35
1.00
Table 5 Significance of estimated factors Estimate
Std. error
t value
Pr(>|t|)
Intercept
–7.8920
2.7677
–2.851
0.0063**
DBHmean
1.2494
0.1558
8.018
0.0000***
AC4
–0.8587
1.2251
–0.701
0.4866
AC5
–1.3237
1.0956
–1.208
0.2326
THIb
3.7631
1.4920
2.522
0.0149*
THIc
5.2550
1.5780
3.330
0.0016**
In general, the bigger number of trees in the initial stand (consisting of more trees with a smaller DBH), the lower the productivity results (Fig. 2c). This »piecesize law« can also be applied when only trees for extraction are considered (AC3 and AC5, Fig. 2d). However, in AC4, a large number of harvested trees had a positive impact on the average productivity.
4. Discussion with conclusion Harvester use for thinning operations in the younger (AC3) and older stands (AC4 and AC5) resulted in different productivities. The stands in AC3, with a considerably larger number of trees in the initial stand, as well as for extracting, gave a lower productivity in comparison with the older stands. It is interesting that in AC3, increased thinning intensity did not raise productivity (Fig. 2a). At this stage of stand development, the trees for harvesting were of a small DBH, which in this case, were 17.5 cm on average, but started from as small as 13.2 cm (Table 1), and the mean volume obtained from each harvested tree amounted to 0.14 m3. Croat. j. for. eng. 37(2016)1
In this case, it was not only the »piece-size law« that had the biggest impact on the results obtained, but also the fact that in a stand with a large density, manoeuvring a crane and positioning the head takes more time than in a stand with a low density. Large density stands also require more careful and accurate work as they are more vulnerable to residual damage (Karaszewski et al. 2013, Bembenek et al. 2013a, 2013b, Stańczykiewicz et al. 2015). In the same figure presenting productivity as a function of THI (Fig. 2a), the curve for AC4 grew most rapidly. In AC4 there were two particular sample plots (14 and 15) with a higher than average number of trees for removal (195 and 300) with large mean DBH (23.4 and 22.0 cm, Table 1). Actually, as the mean statistical productivity showed (Table 2), in this particular case, the stands of AC4 and AC5 were of similar potential. However, in the stands of AC5, the average timber volume obtained from one tree was the highest: 0.29 m3, while in AC4 it was 0.26 m3. A higher mean volume of timber from one harvested tree together with a slightly bigger number of trees harvested per hectare (possibly optimal), eventually led to a slightly larger mean productivity in AC5 (22.60 m3 h–1, Table 2). Also in AC5, in comparison with AC4, there were smaller standard deviation values for: Þ the number of harvested trees, Þ the mean DBH of the harvested trees, Þ the mean THI. Those factors also had a positive impact on the higher mean productivity in AC5. However, this higher mean productivity in AC5 should not be linked with Fig. 2a, where the AC4 trend line presents partially higher productivities than in AC5. The lowest thinning intensities (THIa<30 m3 ha–1) on particular sample plots
33
P. S. Mederski et al.
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
were observed only in AC4 (Fig. 2a). This should be considered carefully as this is not a very common case. Currently in Poland, according to silvicultural prerequisites, a maximum of one thinning treatment per 10 years should be applied, which means that much more than 30 m3 per hectare can be harvested. However, if a very low intensity is achieved, low productivity may be expected. The productivity curves in the present study were linear, which can also be seen in other productivity studies (Sirén and Aaltio 2003, Nakagawa et al. 2007). The authors of the present work decided to use linear curves for three reasons: Þ they fitted best to the data distribution, Þ t he linear model was characterised by the highest determination factor R2, Þ t here was also argumentation based on previous findings: the present study was limited only to thinnings in stands where the harvested trees gave a mean timber volume from 0.14 up to 0.29 m3. This was rather at the low end of potential tree sizes to be harvested. As presented in studies by Spinelli et al. (2010) and Visser and Spinelli (2012), using more complex regression models is more suitable when a broad raw data set including harvested trees of small and large volumes (e.g. from 0.3 up 5.2 m3) is considered. The proposed model (2) also consists of simple factors including DBH, which is easy to obtain. Some researchers have used the volume of harvested tree as a variable of productivity (Spinelli et al. 2010, Visser and Spinelli 2012), however in the case of the data presented in this paper, using the mean volume of harvested tree Vmean, instead of DBHmean, in a model:
positive impact on productivity. Fig. 2d shows that it was only in AC4 that the increasing number of harvested trees had a positive impact on the growing productivity. This curve has to be taken with caution as data dispersion is high and more tests should be conducted to find out how the increasing number of harvested pine trees influences productivity in AC4. Fig. 1 also shows that in AC4, the sample plots 14, 15 and 17÷19 had a particularly large number of trees for harvesting with large mean DBHs and high mean THIs. It can be concluded that the productivity of the Komatsu 931.1 harvester increased along with: Þo lder AC, Þd ecreasing number of trees in the initial stand in each AC, Þ l owering number of trees for harvesting in AC3 and AC5, Þ i ncreasing THI in only AC4 and AC5. Finally, as the model (2) presents, the larger the mean DBH of the trees for harvesting, the greater the productivity. The same model also confirms that within the same mean DBH, the older the AC, the lower the productivity. However, the last factor in the model, THI, cannot be changed freely with the same (fixed) mean DBH. Increasing THI requires the removal of a larger number of trees, resulting in the cutting of thicker trees and, as a consequence, the mean DBH of the harvested trees has to increase. Therefore, the mean DBH has to be considered in conjunction with the number of trees for harvesting (which depends on AC and THI, as variables in the model) when productivity is analysed.
Acknowledgements
Y = 8.607 + 6.990 × Vmean + 3.901d 4 + 2.891d 5 + 7.720d b + 9.401d c This study was completed as part of the project: »Productivity of harvester thinning operations in pine stands of different thinning intensities« supported 90 × Vmean + 3.901d 4 + 2.891d 5 + 7.720d b + 9.401d c (3) by the Regional Directorate of the State Forests in gave a much lower determination factor R2=0.3934. Szczecin, Poland. The authors would like to thank the The number of trees per hectare is also a limiting two anonymous reviewers for their useful comments factor. Fig. 2c shows that the bigger the initial number which greatly improved the final version of the paper. of trees before thinning per hectare, the lower the productivity. In this case, it is linked to the natural stand 5. References condition: the bigger the number of trees within one AC on the unit area, the smaller the diameter of a sinBembenek, M., Giefing, D.F., Karaszewski, Z., Mederski, P.S., gle tree (in the stand and for harvesting). These kinds Szczepańska-Álvarez, A., 2013a: Uszkodzenia drzew w of stands are not only less attractive for thinning opnastępstwie trzebieży wczesnych w nizinnych drzewostanach świerkowych. Sylwan 157(10): 747–753. erations with harvesters giving smaller productivity, but can also give lower income from the timber sold Bembenek, M., Giefing, D.F., Karaszewski, Z., Mederski, P.S., due to a bigger share of timber with small diameters Szczepańska-Álvarez, A., 2013b: Uszkodzenia drzew w (Bembenek et al. 2014). A large number of trees for nizinnych drzewostanach świerkowych podczas zabiegu harvesting together with high mean DBH that have a trzebieży późnej. Sylwan 157(12): 892–898.
34
Croat. j. for. eng. 37(2016)1
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36) Bembenek, M., Karaszewski, Z., Kondradzki, K., Łacka A., Mederski, P.S., Skorupski, M., Strzeliński, P., Sułkowski, S., Węgiel, A., 2014: Value of merchantable timber in Scots pine stands of different densities. Drewno. 57(192): 133–142. Bembenek, M., Mederski, P.S., Karaszewski, Z., Łacka, A., Grzywiński W., Węgiel A., Giefing, D.F., Erler, J., 2015: Length accuracy of logs from birch and aspen harvested in thinning operations. Turkish Journal of Agriculture and Forestry 39: 845–850. Berg, S., Schweier, J., Brüchert, F., Lindner, M., Valinger E., 2014: Economic, environmental and social impact of alternative forest management in Baden-Württemberg (Germany) and Västerbotten (Sweden). Scandinavian Journal of Forest Research 29(5): 485–498. Danilović, M., Tomašević, I., Gačić, D., 2011: Efficiency of John Deere 1470D ECOIII harvester in poplar plantations. Croatian Journal of Forest Engineering 32(2): 533–548. Eliasson, L., 1999: Simulation of thinning with a single-grip harvester. Forest Science 45(1): 26–34. Eliasson, L., Lageson, H., 1999: Simulation study of a singlegrip harvester in thinning from below and thinning from above. Scandinavian Journal of Forest Research 14(6): 589– 595. Evanson, T., McConchie, M., 1996: Productivity measurements of two Waratah 234 hydraulic tree harvesters in radiate pine in New Zealand. Journal of Forest Engineering 7(3): 41–52. Gerasimov, Y., Senkin, V., Väätäinen, K., 2012: Productivity of single-grip harvesters in clear-cutting operations in the northern European part of Russia. European Journal of Forest Research 131(3): 647–654. Glöde, D., 1999: Single and double-grip harvesters – productive measurements in final cutting of shelterwood. Journal of Forest Engineering 10(2): 63–74. Iwaoka, M., Aruga, K., Sakurai, R., Cho, K., Sakai, H., Kobayashi, H., 1999: Performance of small harvester head in a thinning operation. Journal of Forest Research 4(3): 195–200. Karaszewski, Z., Giefing, D.F., Mederski, P.S., Bembenek, M., Dobek, A., Stergiadou, A., 2013: Stand damage when harvesting timber using a tractor for extraction. Forest Research Papers 74(1): 27–34. Kuitto, P.J., Keskinen, S., Lindroos, J., Oijala, T., Rajamäki, J., Räsänen, T., Terävä, J., 1994: Puutavaran koneellinen hakkuu ja metsäkuljetus. Summary: Mechanized cutting and forest haulage. Metsäteho Report 410 p. Kusiak, W., 2008: Tendencje na rynku harwesterów i forwarderów w Polsce. ������������������������������������ In: Bezpieczeństwo pracy w nowoczesnym leśnictwie (Romankow, J., ed.), Katedra Inżynierii Środowiska Pracy UP, Poznań, 24–36 p. Lageson, H., 1997: Effects of thinning type on the harvester productivity and on the residual stand. Journal of Forest Engineering 8(2): 7–14. Croat. j. for. eng. 37(2016)1
P. S. Mederski et al.
Mederski, P.S., 2006: A comparison of harvesting productivity and costs in thinning operations with and without midfield. Forest Ecology and Management 224(3): 286–296. Mederski, P.S., 2013: The potential of harvester use for thinning operations in mixed birch-pine stands. Wydawnictwo UP w Poznaniu. Moskalik, T., 2002: Rozwój technik i technologii maszynowego pozyskiwania drewna. Sylwan 146(10): 31–37. Nakagawa, M., Hamatsu, J., Saitou, T., Ishida, H., 2007: Effects of tree size on productivity and time required for work elements in selective thinning by a harvester. International Journal of Forest Engineering 18(2): 24–28. Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanised cut-to-length harvesting system. Silva Fennica 40(2): 335–363. Ott, L., 1984: An Introduction to Statistical Methods and Data Analysis. Duxbury Press. Picchio, R., Neri, F., Petrini, E., Verani, S., Marchi, E., Certini, G., 2012: Machinery-induced soil compaction in thinning two pine stands in central Italy. Forest Ecology and Management 285: 38–43. Purfürst, F.T., 2010: Learning Curves of Harvester Operators. Croatian Journal of Forest Engineering 31(2): 89–97. Purfürst, F.T., Erler, J., 2011: The Human Influence on Productivity in Harvester Operations. International Journal of Forest Engineering 22(2): 15–22. R Development Core Team, 2013: R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria, http://www.r-project.org. Sirén, M., Aaltio, J., 2003: Productivity and costs of thinning harvesters and harvester-forwarders. International Journal of Forest Engineering 14(1): 39–48. Sowa, J., 2009: Współczesne pożytki z lasu. In: Leśnictwo w górach i rejonach poprzemysłowych (Starzyk, J., ed.), Wydawnictwo UR, Kraków, 129–152 p. Speidel, G., 1952: Das Stückmassegesetz und seine Bedeutung für den internationalen Leistungsvergleich bei der Forstarbeit. Dissertation, University of Hamburg. Spinelli, R., Hartsough, B.R., Magagnoti, N., 2010: Productivity standards for harvesters and processors in Italy. Forest Product Journal 60(3): 226–235. Stampfer, K., 1999: Influence of terrain conditions and thinning regimes on productivity of a track-based steep slope harvester. In: Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium (Sessions, J., Chung, W., ed.), Corvallis, Oregon, 78–87 p. Stańczykiewicz, A., Sowa, J.M., Leszczyński, K., Kulak, D., Szewczyk, G., 2015: Uszkodzenia drzew i odnowienia w wyniku pozyskania drewna z użyciem urządzeń agregowanych z ciągnikami rolniczymi w trzebieżowych drzewostanach świerkowych. Sylwan 159(3): 201–210.
35
P. S. Mederski et al.
Estimating and Modelling Harvester Productivity in Pine Stands of Different Ages ... (27–36)
Suadicani, K., Fjeld, D., 2001: Single-tree and group selection in mountain Norway spruce stands: factors influencing operational efficiency. Scandinavian Journal of Forest Research 16(1): 79–87. Suchomel, C., Spinelli, R., Magagnotti, N., 2012: Productivity of processing hardwood from coppice forests. Croatian Journal of Forest Engineering 33(1): 39–47. Szewczyk, G., Sowa, J.M., Grzebieniowski, W., Kormanek, M., Kulak, D., Stańczykiewicz, A., 2014: Sequencing of harvester work during standard cuttings and in areas with windbreaks. Silva Fennica 48(4): 1–16. Thode Jr., H.C., 2002: Testing for Normality. Marcel Dekker, New York. Visser, R., Spinelli, R., 2012: Determining the shape of the productivity function for mechanised felling and fellingprocessing. Journal of Forest Research 17(5): 397–402.
Visser, R., Stampfer, K., 2003: Tree-length system evaluation of second thinning in loblolly pine plantations. Southern Journal of Applied Forestry 27(2): 77–82. Wang, J., Haarla, R., 2002: Production analysis of an excavator-based harvester: a case study in Finnish forest operations. Forest Product Journal 52(3): 85–90. Więsik, J., 1998: Czynniki decydujące o wyborze maszyn do pozyskiwania drewna w Polsce. Przegląd Techniki Rolniczej i Leśnej 6: 6–9. Zar, J.H., 1999: Biostatistical analysis. Upper Saddle River, N.J.: Prentice Hall. Żabierek, R., Wojtkowiak, R., 2012: The structure and distribution of harvesters and forwarders in individual Regional Directorates of the State Forests in Poland in the early 2010’s. Acta Scientiarum Polonorum. Silvarum Colendarum Ratio et Industria Lignaria 11(4): 67–77.
Authors’ address: Piotr S. Mederski, PhD. * e-mail: piotr.mederski@up.poznan.pl Mariusz Bembenek, PhD. e-mail: mariusz.bembenek@up.poznan.pl Martyna Rosińska, MSc. e-mail: martyna.rosinska@up.poznan.pl Poznań University of Life Sciences Faculty of Forestry Department of Forest Utilisation ul. Wojska Polskiego 71A 60-625 Poznań POLAND Zbigniew Karaszewski, PhD. e-mail: z_karaszewski@itd.poznan.pl Wood Technology Institute Wood Science and Application Department ul. Winiarska 1 60-654 Poznań POLAND
Received: June 1, 2015 Accepted: September 31, 2015
36
Agnieszka Łacka, PhD. e-mail: agnieszka.lacka@up.poznan.pl Anna Szczepańska-Álvarez, PhD. e-mail: anna.szczepanska-alvarez@up.poznan.pl Poznań University of Life Sciences Faculty of Agronomy and Bioengineering Department of Mathematical and Statistical Methods ul. Wojska Polskiego 28 60-637 Poznań POLAND * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System in Early Thinnings Dan Bergström, Fulvio Di Fulvio, Yrjö Nuutinen Abstract The objective of the study was to improve knowledge on effects of harvested tree size and density of undergrowth on the operational efficiency of a bundle-harvester that produces 2.6 m long bundles, with ca. 60–70 cm diameter, in early fuel wood thinnings. In total 26 time study plots were marked out in 30 to 35 year old Scots pine dominated stands with initial density of 2800–9300 trees/ha and stem size range of 15–43 dm3. Ten of the units, randomly chosen, were precleared of undergrowth trees (≤2.5 cm at breast height diameter) prior to harvesting. There were no significant differences between treatments (preclearing vs. no preclearing) in properties or operational efficiency of the harvested and remaining stands. The average height of cut trees and volume of cut stems were 7.4 m and 16.2 dm3, respectively, and on average, 3554 trees/ha were removed. The bundles had a mean fresh mass of 439 kg and the mass was correlated to the proportion of birch trees cut. The productivity was, on average, 3.1 OD t/PM0H (6.6 fresh t/PM0H; 15.1 bundles/PM0H, where PM0H is productive machine hours, without delays) and was modeled with the harvested stem volume (dm3) as a single independent variable. The study provides complementary knowledge to earlier studies of the system’s performance, especially for harvesting stems <30 dm3. Its productivity was limited by the cutting efficiency and could probably be significantly increased by using a felling and bunching head that could cut and accumulate trees during continuous boom movements. Thus, it would be informative to evaluate such a system in various early thinning stand conditions, including assessments of its manoeuvrability in more difficult terrain. Keywords: pre-commercial thinning, productivity, Scots pine, bioenergy
1. Introduction Small-diameter trees in young dense forests are already harvested in the Nordic countries to produce fuels for heat and power generation, and harvests are expected to increase as demand for high quality residual biomasses for biorefining rises (Bergström and Matisons 2014). In Sweden, potential annual extractions of such trees amount to ca. 6.5x106 m3; five times more than current harvests (Routa et al. 2013). Corresponding quantities for Finland are 7.7x106 m3, which would give 33% more than currently (Routa et al. 2013). Small heating plants often require deliveries of comminuted fuels, while large-scale heat and combined heat and power plants can normally comminute material on-site, and thus also receive unprocessed tree parts. The bioCroat. j. for. eng. 37(2016)1
mass can also be delivered unprocessed or comminuted to terminals for reloading, intermediate storage and/ or comminution before further transportation (Kons et al. 2015). This is a potentially important difference, as high payloads during terrain and road transportation are crucial for producing and delivering forest fuels with high cost efficiency. Comminution at roadside landings increases payloads for bulky materials like logging residues from clear cuts and small-diameter whole trees from early thinnings, but it accelerates degradation, so comminuted material must be delivered to industrial sites quickly to avoid significant biomass losses (cf. Jirjis 1995). Another drawback is that specially designed trucks for comminuted materials must be used rather than standard timber trucks in the sup-
37
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
ply chain. Further, comminution is more costly at landings than at terminals or industrial sites as the operational efficiency is affected by the scale of machinery (cf. Kärhä 2011). Alternatively, the biomass could be compressed and bundled into ca 2.5–3.5 m long bundles with densities of 270–780 kg/m3 in the stand, or at roadside, before further handling and transport (Nordfjell and Liss 2000, Pettersson and Nordfjell 2007, Johansson et al. 2006, Kärhä and Vartiamäki 2006, Jylhä and Laitila 2007). Such systems have been studied, and results indicate that they may have sufficient advantages throughout the supply chain, if convenient bundles for transport on conventional timber trucks can be produced (Johansson et al. 2006, Jylhä 2011, Kärhä et al. 2011, Bergström and Di Fulvio 2014a). The bundles are easy to handle when they are reloaded, dry well during storage and can be effectively comminuted using large-scale systems. However, current bundling machinery is costly and new systems with higher cost efficiency are required. Bergström and Di Fulvio (2014a) have shown that with further development bundleharvesting systems for young dense thinnings could be up to 15% more cost efficient (when including transportation in analysis) than conventional tree-part handling systems. In tests of a prototype whole-tree bundle-harvester for small-diameter trees in Finland, reported by Jylhä and Laitila (2007), bundling productivity was limited because simultaneous harvesting and bundling phases accounted for only 8–18% of the monitored effective working time. The cited authors concluded that the studied system was not competitive with conventional harvesting systems but had great potential for future development. Nuutinen et al. (2011) found that a second prototype of the machine had 38–77% higher productivity than the first, due to a higher cuttingaccumulation capacity and better bundling hydraulics, which increased possibilities for simultaneous cutting and bundling. A third version of the bundle-harvester system was launched in 2013, with reported increases in efficiency (time/bundle) of 111–133% compared to the previous version (»Fixteri II«), providing productivities of 9.7–13.8 m3solid /PM0H when thinning Scots pine dominated stands, removing stems with average volumes of 27–84 dm3 (Nuutinen and Björheden 2015). The solid volumes of the produced bundles range from 0.3 to 0.5 m3 (Jylhä and Laitila 2007) and their use increases forwarders and trucks payloads by ca 50%, respectively, in comparison to handling loose materials (Laitila et al. 2009). However, the system’s productivity has not been extensively studied in stands with an average harvested tree volume <30 dm3, in which
38
there may be significant proportions of disturbing undergrowth trees that may reduce cutting productivities (cf. Kärhä 2006, 2015a, 2015b, Jonsson 2015), and hence cost efficiency.
1.1 Objectives The objective of the study presented here was to evaluate effects of harvested tree size and density of undergrowth on the operational efficiency of the third version of the Fixteri bundle-harvester in early fuel wood thinnings.
2. Material and methods 2.1 Treatments A stand containing patches dominated by broadleaves and conifers of various characteristics was selected. The stand area was divided into time study plots, aiming to isolate homogeneous areas in term of dendrometrical features (species composition, size of trees to be removed, density, etc.). Each time study plot was 20 m wide (the distance reachable during harvesting with a 11 m crane from a strip road) and 40–60 m long. In total 26 time study plots were marked out for cutting and bundling work, each with an average area of 1215 m2 (SD 227), covering a total area of 3.2 ha.
2.2 Study site The study site was located in Holmsund (N 63°43’, E 20°25’), near the coast of northern Sweden, in a 30–35 year old stand containing mostly Scots pine (Pinus sylvestris L.), Norway spruce (Picea abies (L.) Karst.) and birch (Betula spp.). Some grey alder (Alnus incana (L.) Moench.) and lodgepole pine (Pinus contorta Douglas) were also present in some patches. The forest had not been previously pre commercially thinned and some parts contained considerable amounts of undergrowth, mainly consisting of birches and Norway spruce (Table 1). The ground generally had good bearing capacity, the surface had no obstacles, the slope was slight and it was classified as 2.1.1 according to Berg´s (1992) terrain classification system. 2.2.1 Time study plot preparation Prior to thinning, strip road center lines were marked out in each time study plot. Two permanent 100 m2 transects (5 m wide and 20 m long) were laid out at 25 m spacing perpendicular to the strip road in each time study plot for inventories of dendrometric features before pre-clearing and thinning work and after thinning work (Table 1). In total 397 trees were sampled and their height and diameter at stump height (15 cm above ground level) were measured. Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
D. Bergström et al.
Table 1 Characteristics of the 26 time study plots before pre-clearing and thinning work. DBH=diameter at breast height, OD=oven-dry, SD=standard deviation Trees ≤2.5 cm DBH
DBH1
DBH basal2
Basal area
Stem volume4
Density
Height
Stem volume4
Biomass volume5
Biomass5
Pine
Spruce
Birch
Density6
Pine
Spruce
Birch
Post density7
Trees >2.5 cm DBH
Stats.
cm
cm
m2/ha
dm3
trees/ha
m
m3/ha
m3/ha
OD t/ha
%3
%3
%3
trees/ha
%3
%3
%3
trees/ha
Mean
7.1
8.0
26.3
26.5
5406
8.2
133.9
189.2
92.3
27.4
23.5
48.4
4523
6.1
33.9
60.0
1516
Min.
5.5
6.3
17.8
15.0
2765
7.0
91.0
124.0
54.0
1.0
0.0
5.0
134
0
0
0
0
Max.
8.5
9.9
36.4
43.0
9302
9.7
206.0
302.0
148.0
95.0
75.0
89.0
11,951
55.0
100
96.0
4289
Median. 7.0
8.0
25.6
24.5
5200
8.1
131.0
173.5
91.0
17.0
16.0
60.5
3648
2.0
25.5
69.0
1165
0.9
1.0
5.3
8.1
1583
0.7
28.9
48.7
24.9
25.8
21.0
28.3
3509
12.1
28.9
29.1
1517
SD 1
Arithmetic mean; 2 Weighted by basal area; 3 In number of trees; 4 Stem volume on bark; 5 Whole tree volume/mass (incl. tops and branches); 6 All plots before pre-clearing; 7 All plots after pre-clearing of ten plots
Trees ≤2.5 cm in diameter at breast height (DBH) were only counted and registered. After inventory, ten of the 26 time study plots were pre-cleared, by cutting undergrowth trees of ≤2.5 cm (DBH) with a cleaning saw and leaving them on the ground before thinning.
2.3 Machine system The machine system studied was a harwarder equipped with a felling crane and a bundling unit capable of bucking the cut trees and bundling them into 2.6 m long cylinders with ca. 60–70 cm diameters (Fig. 1). The base machine was an 8 wheeled Logman 811FC harwarder (Logman Oy) with 125 kW engine power, 15 t mass, 2.8 m width and 65 cm ground clearance. It was equipped with an 11 m reach Logfit FT100 crane (Logfit AB) integrated on a rotating cabin with endless turning. The crane was equipped with a Ni-
Fig. 1 The bundle-harvester system Croat. j. for. eng. 37(2016)1
sula 280E+ (Nisula Forest Oy) accumulating felling head with a mass of 330 kg and maximum cutting diameter of 28 cm. The bundling unit was a Fixteri FX15a machine (mass ca. 6500 kg, width 240 cm, length 410 cm, height 280 cm; www.fixteri.fi). It has two feed rollers, a cutto-length guillotine and a compression and bundling compartment. The bundling chamber has a fixed length of 2.6 m, three sets of chains used for compression and a vertically sliding frame. On one side of the compression chamber, two rolls of plastic net (4000 m long) are mounted. On the opposite side of the chamber, there are two mobile arms for integrating scaling and dropping-off of bundles. 2.3.1 Work sequence Whole trees are cut, accumulated and fed to the bundling unit for processing (Fig. 2). Once the compartment contains sufficient material for producing a bundle (ca. 450–500 kg of fresh mass), the bunch of trees is lifted up to a compaction chamber, where it is compressed by revolving chains and tied up using a plastic net. At the same time, the lower compartment can be fed with other trees. Once the bundle reaches sufficient density, it is automatically unloaded from the compaction chamber to two side arms. The bundle is automatically scaled and information on time of production and mass is recorded on the base machine. The bundle is then dropped on the ground from the arms and a new bundling cycle starts. The bundling process is automated with possibilities for the operator to control the process. The felling, feeding, dropping and weighing work can be performed simultaneously with the bundling process.
39
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
including delay time less than 15 min). The work time consumption was continuously recorded with an Allegro Field PC® running SDI software (Haglöf AB) recording 0.6 second (1 cmin) time-steps (Table 2) (cf. Nuutinen 2013). Delay time was separately recorded. The highest priority in the time recording was given to the crane work, i.e. if the crane work and bundling were performed at the same time, the crane was prioritized. During harvesting, the number of felled trees per crane cycle (DBH>2.5 cm) was also recorded (the DBH threshold was visually estimated). At the same time, the machine computer created a dataset for each time study plot, including the time (hour: minute: second) when each bundle was expelled from the bundler and its fresh mass as acquired from the integrated scale.
Fig. 2 Flow chart of work processes for the studied bundle-harvester (cf. Nuutinen et al. 2011, Nuutinen and Björheden 2015). Fell A, B indicate that the work can include accumulation of several trees 2.3.2 Operator The operator had one year of experience in early thinning operations and had been working for six months with the studied machine in thinnings before our experiment. Before the time studies, the operator had a half day training session at the study site, and was subjectively judged to be »more skilled than the average operator«. 2.3.3 Thinning work method The thinning was carried out selectively from below along strip road systems, with broadleaves prioritized for removal and targeting a residual density of at least 1200–1500 future crop trees/ha (DBH>6–7 cm). Trees with DBH≤2.5 cm were only cut and accumulated if they obstructed the crane from harvesting larger trees.
2.4 Time study The time study was conducted between the 5th and 14 of May 2014, and the total duration of the monitored work was 29.40 PM15H (productive work time th
40
2.5 Stand quality measures After the time study, the DBH and species of all trees, and numbers of undergrowth trees, were recorded again in the inventory transects. The cut area in each time study plot was accurately measured using a Personal Digital Assistant with an external GPS antenna at 1 m precision. The height of all stumps located less than 1 m from the center line of each transect (i.e. along a line perpendicular to the strip road direction) was also measured. The strip road width was measured according to Björheden and Fröding (1986). The distance between strip roads (defined as the sum of the distances on either side of the strip road from the road center to the furthest cut tree, along a line perpendicular to the strip road) was also measured within the transects. The stem volumes of trees with DBH≤5 cm and >5 cm were calculated using functions presented by Andersson (1954) and Näslund (1947), respectively. The oven-dry (OD) biomass content of stems, branches and needles was calculated using functions presented by Marklund (1987), and for conversion to solid volumes, basic density values for crown biomass obtained by Hakkila (1978) were used. Damage was recorded when there was visible harm to sapwood of trees with DBH>2.5 cm, with no restriction on wound size (cf. Wallentin 2007), registering whether the damaged tree was adjacent to a strip road or located inside the stand. 2.5.1 Bundle mass The fresh mass of each bundle was acquired directly from the machine database and converted to oven dry (OD) mass using the moisture content (MC) of material sampled from each time study plot (determined as described below). The bundles’ solid volumes (m3) were also calculated from their dry masses (in OD kg) using average densities of 450, 554 and Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
D. Bergström et al.
Table 2 Definitions of recorded work elements Work element (Abbreviation)
Description
Priority*
Boom out (Crut)
Starts when an empty crane moves towards the first tree to be harvested and stops when the tree has been reached
1
Felling (Fell)
Starts when the first tree has been reached and stops when the last tree in a crane cycle has been felled (moving to successive trees is included)
1
Boom in (Crin)
Starts when the last tree in the crane cycle has been felled and stops when the felling head drops the bunch of trees on the bundler feeding plate/on the ground
1
Arrangement of felled trees on the ground (Artr)
Starts when the felling head drops the bunch of trees on the ground and ends when the cross-cut tree parts are dropped on the ground/on the bundler feeding plate
1
Arrangement of bundles on the ground (Arbu)
Starts when the crane grabs a bundle and ends when the bundle is dropped on the ground
1
Moving (Move)
Starts when the base machine wheels start turning and ends when the base machine stops
2
Bundling (Bundle)
Starts when the crane/base machine wheels are idling and the bundling unit is feeding/compressing trees and ends when the crane/machine starts to move for felling or a bundle is dropped on the scale
3
Scaling and dropping (Drop)
Starts when the crane/base machine wheels are idling and a bundle is dropped on the scale and ends when the bundle is dropped on the ground or the crane/base machine starts moving
3
Miscellaneous (Other)
Other activities e.g. trees are dropped and then picked up again
4
Delays
Time not related to effective work time, e.g. personal breaks, repairing
4
*The lower the number, the higher the priority
536 kg/m3 calculated for pine (10 plots), spruce (3 plots) and birch (13 plots). The calculated basic density (weighted by tree species) was 505 kg/m3. 2.5.2 Moisture content Immediately after harvesting, a 10 cm thick slice (weighing at least 500 g) was cut from half way along a randomly selected bundle from each of the time study plots using a chainsaw. The MC of the sample was determined following standard method CEN/TS 14774-2 (2004), and the average MC for units dominated by pine, spruce and birch was found to be 53.4 (SD 2.5), 58.7 (SD 1.3) and 52.6% (SD 3.0), respectively. 2.5.3 Fuel consumption and energy efficiency Throughout the entire field trial period, from May 5–16, the system consumed 1619 liters of diesel fuel during 98.5 PM15H (including moving between harvesting units, etc.), of which data recorded during 29.40 PM15H were used in the time study. Thus, 69.1 hours of additional time also included unproductive work such as moving between harvesting units, etc. This operational work was performed in the study site under the same conditions, on average, as the average stand conditions during the time study (cf. Table 1). During the total running time (98.5 PM15H), 1444 bundles with a total fresh mass of 305,779 kg were proCroat. j. for. eng. 37(2016)1
duced. The energy efficiency (MJ/OD t) and energy return over energy invested (EROEI) were calculated using the total fuel consumption, total OD mass harvested, heating values of diesel fuel and the biomass of 35.3 MJ/l (cf. Athanassiadis 2000) and 19.2 MJ/kg TS (Ringman 1996), respectively, and a MC of 53.4% (average value obtained for samples collected from the time study plots). 2.5.4 Other measurements The biomass losses during the bundling process were measured in a separate test, as follows. The bundling chamber was emptied and then fed with weighed, cut tree sections until it contained enough to produce a full bundle. The bundle was then weighed, and biomass losses were calculated by simply subtracting its mass from the mass of material used to create it. In total 13 bundles were produced using representative samples of trees with close to average characteristics for their respective stands (Table 1). The average mass of these 13 bundles was 493 kg (SD 115).
2.6 Analysis and statistics The remaining stands properties and the time consumed (s/tree), when harvesting precleared and not precleared time study plots, were compared by analysis of variance (ANOVA). Correlation analysis
41
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
was applied to evaluate correlations between inde ODt Bundle mass = 190.27 + 0.249 × ( share of birch,% No pendent variables using Pearson´s correlation test. bundle Analysis of covariance (ANCOVA) was used for ana and ODt inde lyzing the combined effects of treatments Bundle mass = 190.27 + 0.249 × ( share of birch,% No. of trees cut ) R2 (adj.) = 0.12, p = 0 pendent variables on the productivity (OD t or bun bundle OD t Regression dles/PM H). analysis was used for testing (1) Bundle mass 0 + 0.249 × ( share of birch,% No. of trees cut ) R2 (adj.) = 0.12, p = 0.046) = 190.27 possible significant predictors of the bundles mass bundle (OD kg/bundle). 95% confidence intervals (CI) were 3.2 Time consumption calculated for the biomass losses (fresh kg) of the bundles. A p-level of ≤0.05 was used as a threshold The productive machine hours without delays durfor statistical significance. ing the time study amounted to 26.76 PM0H, and delays accounted for 9.1% of the monitored time. Pre clearance (and thus the density of undergrowth trees 3. Results during thinning) had no significant effect on the harvesting and bundling work time consumption (Table 3.1 Harvest and thinning quality 3). The felling crane was idling 7.4% of the PM0 time, There were no significant differences in properties mainly due to problems with feeding large trees and between time study plots that were precleared and not dropping bundles. On average, 4.1 trees/crane cycle precleared prior to thinning, in either harvested (e.g. were harvested (SD 1.0) and there were no differences tree volume, tree height and density) or remaining between treatments in this respect (p=0.926). On averstands (e.g. basal area, stand density, stem volume, age, each crane cycle took 44.6 sec (SD 4.2), and 5.5 height, damage and strip road spacing). The average crane cycles (SD 0.7) were required to produce a buntree and average stem volumes cut were 23.3 dm3 (SD 7.7, dle. Hence, on average, 4.1 min of work time was rerange 9–28 dm3) and 16.2 dm3 (SD 5.0, range 12–43 dm3), quired per produced bundle (SD 0.7). respectively. The average tree height and numbers of 3.2.1 Bundler work removed stems for all time study plots were 7.4 m (SD 0.7) The number of crane cycles required per bundle and 3554 trees/ha (SD 1184), respectively. was highly correlated to the average harvested tree The remaining stands had, on average, 1852 trees/ha size (R=–0.775; p<0.001). The time required to produce (SD 455, in the range of 1014–2651/ha), of which 39, 20 a bundle was far from significantly correlated to the and 41% were pine, spruce and birch, respectively. The
(
(
number of birch trees per ha was highly correlated to the numbers of both pine trees (p=0.04) and spruce (p=0.01) per ha, but there was no correlation between numbers of pine and spruce trees per ha. On average, 5.5% (SD 4.2) of the remaining trees were damaged and 5.7% (SD 7.9) of the strip-road trees. The striproad width, the distance between strip-roads and the stump height were, on average, 4.5 m (SD 0.3), 19.8 m (SD 1.3) and 18.3 cm (SD 4.5), respectively. The bundles had a mean fresh weight of 439 kg (SD 24.1, in the range of 391–493 kg), and mean dry mass of 203.4 OD kg (SD 17.3). A correlation test showed that the OD mass of the bundles was positively correlated to the proportion of birch trees (R=0.394; p=0.046) and negatively correlated to the proportion of spruce trees (R=–0.409; p=0.038) used to create them. There were negative correlations between both proportions of bundled birch and spruce trees (R=–0.610; p=0.001) and proportions of birch and pine trees (R=–0.548; p=0.004). Therefore, the following prediction model, with OD t/bundle as a dependent variable and proportion of birch trees in the bundle as an independent variable, was constructed:
42
)
)
Table 3 Distribution of effective work time (s/tree) in work elements (mean values for precleared and not precleared time study plots). SD=standard deviation. p-values are given for the treatment effect (preclearing vs. no preclearing) Stats Work element
Mean, s/tree
SD
%
p-value
Move
0.8
0.2
7.4
0.661
Fell
5.8
0.5
51.6
0.229
Crut
1.5
0.4
13.2
0.902
Crin
2.1
0.6
18.6
0.792
Artr
0.1
0.1
1.1
0.901
Arbu
<0.1
0.1
0.2
0.240
Bundle
0.5
0.3
4.5
0.848
Drop
0.3
0.1
2.9
0.660
Other
0.1
0.1
0.4
0.658
Total
11.2
2.0
100
0.864
Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
D. Bergström et al.
fresh weight of the produced bundles (R=0.215; p=0.291), but close to significantly correlated to the number of crane cycles required per bundle (R=0.367; p=0.065) and strongly correlated to the time consumption of the cutting work per bundle (R=0.787, p=<0.001) (Fig. 3). Thus, the time consumption per bundle was modeled as a function of the time consumption per crane cycle (Fig. 3):
Time time min ) = −1.95 + 0.136 cycle,s , R2 (adj.) = 0.60%, p =< 0.001 ( crane bundle
time 0.136 cycle,s , R2 (adj.) = 0.60%, p =< 0.001 crane
(2)
3.3 Productivity The productivity reached, on average, 3.1 OD t/PM0H (SD 0.6) (6.6 fresh t/PM0H, SD 1.2) (Table 4 and Fig. 4). The independent variable harvested stem volume (dm3) provided the highest predictive power, R2(adj) value, and hence was used as a single covariate in the ANCOVA analysis. All combinations of other inde-
Fig. 3 Time consumption (PM0) of the bundle-harvester to produce a bundle as a function of time consumption per crane cycle work. Calculations are based on average values for 26 time study plots
Table 4 ANCOVA table and linear regression model of the bundle-harvester productivity (OD t/PM0H) Source
DF
Seq. SS
Adj. SS
Adj. MS
F
p
R2
R2, adj.
Stem volume, dm3
1
6.625
6.813
6.813
–
<0.001
–
–
Treatment
1
0.197
0.197
0.197
54.36
0.222
–
–
Error
23
2.882
2.883
0.125
1.58
–
–
–
Total
25
9.705
–
–
–
–
0.700
0.68
Coeff.
SE coeff.
T
–
–
–
–
–
1.3865
0.2410
5.75
–
–
<0.001
–
–
0.10556
0.01432
7.37
–
–
<0.001
–
–
Regression terms Constant 3
Stem volume, dm
Table 5 ANCOVA table and linear regression model of the bundle-harvester productivity (bundles/PM0H) DF
Seq. SS
Adj. SS
Adj. MS
F
p
R2
R2, adj.
Stem volume, dm3
1
138.598
142.074
142.074
78.55
<0.001
–
–
Treatment
1
3.560
3.560
3.560
1.97
0.174
–
–
Error
23
41.602
41.602
1.804
–
–
–
–
Total
25
183.760
–
–
–
–
0.77
0.75
Regression term
Coeff.
SE Coeff
T
–
–
–
–
–
Constant
7.3805
0.9154
8.06
–
–
<0.001
–
–
Stem volume, dm3
0.48205
0.05439
8.86
–
–
<0.001
–
–
Source
Croat. j. for. eng. 37(2016)1
43
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
pendent variables gave less good predictions and/or were biased by multicollinearity (Tables 4 and 5). On average, 15.1 bundles/PM0H were produced (SD 2.7, in the range of 10.8–20.3; Fig. 4). The stem volume provided slightly better productivity predictions, R2(adj.) values of 0.75 vs. 0.68, in terms of bundles/PM0H than in terms of OD mass (cf. Tables 4 and 5). 3.3.1 Energy efficiency and biomass losses During the total field trial period (98.5 PM15H), the system consumed 15.87 MWh of diesel fuel and produced 1392 MWh of biofuel, corresponding to an average energy efficiency of 401 MJ/OD t (PM15 time) (441 MJ/OD t in PM0 time) and an EROEI of 80.6 (in PM15 time) (88.7 in PM0 time). On average, a bundle had a fresh weight of 454 kg, corresponding to 0.96 MWh, and fuel consumption averaged 15.1 l/PM0H (16.4 l per hour of machine work time during the whole trial period (approximates to PM15 time)). The tree sections lost, on average, 37 kg (SD 29) mass during the bundling process, as measured in the separate test, 7.1% of their mean mass (±3.0%; in the range of 4.1–10.1%). By visual inspection, this mass consisted mainly of fine branches and needles.
4. Discussion Unexpectedly, the density of undergrowth trees did not significantly affect the efficiency of the cutting work, as found in previous studies (e.g. Kärhä 2006, Jonsson 2015). A possible contributory factor explaining the results in the present study is related to the
nature of the undergrowth, as the study was performed in the beginning of May, when broadleaves have just started to sprout and thus might have only slightly reduced visibility. Accordingly, Jonsson (2015) found that defoliated undergrowth reduces visibility much less than fully leafed trees. Furthermore, the operator used techniques with efficient crane movements, similar to those applied in boom-corridor thinning as described in Bergström et al. (2007), which could have minimized the effects of undergrowth. The undergrowth did not affect the quality of the thinning work either, which is consistent with the hypothesis that the undergrowth did not significantly impair visibility of the operator. However, few of the harvested units in our study had dense spruce undergrowth, which may be significant as spruce has greater branchiness than pine and birch (cf. Kärhä 2006), and thus may have stronger effects. In conventional pulpwood and energy wood thinning, only tree-sizes above ca 7–9 cm DBH are extracted as commercial assortments, meaning that undergrowth trees are defined as trees below ca. 7–9 cm DBH. In that sense, comparing the cutting efficiency in precleared stands not precleared stands would likely give higher effects than found in this study. Thus, it is likely that the effects of undergrowth clearance up to ca. 8 cm DBH would have given significant differences in work efficiency between the treatments. However, such comparison would be irrelevant since we studied whole tree biomass thinning, which here is defined as utilization of all trees, including their tops and branches, above 2.5 cm DBH. The productivity recorded in the present study
Fig. 4 Productivity (to left, OD t/PM0H and to right, bundles/PM0H) of the bundle-harvester system as a function of average size of harvested tree
44
Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
was 23% and 9% lower than values recorded by Nuutinen and Björheden (2015), for harvesting trees of 27 dm3 and 44 dm3, respectively, when the same bundle-harvester system was used for thinning pine dominated stands during winter (Fig. 4). However, the cited authors only used fresh masses in their production estimates, without measuring the MC, and for conversion to solid volumes, they used a density of 855 kg/m3solid, thus for comparison to values presented here, a MC of 53.4% was assumed for their biomass. Furthermore, the uncertainties (deviations of average values) of the data are not stated in the cited study, therefore, no definitive conclusions regarding similarities or differences in productivity can be drawn, but as indicated by the trends shown in Fig. 5, the results seem to be consistent. The results presented here and by Nuutinen and Björheden (2015) show that the cutting efficiency is the limiting factor for the system. The bundle-harvester system monitored in both studies was equipped with a Nisula 280E+ accumulating felling head that cuts trees with sheares/knives. This type of head is robust and requires less hydraulic pressure than heads with saws, but the cutting efficiency is generally limited by the need to keep the head still while cutting a tree, regardless of the tree size cut. However, the Bracke C16 can sweep short distances (ca 1–2 m) during cutting (www.brackeforest.com), and thus cut trees during a continuous movement. This technique can provide improvements in productivity that are negatively related to the size of cut trees and positively related to the degree of continuous crane movement used, as shown by e.g. Bergström et al. (2007), Bergström (2009) and Sängstuvall et al. (2012). Bergström and Di Fulvio (2014a) modeled operations of an optimized bundle-harvester based on the Bracke C16 accumulating felling and bunching head, with no idle time between the crane and bundling work, and obtained simulated productivities (assuming the same conversion rate as used above) of 15 and 21 m3/PM0H for harvesting trees of 27 dm3 and 44 dm3, respectively. These values are in average 55% and 76% higher, for trees of corresponding sizes, than those recorded in the present study and by Nuutinen and Björheden (2015). The modeling indicates that there is significant potential to increase productivity if efficient felling and bunching technologies are integrated with bundling systems. Bergström and Di Fulvio (2014a) also considered new cutting technologies especially designed for continuous cutting and accumulation in boom-corridors combined with optimized bundling systems (i.e. with no idle time between cutting and bundling). Such Croat. j. for. eng. 37(2016)1
D. Bergström et al.
bundle-harvester systems could significantly reduce costs in stands where the average size of cut trees is <30 dm3. They also show that a conventional bundleharvester system, such as the Fixteri system equipped with (for instance) a Bracke C16 head, is less costly when cutting trees from ca. 30 to ca 70 dm3. It should be noted that biomass losses during the bundling process lead to proportional losses in productivity, and are probably correlated to the sizes of cut trees, and ratios of conifers to broadleaves. However, in a study of a test-bench for compression-processing of bunched small-diameter conifer trees, Bergström et al. (2010) found that processing fresh bunches resulted in mass losses of about 10% to 15% (for trees of 5–8 and 12–15 cm DBH, respectively), with 35–50% reductions in ash contents and 80–160% increases in bulk and net energy density. In the present study mass losses of 4–10% were recorded, consisting mainly of nutrient-rich fractions (according to visual observations), indicating that the ash content in stands, where mostly conifers are cut, could be decreased by up to ca 35%. Whether the losses due to bundling should be minimized or set at certain levels is a question of prioritizing productivity or nutrient removal and fuel quality. This question is highly relevant when considering stands that are sensitive to nutrient removal as the studied bundle-harvester cut and bundle the whole tree above ground. For instance, in Finland the whole-tree harvesting guidelines for early thinnings report that ca. 30% of the biomass cut after whole-tree harvesting should be left at the felling site (cf. Kärhä 2015a), e.g. by preclearing trees below 7–9 cm DBH and by delimbing the trees cut by harvester. In studies of the Bracke MAMA prototype head designed for compression-processing, Bergström and Di Fulvio (2014b) found that biomass losses during processing reduced harvesting yields by 10–23%. Thus, bundling using the Fixteri FX15 system seems to result in relatively low biomass losses and to be less »aggressive« than the feed-roller-based compression-processing techniques studied by Bergström et al. (2010) and Bergström and Di Fulvio (2014b). However, the magnitude of losses due to compression/bundling should be controlled, regardless of the technology and system used, to optimize the balance between productivity and losses in accordance with stand conditions and economic goals. The additional tests of biomass losses during processing of bundles did not cover all possible types of tree mixtures that can occur in thinnings, but the results indicate possible losses for pine-dominated stands. Losses are likely to be similar for spruce-dominated stands and lower for birch-dominated stands,
45
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
but differences in losses between seasons are also likely, because (for instance) branches are more brittle when frozen. The study provides predictive models for early thinnings from below in dense stands, in which trees with 9–28 dm3 stem volumes are cut, and production data in both fresh and OD masses per effective hour of work, derived using moisture contents of representative samples. The productivity prediction models are highly significant and provide high precision estimates. The overall productivity for the whole trial time is very similar to values obtained from the time study, corroborating the robustness of the models obtained from the time study plots. However, users should be aware of the limited numbers of operators and stand types that the models are based upon. The operational fuel consumption during the field trial period is consistent with earlier measurements under somewhat different conditions (Jylhä 2011), indicating that the system consumes ca. 16 l diesel/PM15H. However, effects on fuel consumption of variations in sizes of cut trees (and hence productivity) were not measured due to limitations in resources. The conditions at the study site are representative of large tracts of forest in Sweden, especially in northern parts (cf. Fernandez-Lacruz et al. 2015). To cover forests more representative of southern parts, spruce-dominated stands with high proportions of suppressed birch trees and spruce undergrowth should be included in
Fig. 5 Productivity as a function of harvested whole tree size (stem+branch volume) recorded in the present study and according to findings by Nuutinen and Björheden (2015)
46
Fig. 6 Time consumption to produce a bundle as a function of the bundles mass. The dotted line indicates the maximum capacity of 1.2 min/bundle, reached for bundles with masses between 400 and 500 kg in the study. Values of time consumption lower than ca. 7 min are considered as PM0 time, i.e. work with no delays
further studies. The study was conducted in forest sites with good bearing capacity, low roughness and limited slopes. Due to the system’s high mass, it could potentially be limited by difficult soil conditions, thus further studies are also needed to assess effects of soil properties on its operational efficiency. The machine’s center of gravity was not measured, but it is likely to be higher than for a standard harvester, due to the addition of the bundling unit. This might also somewhat restrict the machine’s operational maneuverability on slopes, and warrants investigation. The operator’s effect was kept constant during the trials, although the time study covered all weekdays and daylight hours. In order to obtain more comprehensive results, reflecting variations in operator skills, using several drivers would be advantageous, as operators strongly affect the performance of harvesters in thinning (cf. ������������������������������������������� Väätäinen et al. 2005, �������������������� Lindroos 2010). However, the main aims of the study were to study the effects of undergrowth density on productivity, as well as productivity levels per se. Thus, the study design included compromises intended to meet these twin aims, and keeping the operator constant reduces both costs and management complexities. During the whole trial period, between the time studies another operator also drove the machine. The mean productivity during this period was 14.7 bundles/PM15H or 3.1 OD t/PM15H Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
(very similar to PM0H values since delays were minor), in line with the time study, which indicates that differences in productivity levels between operators were minor. However, the cutting work with the head used here is relatively slow and straightforward. Thus, if a felling and bunching head affording higher cutting efficiency (and hence more complex movements) was used, there would probably be greater differences in the efficiency between operators. In such cases crane manoeuvrability can be supported to a higher extent by shared control/semi-automation (cf. Jundén et al. 2012). The average mass of bundles produced in the time study was 439 kg, and their masses were correlated to the proportions of birch trees in the initial stands, as birch wood is generally denser than pine and spruce wood. The dotted line in Fig. 6 indicates the minimum bundling time that was approximately reached for bundles with a fresh weight between 400 to 500 kg, showing that the maximum capacity of the system was ca. 1.2 min per bundle. This would correspond to a crane cycle time of 23.2 sec, according to the function in Fig. 3, and provide productivity of 20 fresh t/PM0H. Assuming that the factor for conversion to PM15H is 1.3, the productivity would be ca. 100% higher when harvesting trees with an average volume of 23 dm3 than reported by Bergström and Di Fulvio (2014a). During the time studies, it was noted that some extra »planning time« occurred during unloading/dropping of bundles, as the operator had to check that there was enough space in the place allocated for unloading in order to avoid damaging the remaining trees and ensure that bundles were located off-road, which is highly important for efficient forwarding work. For example, sometimes dropped bundles fell into the strip road area and had to be relocated before forwarding, since the allocated forwarder crane work area is in the opposite direction to the driving direction.
5. Conclusions The efficiency of the studied bundle-harvester was not affected by the density of undergrowth trees, but highly correlated with the size of harvested trees. The study provides information about the system’s performance that complements earlier findings, especially when handling relatively small trees, and the recorded productivity is consistent with previous reports. The system’s time consumption per bundle was not affected by either tree size or the mixture of tree species harvested, but the mass of the bundles was positively correlated with the proportion of birch trees cut. The bundling unit maximum efficiency was not reached during the trial, but estimates indicate that it could be Croat. j. for. eng. 37(2016)1
D. Bergström et al.
significantly (perhaps up to 100%) higher. However, to reach such efficiency, the system would have to be equipped with a felling and bunching head that can cut trees during continuous boom movements. In the near future it should be equipped with a head with higher cutting efficiency, e.g. the Bracke C16 head, and its productivity, manoeuvrability and quality of bundles should be further investigated in various forest conditions.
Acknowledgements The research leading to these results has received funding from the European Union Seventh Framework Programme (FP7/2012-2015) [grant agreement no. 311881] and the SKM project was funded, inter alia, by the Swedish Energy Agency.
6. References Andersson, S.O., 1954: Funktioner och tabeller för kubering av småträd [Functions for stem volume prediction of small trees]. Meddelanden från Statens Skogsforskningsinstitut, Band 44 nr 12. (In Swedish). Athanassiadis, D., 2000: Energy consumption and exhaust emissions in mechanized timber harvesting operations in Sweden. Sci Total Environ. 255(1–3):135–143. Berg, S., 1992: Terrain classification system for forestry work. Kista: The Forest Operations Institute of Sweden. 28 p. Bergström, D., 2009: Techniques and systems for boom-corridor thinning in young dense forests. Doctoral thesis. Acta Universitatis Agriculturae Sueciae, 87 p. Bergström, D., Bergsten, U., Nordfjell, T., Lundmark, T., 2007: Simulation of geometric thinning systems and their time requirements for young forests. Silva Fennica 41(1): 137–147. Bergström, D., Nordfjell, T., Bergsten, U., 2010: Compression processing and load compression of young Scots pine and birch trees in thinnings for bioenergy. International Journal of Forest Engineering 21(1): 31–39. Bergström, D., Matisons, M., 2014: Forest Refine 2012–2014 – Efficient forest biomass supply chain management for biorefineries. Synthesis report. Swedish University of Agricultural Sciences, Department of Forest Biomaterials and Technology, Work report 2014:18. Bergström, D., Di Fulvio, F., 2014a: Comparison of the cost and energy efficiencies of present and future biomass supply systems for young dense forests. Scandinavian Journal of Forest Research 29(8): 793–812. Bergström, D., Di Fulvio, F., 2014b: Studies on the use of a novel prototype harvester head in early fuel wood thinnings. International Journal of Forest Engineering, 25(2): 156–170.
47
D. Bergström et al.
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
Björheden, R., Fröding, A., 1986: A new routine for checking the biological quality of thinning in practice. The Swedish University of Agricultural Sciences, Department of Operational Efficiency Research Notes 48, 14 p. CEN/TS 14774-2, 2004: Solid biofuels – Methods for the determination of moisture content – Oven dry method – Part 2: Total moisture – Simplified method. Fernandez-Lacruz, R., Di Fulvio, F., Athanassiadis, D., Bergström, D., Nordfjell, T., 2015: Distribution, characteristics and potential of biomass-dense thinning forests in Sweden. Silva Fennica 49 (5): article id. 1377, 17 p. Hakkila, P., 1978: Harvesting small-sized trees for fuel. Folia Forestalia, 342, 38 p. Jundén, L., Bergström, D., Servin, M., Bergsten, U., 2013: Simulation of boom-corridor thinning using a double-crane system and different levels of automation. International Journal of Forest Engineering 24(1): 16–23. Jirjis, R., 1995: Storage and drying of wood fuel. Biomass and Bioenergy 9 (1–5): 181–190. Jylhä, P., Laitila, J., 2007: Energy wood and pulpwood harvesting from young stands using a prototype whole-tree bundler. Silva Fennica 41(4): 763–779. Johansson, J., Liss, J.E., Gullberg, T., Björheden, R., 2006: Transport and handling of forest energy bundles–advantages and problems. Biomass and Bioenergy 30(4): 334–341. Jonsson, F., 2015: Hur påverkar avlövad underväxt kvaliteten och drivningskostnaden i gallring? [Effects of defoliated undergrowth trees on harvesting cost in thinnings]. Master’s thesis. Swedish University of Agricultural Sciences, Department of Forest Biomaterials and Technology, Work report 2015:08. (In Swedish). Jylhä, P., 2011: Harvesting undelimbed Scots pine (Pinus sylvestris L.) from first thinnings for integrated production of kraft pulp and energy. Academic dissertation. Dissertationes Forestales 133. University of Helsinki, 73 p. Jylhä, P., Latila, J., 2007: Energy wood and pulpwood harvesting from young stands using a prototype whole-tree bundler. Silva Fennica 41(4): 763–779. Kons, K., Bergström, D., Eriksson, U., Athanassiadis, D., Nordfjell, T., 2014: Characteristics of Swedish forest biomass terminals for energy. International Journal of Forest Engineering 25(3): 238–246. Kärhä, K., 2006: Effect of undergrowth on the harvesting of first-thinning wood. Forestry Studies 45: 101–117.
Kärhä, K., 2015b: Towards better pre-clearance guideline of undergrowth in first thinnings: Case study Stora Enso Wood Supply Finland. Proceedings of the 48th FORMEC Symposium 2015 – Forest Engineering: Making a positive contribution, October 4–8, Linz, Austria. Kärhä, K., Vartiamäki, T., 2006: Productivity and costs of slash bundling in Nordic conditions. Biomass and Bioenergy 30(12): 1043–1052. Kärhä, K., Jylhä, P., Laitila, J., 2011: Integrated procurement of pulpwood and energy wood from early thinnings using whole-tree bundling. Biomass and Bioenergy 35(8): 3389– 3396. Laitila, J., Kärhä, K., Jylhä, P., 2009: Time consumption models and parameters for off and on-road transportation of whole-tree bundles. Baltic Forestry 15(1): 105–114. Lindroos, O., 2010: Scrutinizing the theory of comparative time studies with operator as a block effect. International Journal of Forest Engineering 21(1):20–30. Marklund, L.G., 1987: Biomassafunktioner för tall, gran och björk i Sverige [Biomass functions for pine, spruce and birch in Sweden]. Sveriges lantbruksuniversitet, Institutionen för skogstaxering, Rapport 45, 79 p. (In Swedish). Nordfjell, T., Liss, J.E., 2000: Compressing and drying of bunched trees from a commercial thinning. ������������������ Scandinavian Journal of Forest Research 15(2):284–290. Nuutinen, Y., Björheden, R., 2015: Productivity and work processes of small-tree bundler Fixteri FX15a in energy wood harvesting from early pine dominated thinnings. International Journal of Forest Engineering, http://dx.doi.org /10.1080/14942119.2015.1109175 Näslund, M., 1947: Functions and tables for computing the cubic volume of standing trees. Pine, spruce and birch in southern Sweden and in the whole of Sweden. Reports of the Forest Research Institute of Sweden, 36, 1–41. Nuutinen, Y., Kärhä, K., Laitila, J., Jylhä, P., Keskinen, S., 2011: Productivity of whole tree bundler in energy wood and pulpwood harvesting from early thinnings. Scandinavian Journal of Forest Research 26(4): 329–338. Nuutinen, Y., 2013: Possibilities to use automatic and manual timing in time studies on harvester operations. Doctoral thesis. Dissertationes Forestales 156, 68 p. Pettersson, M., Nordfjell, T., 2007: Fuel quality during seasonal storage of compacted logging residues and young trees. Biomass & Bioenergy 31(11): 782–792.
Kärhä, K., 2011: Industrial supply chains and production machinery of forest chips in Finland. Biomass and Bioenergy 35(8): 3404–3413.
Ringman, M., 1996: Wood fuel assortments – definitions and properties. Department of Forest Products. Report No. 250. The Swedish University of Agriculture Sciences. Uppsala. ISSN 0348-4599.
Kärhä, K., 2015a: Alikasvoksen ennakkoraivaus ja ensiharvennuspuun korjuu [Preclearing of undergrowth and harvesting of pulpwood in first thinnings]. TTS:n tiedote: Metsätyö, –energia ja – yrittäjyys 1/2015 (781). 8 p. (In Finnish).
Routa, J., Asikainen, A., Björheden, R., Laitila, J., Röser, D., 2013: Forest energy procurement: State of the art in Finland and Sweden. Wiley Interdisciplinary Reviews: Energy and Environment 2(6): 602–613.
48
Croat. j. for. eng. 37(2016)1
Effect of Forest Structure on Operational Efficiency of a Bundle-Harvester System ... (37–49)
D. Bergström et al.
Sängstuvall, L., Bergström, D., Lämås, T., Nordfjell, T., 2012: Simulation of harvester productivity in selective and boomcorridor thinning of young forests. Scandinavian Journal of Forest Research 27(1): 56–73.
tulokseen työpistetasolla [The significance of the harvester operator’s tacit knowledge on cutting with a single grip harvester]. Metsäntutkimuslaitoksen tiedonantoja 937; 100 p. (In Finnish).
Väätäinen K, Ovaskainen H, Ranta P, Ala-Fossi A. 2005: Hakkuukoneenkuljettajan hiljaisen tiedon merkitys hakkuu-
Wallentin, C., 2007: Thinning of Norway spruce. Acta Universitatis Agriculturae Sueciae, 29. Doctoral Thesis. ISBN 978-91-576-7328-2.
Authors’ address: Researcher, Assoc. prof. Dan Bergström, D. Tech.* e-mail: dan.bergstrom@slu.se Department of Forest Biomaterials and Technology Swedish University of Agricultural Sciences SE-901 83 Umeå SWEDEN Researcher, Fulvio Di Fulvio, PhD. e-mail: fulvio.di.fulvio@slu.se; difulvi@iiasa.ac.at Department of Forest Biomaterials and Technology Swedish University of Agricultural Sciences SE-901 83 Umeå SWEDEN and International Institute for Applied Systems Analysis Ecosystems Services for Applied Systems Analysis A-2361, Laxemburg AUSTRIA
Received: May 29, 2015 Accepted: November 30, 2015 Croat. j. for. eng. 37(2016)1
Researcher, Yrjö Nuutinen, PhD. e-mail: yrjo.nuutinen@luke.fi Natural Resources Institute Finland FI-801 01 Joensuu FINLAND * Corresponding author
49
Original scientific paper
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy Wood Supply Network Christian Kanzian, Martin Kühmaier, Gernot Erber Abstract The supply of wood for energy is challenging due to high supply costs and rapidly increasing demand. As an important quality criterion, moisture content (MC) influences the revenues, demand and supply costs. For transport, the limiting factor is payload, if the MC is high. The effects of MC on costs and greenhouse gas (GHG) emissions for an optimized supply network have been analyzed using a previously developed multi-criteria optimization model by using different MCs in the range from 50 to 20%. The weighted sum scalarization approach was used to derive Pareto optimal points by changing weights stepwise from maximum profit to minimal GHG on a relatively large scale network of 356 storage locations, 119 freight stations and 228 plants. A decrease of 10% in MC from 40 to 30% will double the profit from 5.10 to 12.00 EUR × t–1. In the case of MC independent revenues, the sensitivity of the model is lower but clearly visible, with a profit increase from 6.00 EUR × t–1 at the MC of 40% to 10.00 EUR × t–1 at the MC of 30%. As expected, emissions will decrease with a decreasing MC. However, the effect on emissions is less prominent than the effect on profit. Reducing MC from 40 to 30% will save approximately 4% of the GHG per dry t. Keywords: supply network, moisture content, forest biomass, chips, transport, multi-objective optimization
1. Introduction The European Union (EU) has set an ambitious target for renewables to represent 20% of the overall energy supply by 2020 (EU, 2009). Based on the initial position of a country, various targets were set. In Austria, 34% of the energy in gross final consumption should originate from renewable sources by 2020. Limits for greenhouse gas (GHG) emissions were also set. Additionally, the Energy Efficiency Plan 2011 aims to decrease energy consumption by 20% by 2020 through increasing the energy efficiency at all stages of the energy supply chain (EU, 2012). The replacement of fossil fuels by forest biomass should help to mitigate GHG. However, the supply of wood for energy is challenging due to high supply costs and rapidly increasing demand. Croat. j. for. eng. 37(2016)1
In contrast to other timber products, the quality of the product, expressed by a higher calorific value, can be increased through storage for the energy supply (Brand et al. 2011). During storage, natural drying reduces the moisture content (MC), which leads directly to a higher calorific value. However, depending on material type and weather conditions, different results have been obtained during precise measurement of natural drying (Erber et al. 2014; Routa et al. 2015). Supply and demand fluctuate between heating and non-heating season. To balance supply and demand and to enhance the fuel quality, storage of wood to be used for energy is preferable in most cases. Suitable storage or terminal locations within a supply network for energy wood can be determined with different approaches, where mathematical optimiza-
51
C. Kanzian et al.
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
tion is often applied. Driven by the need to make it more economical, research started to focus on biomass supply over the last decade. The number of research papers dealing with biomass supply chain models is rising exponentially, with linear-, integer-, mixed integer-, and nonlinear programming, heuristics and multi-criteria decision analysis as common methods (Meyer et al., 2014). Examples of strategic supply chain optimization models can be found in Gunnarsson and Rönnqvist (2008), D’Amours et al. (2008) or Flisberg et al. (2012) and for operational optimization, such as truck routing and scheduling, in Flisberg et al. (2012), Hirsch (2011) or Oberscheider et al. (2013). ZamoraCristales et al. (2015) used a simulation model to calculate costs for different supply chains on the pile level and used the results as input for a mixed integer optimization model to select the best supply options. Of course, most of the optimization studies focus on economics. However, there is an increasing interest in optimization of supply chain sustainability, taking into account the three dimensions of economy, environment and social issues (Eskandarpour et al. 2015). Especially for the biomass supply, environmental impacts such as GHGs are of interest and need to be minimized. The impact of moisture content on supply costs and emissions for an energy wood supply network has not been studied extensively. For example, Acuna et al. (2012) developed a multi-period optimization model to analyze the effect of MC on storage, chipping and transport for three different supply chains over a twoyear period delivering forest energy to a single plant. Results show that proper storage and drying results in saving 33% of the harvested volume. Sosa et al. (2015) applied a quite similar optimization model in an Irish case study. Interestingly, a constraining MC for delivered material led to higher costs compared to an unconstrained MC scenario due to the higher transport distance to gather only material with lower MC. Changing moisture content has an impact on the whole supply chain. To investigate the effect of MC on supply costs and emissions for an energy supply network, the multi-objective mixed integer optimization model presented in Kanzian et al. (2013) was extended. The model considers two objectives: the first is to maximize the profit, and the second is to minimize CO2 emissions. By employing the weighted sum scalarization approach (Ehrgott 2000), where the sum of two scaled objectives has to be minimized, Pareto optimal solutions for different weighting combinations were determined. Staying on the one hand within the forest resource limit and on the other hand fulfilling the de-
52
mand is done by constraints. Using flow and capacity constraints for the terminals and shipping stations ensured that flow over terminals is kept within given limits. To reduce chipper movement and prevent transport capacity underutilization, forest resource points (road side storage) were either classified »material dedicated to chipping« or »material not dedicated to chipping«. Thus, splitting of a resource point material into two different transport forms was avoided. As case studies have shown, demand and resources do not always meet. Based on these experiences, three conditional model constraints were added to enhance the model robustness in the case of limited resources. A total of 90% of the resources should be allocated to demand points and at least 50% of each demand must be fulfilled. Detailed information on the model formulations is provided in Kanzian et al. (2013). For the study, the MC has been assumed to be constantly at an average of 37.5% at first. In a further setup, to figure out the sensitivity on the given supply network, different MCs will be added as additional model parameter. Different sensitivity analyses were carried out to determine and show the impact of this parameter.
2. Materials and methods 2.1 Supply chain assumptions As in Kanzian et al. (2013), five different supply chains have been considered in the study (Fig. 1). Each chain starts at the forest road after harvesting. Depending on the chipping location, different types of materials need to be transported. Supply chain 1 (SC 1): Energy wood is chipped directly at the forest road or landing and transported, chipped, to the plant by trucks.
Fig. 1 Supply chains considered in the optimization model Croat. j. for. eng. 37(2016)1
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
Supply chain 2 (SC 2): Log trucks transport unchipped material directly to the plant, where it is chipped. Supply chain 3 (SC 3): Log trucks transport unchipped material to an intermediate storage area, where it is chipped directly onto trucks and then transported to the plant after seasoning. Supply chain 4 (SC 4): Energy wood is chipped at the forest road and transported to and unloaded at an intermediate storage area. The chips are later loaded onto trucks again and transported to the plant. Supply chain 5 (SC 5): Log trucks transport the unchipped material to a storage area for seasoning. After seasoning, log trucks transport the unchipped material to the plant for chipping. Supply chain 5a (SC 5a): Log trucks transport the unchipped material to a shipping station and load the unchipped material onto wagons. The material is taken to the plant by railroad for chipping.
2.2 Mathematical model parameters and moisture content To optimize the selected supply chains, the parameter calculation in the model code from Kanzian et al. (2013) was reworked to enable studying changes in MC. To enhance the model flexibility and facilitate studying the model sensitivity, cost and emission data calculations were implemented in the model code. The MC of the energy wood mainly influences demand, transport costs and revenues. The lower the MC, the lower the demand will be, due to the higher heating value of the wood. Furthermore, most pricing schemes for energy wood depend on the MC of the delivered material. The higher the MC, the lower the price will be. The revenues at the plant for solid and chipped material originate from a tariff list of the biggest plant within the testing area, which considers the MC for pricing. Other pricing data were not available and thus, this list was assumed to be representative for the study area. Based on this list, using regression analysis, functions for predicting the revenue at a given MC have been generated (1–2).
rj,k = 0 = 60.273 + 36.105 MC − 99.415 MC
(1)
rj,k =1 = 81.346 + 30.558 MC − 86.271MC 2
(2)
2
Road transport costs per entity (cijk) depend on time associated with transport, loading, unloading and operational delays. Time consumption for driving empty and loaded was assumed to be equal. Total transport time is multiplied by the hourly costs, and road Croat. j. for. eng. 37(2016)1
C. Kanzian et al.
charges are added. Finally, the costs for one trip are multiplied by the number of trips needed for completing the job and then divided by the volume per resource point to determine the costs per entity (3). The number of trips (nik) was determined by dividing the volume per resource point and the payload (lvk). This number was then rounded up to the next integer, as there is always one truck trip needed, regardless of the amount transported (4).
cijk =
(( t
L k
)
+ 2t D + t U + 2 pkW + tkD )ckh + 2cijtoll nik si s nik = i lvk
(3) (4)
The transport capacity of trucks and wagons is limited by either maximum payload (tons) or volume (m³). Load limits for the truck and trailer used in the analysis can be found in Table 2. To determine the maximum payload at a given MC, a simple routine was devised for checking whether the payload or the volume was the limiting value. The conversion factor from m³ loose to dry ton was set at 5.26 based on an average wood density of 475 kg m–3 and a bulking factor of 2.5 from solid to chipped wood. Time consumption for different working steps was obtained from our own studies (Holzleitner et al. 2011, 2013). The shortest drive times between the network nodes were calculated in a GIS and stored in a geodatabase. Unluckily, for CO2 emissions we do not have such detailed analysis of different work phases, and we needed to fall back on distance-based emission calculations, using average values of fuel consumption per km. The transport cost by rail for timber freight cars was derived by joining the tariff list scaled by distance and the shortest railroad distance from GIS analysis. Emissions for rail transport per distance and weight were taken from the database of the Global Emission Model for Integrated Systems (GEMIS) and apportioned to the transported energy wood.
2.3 Test case For testing purposes, we used the same project area as was chosen in Kanzian et al. (2013). This project area has a total size of 47,200 km² and is split into 38 forest administrative districts (FADs). Gronalt and Rauch (2008) estimated the total available volume of energy wood for these districts to be 882,170 oven dry tons (t). A square grid of 1.5 by 1.5 kilometers, laid over the forest area and resulting in 9,984 possible resource points, represents the resources. Depending on the forest area and the resources of each FAD, between 31
53
C. Kanzian et al.
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
and 518 points with a single resource volume between 32.7 and 139.9 t × a–1 were generated. The number of heating plants across the study area increased over the last two decades, so we selected a total of 228 heating and combined heating plants with a heating capacity of more than one MW per plant by merging the data provided by different provinces. Smaller plants with a heating capacity lower than one MW were excluded, as their catchment area was assumed to be too small for our chosen scale. The total energy wood consumption of the selected plants is 982,000 t × a–1, which exceeds the forest resource potential of the supply region. The energy wood demand is not uniformly distributed because larger heating plants with a demand of more than 20,000 t × a–1 are mainly located in the east and north, close to the borders of the study area (Fig. 2). A survey of 72 plants performed in 2010 discovered an average MC of 36.8% for the energy wood delivered. A clear and significant trend was detected showing that larger plants take fuel wood with a higher MC (Matz-
inger 2010). Using the prediction of Matzinger (2010), the average MC for the test case was set to 37.5%. As the MC influences the heating value, the demand in dry t was also assumed to change. Based on the actual MC, the demand was adjusted by its heating value, using the average MC as the reference value. A GIS procedure helped to find possible storage and terminal sides, respectively (Kühmaier et al. 2014). Due to the risk of bark beetle infestation in coniferous stands in Austria, storing of softwood for longer periods should be avoided in spring and summer. For different criteria, such as distance to settlements or coniferous forests, the public road network and a minimum slope, grid layers were calculated in GIS (Kühmaier et al. 2014). Areas suitable for terminals were generated by weighting and combining these layers. Within these areas, terminal points with a ten km radius each were located, totaling 356 terminals (Fig. 2). Equipped with a 60 cm thick gravel layer and an expected lifetime of ten years, fixed cost for these terminals amounted to 1,100 EUR × a–1. Variable costs per t depended on the
Fig. 2 Study area with 38 forest administrative districts, forest land cover, locations of heating plants categorized into three sizes, terminal locations for storage and shipping stations for railway transport (Kanzian et al. 2013)
54
Croat. j. for. eng. 37(2016)1
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
Table 1 Indices, decision and data variables used within the multiobjective optimization model +
Description Sets
P
Set of forest resource points (roadside stocks)
L
Set of terminals
S
Set of shipping stations
H
Set of plants
K
Transport mode – (0) solid or (1) chipped Parameters
si
Volume of energy wood at i;
rjk
Revenue at j for fuel type k, j ∈ H; k ∈ K
cijk
Transport costs from i to j for fuel type k, i ∈ P ∪ L ∪ I; j ∈ H ∪ L; k ∈ K Variables
xijk
Volume to be transported from i to j at mode k, i ∈ P ∪ T ∪ S; j ∈T ∪ S ∪ H; k ∈ K
type of energy wood (solid or chipped) and ranged between 10.4 and 9.5 EUR × t–1. Emissions for the terminal construction were estimated to be 90 kg CO2 × a–1 and
C. Kanzian et al.
0.45 kg CO2 × t–1 per t transferred via terminals (Kanzian et al. 2013).
2.4 Model implementation A commercial solver platform (XpressMP, Fair Isaac Cooperation) was chosen for the implementation of the optimization model. Input and output data were stored and managed within personal databases partly using SQL queries. Origin destination matrices were calculated in a GIS using network analysis extensions. Output data were analyzed with the statistical software R (R Development Core Team 2014) and the packages RODBC (Ripley and Lapsley 2015), reshape2 (Wickham and Hadley 2007) and plotrix (Lemon 2006).
3. Results and discussion 3.1 Pareto analysis and moisture content sensibility Supposing that the Pareto optimal solutions follow the convexity assumption (Ehrgott 2000), a finite number of combinations was selected to generate Pareto curves for decision makers. Weighting values for profit (λp) and emissions (λe) were set in a range from 0 to 1 with an increment of 0.5 and 0.1, respectively. By plotting the results for each model run, the Pareto curve provides a starting point for interpretation. As
Fig. 3 The result of 20 model runs with changing weights from 0 to 1 in steps of 0.05, showing the trade-offs between profit and GHG emission, was a typical Pareto curve (left). The share of energy wood supplied to the plants split into solid or chipped and its origin depending on the weighting between profit and emissions (right). A profit weight of 0 results in a minimum of emissions, whereas a weight of 1 results in a maximum of profit (Kanzian et al. 2013) Croat. j. for. eng. 37(2016)1
55
C. Kanzian et al.
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
forest to plant. To maximize the profit by changing the weight, GHG emissions will only rise by 4.5%, whereas the profit more than doubles from 3.0 to 7.4 EUR × t–1. Thereafter, close to 90% have to be supplied chipped at the terminal because transport of chips is cheaper than transport of solid energy wood by log trucks. Furthermore, chipping costs at the terminal or storage place were estimated to be lower than costs of chipping at the forest roadside. Collecting energy wood at terminals will increase the transport distance and therefore increase GHG emissions under the given assumptions (Fig. 3, right). Actually, the legal gross vehicle weight limit is 42 t for log trucks and 40 t for chip trucks. If the values for both trucks were harmonized and set to 42 t, transport of chipped material would be cheaper, and the share would increase even more (Kanzian et al. 2013).
Fig. 4 Pareto curves for different levels of moisture content shown by Kanzian et al. (2013), to minimize the GHG emissions, 30% of the woody biomass should be delivered chipped from the terminals and more than 50% should be chipped directly from forest (Fig. 3, right), which causes emissions of 24.3 kg CO2 × t–1 and results in gaining a profit of 3.0 EUR × t–1 (Fig. 3, left). The rest has to be delivered as solid energy wood directly from
Using weights from 0 to 1 in steps of 0.1 for profit and emissions and three different MC levels of 30, 37.5 and 45%, the resulting Pareto curves are shifted as expected. Higher MC induces less profit and higher overall emissions (Fig. 4). In general, the profit is very sensible to changes in MC. At equal weights for profit and emissions, the profit will be close to 0 EUR × t–1 at emissions of 26.15 kg CO2 × t–1 at the highest MC of 45%. If the MC drops from 45% to 37.5%, the increase in profit will be 6.80 EUR × t–1 and hence higher than during a further MC drop from 37.5% to 30% (4.80 EUR × t–1).
Fig. 5 Sensitivity analysis of profit (left) and CO2 emissions (right) on changing moisture content for two different revenue scenarios. In the first scenario – business as usual – the revenue is based on moisture content, whereas in the second scenario, the revenue is constant/fixed for an average moisture content of 37.5%
56
Croat. j. for. eng. 37(2016)1
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
C. Kanzian et al.
Fig. 6 Sensitivity of supplied energy wood and number of truckloads needed to forward the energy wood at different MCs
Fig. 7 Share of supply sensitivity against changes in MC for equal profit and emission weighting value of 0.5
Of course, the profit is affected twice by the MC change, on one hand by the costs and on the other by the revenues. To rule out the effect of changing revenues and to show how transport costs are affected, another simulation with fixed revenues was conducted. In this case, the sensitivity of the model was lower but still clearly visible, with a profit increase from 6.00 EUR × t–1 to 10.00 EUR × t–1 by reducing MC from 40% to 30% (Fig. 5, left). With variable revenues, a decrease of 10% MC from 40 to 30% will double the profit from 5.10 to 12.00 EUR × t–1. As expected, the emission will decrease with a decreasing MC, and there seems to be no dependence on the revenue. However, the effect of the decreasing MC is less prominent compared to the profit. Reducing MC from 40 to 30% will save approximately 4% of CO2 emissions per t (Fig. 5, right). Based on an average MC of 37.5%, the demand was adjusted before the optimization. For the test data, the supply decreased slightly if the MC was set to a lower value, but the change was small. Approximately 728,150 t would be supplied at 37.5% MC, while 2.2% less (711,020 t) will be supplied at 30% MC (Fig. 6). Of course, this slight change is caused by the soft constraints, which were added because of insufficient energy wood resources. Considering the transport modes and how the material should be supplied, MC has an influence on the results. At the base MC level close to 17%, wood should be supplied chipped directly from the forest to the plants. The major share of 75% has to be delivered
chipped via terminal. The amount of direct transport of chips decreases to 10% if the MC was set to 30%, whereby the delivery as chips via terminal increases to 82% (Fig. 7). The average volume weighted road transport distances seem not to be very sensitive to different MCs
Croat. j. for. eng. 37(2016)1
Fig. 8 Average transport distances for solid and chipped material and its dependence on MC for equal profit and emission weighting values of 0.5
57
C. Kanzian et al.
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
because the distance is more or less the same if an MC of 37.5% and 30% is tested at distances of 47.0 to 46.6 km. Of course, there is a relation between the share of supply and transport distances for different modes. The more material to be delivered chipped directly, which is the case at higher MC, the longer the transport distances of this mode (Fig. 8). Increasing demand at higher MC results in a higher number of truckloads and a larger relative increase of truckloads than in actual supply (Fig. 6). There are 7% fewer truckloads needed to haul the energy wood at 30% MC compared to 37.5%, when the supply will be only 2.2% lower. For fresh material of 45 to 50% MC, the supply from the forests increases by 3.3% but requires 10% more truckloads. Considering specific road transport distance and emissions per t, the effect of the MC becomes even more visible. The specific distance will decrease from 2.9 to 2.5 km × t–1, if the MC is set to 30% instead of 37.5%. In addition, the specific emissions decrease from 0.33 to 0.27 kg CO2 × t–1.
4. Conclusions In this study, the multi-objective optimization model developed by the authors and presented in Kanzian et al. (2013) has been extended to study the impact of MC on profit and GHG for a supply network of energy wood. Clearly, the MC has an influence on the efficiency of the whole supply chain network for several evaluation criteria. The weighted sum scalarization approach gives the possibility of including several objectives and to figure out the effects on profit and emissions quite quickly. Due to the nature of the study data, profit is more sensitive to changing MC than GHG. This result is less pronounced but still traceable via lower transport costs. MC-related revenues are excluded from the analysis. Lower MC means reducing truckloads, which is beneficial both in terms of economy and environment. Interestingly, the specific emissions take more than proportional advantage of a lower MC. The effect of MC was expected to be more present, especially on the demand side and the transport distance, probably a result of the applied demand constraints that do not balance the demand. Interestingly, fresh material with a high MC is more likely to be transported directly from forest to plant. Currently the model considers only a period of one year and no change in MC. A further development into a multi-period optimization model opens up the pos-
58
Table 2 List of parameters for transport cost calculations for solid and chipped material Unit
Solid k=0
Chipped k=1
Loading time for mode k, tkL
h
0.8
1.08
Driving time from i to j k, tijkD
h
Definition, terms
U k
GIS
Unloading time k, t
h
0.53
0.62
Waiting time as percentage of driving time, pkW
%
0
20
EUR× h–1
78
65
Hourly costs k, ckh Road charge from i to j, cijtol
EUR× km–1
Load volume for mode k, lkV
m3
85
90
Gross legal weight limit
t
42
40
Payload for mode k ()
t
22
22
CO2 conversion factor from l to kg
2.64
CO2 emissions transport
kg CO2 km–1
CO2 emissions chipping
kg CO2 t–1
Conversion factor from loose m3 to dry t
–3
2.05
1.32 8.4
kg m
5.26
Average wood density
–
475
Conversion factor solid to loose m3
–
2.5
Integer number of trips needed to transport the total volume from i for mode k (nik)
–
–
sibility of considering a change in MC, e.g., by including the storage effect, estimated by natural drying models that have been developed and published recently. Transport costs are the driving force in fuel wood supply and need to be estimated as accurately as possible. In addition to MC, the bulk density plays an important role in determining the actual payload. The lower the bulk density of the raw material, especially the case for harvesting residues, the more likely is chipping in the forest to increase the density. Energy wood properties also affect chipper performance (Spinelli et al. 2011), which were assumed to be constant in the present study. Although the results give a better understanding of the interactions between MC and energy wood supCroat. j. for. eng. 37(2016)1
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
C. Kanzian et al.
Fig. 9 Specific emissions and transport distance per dry t depending on MC
ply network design, there are abundant possibilities for a further development of the optimization approach presented and a model to enhance the practicality and the decision quality.
Acknowledgments This study has received funding from the European Union Seventh Framework Programme (FP7/2012–2015) under Grant Agreement No. 311881 (INFRES Project).
5. References Acuna, M., Anttila, P., Sikanen, L., Prinz, R., Asikainen, A., 2012: Predicting and Controlling Moisture Content to Optimise Forest Biomass Logistics. Croatian Journal of Forest Engineering 33(2): 225–238. Brand, M.A., de Muñizb, G.I.B., Quirino, W.F., Brito, J.O., 2011: Storage as a tool to improve wood fuel quality. Biomass and Bioenergy 35(7): 2581–2588. D’Amours, S., Rönnqvist, M., Weintraub, A., 2008: Using Operational Research for supply chain planning in the forest products industry. INFOR 46(4): 265–281.
Eskandarpour, M., Dejax, P., Miemczyk, J., Péton, O., 2015: Sustainable supply chain network design: An optimizationoriented review. Omega 54: 11–32. EU, 2009: Directive 2009/28/EC of the European Parliament and of the Council of 23 April 2009 on the promotion of the use of energy from renewable sources and amending and subsequently repealing Directives 2001/77/EC and 2003/30/ EC. Official Journal of the European Union L 140, 05/06/2009, 16–62. EU, 2012: DIRECTIVE 2012/27/EU OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 25 October 2012 on energy efficiency, amending Directives 2009/125/EC and 2010/30/EU and repealing Directives 2004/8/EC and 2006/32/ EC. Official Journal of the European Union 55, 1–56. Flisberg, P., Frisk, M., Rönnqvist, M., 2012: FuelOpt: A decision support system for forest fuel logistics. Journal of the Operational Research Society 63(11): 1600–1612. Gronalt, M., Rauch, P., 2008: BioLog II – Interregional Logistics and Procurement Network for Forest Fuel in Austria (FFG 812774). Technical report, Institute of Production and Logistics, University of Natural Resources and Life Sciences, Vienna, Vienna, Austria, 104 p.
Ehrgott, M., 2000: Multicriteria optimization. Springer, Berlin [u.a.].
Gunnarsson, H., Rönnqvist, M., 2008: Solving a multi-period supply chain problem for a pulp company using heuristics – An application to Södra Cell AB. International Journal of Production Economics 116(1): 75–94.
Erber, G., Routa, J., Kolström, M., Kanzian, C., Sikanen, L., Stampfer, K., 2014: Comparing two different approaches in modeling small diameter energy wood drying in logwood piles. Croatian Journal of Forest Engineering 35(1): 15–22.
Hirsch, P., 2011: Minimizing empty truck loads in round Timber Transport with Tabu Search strategies. International Journal of Information Systems and Supply Chain Management 4(2): 15–41.
Croat. j. for. eng. 37(2016)1
59
C. Kanzian et al.
Effects of Moisture Content on Supply Costs and CO2 Emissions for an Optimized Energy ... (51–60)
Holzleitner, F., Kanzian, C., Höller, N., 2013: Monitoring the chipping and transportation of wood fuels with a fleet management system. Silva Fennica 47(1): 1–11.
house gas emissions in timber transport with a near-exact solution approach. Scandinavian Journal of Forest Research 28(5): 493–506.
Holzleitner, F., Kanzian, C., Stampfer, K., 2011: Analyzing time and fuel consumption in road transport of round wood with an onboard fleet manager. European Journal of Forest Research 130(2): 293–301.
Team, R.C., 2014: R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria, 2012.
Kanzian, C., Kühmaier, M., Zazgornik, J., Stampfer, K., 2013: Design of forest energy supply networks using multi-objective optimization. Biomass and Bioenergy 58: 294–302. Kühmaier, M., Kanzian, C., Stampfer, K., 2014: Identification of potential energy wood terminal locations using a spatial multicriteria decision analysis. Biomass and Bioenergy 66: 337–347. Lemon, J., 2006: Plotrix: a package in the red light district of R. R-News 6(4): 8–12. Matzinger, H., 2010: Versorgungsanalyse von Bioenergieanlagen (Supply analysis of biomass plants)’, Master’s thesis, Institute of Forest Engineering, University of Natural Resources and Life Sciences, Vienna. Meyer, A.D., Cattrysse, D., Rasinmäki, J., Orshoven, J.V., 2014: Methods to optimise the design and management of biomass-for-bioenergy supply chains: A review. Renewable and Sustainable Energy Reviews 31: 657–670. Oberscheider, M., Zazgornik, J., Henriksen, C.B., Gronalt, M., Hirsch, P., 2013: Minimizing driving times and green-
Ripley, B., Lapsley, M., 2015: RODBC: ODBC Database Access. R package version 1.3–11. Routa, J., Kolström, M., Ruotsalainen, J., Sikanen, L., 2015: Precision measurement of forest harvesting residue moisture change and dry matter losses by constant weight monitoring. International Journal of Forest Engineering 26(1): 71–83. Sosa, A., Acuna, M., McDonnell, K., Devlin, G., 2015: Managing the moisture content of wood biomass for the optimisation of Ireland’s transport supply strategy to bioenergy markets and competing industries. Energy 86: 354–368. Spinelli, R., Magagnotti, N., Paletto, G., Preti, C., 2011: Determining the impact of some wood characteristics on the performance of a mobile chipper. Silva Fennica 45(1): 85–95. Wickham, H., 2007: Reshaping data with the reshape package. Journal of Statistical Software 21(12): 1–20. Zamora-Cristales, R., Sessions, J., Boston, K., Murphy, G., 2015: Economic Optimization of Forest Biomass Processing and Transport in the Pacific Northwest USA. Forest Science 61(2): 220–234.
Authors’ addresses:
Received: June 25, 2015 Accepted: August 04, 2015
60
Christian Kanzian, MSc. * e-mail: christian.kanzian@boku.ac.at Martin Kühmaier, PhD. e-mail: martin.kuehmaier@boku.ac.at Gernot Erber, MSc. e-mail: gernot.erber@boku.ac.at University of Natural Resources and Life Sciences, Vienna Department of Forest and Soil Sciences Institute of Forest Engineering Peter Jordan Strasse 82 A-1190 Vienna AUSTRIA * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar (Populus spp.) Plantations Alberto Assirelli, Enrico Santangelo, Raffaele Spinelli, Luigi Pari Abstract In Italy, there has been a significant increase of the areas cultivated with short-rotation forestry (SRF) poplar (Populus spp.) for the production of lignocellulosic biomass. This species has been generally introduced on soils managed with conventional farming practices that led to the formation of a hardpan. This constitutes a serious obstacle for root development and water availability, which affect the successful establishment of the plantation. To this end the Unit of Agricultural Engineering of the Agricultural Research Council (CRA-ING) has developed a new system for reduced tillage (RT), to be used during the establishment of SRF poplar. This new system aims at breaking the tillage pan and at reducing both traffic intensity and site preparation cost. A new machine has been developed, which is based on a commercial rotary plough, suitably modified by adding a shank subsoiler. This machine can perform both deep soil ripping and surface ploughing in a single pass, treating narrow strips where poplar cuttings are to be planted. The study compared conventional tillage (CT) with RT, showing that latter allowed a dramatic reduction of the number of field operations and of all related problems, while creating better conditions for poplar rooting without meaningful effects on yield. Keywords: poplar, short rotation forestry, soil tillage, hardpan
1. Introduction Over the past 20 years, the supply of energy wood from Short Rotation Forestry (SRF) has increased dramatically. SRF stands are very dense (6000â&#x20AC;&#x201C;7000 trees/ha) tree plantations intensely managed and harvested every 2 to 5 years. Among the species suitable for SRF, poplar (Populus ssp.) has shown a very high potential for biomass production, favoured by easy propagation and establishment, fast growth and a low demand for fertilizer and pesticide applications, compared to conventional farming. In Italy, the establishment of SRF poplar has been promoted by Regional Governments through attractive grant schemes, which has led to a significant increase of the areas dedicated to SRF. The new plantations have spread from Northern Italy (where Lombardia is the region with the largest acreage) to the Central regions, such as Emilia Romagna, Toscana, Croat. j. for. eng. 37(2016)1
Umbria, Marche, Abruzzo and Lazio (Salvati et al. 2007). Since the Regional Programs for Rural Development favour the local supply of thermoelectric power plants, SRF plantations have been established as close as possible to the user plants, in order to reduce the cost of delivery. This has led to the introduction of poplar to areas that are somewhat less suited to its cultivation, where previous farming practices involved the use of heavy machines and the recurring application of deep ploughing. Such soil management techniques generally determine the formation of a tillage pan, which represents a serious obstacle to the development of poplar root systems. The presence of a compacted soil horizon hinders root penetration and represents a severe risk factor, with adverse effects on both survival and growth. In addition, the presence of a hardpan reduces water availability, which is especially harmful during dry periods, and when irrigation is not a cost-
61
A. Assirelli et al.
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
effective option. Water availability is one of the main factors affecting the success of SRF poplar in the Po valley and in Central Italy, where irrigation should be considered, when trying to avoid yield losses during dry years (Bergante et al. 2010). Beside all agronomic considerations, economic and energy issues should be analyzed, since site preparation accounts for about one third of the total production cost of a SRF poplar in Italy (Manzone et al. 2009). The establishment phase begins with site preparation and extends until full growth is achieved, and it is critical for the success of SRF investments. A good start strongly affects long-term production and sustainability. If survival rates at the end of the first year are below 80%, then the establishment is considered unsuccessful, and a low biomass production in the first year will detract from the overall performance of the crop. Financial success can be achieved by reducing establishment cost, without compromising initial survival and growth (Volk et al. 2002). This is especially true for crops like SRF poplar, where financial performance is conditioned by the low price of the final product, which is capped by global competition. For this reason, Italian poplar growers need to reduce production costs, maximize biomass yields and minimize work and material inputs. SRF poplar plantations require a different approach than developed for conventional farming and especially a dramatic reduction of soil tillage operations (Khale et al. 2010, Culshaw and Stokes 1995). Several technical operations are commonly carried out for soil preparation before establishing a poplar plantation. Currently, the most popular are: ploughing followed by two disc passes; deep ploughing (0.7–1 m) followed by one or two harrowings; and surface ploughing (0.3–0.4 m) followed by deep ripping (Anderson et al. 1983, Zoralioglu and Kocar 1996, Facciotto 1998). All of these techniques require several passes, with a deep impact on soil structure (compaction), a substantial increase of the production costs, a high energy demand and the accelerated wear of available machinery. Earlier studies have already proposed reduced (conservation) tillage techniques combined with herbicide treatments (Hansen et al. 1984, Mitchell et al. 1999), but these solutions incur the disadvantage of requiring additional chemical inputs for weed control. On the other hand, SRF have been demonstrated to have a positive effect on agricultural soils, by reducing water runoff rates and the amount of soil losses due to water erosion, and by enhancing CO2 sequestration and the soil organic matter content (Khale et al. 2010, Kort et al. 1998). Therefore, combining SRF es-
62
tablishment with new soil preparation techniques that are even more respectful of soil development could help meeting the EU soil protection objectives (Commission of the European Communities 2006). New soil preparation techniques that minimize traffic-induced soil compaction would improve the soil physical properties and stimulate root development. For this reason, CRA-ING developed a prototype tiller for reducing both soil impact and site preparation cost. The new machine was obtained by modifying a commercial rotary plough, a machine already known for its good work performance (Pezzi 2004). This study presents the comparative tests of conventional tillage (CT) performed with multiple machines and passes, and reduced tillage (RT) performed in a single pass with the new machine. With the new machine, shallow ploughing and deep ripping are applied in a single pass to narrow bands, where poplar cuttings are going to be planted. In contrast, CT requires three separate passes (ripping, ploughing and harrowing) and is applied to all the field surface. Therefore, RT obtains a dramatic reduction of field traffic, soil loading and time consumption. On the other hand, the results on plant survival and growth are uncertain. Therefore, the goal of this study was to compare CT and RT for these concerns: Þ t ime consumption and soil preparation cost; Þp lant survival; Þp lant growth.
2. Material and methods 2.1 Site description and experimental design The study was conducted between September 2007 and October 2008 in Vergiano (44°02’N–12°29’E, 36 m a.s.l.), near the city of Rimini, Italy. The soil was a silty clay loam type with physical and chemical properties: 21% sand, 39% silt, 40% clay; pH 8.1; 2.42% organic matter; 1.53% total nitrogen; 11 ppm available P; 489 ppm exchangeable K and a C:N ratio of 9.17. A cultivation of Italian chicory for seed production preceded the plantation of poplar. As a result, before performing the tillage (conventional or reduced), the soil was uniformly covered by this specie. Planting was carried out on 28 February 2008 using a semi-automatic planter. Planting material consisted of 22 cm long poplar cuttings, with diameter ranging between 10 and 25 mm. The cuttings had been stored in a fridge at 3–4°C before planting. Two poplar clones were tested: AF2 and Monviso®. The AF2 clone belongs to the hybrid species Populus x canadensis Moench, deriving from a cross between P. deltoids and P. nigra. Croat. j. for. eng. 37(2016)1
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
A. Assirelli et al.
Table 1 Dimensional features of the machines used in the study Operation
Machine
Make/Model
Power, CV
Total weight, Mg
Working data Depth, m
Width, m
Conventional Tillage Subsoiling Ploughing First harrowing Second harrowing
Tractor
John Deere 7810
185
6.74
0.5
–
Subsoiler
Rinieri
–
0.87
–
–
Tractor
John Deere 7810
185
6.74
–
–
Reversible two-share plough
Mattioli
–
1.30
0.4
–
Tractor
John Deere 6400
100
4.63
–
–
Rotary harrow
Maschio 3000
–
1.01
–
3
Tractor
John Deere 7810
185
6.74
–
–
Spring harrow
Fraternali
–
1.08
0.25
3
Reduced Tillage Tractor
John Deere 3050
115
5.54
–
–
Falc Land 1500 modified
Rotary lances plus central ripping tool
–
1.90
0.84
1.5 (3)
The Monviso clone is a more complex three-way crossing of P. x generosa A. Henry x P. nigra L., where P. x generosa is a crossing of P. deltoides x P. trichocarpa. For each clone, two different soil preparation techniques (CT and RT) were compared. Chemical weeding was applied shortly after planting using oxyfluorfen (at a rate of 1 l/ha), pendimethalin (2.5 l/ha) and glufosinate ammonio (3 l/ha). Since during the experiment no crop diseases were detected, no pesticide was applied. No fertilizers were applied, either. The experiment was a split-plot design with the main plot being the two soil preparations and the subplot the clones. Cuttings were planted with an interrow distance of 3 m and a space of 0.57 m along the
rows (5800 cuttings/ha). Subplots consisted of two rows with a length of 350 m, containing 614 plants. The whole experimental design was completed with two border rows. The total area invested was 1.05 ha.
2.2 Machine characteristics and operations The main characteristics of the machines used for the comparison are reported in Table 1. The prototype was developed in cooperation between CRA-ING and Falc, a local machine manufacturer based in Faenza (Ravenna, Italy). The new machine was obtained by modifying the standard Land model rotary plough, produced by Falc. The machine featured 12 lances mounted on a horizontal axis with
Fig. 1 Falc/CRA-ING prototype: rotary apparatus and ripping tool (left); the rotary plough during the work (right) Croat. j. for. eng. 37(2016)1
63
A. Assirelli et al.
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
Fig. 2 Diagram showing the position of the vertical blade performing the soil refinement three sectors, spaced 35 cm, each comprising four lances. The working width of the axis was 1.5 m, the central and the lateral sectors reaching a depth of 55 and 45 cm, respectively. CRA-ING modified the rotary plough by applying a vertical blade subsoiler in front of the support mast, for ripping the worked area to a depth of 80 cm. The blade in front of the work area of the rotary lances was fixed with a trunnion that allowed folding it during transportation. The blade is adjustable to allow for setting the working depth as needed. In this way, the original work of the rotary plough was integrated by a subsoiling action for breaking the tillage hardpan (Fig. 1 and Fig. 2). The site preparation treatments were applied on the following dates: Þ RT: one pass on September 16th 2007; Þ CT: subsoiling on August 30th 2007; ploughing on September 4th 2007; first harrowing on September 10th 2007; second harrowing on February 25th 2008; ÞO n February 28th 2008, the field was planted with poplar cuttings.
2.3 Climatic conditions, work productivity, plant survival and growth The climatic conditions were monitored throughout the study by means of a weather station placed on site and properly equipped for the purpose. All operations were timed according to the CIOSTA (Comitè International d’Organisation Scientifique du Travail en Agriculture) rules, and to the recommendations issued by the Italian Association of Agricultural Engineering (AIIA). Machines cost was estimated with the method reported by Baraldi and Capelli (1973). The effect of soil preparation technique was determined by counting the number of successfully established cuttings on the whole experimental field, and by measuring plant height on a sub-sample of 35
64
plants per row after 4 and 8 months from planting (June and October 2008). To avoid edge effects, they were excluded from the reliefs of the first 27 plants for each side choosing the 35 plants per row by measuring 1 plant every 16.
3. Results and discussion 3.1 Rooting, growth and survival At plantation, the poplar cuttings must find the most suitable physical conditions for a fast rooting and suitable availability of water and nutrients. The promotion of a fast deepening of the roots in the soil layers below the hardpan (frequently present in the heavy soils of North and Central Italy), increases the rate of plant survival and allows growth when irrigation is not an option. After spring planting, the cuttings benefited from mild temperatures that grew progressively higher as summer approached. The moisture condition of the soil was good due to the supply of winter rainfall (Fig. 3). The cuttings rooted well and developed homogenously as can be inferred by the plant height detected four (June) and eight (October) months after establishment (Fig. 4). The lack of statistical differences between the two treatments demonstrates their equivalence in terms of plant development. Therefore, it can be safely stated that RT does not cause any reduction of plant survival or growth compared with CT. Four months after the transplant when the poplar cuttings were rooted, a check of the weeds present in the experimental field revealed the predominance of three species: perennial ryegrass (Lolium perenne L.), field bindweed (Convolvulus arvensis L.) and creeping thistle (Cirsium arvense (L.) Scop.). The ryegrass and the field bindweed grew preferably along the poplar rows; in some cases, the latter extended its procumbent stems in the inter-row. On the other hand, in some Croat. j. for. eng. 37(2016)1
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61â&#x20AC;&#x201C;69)
A. Assirelli et al.
Fig. 3 Rainfall and average temperatures at the site of the experimental trial during period of observation scattered areas, the creeping thistle had colonized the space along and between the rows. However, compared to the pre-planting state, where the field was a continuous lawn, after the treatments (CT and RT), the presence of weed was localized almost exclusively along the rows, without significant differences between the CT and the RT. A few months after planting, the genotype effect was very clear. Monviso sprouts were generally short-
Fig. 4 Height of poplar AF2 and Monviso clones measured four (June) and eight (October) months after transplant and treated with RT (reduced tillage) or CT (conventional tillage); data are the mean of thirty-five trees ÂąS.E. Croat. j. for. eng. 37(2016)1
er than 1 m, whereas about 25% of the AF2 sprouts were taller than 1 m (Fig. 5a and Fig. 5b). In October, the effect of tillage also started to show: Monviso sprouts, grown on the plots treated with RT, were distributed for almost 90% in the upper height classes, against 58% of the sprouts treated with CT. Even if less evident, the same difference was also visible for clone AF2, where 100% of the sprouts, treated with RT, were distributed in the four upper classes, while those managed with CT reached these same classes for 92% only (Fig. 5c and Fig. 5d). A very important result concerns survival rate (Fig. 6), because the cuttings of both clones showed a higher percentage of survival when RT was used, and these differences reached statistical significance for the clone Monviso. The structural conditions created by the repeated passages operated with CT were not favourable for successful rooting. In contrast, the concurrent actions of soil breaking, soil turning and ripping obtained with the prototype resulted in a breaking of the hardpan, a more homogeneous structuring of topsoil and an opening of the subsoil. Such results are in contrast with those published by Balbuena et al. (2002), who studied the effect of traffic intensity on poplar plantations and showed that the height of poplar shoots was not affected by the traffic intensity during harvesting. However, the same Authors showed that traffic intensity did affect shoot mortality, which increased by 42% and 51%, respectively, on soils receiving 5 and 10 machine passages. (Balbuena et al. 2002). It should be noted that the influence of both subsoiling and surface ploughing on plant survival during dry periods is favourable, if irrigation is not supplied. This
65
A. Assirelli et al.
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61â&#x20AC;&#x201C;69)
Fig. 5 Class frequency of the heights of poplar AF2 and Monviso clones treated with RT (reduced tillage) or CT (conventional tillage) measured in June (a and b) and in October (c and d) is confirmed by the fact that, at the end of rotation, the two experimental plantations yielded 27.2 t/ha and 28.4Â t/ha, respectively, for RT and CT, and the difference was not significant according to the ANOVA test.
3.2 Working capacity and economic evaluation Working speed for the prototype was halfway between that of the deep plough and of the rotary plough (Table 2). Considering the braking effect of the vertical ripper blade, one could expect the observed reduction
66
of speed, compared with the rotary plough. That brought no substantial reduction of work efficiency. In fact, although performing the functions of a rotary plough and a subsoiler at one time, the prototype showed an effective work time that was lower than that of a rotary harrow alone, and registered reduced delays for turning (Fig. 7). Considering the speed and the actual working width given from the inter-row space of transplants (3 m), the effective working capacity was 0.54 ha/h. Croat. j. for. eng. 37(2016)1
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
A. Assirelli et al.
Table 2 Operative data of the machines employed Reduced Tillage Falc Land 1500 modified
Conventional Tillage Subsoiler
Reversible two-share plow
Rotary harrow
Spring harrow
Effective speed, km/h
1.91
2.66
1.51
2.20
4.28
Operative speed, km/h
1.80
2.48
1.37
2.09
4.00
Working capacity, ha/h
0.54
0.75
0.41
0.63
1.20
Effective work time, %
94.93
94.12
91.38
96.17
93.10
5.07
5.88
8.62
3.83
6.90
100.00
100.00
100.00
100.00
100.00
Delays, % Total work time, %
By analyzing in detail the figures for delays (Fig. 7), the prototype showed good flexibility and manoeuvrability in the reduction of turning time (3.5% of the total working time). On the other hand, it also required more time for field adjustments than the two harrows and the reversible plough. From the standpoint of environmental impact, it is clear that RT produced the same agronomic result with a reduction of fuel use and soil loading (Fig. 8). RT allowed saving 80.4 l of diesel per ha (103.8 l/ha being the sum of fuel consumption for CT vs 23.5 l/ha for RT). Therefore, RT allowed reducing fuel consump-
Fig. 7 Detailed delays registered for soil preparation of poplar planting by CT and RT
Fig. 6 Percentage of cuttings survival (%±SE) of AF2 and Monviso clones registered two months (April) after the transplant and treated with RT (reduced tillage) or CT (conventional tillage); significant differences between treatment within each clone were determined by Student’s t-test; **p<0.001 probability levels, is not significant Croat. j. for. eng. 37(2016)1
tion by more than 4 times. Likewise, soil loading amounted to 7.4 Mg for RT and 29.0 Mg for CT (Table 3). RT showed a ratio trampled/worked area equal to the first harrowing and, although with a relatively high weight of the unit, it achieved a noteworthy reduction of traffic intensity. The hourly cost of the prototype appeared in line with the values calculated for the other treatments, with the only exception of the rotary harrow (Table 4). Taking into account its working capacity, the cost of RT was rather high (109 €/ha), but lower than that observed for ploughing (139 €/ha). However, since the RT soil preparation only involves the use of the prototype, the cost per unit area is limited to this intervention, with a reduction of 65% compared to CT.
67
A. Assirelli et al.
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
Table 3 Soil surface exposed to trampling by tilling machines used during the trial Tyre width of the tractor, mm Front
Rear
Total weight of the unit, Mg
Width Trampleda, m
Worked, m
Ratio trampled/worked, %
Conventional Tillage Subsoiling
540
710
7.60
1.42
2.0
71.0
Ploughing
540
710
8.04
1.42
1.2
118.3
First harrowing
380
420
5.64
0.84
3.0
28.0
Second harrowing
540
710
7.82
1.42
3.0
47.3
0.84
3.0
28.0
Reduced tillage Falc Land 1500 a
380
420
7.44
Calculated for the rear tires
Table 4 Economic aspects of the tilling machines analyzed during the trial Hourly cost, €/h
Cost per unit area, €/ha
Conventional Tillage Subsoiling
54.23
72.31
Ploughing
57.17
139.44
First harrowing
40.20
63.81
Second harrowing
53.87
44.89
Reduced tillage Falc Land 1500 modified
Fig. 8 Fuel consumption (l/ha) and machine mass (Mg) ascribed to RT and CT
4. Conclusions The two tillage systems (conventional vs reduced tillage) compared in this work showed deep differences in terms of machine productivity and economic performance. The working capacity of the Falc/CRAING prototype was halfway between the ploughing and the first harrowing, but its strength lies in the possibility of reducing the number of operations in the field, and all related problems, such as: soil compaction, need of finding the soil in good conditions for machine accessibility, reduction of emission due to cultural techniques, and fuel consumption.
68
54.97
109.46
The prototype revealed that it was worth combining a deep subsoiler, requiring considerable traction effort, with a rotary plough, which used the energy supplied by the power take-off and gave a helpful boost to machine advancement. Compared to CT, RT allowed a similar growth of the cuttings while improving their survival. This was probably due to the easier root development that derived from positioning the cuttings on the deep cracking, which is not always feasible when adopting CT. The utilization of this new technique makes it possible to introduce SRF poplar to new areas of North and Central Italy where the conventional farming practices have caused the formation of a hardpan, which may negatively affect the plant establishment and biomass yields. Single pass soil preparation is advantageous from the agronomic, economic and environmental point of view, and represents a valid alternative to time- and energy-consuming traditional practices. Croat. j. for. eng. 37(2016)1
A Single-pass Reduced Tillage Technique for the Establishment of Short-Rotation Poplar ... (61–69)
Acknowledgements This work was funded by the research project of Italian Ministry of Agriculture »SUSCACE: Supporto Scientifico alla Conversione Agricola verso le Colture Energetiche«.
5. References Anderson, H., Papadopol, C., Zsuffa, L., 1983: Wood energy plantations in temperate climates. Forest Ecology and Management 6(3): 281–306. Balbuena, R., Mac Donagh, P., Marquina, J., Jorajuria, D., Terminiello, A., Claverie, J., 2002: Wheel traffic influence on poplar regeneration and grass yield. Biosystems Engineering 81(4): 379–384. Baraldi, G., Capelli, G., 1973: Elementi tecnici per il calcolo del costo di esercizio delle macchine agricole. Genio Rurale 36: 37–76. Bergante, S., Facciotto, G., Minotta, G., 2010: Identification of the main site factors and management intensity affecting the establishment of Short-Rotation-Coppices (SRC) in Northern Italy through Stepwise regression analysis. Central European Journal of Biology 5(4): 522–530. Commission of the European Communities, 2006: Thematic Strategy for Soil Protection. COM(2006)231 final, Brussels, September 22th, 12 p. Culshaw, D., Stokes, B., 1995: Mechanisation of short rotation forestry. Biomass and Bioenergy 9(1–5): 127–140. Facciotto, G., 1998: Le lavorazioni del suolo in pioppicoltura. Sherwood: 39–44. Hansen, E., Netzer, D., Reitveld, W., 1984: Site preparation for intensively cultured hybrid poplar plantations. Research
A. Assirelli et al.
Note NC-320. St. Paul, MN: U.S. Dept. of Agriculture, Forest Service, North Central Forest Experiment Station. Khale, P., Baum, C., Boelcke, B., Kohl, J., Ulrich, R., 2010: Vertical distribution of soil properties under short-rotation forestry in Northern Germany. Journal of Plant Nutrition and Soil Science 173(5): 737–746. Kort, J., Collins, M., Ditsch, D., 1998: A review of soil erosion potential associated with biomass crops. Biomass and Bioenergy 14(4): 351–359. Manzone, M., Airoldi, G., Balsari, P., 2009: Energetic and economic evaluation of a poplar cultivation for the biomass production in Italy. Biomass and Bioenergy 33(9): 1258–1264. Mitchell, C.P., Stevens, E.A., Watters, M.P., 1999: Short-rotation forestry – operations, productivity and costs based on experience gained in UK. Forest Ecology and Management 121(1–2): 123–126. Pezzi F., 2004: Regolazioni e prestazioni di un coltivatore rotativo. Rivista di Ingegneria Agraria 35(1): 43–50. Salvati, R., Chirici. G., Corona, P., 2007: Modello di valutazione dell’attitudine fisica del territorio per la realizzazione di impianti cedui da biomassa in Italia. L’Italia forestale e Montana 62(5/6): 399–410. Volk, T., Abrahamson, L., White, E., Robison, D., 2002: Alternative methods of site preparation for Short-Rotation willow and poplar biomass crops. Biomass Power for Rural Development, Technical Report (Neuhauser, E., ed.), State University of New York, College of Environmental Science and Forestry, New York: 1–142. Zoralioglu, T., Kocar, S., 1996: Mechanization techniques for poplar development in Turkey. Biomass and Bioenergy 10(5–6): 261–265.
Authors’ address: Alberto Assirelli, PhD. * e-mail: alberto.assirelli@entecra.it Enrico Santangelo, PhD. e-mail: enrico.santangelo@entecra.it Luigi Pari, PhD. e-mail: luigi.pari@entecra.it Consiglio per la ricerca e la sperimentazione in agricoltura CRA ING – Unità di Ricerca per l’Ingegneria Agraria Via della pascolare 16 00016 Monterotondo (Rm) ITALY
Received: October 13, 2014 Accepted: March 6, 2015 Croat. j. for. eng. 37(2016)1
Raffaele Spinelli, PhD. e-mail: spinelli@ivalsa.cnr.it CNR IVALSA Via Madonna del Piano 10 50019 Sesto Fiorentino (FI) ITALY * Corresponding author
69
Original scientific paper
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China with Fuzzy Clustering Methods Aihua Yu, Tom Gallagher, Chen Zhao, Yao Zhao Abstract Over the years, China has shown a significant reduction in natural forest resources, while the increasing area of plantations has made greater contributions to the huge demand for wood. In southern China, these new plantations have produced some problems such as environmental hazards of logging operations and the most reasonable use of forest resources. A new management process called Âťcleaner productionÂŤ is defined as reducing pollution from its source, increasing the rate of utilization of resources, and preventing the generation of pollutants in the production of services and products. In recent years, cleaner production has been widely applied to industrial processes such as agriculture and other environmental industries. In order to make rational use of plantation resources, to achieve maximum economic efficiency and to reduce or remove the environmental hazards of logging operations, it is necessary to carry out an in-depth study of cleaner production on the process of logging operations. This paper aims to establish an index system for cleaner production evaluation of plantation logging. The fuzzy clustering method was used to initially screen twenty-nine indices. After screening by the fuzzy clustering method, six first-grade indices and twelve second-grade important indices were selected as formal evaluation indices. The six first-grade indices are 1) cutting area design index, 2) logging operation techniques index, 3) ecological environmental impact index, 4) utilization of resource and energy index, 5) sustainable development index, and 6) safety production management and protection index. A maximum and minimum matrix method and a correlation coefficient matrix method were used to establish the similar matrix in the fuzzy clustering method. The screening results were then compared. The comparison shows that out of the twelve second-grade indices, ten are similar and two are different. The results suggest that the fuzzy clustering method is reliable for screening indices. Keywords: plantation, logging operation, cleaner production (CP), evaluation indices, fuzzy clustering method.
1. Introduction With a sharp reduction in the natural forest over the past decade in China, a greater emphasis has been put on plantations to meet lumber demand. The recently completed Eighth National Forest Inventory supports this assumption. Specifically, the proportion of plantation harvesting has increased from 39% to 46%, and plantation harvesting has increased to 155 million cubic meters annually (the results of the eighth naCroat. j. for. eng. 37(2016)1
tional forest resources inventory in China (2009â&#x20AC;&#x201C;2013). While the additional plantation harvesting has many economic benefits, it has also produced some new problems in southern China, such as soil erosion and waste timber in the land. Some criteria needed to be developed to measure what is important in the performance of these plantations. This paper focuses on the index system for cleaner production (CP), which is an evaluation of plantation logging in order to rationally use plantation resources, achieve the maximum eco-
71
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
nomic benefit, and reduce the effects of logging on the environment. CP has been defined in »The Law of the People’s Republic of China on Promotion of Cleaner Production« by continuously adopting measures to improve design, use cleaner energy and raw materials, introduce advanced techniques and equipment, improve management and make comprehensive use of resources as well as other measures. CP is also defined by reducing pollution from its source, increasing the utilization ratio of resources, and reducing or preventing the generation and discharge of pollutants during production and in providing services. The goal is to alleviate or eliminate harm to human health and the environment (Zhao and Zhang 2003). This not only applies to industrial processes, but also to agriculture, planning, construction, and services as this is a general issue. Based on the industry definition of CP and the characteristics of plantation logging, CP involves improving cutting area design with the local conditions, using cleaner energy, developing reasonable logging technologies and equipment, improving management practices, minimizing pollution, saving, and protecting the forestry environment and workers. The results should be to achieve sustainable plantations. The methodology is to produce a model for plantation resources by using a whole environmental strategy for the processes and productions of plantation logging in order to reduce or eliminate harm to human health and the environment while fully satisfying human needs. CP in plantation logging is an important means for achieving sustainability of plantations. The final goal is to maximize the balance between natural resources, energy use and economic benefits and minimize the harm to humans and the environment, all in order to save resources, reduce waste and protect the environment while harvesting timber.
2. Contents of CP evaluation for plantation logging Cutting area design, logging technology, ecological environment impact, resource and energy use, and sustainable businesses of forest protection and management for workers, all contribute to the CP evaluation for plantation logging. So the evaluation not only considers the longevity, gradualness, and complexity of the impacts of logging on forest ecological systems, but also the method and intensity of logging, scale of production, and the stability and recovery of the ecological environment.
72
2.1 Evaluations for cutting area design CP for cutting area design should include the following: designing energy saving processes, promoting low-energy techniques, shortening the working time, simplifying equipment, and paying attention to energy management. In addition, using logging equipment to benefit the ecological environment, reducing logging waste and combustible materials of forest fires, and improving the benefits of logging by-products were also considered when evaluating the importance of CP. In designing of an index for CP, the following should be considered: the rationality of logging processes (including the type of logging system, index of logging intensity, index of outturn percentage and output, index of road design, index of bucking, index of skidding and transportation, organization of logging team and so on), advancement of logging technology, reliability of preventative measures for cutting renewable areas, and operation management considerations.
2.2 Evaluation of logging technology Cutting, skidding, and transportation are three important processes in timber production technology in a logging operation. Various working methods and equipment types are used in each process, so that many different timber production models and economic benefits can be compared. Based on the investigation of plantation logging in the Fujian Province, clear-cutting was the main type of cutting method with the exception of the cases provided in »Regulation of Forest Harvesting and Renew«. Chainsaw operation was the main equipment method of felling. Angular saw was only used in a few China fir collective forest areas with mostly small diameter logs (diameter <8 cm). Methods of skidding included aerial cableway, walking tractor, push cart, dirt chute, and manpower. Transportation included truck, farm vehicle, shipping, and rafting (manpower) (Zhang et al. 2008). CP of logging mainly reflected the interference and acceptability to the environment. Reduced Impact Logging (RIL) was based on principles of science and engineering and a combination of education and training (Dykstra 2001). RIL required specific forest investigation procedures before logging, including a plan and construction of logging roads and landings, reliable ways for cutting and bucking, skidding felled timbers along skid roads, skidding systems for protecting soil and vegetation, and evaluation after logging. In addition, RIL also included the impacts of logging on landscapes, biodiversity, vegetation, water, and soil (Long 2006). Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71â&#x20AC;&#x201C;87)
Evaluation for logging mainly includes the following items: advance of technology, rationality of logging processes, coordination, economic benefits, security and work efficiencies of a human-equipment-environment system. The economic benefit was based on reducing working costs while not destroying the forested ecological environment.
2.3 Evaluation for ecological environment impacts For a small cutting area, soil and vegetation in the ecological environment were the most directly impacted. When evaluating logging practices such as cutting, skidding, and transportation, it becomes increasingly important to always consider the effects on the ecological environment. The unreasonable cutting, skidding and ground disturbance impact many parts of the ecological environment when logging, including the effects on soil, reserve, rivers, biodiversity, loads of CO2 in air, landscape, and regional climate. Therefore, a cleaner logging model must minimize the above effects, in order to maximize the ecological environmental benefits of logging in the forest.
2.4 Evaluation for resources and energy use Plantations were resource used for an evaluation of logging. The forest resource in China is considered a scarce raw material when evaluating its biological and economic benefit. So, in order to save on the use of raw materials as well as maintain higher working ability and level, larger outputs and fewer resource wastes should be expectated of plantation logging. Energy use in the logging processes mainly means that fuel and lubricants were consumed by equipment that was used in cutting, skidding and transportation. So, low-energy and clean energy equipment should be used. Also, the index of evaluation for resource and energy use included standing tree utilization, volume of waste timber in cutting area, volume of waste timber in loading bay, logging slash utilization, and the fuel consumption of equipment used during logging.
2.5 Evaluation for sustainable businesses of plantation China has implemented new practices in forestry development, though mostly for afforestation, because of many factors. These factors include harvesting in areas where forest harvesting has followed the traditional methods and patterns. It can be said that in these areas we do not see the forest for the trees. In afforestation areas, some areas were deforested without reforCroat. j. for. eng. 37(2016)1
A. Yu et al.
estation and had false reporting, which led to inaccurate forest resource database. These factors have led to low-quality afforestation, reforestation missing from the forest each year, and an increased impact to the environment and ecological balance, which have caused irreversible damage to woodland sites. Logging has an important effect on the ecological environment, such as reducing the depth of litter fall and destroying the surface soil. Especially in spring and summer, increased surface runoff usually results in heavy soil erosion. Moreover, pushing timber by skidding equipment negatively impacted the surface soil and damaged the young trees and reserves (Wang 1997). Evaluation of plantations for sustainable businesses included the utilization rate of wood, renewable rate of cutover land, and the survival rate of regeneration. The aim of CP for plantation logging was sustainable businesses.
2.6 Evaluation of safety production management and protection To make plantation logging cleaner, production must be cleaner first. This includes cleaner awareness and action of the producers. This was particularly reflected in the aspects of safety on production and protection. So, better education of workers was the basis for achieving CP. If working conditions were improved, employees could benefit from increased safety and a better overall working environment. CP could minimize the injury rate of operators in working processes. Meanwhile, CP was also found to ensure worker rights and safety. Only labor protection was able to become the essence of CP. So for evaluation of safety production, one aspect of the evaluated content consisted of reducing operation damage of workers and ensuring worker rights and safety.
3. Evaluation indices 3.1 The principles for determining evaluation indices The index system for CP evaluation of plantation logging consists of many structured and graduated indices that connect and complement each other. It is a combination of the evaluation of sustainable and high efficient utilization levels of resources and its index, and how it directly affects the results of forest resources management and utilization levels. Correctly implemented, the result is an accurate representation of CP for plantation logging. Sustainable development theory and ecology-economy-society theory
73
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71â&#x20AC;&#x201C;87)
were the guides for setting up the index system for the CP evaluation of plantation logging, and they were also based on the following five principles: 1) Methods combined qualitative and quantitative analysis. In order to ensure the accuracy and scientific nature of the results for evaluating the CP of plantations, the quantity index should be selected to set up a quantitatively evaluable model. On the other hand, the evaluated objects were completed production processes that involved various quarters. Therefore, the index system of CP was a completed and intrinsic closed contact system. Absolute quantity, relative quantity, averages and other indices could be used, and some indices could also be used as quality if they could not be managed by quantitative analysis. 2) Independence: the status of the system could be described by a number of indices, but intercrossing of information always occurs. In order to establish an index system, representative and independent indices must be selected using the scientific method to improve its accuracy and scientific nature. 3) Evaluation through the whole process: the index system not only includes the whole production process, but also the product itself. In other words, CP evaluations of plantation logging aimed to analyze and evaluate the raw material and energy consumption, as well as pollutant creation and its toxicity in the whole process of design, production, storage, and transportation. 4) Improving sustainability: CP is a sustainable improving process, and it requires the company to continuously achieve higher environmental objectives on the basis of existing economic, technological, and environmental indices. Therefore, for the purpose of promoting CP, different CP objectives should be selected to promote sustainable development on the basis of the existing situation. 5) Simple and focused: the index system for CP could not cover all the processes in the plantation logging. Generally, the most simple and focused indices are implemented effectively.
3.2 Selection of the evaluation index The process begins with selecting some evaluation indices on the basis of contents and principles (as shown in Table 1) (Yu et al. 2009). Such indices should not be combined together directly because of information interference and the fact that this could have an impact on the evaluation results (Chen 2003). Cluster analysis divides data into groups (clusters) such that similar data objects belong to the same cluster and dissimilar data objects to different clusters. The
74
resulting data partition improves data understanding and reveals its internal structure. Partitional clustering algorithms divide up a data set into clusters or classes, where similar data objects are assigned to the same cluster, whereas dissimilar data objects should belong to different clusters. In real applications, there is very often no sharp boundary between clusters so that fuzzy clustering is often better suited for the data. Membership degrees between zero and one are used in fuzzy clustering instead of crisp assignments of the data to clusters. Fuzzy clustering is the method that can capture the uncertainty situation of real data and it is well known that the fuzzy clustering can obtain a robust result as compared with conventional hard clustering (Silviu 2013). Conventional clustering means classifying the given observation as exclusive clusters. It can be clearly seen whether an object belongs to a cluster or not. However, such a partition is insufficient to represent many real situations. Therefore, a fuzzy clustering method is offered to construct clusters with uncertain boundaries, so this method allows that one object belongs to some overlapping clusters to some degree. In others words, the essence of fuzzy clustering is to consider not only the belonging status to clusters, but also to consider to what degree the objects belong to the clusters (Sato-Ilic et al. 2006). One of the main advantages of fuzzy clustering is the ability to express ambiguity in an assignment of objects to clusters (Silviu 2013). A corresponding fuzzy set was used to describe the uncertainties (Han et al. 2011). However, apart from this, experimental results prove that fuzzy clustering seems also to be more robust in terms of local minima of the objective function (Klawonn 2004). Another distinct advantage of fuzzy clustering over its crisp counterpart is that the continuous range of the combinatorial functions turns into smooth functions. This makes it possible to design algorithms that are more likely to attain a global solution, whereas crisp techniques often wind up in the local solution. (Rousseeuw 1995). The fuzzy relation between samples is quantified in fuzzy cluster, so fuzzy cluster is more objective and accurate (Yang 2011). In recent years, the fuzzy clustering has been widely studied and applied in a variety of key areas (Wang and Zhang 2011) such as in data mining, economic analysis, and selection evaluation indices (Xu et al. 2005). Li et al. (2008) used this method to select the evaluation indices of stock investment value, providing an empirical basis for artificial intelligence methods in the stock value of investment. Guan et al. (2009) applied the fuzzy clustering approach in constructing the evaluation index system of core competence of Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Table 1 CP evaluation indicators system for forest plantation loggi First level index Second level index Cutting area division (x1) Cutting area design
Cutting area survey (x2) Engineering design (x3) Production process design (x4) Rationality of operation process (x5)
Primary index system for cp evaluation of plantation logging
Logging technology
Advanced of operation technology (x6) Efficiency (x7) Human-machine-environment harmony (x8) Economics and safety of ways to work (x9) Soil physical properties (x10) Soil chemical properties (x11)
Impacts on ecological environment
Soil and water conservation (x12) Injury rate of retention tree in slash (x13) Average wind speed and temperature (x14) Biomass (x15) Biodiversity (x16) The rate of slash soil erosion area (x17) The number of discarded wood in cutting area (x18) Utilization rate of wood (x19)
Resources and Utilization rate of slash (x20) energy use The number of discarded wood in the landing place (x21) Logging equipment fuel consumption (x22) Sustainable development
Improvement of safety and management
Survival ratio of renew (x23) Renew ratio of cutting area (x24) Wood renewal utilization (x25) Safety management (x26) Equipment safety (x27) Labor protection (x28) Labor intensity (x29)
corporation and achieved very good results. Yan et al. (2008) forecasted the heavy rainstorm based on the fuzzy cluster type in Jiangsu province and reduced the storm empty reported rates. Chiba et al. (2012) analyzed the web survey data with the similarity fuzzy cluster and showed a better performance by using numerical examples. A secondary selection should be used by the fuzzy cluster method. The basic idea of fuzzy cluster was to construct the fuzzy matrix according to attributes of a research object, and on this basis the clustering relation was determined according to certain subordinate relations (Peng 2003). Croat. j. for. eng. 37(2016)1
A. Yu et al.
The process of the fuzzy cluster method is described by the following 3 steps: 1) Set up a fuzzy similarity matrix by calculating similar coefficients between samples and variables. 2) Transform the fuzzy similarity matrix into a fuzzy equivalence matrix with the help of fuzzy operation. 3) Classify the fuzzy equivalence matrix according to a different fuzzy graph λ.
3.3 Data processing The fuzzy cluster process begins by setting n items as a classification, the indicator set X = {X1, X2,..., Xn}, with available m-dimensional vector describing the sample, Xi = {X1i, X2i,..., Xmi}, i = 1,2,..., n. As there were some qualitative analysis indices in CP evaluation of plantation harvesting, this paper used a five grade Likert scale to evaluate indices, ranging from »very important (5)« to »not care (1)«, to determine a better measure of each index. The membership function was then defined: t was the number of classifications of certain properties, Cp for p class, x was an attribute value of Cp, u (x) = N (Cp) m, p = 1,2,..., t. N (Cp) was the number of attribute values included in the class Cp. Following the above process, data matrix was initialized, if yij represents the property value of the ith row and column j, then 0≤yij≤1, and yij size reflects the dependence of the property value for the property. Establishing the fuzzy similar matrix: The domain U = {y1, y2,..., yn} was concerned, yi, yj relationship was described by R (yi, yj). There were many methods to help set up a fuzzy similarity matrix, such as the distance method, correlation coefficient method, and maximum and minimum method. Both the maximum and minimum method and correlation coefficient method were used in this paper to set up a fuzzy similarity matrix for the CP evaluation of plantation logging, and also to compare the difference between them. (1) Maximum and minimum method A value is calculated by the eq. (1):
1, i = j m R yi , y j = ∑ k =1min yik , y jk , i ≠ m ∑ k =1max yik , y jk
(
)
( (
) )
j
(1)
Where, yik was the attribute value of row i column k; yjk was the attribute value of row j column k.
75
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
A fuzzy similarity matrix R was set up as follows: R y ,y R ( y1 , y2 )… R ( y1 , yn ) ( 1 1) R = R ( y2 , y1 ) R ( y2 , y2 )… R ( y2 , yn ) R ( y , y ) R ( y , y ) R ( y , y ) n 1 n 2 n n
(2)
As shown in the above matrix, reflexivity and symmetry are satisfied due to R(yi，yi)=1; R(yi，yj)=R(yj，yi), but do not meet the transitivity. Therefore, the values cannot be classified directly. Squares method was used to calculate transitive closure of the fuzzy matrix, so that a fuzzy equivalence matrix was set up from the fuzzy similarity matrix R (as shown below).
R→ R 2→ R 4→ R 2k→
(3)
R was the transitive closure of the fuzzy matrix when Rk0 Rk = R2k. In other words, Rk was the fuzzy equivalence matrix. The cube operation was calculated as follows: k
k =1
(
)
A = R0 R ↔ Aij = ∨ Rik ∧ Rjk
m
(4)
λ was the fuzzy graph： where:
( )
R = rij
n×n
∀l ∈0,1 Rl = (rij ( l ))n × n
(5)
1, rij ≥ l , so Rλ was the fuzzy graph of matrix Ro. rij ( l ) = 0, rij < l
Different values would be assigned to λ from large to small after the fuzzy equivalence matrix was set up, and different classifications would be gained by calculating l. In other words, l was assigned by the actual needs, and classification was selected by l (Li et al. 2003). A representative index was selected from each classification as the typical index after classifying. The specific method was the following: first, correlation coefficients of each classification were calculated; then, the averages of the squares of the correlation coefficient between one index and the other were calculated, and the maximum was the typical index. The index could be put into the index set, if only one index was classified and one index of two indices existed in classification (Sârbu and Einax 2008). The main factors that impacted the CP of plantation logging would be the index of cluster on the basis of literature review and investigations. We designed the questionnaire according to the content and features of cleaner production in plantation logging, issued over 100 questionnaires to almost twenty for-
76
estry companies or universities by email, including colleges of forestry, forestry research institutes, forest engineering enterprise and so on, and recovered 30 valid questionnaires. A five grade Likert scale was used to evaluate indices. All thirty samples were used for the analysis and the raw data is shown in Table 2. The membership function was calculated from the definition itself, and an initialization process was used for these data. According to the raw data, membership function of each attribute is calculated and shown in Table 3. In the first attribute, x1 means the first index (cutting area division); {1–5} means the important degree (from »very important {5}« to »not care {1}«); 0.37, 0.40, 0.23 were the proportion of rating {4}, {5} and {3} in 30 samples, respectively. After mapping each membership function, the resulting initialized data is presented in Table 4. According to the eq. (1) and combining the initialized data, fuzzy similar matrices R were established. Matrices of cutting area design (6), logging technology (7), impacts on ecological environment (8), resources and energy use (9), sustainable development and safety production (10), and labor protection (11) were calculated with the maximum and minimum method, as shown in eq. (6)–(11). In these similar matrices, each value means the correlation of two indices; the higher value has the stronger correlation, thus value 1 means fully correlated. For example, in matrix (6), the first row, the values 1, 0.6693, 0.8085, 0.7380 mean the correlation between index of the first and first, the first and the second, the first and the third, the first and the fourth, respectively. In the first column, the values meaning as the first row, and so on.
1.0000 0.6693 0.8085 0.7380
1.0000 0.7408 0.6813 0.5171 0.6441
0.6693 1.0000 0.6904 0.6544
0.7408 1.0000 0.6429 0.6699 0.6103
0.8085 0.6904 1.0000 0.6791
0.6813 0.6429 1.0000 0.5211 0.5203
0.7380 0.6544 0.6791 1.0000
0.5171 0.6699 0.5211 1.0000 0.5413
0.6441 0.6103 0.5203 0.5413 1.0000
(6)
(7)
1.0000 0.5125 0.6285 0.5227 0.6466 0.6667 0.7293 0.6033 0.5125 1.0000 0.6384 0.5586 0.5472 0.5206 0.5145 0.5269 0.6285 0.6384 1.0000 0.6522 0.6209 0.6831 0.6486 0.6603 0.5227 0.5586 0.6522 1.0000 0.5283 0.5560 0.5443 0.5731 0.6466 0.5472 0.6209 0.5283 1.0000 0.7061 0.6879 0.6177 0.6667 0.5206 0.6831 0.5560 0.7061 1.0000 0.7221 0.6731 0.7293 0.5145 0.6486 0.5443 0.6879 0.7221 1.0000 0.6522 0.6033 0.5269 0.6603 0.5731 0.6177 0.6731 0.6522 1.0000
(8)
Croat. j. for. eng. 37(2016)1
R2
4
3
3
4
3
5
3
5
2
4
4
5
2
3
4
3
1
4
4
4
3
4
5
2
2
5
4
2
2
R1
4
2
3
2
5
4
5
2
3
4
4
4
4
5
5
3
4
5
5
5
5
2
4
2
3
4
5
4
5
Index
X1
X2
X3
X4
X5
X6
Croat. j. for. eng. 37(2016)1
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
X22
X23
X24
X25
X26
X27
X28
X29
4
4
4
5
4
2
5
4
5
4
4
4
1
4
3
2
4
5
3
4
3
4
5
5
3
4
3
3
5
R3
3
1
4
4
4
2
5
4
5
4
3
5
4
1
3
4
3
3
5
5
3
3
4
3
4
2
4
4
4
R4
3
3
4
5
4
4
5
4
4
3
5
5
3
4
4
4
1
4
4
4
4
5
4
4
5
4
5
4
4
R5
2
2
5
3
4
2
5
4
3
3
5
5
1
5
1
1
3
3
4
3
5
3
5
3
3
4
5
3
5
R6
Table 2 Raw data from the questionnaire
3
3
5
4
5
4
5
5
4
5
4
5
4
5
4
3
5
4
4
2
5
5
4
4
4
3
5
4
3
R7
3
2
4
5
3
1
3
4
3
1
4
4
1
4
4
5
3
4
5
4
4
3
5
3
5
4
3
3
5
R8
3
1
3
4
1
2
3
3
4
2
5
4
2
2
3
2
1
4
4
3
3
3
3
3
4
3
5
4
5
R9
2
5
5
5
3
1
4
3
5
3
5
5
3
4
4
3
3
3
4
1
3
4
4
4
4
4
5
4
4
R10
4
4
5
3
1
2
5
4
4
4
4
3
4
2
4
2
3
4
4
2
5
2
5
4
4
2
4
3
3
R11
5
4
4
4
4
3
3
3
3
4
5
4
1
3
4
4
4
5
5
4
5
3
4
3
5
2
4
2
5
R12
4
1
4
4
4
5
5
2
5
3
4
5
3
2
3
2
4
5
4
4
3
1
3
3
5
4
3
2
3
R13
2
4
4
4
3
1
4
5
5
4
3
3
3
3
3
3
4
3
3
1
2
3
4
4
5
2
5
3
5
R14
5
4
5
4
3
2
4
3
4
3
4
5
3
2
3
3
3
4
5
3
3
3
5
5
4
2
4
3
4
R15
3
3
4
5
3
1
3
5
3
4
5
4
5
4
2
4
3
3
5
3
3
2
4
4
4
2
4
1
5
R16
4
3
3
3
2
2
5
5
3
2
3
4
4
3
2
3
2
4
5
2
3
1
5
5
5
4
3
1
5
R17
4
2
5
5
5
3
4
2
4
4
5
3
1
1
3
5
3
4
5
3
4
5
3
5
5
4
4
5
5
R18
2
4
3
5
2
4
4
4
3
4
5
3
3
4
3
4
3
4
4
2
2
4
4
4
5
4
5
2
3
R19
4
4
4
5
4
4
3
5
5
5
5
5
1
3
3
5
4
4
3
4
4
5
4
3
5
3
4
3
4
R20
3
3
2
3
3
3
1
1
3
2
3
3
3
2
5
1
3
3
3
4
3
3
3
4
4
4
4
4
4
R21
3
3
3
2
2
2
2
3
3
3
3
3
4
4
2
2
2
5
5
4
3
4
2
3
5
5
5
4
5
R22
3
4
4
5
5
5
5
5
4
4
4
4
4
5
5
5
5
5
3
5
4
5
4
5
5
5
5
5
5
R23
4
3
4
5
5
5
4
4
4
4
4
4
3
5
5
4
3
5
4
5
3
5
3
4
5
5
5
5
5
R24
3
4
4
3
3
3
4
2
4
3
3
3
3
2
4
2
2
4
4
3
4
2
4
3
4
3
4
3
4
R25
3
4
4
3
2
2
4
2
3
3
4
4
3
2
3
2
2
3
4
3
3
3
4
5
4
3
4
3
3
R26
2
4
4
3
2
2
4
2
4
2
4
4
2
3
2
2
2
3
4
2
3
2
4
4
4
3
3
2
4
R27
3
3
4
3
2
2
3
2
4
2
4
4
3
2
2
2
3
3
4
2
2
2
4
3
4
2
4
3
3
R28
3
4
4
3
4
4
5
4
5
4
5
4
5
4
3
2
3
4
4
3
5
4
4
5
4
5
4
4
3
R29
3
4
3
2
2
2
3
2
4
4
4
3
2
3
2
2
3
3
4
4
3
3
4
3
5
3
5
3
4
R30
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71â&#x20AC;&#x201C;87) A. Yu et al.
77
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Table 3 Membership function of each attribute
0.37 … x1 ∈{4} u ( x1 ) = 0.40… x1 ∈{5} 0.23… x ∈{3} 1
0.10… x5 ∈{3} u ( x5 ) = 0.43… x5 ∈{4} 0.47 … x ∈{5} 5
0.13… x9 ∈{2} 0.50 … x9 ∈{3} u ( x9 ) = 0.20 … x9 ∈{4} 0.17 … x9 ∈{5}
0.27 … x18 ∈{3} u ( x18 ) = 0.43… x18 ∈{4} 0.30… x ∈{5} 18
0.07 … x10 ∈{1} 0.20… x10 ∈{2} u ( x10 ) = 0.27 … x10 ∈{3} 0.37 … x ∈ 4 {} 10 0.10… x10 ∈{5}
0.03… x15 ∈{1} 0.20… x15 ∈{2} u ( x15 ) = 0.37 … x15 ∈{3} 0.27 … x ∈ 4 {} 15 0.13… x15 ∈{5}
0.03… x26 ∈{2} 0.17 … x26 ∈{3} u ( x26 ) = 0.50… x26 ∈{4} 0.30… x26 ∈{5}
78
∈{2}
∈{3} ∈{4} ∈{5}
0.27 … x4 ∈{2} 0.23… x4 ∈{3} u ( x4 ) = 0.37 … x4 ∈{4} 0.13… x4 ∈{5} 0.07 … x8 ∈{1} 0.20… x8 ∈{2} u ( x8 ) = 0.33… x8 ∈{3} 0.17 … x ∈ 4 {} 8 0.23… x8 ∈{5}
0.33… x12 ∈{3} u ( x12 ) = 0.43… x12 ∈{4} 0.24… x ∈{5} 12
0.23… x17 ∈{1,4} 0.07 … x16 ∈{1} 0.10… x17 ∈{2} u ( x16 ) = 0.27 … x16 ∈{2,3,4} u ( x17 ) = 0.13… x ∈{5} 0.37 … x17 ∈{3} 16 0.07 … x17 ∈{5}
0.20… x19 ∈{3} u ( x19 ) = 0.43… x19 ∈{4} 0.27 … x ∈{5} 19
0.23… x27 ∈{3} u ( x27 ) = 0.53… x27 ∈{4} 0.24… x ∈{5} 27
0.03… x7 0.20… x7 u ( x7 ) = 0.53… x7 0.23… x7
0.17 … x11 ∈{3} u ( x11 ) = 0.57 … x11 ∈{4} 0.26… x ∈{5} 11
0.03… x23 ∈{1,2} 0.23… x23 ∈{3} u ( x23 ) = 0.33… x23 ∈{4} 0.37 … x23 ∈{5}
0.23… x3 ∈{3} u ( x3 ) = 0.40… x3 ∈{4} 0.37 … x ∈{5} 3
0.03… x22 ∈{1} 0.27 … x22 ∈{2} u ( x22 ) = 0.17 … x22 ∈{3} 0.33… x ∈ 4 {} 22 0.20… x22 ∈{5}
0.37 … x6 ∈{3} u ( x6 ) = 0.37 … x6 ∈{4} 0.27 … x ∈{5} 6
0.07 … x13 ∈{2,5} u ( x13 ) = 0.20… x13 ∈{2,4} 0.47 … x ∈{3} 13
0.07 … x2 ∈{1} 0.17 … x2 ∈{2} u ( x2 ) = 0.40… x2 ∈{3} 0.27 … x ∈ 4 {} 2 0.10… x2 ∈{5}
0.03… x20 ∈{1} 0.17 … x20 ∈{2} u ( x20 ) = 0.27 … x20 ∈{3} 0.43… x ∈ 4 {} 20 0.10… x20 ∈{5}
0.13… x24 ∈{1,3} 0.47 … x24 ∈{2} u ( x24 ) = 0.17 … x24 ∈{4} 0.10… x24 ∈{5} 0.10… x28 ∈{1} 0.17 … x28 ∈{2} u ( x28 ) = 0.40… x28 ∈{3} 0.30… x ∈ 4 {} 28 0.03… x28 ∈{5}
0.33… x21 ∈{3} u ( x21 ) = 0.40… x21 ∈{4} 0.27 … x ∈{5} 21
0.07 … x25 ∈{1} 0.17 … x25 ∈{2} u ( x25 ) = 0.40… x25 ∈{3} 0.23… x ∈ 4 {} 25 0.13… x25 ∈{5}
0.20… x29 ∈{2} u ( x29 ) = 0.33… x29 ∈{3,4} 0.14… x ∈{5} 29
Croat. j. for. eng. 37(2016)1
R2
0.37
0.40
0.23
0.37
0.10
0.27
0.20
0.23
0.13
0.37
0.57
0.23
0.20
0.20
0.27
0.27
0.23
0.43
0.43
0.43
0.33
0.33
0.37
0.47
0.17
0.30
0.53
0.17
0.20
R1
0.37
0.17
0.23
0.27
0.47
0.37
0.23
0.20
0.50
0.37
0.57
0.43
0.20
0.17
0.13
0.27
0.23
0.30
0.37
0.10
0.27
0.27
0.33
0.47
0.40
0.50
0.23
0.30
0.13
Index
X1
X2
X3
X4
X5
X6
Croat. j. for. eng. 37(2016)1
X7
X8
X9
X10
X11
X12
X13
X14
X15
X16
X17
X18
X19
X20
X21
X22
X23
X24
X25
X26
X27
X28
X29
0.33
0.30
0.53
0.30
0.23
0.47
0.37
0.33
0.27
0.43
0.43
0.43
0.23
0.27
0.37
0.37
0.20
0.23
0.17
0.37
0.50
0.17
0.23
0.27
0.10
0.37
0.23
0.40
0.40
R3
Table 4 Initialized data
R4
0.33
0.10
0.53
0.50
0.23
0.47
0.37
0.33
0.27
0.43
0.20
0.30
0.23
0.07
0.37
0.20
0.47
0.33
0.27
0.10
0.50
0.33
0.53
0.37
0.43
0.27
0.40
0.27
0.37
R5
0.33
0.40
0.53
0.30
0.23
0.17
0.37
0.33
0.40
0.27
0.37
0.30
0.37
0.27
0.27
0.20
0.07
0.43
0.57
0.37
0.20
0.23
0.53
0.37
0.47
0.37
0.37
0.27
0.37
R6
0.20
0.17
0.23
0.17
0.23
0.47
0.37
0.33
0.33
0.27
0.37
0.30
0.23
0.13
0.03
0.07
0.47
0.33
0.57
0.27
0.17
0.33
0.23
0.37
0.10
0.37
0.37
0.40
0.40
R7
0.33
0.40
0.23
0.50
0.13
0.17
0.37
0.20
0.40
0.10
0.43
0.30
0.23
0.13
0.27
0.20
0.07
0.43
0.57
0.20
0.17
0.23
0.53
0.37
0.43
0.23
0.37
0.27
0.23
R8
0.33
0.17
0.53
0.30
0.40
0.13
0.23
0.33
0.33
0.03
0.43
0.43
0.23
0.27
0.27
0.17
0.47
0.43
0.27
0.37
0.20
0.33
0.23
0.37
0.47
0.37
0.23
0.40
0.40
R9
0.33
0.10
0.23
0.50
0.07
0.47
0.23
0.17
0.40
0.17
0.37
0.43
0.10
0.27
0.37
0.37
0.07
0.43
0.57
0.27
0.50
0.33
0.20
0.37
0.43
0.23
0.37
0.27
0.40
R10
0.20
0.03
0.23
0.30
0.40
0.13
0.33
0.17
0.27
0.27
0.37
0.30
0.37
0.27
0.27
0.20
0.47
0.33
0.57
0.07
0.50
0.17
0.53
0.37
0.43
0.37
0.37
0.27
0.37
R11
0.33
0.30
0.23
0.17
0.07
0.47
0.37
0.33
0.40
0.43
0.43
0.27
0.23
0.27
0.27
0.37
0.47
0.43
0.57
0.20
0.17
0.20
0.23
0.37
0.43
0.27
0.40
0.40
0.23
R12
0.13
0.30
0.53
0.50
0.23
0.13
0.23
0.17
0.33
0.43
0.37
0.43
0.23
0.27
0.27
0.20
0.20
0.23
0.27
0.37
0.17
0.33
0.53
0.37
0.47
0.27
0.40
0.17
0.40
R13
0.33
0.10
0.53
0.50
0.23
0.10
0.37
0.27
0.27
0.27
0.43
0.30
0.37
0.27
0.37
0.37
0.20
0.23
0.57
0.37
0.50
0.07
0.20
0.37
0.47
0.37
0.23
0.17
0.23
R14
0.20
0.30
0.53
0.50
0.40
0.13
0.33
0.20
0.27
0.43
0.20
0.27
0.37
0.27
0.37
0.20
0.20
0.33
0.17
0.07
0.13
0.33
0.53
0.37
0.47
0.27
0.37
0.40
0.40
R15
0.13
0.30
0.23
0.50
0.40
0.47
0.33
0.17
0.40
0.27
0.43
0.30
0.37
0.27
0.37
0.20
0.47
0.43
0.27
0.27
0.50
0.33
0.23
0.27
0.43
0.27
0.40
0.40
0.37
R16
0.33
0.40
0.53
0.30
0.40
0.13
0.23
0.20
0.33
0.43
0.37
0.43
0.07
0.27
0.20
0.20
0.47
0.33
0.27
0.27
0.50
0.20
0.53
0.37
0.43
0.27
0.40
0.07
0.40
R17
0.33
0.40
0.23
0.17
0.17
0.47
0.37
0.20
0.33
0.17
0.20
0.43
0.23
0.27
0.20
0.20
0.20
0.43
0.27
0.20
0.50
0.07
0.23
0.27
0.47
0.37
0.23
0.07
0.40
R18
0.33
0.17
0.23
0.30
0.13
0.13
0.33
0.27
0.40
0.43
0.37
0.27
0.23
0.07
0.37
0.17
0.47
0.43
0.27
0.27
0.20
0.23
0.20
0.27
0.47
0.37
0.40
0.10
0.40
R19
0.20
0.30
0.23
0.30
0.17
0.17
0.33
0.33
0.33
0.43
0.37
0.27
0.37
0.27
0.37
0.20
0.47
0.43
0.57
0.20
0.13
0.17
0.53
0.37
0.47
0.37
0.37
0.17
0.23
R20
0.33
0.30
0.53
0.30
0.23
0.17
0.23
0.20
0.27
0.10
0.37
0.30
0.23
0.27
0.37
0.17
0.20
0.43
0.17
0.37
0.20
0.23
0.53
0.37
0.47
0.23
0.40
0.40
0.37
R21
0.33
0.40
0.23
0.03
0.40
0.13
0.03
0.03
0.33
0.17
0.20
0.27
0.37
0.27
0.13
0.07
0.47
0.33
0.17
0.37
0.50
0.33
0.20
0.37
0.43
0.37
0.40
0.27
0.37
R22
0.33
0.40
0.23
0.17
0.17
0.47
0.03
0.17
0.33
0.27
0.20
0.27
0.23
0.27
0.20
0.37
0.20
0.23
0.27
0.37
0.50
0.17
0.03
0.37
0.47
0.13
0.37
0.27
0.40
R23
0.33
0.40
0.53
0.50
0.13
0.10
0.37
0.20
0.40
0.43
0.43
0.43
0.23
0.13
0.13
0.17
0.07
0.23
0.17
0.10
0.20
0.23
0.53
0.27
0.47
0.13
0.37
0.10
0.40
R24
0.33
0.30
0.23
0.50
0.13
0.10
0.33
0.33
0.40
0.43
0.43
0.43
0.37
0.13
0.13
0.20
0.47
0.23
0.57
0.10
0.50
0.23
0.20
0.37
0.47
0.13
0.37
0.10
0.40
R25
0.33
0.40
0.53
0.50
0.40
0.13
0.33
0.27
0.40
0.27
0.20
0.27
0.37
0.27
0.27
0.37
0.20
0.43
0.57
0.27
0.20
0.20
0.53
0.37
0.43
0.23
0.40
0.40
0.37
R26
0.33
0.40
0.53
0.50
0.40
0.47
0.33
0.27
0.33
0.27
0.43
0.43
0.37
0.27
0.37
0.37
0.20
0.33
0.57
0.27
0.50
0.33
0.53
0.27
0.43
0.23
0.40
0.40
0.23
R27
0.33
0.17
0.53
0.50
0.40
0.47
0.33
0.27
0.40
0.17
0.43
0.43
0.10
0.27
0.20
0.37
0.20
0.33
0.57
0.20
0.50
0.20
0.53
0.37
0.43
0.23
0.23
0.17
0.37
R28
0.20
0.40
0.23
0.50
0.40
0.47
0.23
0.27
0.40
0.17
0.43
0.43
0.37
0.27
0.20
0.37
0.47
0.33
0.57
0.20
0.13
0.20
0.53
0.37
0.43
0.27
0.40
0.40
0.23
R29
0.13
0.40
0.53
0.50
0.40
0.17
0.37
0.33
0.27
0.43
0.37
0.43
0.07
0.27
0.37
0.37
0.47
0.43
0.57
0.27
0.17
0.17
0.53
0.27
0.43
0.13
0.40
0.27
0.23
R30
0.33
0.40
0.53
0.17
0.17
0.47
0.23
0.27
0.40
0.43
0.43
0.27
0.10
0.27
0.20
0.37
0.47
0.33
0.57
0.37
0.50
0.33
0.53
0.37
0.47
0.23
0.37
0.40
0.37
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71â&#x20AC;&#x201C;87) A. Yu et al.
79
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
1.0000 0.8198 0.7030 0.8040 0.6876
0.8198 1.0000 0.6828 0.8082 0.6637
0.7030 0.6828 1.0000 0.6760 0.6553
0.8040 0.8082 0.6760 1.0000 0.7050
0.6 876 0.6637 0.6553 (9) 0.7050 1.0000
1.0000 0.5541 0.6030 0.5541 1.0000 0.5131 0.6030 0.5131 1.0000
1.0000 0.6837 0.5846 0.5863
0.6837 1.0000 0.5798 0.6016
0.5846 0.5798 1.0000 0.6667
(10)
0.5863 0.6016 0.6667 1.0000
(11)
As shown in the above matrices 6–10, reflexivity and symmetry are satisfied due to R(yi，yi)=1; R(yi，yj)=R(yj，yi), but do not meet the transitivity. Therefore, values cannot be classified directly. So, fuzzy equivalent matrices were established according to the eq. (3). Six fuzzy equivalent matrices of cutting area design (12), logging technology (13), impacts on ecological environment (14), resources and energy use (15), sustainable development and safety production (16), and labor protection (17) should be calculated by the squares method, as shown in eq. (12)–(17). Matrices (12)–(17), reflexivity, symmetry and transitivity are all satisfied, so classification will be done directly. Each value in the matrices means the l value. 1.0000 0.6904 0.8085 0.7380
1.0000 0.7408 0.6813 0.6699 0.6441
0.6904 1.0000 0.6904 0.6904
0.7408 1.0000 0.6813 0.6699 0.6441
0.8085 0.6904 1.0000 0.7380
0.6813 0.6813 1.0000 0.6699 0.6441
0.7380 0.6904 0.7380 1.0000
0.6699 0.6699 0.6699 1.0000 0.6441
0.6 441 0.6441 0.6441 0.6441 1.0000
(12)
(13)
0.7030 0.7030 1.0000 0.7030 0.7030
0.8082 0.8082 0.7030 1.0000 0.7050
0.7050 0.7050 0.7030 (15) 0.7050 1.0000
1.0000 0.5541 0.6030 0.5541 1.0000 0.5541 0.6030 0.5541 1.0000
1.0000 0.6837 0.6016 0.6016
0.6837 1.0000 0.6016 0.6016
0.6016 0.6016 1.0000 0.6667
0.6016 0.6016 0.6667 1.0000
(16)
(17)
According to the eq. (4) and (5), λ cut matrices were calculated and different classifications were gained. Six classified results of cutting area design (Table 5), logging technology (Table 6), impacts on ecological environment (Table 7), resources and energy use (Table 8), sustainable development and safety production (Table 9), and labor protection indices (Table 10) can be calculated by assigning to λ from large to small and classification, as shown in Tables 5–10. A total of twelve indices were chosen for evaluation. Two indices were selected from each of the following conditions: cutting area design, logging tech-
l value
Classification number
Specific category
1
4
{X1},{X2},{X3},{X4}
0.81
3
{X1,X3},{X2},{X4}
0.74
2
{X1,X3,X4},{X2}
0.69
1
{X1,X2,X3,X4}
Table 6 The cluster result of logging technology
0.6522 0.6384 0.6522 1.0000 0.6522 0.6522 0.6522 0.6522 0.7061 0.6384 0.6831 0.6522 1.0000 0.7061 0.7061 0.6731 0.7221 0.6384 0.6831 0.6522 0.7061 1.0000 0.7221 0.6731 0.7293 0.6384 0.6831 0.6522 0.7061 0.7221 1.0000 0.6731 0.6731 0.6384 0.6731 0.6522 0.6731 0.6731 0.6731 1.0000
80
0.8198 1.0000 0.7030 0.8082 0.7050
Table 5 The cluster result of cutting area design
1.0000 0.6384 0.6831 0.6522 0.7061 0.7221 0.7293 0.6731 0.6384 1.0000 0.6384 0.6384 0.6384 0.6384 0.6384 0.6384 0.6831 0.6384 1.0000 0.6522 0.6831 0.6831 0.6831 0.6731
1.0000 0.8198 0.7030 0.8082 0.7050
(14)
l value
Classification number
Specific category
1
5
{X5},{X6],{X7},{X8},{X9}
0.74
4
{X5,X6},{X7},{X8},{X9}
0.68
3
{X5,X6,X7},{X8},{X9}
0.67
2
{X5,X6,X7,X8},{X9}
0.64
1
{X5,X6,X7,X8,X9}
Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Table 7 The cluster result of ecological environmental impact l Classification value number
Specific category
1
8
{X10} {X11},{X12}.{X13},{X14},{X15},{X16},{X17}
0.73
7
{X10,X16} {X11},{X12}.{X13},{X14},{X15},{X17}
0.72
6
{X10,X15,X16},{X11},{X12}.{X13},{X14},{X17}
0.71
5
{X10,X14,X15,X16},{X11},{X12}.{X13},{X17}
0.68
4
{X10,X12,X14,X15,X16},{X11},{X13},{X17}
0.67
3
{X10,X12,X14,X15,X16,X17},{X11},{X13}
0.65
2
{X10,X12,X13,X14,X15,X16,X17},{X11},
0.64
1
{X10,X12,X13,X14,X15,X16,X17,X11},
value of l=0.74, where {X2} was a part directly put in the index set, but one typical index should be selected from {X1，X3，X4}. For logging technology, specific categories of {X5，X6，X7，X8} and {X9} yielded a value of l=0.67, where {X9} was a part directly put in the index set, but one typical index should be selected from {X5, X6, X7, X8}, similar to the others. Similar indices have to selected by the correlation index method. Here, the logging technology is taken as an example. The correlation indices r of X5，X6，X7，X8 should be calculated first, as shown in Table 11. Table 11 Correlation coefficients of X5, X6, X7, X8 rij
Table 8 The cluster result of utilization of resources and energy l value
Classification number
Specific category
1
5
{X18},{X19},{X20},{X21},{X22}
0.82
4
{X18,X19},{X20},{X21},{X22}
0.81
3
{X18,X19,X21},{X20},{X22}
0.71
2
{X18,X19,X21,X22},{X20}
0.7
1
{X18,X19,X21,X20,X22}
Classification number
Specific category
1
3
{X23},{X24},{X25}
0.6
2
{X23,X25},{X24}
0.5
1
{X23,X25,X24}
Table 10 The cluster result of safety production management and protection l value
Classification number
Specific category
1
4
{X26},{X27},{X28},{X29}
0.68
3
{X26,X27},{X28},{X29}
0.67
2
{X26,X27},{X28,X29}
0.6
1
{X26,X27,X28,X29}
nology, impacts on ecological environment, resources and energy use, sustainable development and safety production, and labor protection. For the cutting area, specific categories of {X1，X3，X4} and {X2} yielded a Croat. j. for. eng. 37(2016)1
X5
X5
X5
X5
X5
1.0000
0.2999
0.2723
0.0539
X6
0.2999
1.0000
0.1531
0.1322
X7
0.2723
0.1531
1.0000
0.1087
X8
0.0539
0.1322
0.1087
1.0000
The correlation indices R are calculated as follows.
R5 =
r562 + r572 + r582 = 0.0557 3
(18)
R6 =
r652 + r672 + r862 = 0.0436 3
(19)
R7 =
r752 + r762 + r782 = 0.0684 3
(20)
R8 =
r852 + r862 + r872 = 0.0107 3
(21)
Table 9 The cluster result of sustainable development l value
A. Yu et al.
In this instance, the results indicate that R5>R6>R7>R8. Therefore, R5 was put into the category because it was the maximum. The results of the above calculations showed that the indices of logging technology were the rational processes and economics and safe ways to work. So, the indices of cutting area design, impacts on ecological environment, environmental benefits, as well as sustainable development and labor protection could be calculated by the above methods, and the results are shown in Table 12. The indices shown in Table 12 yielded the following results: first, the relative fuzzy similarity matrix was set up by the maximum and minimum method; then, the fuzzy equivalence matrix was calculated from squares and transitive closure; finally, the indices were selected by the correlation coefficient method.
81
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Table 12 CP assessment indicators of plantation logging – six first grade indices and twelve second grade indices (max and min matrix method) Cutting area design
Logging technology
Impacts on ecological environment
Resources and energy use
Sustainable development
Safety production management and protection
Cutting area survey
Rational processes
Soil physical properties
Utilization ratio of wood
Survive ratio of renew
Safety management
Engineering design
Economics and safety of ways to work
Biodiversity
Utilization ratio of slashes
Renew ratio of cutting area
Labor protection
(2) Relation coefficient method A fuzzy similarity matrix was set up by the relation coefficient method, and then the indices of CP evaluation for plantation logging can be calculated by the steps that were similar to the maximum and minimum method. After the raw data was initialized, the relation coefficients between each index were calculated, and the absolute value of the relation coefficients can be the elements for determining the fuzzy similarity matrix. Six fuzzy similarity matrices of cutting area design (22), logging technology (23), impacts on the ecological environment (24), resources and energy use (25), sustainable development and safety production management (26) and labor protection (27) were calculated by relation coefficient method, as shown in eq. (22)–(27). 1.0000 0.1606 0.1473 0.0574
1.0000 0.2999 0.2723 0.0539 0.1084
0.1606 1.0000 0.1168 0.0852
0.2999 1.0000 0.1531 0.1322 0.0067
0.1473 0.1168 1.0000 0.3724
0.2723 0.1531 1.0000 0.1087 0.2941
0.0574 0.0852 0.3724 1.0000
0.0539 0.1322 0.1087 1.0000 0.0745
0.1084 0.0067 0.2941 0.0745 1.0000
(22)
(23)
82
(24)
0.3713 1.0000 0.0218 0.2802 0.3744
0.0046 0.0218 1.0000 0.0237 0.2917
0.0516 0.2802 0.0237 1.0000 0.0076
0.1345 0.3744 0.2917 (25) 0.0076 1.0000
1.0000 0.0346 0.1412 0.0346 1.0000 0.1019 0.1412 0.1019 1.0000
(26)
1.0000 0.2620 0.1077 0.2588 0.2620 1.0000 0.0470 0.1584
(27)
Six fuzzy equivalence matrices of cutting area design (28), logging technology (29), impacts on the ecological environment (30), resources and energy use (31), sustainable development and safety production management (32) and labor protection (33) were calculated by the squares method, as shown in eq. (28)– (33).
1.0000 0.0460 0.0025 0.1657 0.0903 0.0186 0.4263 0.1835 0.0460 1.0000 0.1677 0.0214 0.3787 0.1073 0.1053 0.0310 0.0025 0.1677 1.0000 0.1179 0.0916 0.2481 0.0852 0.0748 0.1657 0.0214 0.1179 1.0000 0.1741 0.0976 0.1401 0.0297 0.0903 0.3787 0.0916 0.1741 1.0000 0.3394 0.3937 0.1963 0.0186 0.1073 0.2481 0.0976 0.3394 1.0000 0.1546 0.0869 0.4263 0.1053 0.0852 0.1401 0.3937 0.1546 1.0000 0.0359 0.1835 0.0310 0.0748 0.0297 0.1963 0.0869 0.0359 1.0000
1.0000 0.3713 0.0046 0.0516 0.1345
1.0000 0.1606 0.1473 0.1473 1.0000 0.2999 0.2723 0.0539 0.1084
0.1606 1.0000 0.1473 0.1473
0.2999 1.0000 0.1531 0.1322 0.0067
0.1473 0473 1.0000 0.3724
0.2723 0.1531 1.0000 0.1087 0.2941
0.1473 0.1473 0.3724 1.0000
0.0539 0.1322 0.1087 1.0000 0.0745
(28)
0.1084 0.0067 (29) 0.2941 0.0745 1.0000
Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71â&#x20AC;&#x201C;87) 1.0000 0.3787 0.2481 0.1741 0.3937 0.3394 0.4263 0.1963 0.3787 1.0000 0.2481 0.1741 0.3787 0.3394 0.3787 0.1963 0.2481 0.2481 1.0000 0.1741 0.2481 0.2481 0.2481 0.1963 0.1741 0.1741 0.1741 1.0000 0.1741 0.1741 0.1741 0.1741 0.3937 0.3787 0.2481 0.1741 1.0000 0.3394 0.3937 0.1963 0.3394 0.3394 0.2481 0.1741 0.3394 1.0000 0.3394 0.1963 0.4263 0.3787 0.2481 0.1741 0.3937 0.3394 1.0000 0.1963 0.1963 0.1963 0.1963 0.1741 0.1963 0.1963 0.1963 1.0000
1.0000 0.3713 0.2917 0.2802 0.3713
0.3713 1.0000 0.2917 0.2802 0.3744
0.2917 0.2917 1.0000 0.2802 0.2917
0.2802 0.2802 0.2802 1.0000 0.2802
(30) 0.3713 0.3744 0.2917 . (31) 0.2802 1.0000
1.0000 0.1019 0.1412 0.1019 1.0000 0.1019 0.1412 0.1019 1.0000 1.0000 0.2620 0.1077 0.2588
0.2620 1.0000 0.1077 0.2588
0.1077 0.1077 1.0000 0.1077
0.2588 0.2588 0.1077 1.0000
(32)
(33)
Later steps were quite the same using the maximum and minimum method, and the indices are shown in Table 13. Results from Table 12 and Table 13 were quite similar. Only two indices were different, which indicates that both of the methods were reliable for calculating the index. Using both methods strengthens the analysis.
4. Discussion In this study, we have selected some evaluation indices of CP on the basis of plantation logging and set principles. CP has been widely applied in all kinds of industry, but seldom used in forestry. In recent years, the theories and concepts of CP have been considered for the harvesting area in China, but they only discussed the concept, role of target, and no real analysis of the products. While the production process was generally found to be environmentally acceptable during timber harvesting, a more formal process is needed for an accurate evaluation (Qiang 2000, Zhao 2000, Zhao 2008, Yu 2009). There was a lack of specifics and targeted results in practice, and guidance was not strong in the forest enterprise. The findings of this Croat. j. for. eng. 37(2016)1
A. Yu et al.
study indicate implementation of cleaner production in China could have a guiding role in forest engineering companies. To evaluate the contents of cleaner production in plantation harvesting operations, we distributed 100 questionnaires to professors in university, researchers in forestry research institute and loggers or workers in forestry enterprises and recovered thirty valid questionnaires. The low participation may indicate that most of them put less emphasis on cleaner production, especially the forest engineering enterprises in China. It has been proposed that the Chinese government should increase publicity and education in this area. Like other industries, implementation could include the introduction of incentives or relevant laws and regulations to ensure that more attention is paid on the cleaner production in the forestry enterprises. According to the investigation results based on the thirty submitted responses, the fuzzy clustering method was used to initially screen twenty-nine secondgrade indices. Six first-grade indices and twelve second-grade indices were selected as formal evaluation indices. As we all know, the other second-grade indices not selected are also important, but too many indices will be difficult to analyze in practice in forest companies. According to the principles for determining evaluation indices, the index system for CP could not cover all the processes in the plantation logging. Generally the most simple and focused indices are implemented effectively. The purpose of the questionnaire was issued for the forestry enterprises, with the more important indices related to the implementation of cleaner production in plantation logging. According to the 30 valid questionnaires, the fuzzy cluster methods were used to select the more important indices. This indicates that the selected second-grade indices are more important than other second-grade indices not selected according to the results of questionnaires and fuzzy cluster methods. At the same time, several methods were used to screen evaluation indices, such as the gray correlation method, analytic hierarchy process, etc., (Shen 2002, Xu 2007) but the fuzzy clustering analysis was widely used for objectivity and accuracy (Yang 2011). The results showed that the fuzzy clustering method is reliable for screening indices. CP of plantation logging had its own advantages compared with other CP objectives that were combined, such as saving resources, reasonable use of renewable energy in forests, protecting the environment in forests and educating personal protection for workers. The basic principle with CP is Âťprevention must come first, treatment should only
83
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Table 13 CP assessment indicators of plantation logging – six first grade indices and twelve second grade indices (correlation coefficient matrix method) Cutting area design
Logging technology
Impacts on the Resources and energy ecological environment use
Sustainable development
Safety production management and labor protection
Cutting area survey
Rational processes
Soil physical properties
Utilization ratio of wood
Survive ratio of renew
Safety management
Production process design
Economics and safety of ways to work
Biodiversity
Logging equipment fuel consumption
Renew ratio of cutting area
Labor protection
be needed in special circumstances« in all the stages of the logging process. Our initial findings indicate the methods of using advanced technology and improving process and equipment. Different logging methods (included cutting, bucking, skidding and transportation) were used in cutting areas due to different terrain conditions, and even the working efficiencies differed when using the same logging methods in the same cutting area. For example, skidding with cableway was suitable for hills or mountain communities, but skidding with tractors was suitable for flat communities. The harvesting machine has been used due to its high work efficiency; however, it is not suitable for all cutting areas because of working conditions. Therefore, it is better to use advanced technology and improve processing and equipment based on local conditions. Secondly, in the ways of saving and rationally using resources and energy, the tree was the raw material for CP of logging. The best time to log was when the tree biomass has been at its maximum. On the other hand, the special characteristics of CP for logging were the complex working conditions, fertility of the slash, overlapping of the working system and ecological system, complexity and fuzziness of multiple factors. In recent years, two areas have been studied on plantation logging. One area was the impacts of logging on the ecological environment, whic included the soils, vegetation, tree survival, wild animals, biological diversity, landscapes, and use of slash (Wang 2005, Rab 1994a and 1996b, Li 1994, Qiu 1998, Chen 1999, Croke et al. 2001a and 2006b, Blumfield et al. 2003, Hartanto et al. 2003, Bird et al. 2004, Pennington et al. 2004, Radel et al. 2006, Demir et al. 2007, Langer et al. 2008, Frey et al. 2009, Walmsley et al. 2009, Stoffel et al. 2010) The other study was done on the ecological logging processes and technology, which included the evaluation of ecological, economic and combined benefits (Huang 1995, Deng 2005, Zhang 2005, Spinelli et
84
al. 2012, Berhongaray et al. 2013). Many evaluation systems for CP have been widely used in chemical engineering, services, and agricultures all over the world. Studies on CP for forests pointed out the conception, overall goals, guidelines, and the implementation of the measures without further investigations. In China, some rules and regulations of CP have been accomplished in a few large logging enterprises, as indicated in the article »The Measures for CP by Genhe Forestry Bureau«. Unfortunately, there was not any further analysis involving contents, index system for evaluation, and methods for evaluation. Therefore, more studies will be needed on CP for plantation logging.
5. Conclusions In this paper, an index system of CP evaluation for plantation logging was set up on the basis of CP, according to the characteristics of plantation logging. The contents of such an index system included the following six first-grade criteria: cutting area design, logging technology, impacts on the ecological environment, resources and energy use, sustainable development and safety production management, and labor protection. The fuzzy cluster method was used to select the indices. Finally, twelve second-grade indices were selected as the evaluation indices. A comparison using both maximum and minimum analysis and correlation analysis showed that 10 of the 12 indices were acceptable. The twelve second-grade indices selected are just the first step for CP of logging. According to the results of selecting, these made the standard for CP evaluation of plantation logging. Then, according to the standards, cleaner production audit will be implemented to check the forest companies, expecting forest companies to meet the standards. The implemented guidelines of CP in plantation logging will give them the right directions. So, the results of selecting indices will be beneficial for sustainable production and management of plantation forests in China. Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Acknowledgments This research received support from International Science and Technology Cooperation Projects of China (China–Finland) (2006DFA32840): Study on Harvesting Model for Plantation Forest Based on Industrial Ecology. The project was funded by the »Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD)«.
6. References Anon.: The results of the eighth national forest resources inventory in China (2009–2013). Available on http://www. forestry.gov.cn/main/65/content-659670.html Long, A.J., 2006: Environmentally sound forest harvesting. University of Florida，IFAS Extension, SS-FOR-6: 1–11. Frey, B., Kremer, J., Rüdt, A., Sciacca, S., Matthies, D., 2009: Compaction of forest soils with heavy logging machinery affects soil bacterial community structure. European Journal of Soil Biology 45(4): 312–320. Berhongaray, G., Kasmioui, E.O., Ceulemans, R., 2013: Comparative analysis of harvesting machines on an operational high-density short rotation woody crop (SRWC) culture: One-process versus two-process harvest operation. Biosystems Engineering 58: 333–342. Bird, S.B., Coulson, R.N., Fisher, R.F., 2004: Changes in soil and litter arthropod abundance following tree harvesting and site preparation in a loblolly pine (Pinus taeda L.) plantation. Forest Ecology and Management 202(1–3): 195–208． Blumfield, T.J., Xu, Z.H., 2003: Impact of harvest residues on soil mineral nitrogen dynamics following clearfall harvesting of a hoop pine plantation in subtropical Australia. Forest Ecology and Management 179(1–3): 55–67. Chen, H.J., 1999: Ecological environment analysis of Abies holophylla plantations under different cutting systems. Journal of Forestry Research 10(3): 181–182. Chen, K.Y., 2003: Ecosystem health: ecological sustainability target of strategic environment assessment. Journal of Forestry Research 14(2): 146–150. Chiba, R., Sato-llic, Mika, 2012: Analysis of Web Survey Data based on Similarity of Fuzzy Clusters. Procedia Computer Science 12: 224–229. Croke, J., Peter, H., Peter, F., 2001: Soil recovery from track construction and harvesting changes in surface infltration, erosion and delivery rates with time. Forest Ecology and Management 143(1–3): 3–12.
A. Yu et al.
and surface soil properties on skid road in an oak (Quercus petrea L.) stand. Building and Environment 42(3): 1194–1199. Deng, S.M, 2005: Evaluation on Economic Benefits of Operation Modes of Logging Commonly Used in an Plantation Forest. Forest Engingeering 21(1): 57–59. Dykstra, D.P., 2001: Reduced impact logging: concepts and issues．In: Applying Reduced Impact Logging to Advance Sustainable Forest Management. Asia-Pacific Forestry Commission International Conference Proceedings, Kuching, Malaysia, Feb 26–Mar 1, 15–31. Gillman G.P., Sinclair, D.F., Knowlton, R., Keys, M.G., 1985: The effect on some soil chemical properties of the selective logging of a north Queensland Rainforest. Forest Ecology and Management 12(3–4): 195–214. Guan, Y.X., Wang, Y.L., Guo, X.Y., 2009: Apply fuzzy clustering approach in construct the evaluation index system of core competence of corporation. Mathematics in Practice and Theory 3: 1–7. Han, P., Wang, J.H., Ma, Z.H., Lu, A.X., Gao, M., Pan, L.G., 2011: Application of Fuzzy Clustering Analysis in Classification of Soil in Qinghai and Heilongjiang of China．Computer and Computing Technologies in Agriculture IV; IFIP Advances in Information and Communication Technology 344: 282–289. Hartanto, H., Prabhu, R., Widayat, A.S.E., Asdak, C., 2003: Factors affecting runoff and soil erosion: plot-level soil loss monitoring for assessing sustainability of forest management. Forest Ecology and Management 180(1–3): 361–374. He, H.S., Mladenoff, D.J., Gustafson, E.J., 2002: Study of landscape change under forest harvesting and climate warming-induced fire disturbance. Forest Ecology and Management 155(1–3): 257–270. Huang, G.P., 1995: Optimal solution of cargo timber production in cutting area. Forest Engineering 11(3): 21–24. Klawonn, F., 2004: Fuzzy clustering: Insight and a new approach. Mathware & softcomputing 11:125–142. Lisa Langer, E.R., Steward, G.A., Kimberley, M.O., 2008: Vegetation structure, composition and effect of pine plantation harvesting on riparian buffers in New Zealand. Forest Ecology and Management 256(5): 949–957. Lee, C.S., Leitmann, G., 1994: A stabilizing harvesting strategy for an uncertain model of an ecological system. Computers & Mathematics with Applications 27(9–10): 199–212. Li, Y.F., Li, P.Y., 2008: Evaluation indexes select ion of stocks investment value based on fuzzy clustering. Journal of Yanshan University 32(6): 551–556.
Croke, J., Nethery, M., 2006: Modeling runoff and soil erosion in logged forests: Scope and application of some existing models. Catena 67(1): 35–49.
Li, M., Gu, Y.D., Feng, Y.B., Wang, J.Y., 2003: A Method to Decrease the Distortion of Fuzzy. Journal of Beijing Normal University (Natural Science) 39(5): 601–605.
Demir, M., Makineci, E., Yilmaz, E., 2007: Investigation of timber harvesting impacts on herbaceous cover, forest floor
Peng, Z.Z., 2003: Fuzzy mathematics and its application. Wuhan: Wuhan University press.
Croat. j. for. eng. 37(2016)1
85
A. Yu et al.
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
Pennington, P.I., Laffan, M., 2004: Evaluation of the use of pre- and post-harvest bulk density measurements in wet Eucalyptus oblique forest in Southern Tasmania. Ecological Indicators 4(1): 39–54.
Walmsley, J.D., Jones, D.L., Reynolds, B., Price, M.H., Healey, J.R., 2009: Whole tree harvesting can reduce second rotation forest productivity. Forest Ecology and Management 257(3): 1104–1111.
Qiang, X.B., Xing, L.P., Zhao, H., 2000: The theory and method of cleaner production in harvesting area. Forest Engineering 16(6): 6–8.
Wang, L.H., 1997: Rational harvesting policy and sustainable development of community in forest regions. Journal of Forestry Research 8(1): 50–53.
Qiu, R.H., 1998: Theory and research of Ecological Forestry Operations. Forestry Science and Technology 1: 7–10.
Wang, L.H., Yang, X.C., 2005: Harvesting and Environment of Forestry. Harbin: Northeast Forestry University Press.
Rab, M.A., 1994: Changes in physical properties of a soil associated with logging of Eucalyptus regnans forest in southeastern Australia. Forest Ecology and Management 70(1–3): 215–229.
Wang, W.N., Zhang, Y.J., 2007: On fuzzy cluster validity indices. Fuzzy Sets and Systems 158(19): 2095–2117.
Rab, M.A., 1996: Soil physical and hydrological properties following logging and slash burning in the Eucalyptus regnans forest of southeastern Australia. Forest Ecology and Management 84(1): 159–176.
Xu, H., Luo, C., Liu, Z.G., 2007: Study on the selection method of evaluation indices by AHP. China offshore oil and gas 19(6): 415–418. Xu, H.Y., Wang, G.A., Wang, W.S., 2005: A study of fuzzy clustering analysis in data mining. Computer Engineering and Applications 41(17): 177–179.
Radel, V.C., Mladen, D.J., Gustafson, E.J., 2006: Modeling forest harvesting effects on landscape pattern in the Northwest Wisconsin Pine Barrens. Forest Ecology and Management 236(1): 113–126.
Yan, M.L., Wang, M., Yu, B., 2008: A heavy rainfall forecast method based on fuzzy cluster typing by using application and interpretation of NWP. Scientia Meteorologica Sinica 28(5): 581–585.
Rousseeuw, P.J., 1995: Discussion: Fuzzy Clustering at the Intersection. Technometrics 37(3): 283–286.
Yang, C.L., Yang, C.Y., 2011: Application of fuzzy cluster in prediction coal and rock dynamic disasters. Procedia Engineering 26: 1541–1546.
Sato-Ilic, M., Jain, C.L., 2006: Innovations in Fuzzy Clustering-Theory and Applications. Studies in Fuzziness and Soft Computing. Springer, 205 p. Sârbu, C., Einax, J.W., 2008: Study of traffic-emitted lead pollution of soil and plants using different fuzzy clustering algorithms. Anal Bioanal Chem 390(5): 1293–1301. Silviu. B., 2010: Fuzzy clustering. Details see at http://science5.net/f/fuzzy-clustering---ubb,-cluj-w14253.html Shen, Z.Y., Yang, Z.F., 2002: Gray associate analysis method in screening of index system. Mathematics in practice and theory 32(5): 728–732. Spinelli, R., Schweier, J., Francesco, F.D., 2012: Harvesting techniques for non-industrial biomass plantations. Biosystems Engineering 113(4): 319–324. Stoffel, J.L., Gower, S.T., Forrester, J.A., Mladenoff, D.J., 2010: Effects of winter selective tree harvest on soil microclimate and surface CO2 flux of a northern hardwood forest. Forest Ecology and Management 259(3): 257–265.
86
Yu, A.H., Zhao, C., Huang, Y., 2009: A Study on Cleaner Production Evaluation of Forest Plantation Logging. Acta Agriculturae Universitatis Jiangxiensis 31(2): 311–316. Zhang, Z.X., Zhou, X.N., Zhao, C., Chen. Y.F., 2008: Selecting of the Optimum Operation Model of Ecological Harvesting and Transportation in Southern Artificial Forest Area in China. Sciential Silvae Siniae 44(5): 128–134. Zhang, Z.X., 2005: Study on plantation harvesting operations System in northwestern Fujian, Nanjing Forestry University. Zhao, J.R., Zhang, D.L., 2003: Cleaner Production Promotion Law Q & A [M]. Beijing: Academy Press, 15–20. Zhao, H., Li, X.L., Qiang, X.B., 2000: Summary of the cleaner production of Silvicultural enterprise. Forestry Enterprise of China (4): 42–43. Zhao, K., Zhao, C., 2008: Cleaner production applied in timber harvesting and utilization of fast growing forest. Forest resources management 2: 47–50.
Croat. j. for. eng. 37(2016)1
Selecting Evaluation Indices for Cleaner Production of Plantation Logging in Southern China ... (71–87)
A. Yu et al.
Authors’ address: Assoc. prof., Aihua Yu, PhD. e-mail: azy0008@auburn.edu Prof., Chen Zhao, PhD. e-mail: czh@njfu.edu.cn Lecturer, Yao Zhao, PhD. e-mail:zhaoyaonfu@163.com Department of Forest Engineering Nanjing Forestry University 159 RongPan Road 210037 Nanjing, Jiangsu province P.R. CHINA
Received: October 15, 2014 Accepted: April 2, 2015 Croat. j. for. eng. 37(2016)1
Assoc. prof., Tom Gallagher, PhD.* e-mail: tgallagher@auburn.edu School of Forestry and Wildlife Sciences Auburn University 3425 Forestry and Wildlife Sciences Building 36830 Auburn, Alabama USA * Corresponding author
87
Original scientific paper
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds (Fagus Sylvatica l.) with the Use of Pneumatic Sowing Set Józef Walczyk, Maria Walczykova Abstract The aim of the present study was to adapt the pneumatic seeder designed for thick seeds to the sowing of beech seeds and to select appropriate parameters of its operation. In order to evaluate the possibility of sowing particular seeds by the single-seed seeder, one must examine the process of their gathering by the sowing set, stability of the outlet and the flight trajectory. All these parameters can be evaluated simultaneously by applying the accelerated motion film technology. This task was performed using the GoPro Black Edition film camera. The research was performed at a laboratory measurement site, for four rotational speeds of the spreading disc, four values of the vacuum and three dimensions of the spreading disc holes. As a result, the most appropriate parameters for sowing beech seeds were determined. The research found that, after modification, the disc pneumatic seeder of the Agricola Italiana PK can be applied as a machine for sowing the seeds of European beech. The best sowing parameters are obtained using the disk with holes having a diameter of 5 mm, under a vacuum of 800 mm H2O; and the frequency of unfilled openings in the spreading disc in the case of beech depends more on the value of the applied vacuum than on the rotational speed of the spreading disc. It was also found that the calculated theoretical trajectory of the seed flight is similar to the trajectory obtained from the analysis. Keywords: beech, single-seed sowing, film method, sowing parameters
1. Introduction In Poland, between 50 and 60 thousand hectares of forest are regenerated annually, of which only about 1.5 thousand hectares is natural regeneration. For the remaining area, seedlings are needed. The area of forest nurseries in Poland amounts to 2282 ha; these nurseries produce over 800 million of seedlings per year (Forests in Poland 2011). It follows that such a high demand for the planting material requires continuous investment in forest nurseries and development of new technologies that will reduce workload, reduce the difficulty of the work, shorten the periods of sowing, improve the use of the sowing space and yield planting material of better quality (Wesoły and Hauke 2009). One way to improve the quality of sowing and reduce its workload is the use of single-seed sowing (Zhao-qian et al. 2005). This kind of sowing is commonly used in agriculture and horticulture but in forest nurseries it is still Croat. j. for. eng. 37(2016)1
uncommon; however, the practice of forest management clearly feels its absence. The specificity of sowing in forest nurseries, mainly under controlled conditions, requires a good use of the sowing surface. Under these conditions, the substrate usually consists of peat, which is imported not only in Poland but in many countries; controlled mycorrhization is also often used (Alexandrowicz-Trzcińska et al. 2013). In addition, expensive granular fertilizers with a prolonged period of decomposition are used, and the construction of troughs, greenhouses or plastic tents is also expensive. For these reasons, well performed sowing has special significance (Walczyk and Słowiński 2013). So far, due to the lack of appropriate equipment, the sowing of beech under controlled conditions has been performed manually or with the use of less accurate row seeders. This way of sowing does not guarantee a uniform vertical and horizontal distribution of seeds and proper contact of the sown seeds with the soil. The result is worse germination and the seedlings
89
J. Walczyk and M. Walczykova
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
are characterized by a great diversity of morphological parameters (Barzdajn 1981). These problems can be avoided by using the single-seed sowing method. However, there are technical problems that need to be dealt with. Thus in the case of beech seeds, due to their shape, it is difficult to gather the sowing sets. Sowing these seeds requires a seeder with a large seed chamber, since they tend to hang up in the container, which particularly concerns the sprouted seeds. Thus, for the purposes of this study and modification, the Agricola Italiana, PK model of the disc pneumatic seeder for sowing single large seeds was selected as potentially suitable for sowing beech seeds. In order to fully evaluate its usefulness in this application, the scope of the research included: Þ analysis of gathering the seeds by the spreading disc in terms of number of seeds sucked to the disc, determination of the point of outlet of the seeds from the spreading disc and the trajectory of the sown seeds; Þ selection of the vacuum on the spreading disc, the diameter of the disc holes and the rotational speed of the spreading disc necessary to obtain a single seed in a hole of the spreading disc.
2. Materials and methods The study was performed at a measurement set-up (Fig.1) for: Þ four speed ranges of the spreading disc, which corresponds to four seeder driving speeds: 18, 35, 50, 70 rev min–1, Þ three diameters of the spreading disc holes: 3, 5 and 6 mm, Þ four vacuums: 600, 700, 800 and 900 mm H20.
2.1 Description of the measurement setup The basic elements of the measurement setup (Fig. 1) are a section of the tested Agricola Italiana PK seeder and a film camera. Other components of the equipment, such as the propulsion system with an electric motor with a reducer and a multi-chain transmission, a fan with a stepless expenditure regulation and a pressure gauge allow for changing the relevant parameters of the seeder operation, i.e. the vacuum value and the rotational speed of the spreading disc. The Agricola Italiana PK seeder belongs to a group of vacuum single-seed seeders. It is designed for precise sowing of thick seeds, such as maize, faba beans, kidney beans, and it is constructed as a typical seeder. It has a sectional structure with the possible number of sections from 2 to 12, seed chambers with a capac-
90
Fig. 1 A view of the measurement site ity of 20 dm3, a central vacuum fan, elastic air hoses, coulters, kneading wheels, wheels pressing the seeds in the furrow, sweeps, a system for regulating the sowing depth and the drive system. Adjustment of the seeder to the sowing of a given seed type is done by using the spreading disc with a different diameter and number of holes as well as by vacuum selection. Adjusting the distance of seeds in a row is done by changing the gear ratio of the drive system or the use of the spreading disc with a different number of holes. The sowing of beech seeds was done using the spreading discs with a single row of holes on their circumference. The first attempts to sow at the measurement setup showed a high non-uniformity of seeder operation. The reason for this was frequent suspension of the seeds in the seed chamber of the sowing set. To prevent this phenomenon, it was necessary to design a seed mixer. It was assumed that the mixer should be easily removable and should not interfere with the structure of the seeder design. The mixer constructed in this way (Fig. 2) meets these requirements and effectively prevents suspending the seeds in the container, thus allowing stable operation of the sowing unit (Walczyk and Tylek 2014). The mixer (Fig. 2) consists of a ring, to which the following items are attached: a stirring rod, drive spring, return spring and return movement limiter. It is mounted on the hub of the spreading disc axle in the seed chamber of the tested seeder and its operation does not require any other equipment or structural changes of the seeder. It is operated by the projections located on the spreading disc, which, by catching on Croat. j. for. eng. 37(2016)1
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
J. Walczyk and M. Walczykova
Fig. 2 The seeder mixer constructed in the present study the drive spring, cause the rotation of the ring on the hub, and thus also the movement of the seed mixer in the seed chamber.
Fig. 3 Diagram of the analysis of the seed flight trajectory Croat. j. for. eng. 37(2016)1
2.2. Recording seeder operation and determining the outlet point and seed flight trajectory The seeds used for sowing were beech (Fagus sylvatica L.), with the weight of a thousand grains amounting to 306 g. Recording and analysing of the sowing process in the range indicated above is not possible without applying the accelerated motion film technology, as evidenced by the studies of other authors (Karaya et al. 2006). Therefore, the filming of the course of sowing was done with the use of the Black Edition GoPro camera, filming the course of sowing with a frequency of 240 frames s-1. During the rendering in the Adobe Premier Pro CS6 editing software, the film was decelerated 10 times and subjected to qualitative and quantitative analysis (Walczyk 2005). Under the research conditions at the measurement setup, the circumferential speed vo, acting on a seed sown at the outlet point S, has a horizontal component vx and a vertical component vy (Fig. 3). In order to calculate a route s travelled by the seed in the horizontal direction in a given time, the horizontal speed component vx (eq. 1) must be multiplied by the seed flight time (eq. 2):
Vx = Vo × cosa
(1)
S = t × Vx
(2)
91
J. Walczyk and M. Walczykova
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
Table 1 The effect of the spreading disc hole-diameter on gathering of seeds by the spreading disc holes for the vacuum of 800 mm H2O and 35 disc revolutions per minute Voids
Single seed gathering
Double seed gathering
Spreading disc dimensions
Mean
Standard deviation
Mean
Standard deviation
Mean
Standard deviation
45* x 3.0**
33.0
2.2
67.0
2.2
0.0
0.0
35* x 5.0**
2.7
1.3
95.2
1.9
2.1
1.4
35* x 6.0**
3.7
1.5
86
2.7
10.3
2.1
*number of holes, **disc hole diameter (mm)
The analysis of the film allowed for simultaneous determination of the number of seeds in individual holes of the spreading disc. The angle α of the outlet of seeds from the disc was determined for the x and y coordinates of the point of a seed at the moment of its falling from the hole of the disc. It was also the zero point for the analysis of the trajectory of a falling seed. The coordinates of the position of successive points of the falling seed trajectory were determined at every 6 frames of the film, i.e. at every 0.025 s. Measurement of the seed trajectory was performed over a period of 0.1 s, i.e. for 4 points of the trajectory. On the basis of the measurements, the path travelled by the seed in the direction of the x axis (horizontal) was determined for each circumferential speed of the spreading disc and it was compared with the theoretical path obtained by using eq. (1). In order to limit the number of repetitions performed during the study, which would
disturb the transparency of the obtained results, first the effect of the spreading disc hole diameter on the gathering of the seeds was determined at the vacuum, recommended by the manufacturer, for the sowing of large seeds. Then, for the diameter of the holes, for which the best results were obtained, an analysis of the effect of the vacuum and the rotational speed of the disc on sowing quality was performed.
3. Results 3.1. Analysis of seed gathering Results of research concerning the selection of hole diameters of the spreading disc, appropriate for seeds of beech of the given weight, showed that the best results were obtained for the disc with 35 holes having a diameter of 5.0 mm (Table 1). The operation of this
Table 2 The effect of vacuum and spreading disc rotation on gathering of seeds, for a disc hole-diameter of 5 mm Vacuum on spreading disc holes, mm H20 600
Disc rotations rev/min–1
18
35
50
70
92
700
800
900
Number of seeds gathered through disc holes 0–voids, 1–single seed gathering, 2–double gathering 0
1
2
0
1
2
0
1
2
0
1
2
Mean
10.7
89.3
0.0
9.2
83.0
7.7
4.1
87.5
8.5
10.8
76.2
13.0
Std. dev.
2.1
2.8
0.0
2.3
2.7
2.2
1.9
2.7
2.3
2.7
3.1
3.2
Mean
12.4
87.6
0.0
11.2
86.9
2.0
3.7
91.6
4.7
5.0
88.5
6.4
Std. dev.
2.3
2.7
0.0
2.4
3.1
1.8
1.8
2.1
1.6
1.7
3.5
2.1
Mean
14.5
85.5
0.0
12.4
87.6
0.0
4.8
95.2
0.0
2.9
94.1
3.0
Std. dev.
3.1
2.8
0.0
2.5
2.4
1.8
1.6
1.9
0.0
1.3
1.9
1.2
Mean
18.1
81.9
0.0
12.9
87.1
0.0
5.6
94.4
0.0
2.4
97.6
0.0
Std. dev.
3.5
3.7
0.0
2.7
2.2
0.0
2.1
1.8
0.0
1.2
1.7
0.0
Croat. j. for. eng. 37(2016)1
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
J. Walczyk and M. Walczykova
Table 3 Results of the statistical analysis for the disc hole diameter of 5 mm Disc rotations rev/min–1
Single seed gathering Voids % %
Double gathering %
Execution of the sowing standard,%
18
87.5
4.1
8.5
112.9
35
91.6
3.7
4.7
105.7
50
95.2
4.8
0.0
95.2
70
94.4
5.6
0.0
94.4
Mean
91.5
4.5
4.0
103.5
Standard deviation
3.4
0.7
3.9
8.4
R between spreading disc speed and share of correctly gathered seeds
0.9
–
–
–
R* between spreading disc speed and share of voids
–
0.84
–
–
R* between spreading disc speed and share of double gathering
–
–
–0.94
–
R* between spreading disc speed and share of sowing standard execution
–
–
–
–0.95
R* – coefficient of correlation
disc has the highest (average 95.2) percentage of single seeds sucked to the sowing holes. The number of empty holes (voids), as well as the holes filled with two seeds, amounted to only about 2%. The discs with 35 holes, having a diameter of 6.0 mm, and the disc with 45 holes, having a diameter of 3.0 mm, showed much worse results (Table 1). For correct sowing, it is important that the seeds sucked into the holes do not come into contact with each other. This condition is fulfilled by the discs with 35 holes. Analysis of the impact of the rotational speed and vacuum of the disc on the correct gathering of the seeds allows for the conclusion that the best results were achieved for the rotational speed of the disc amounting to 50 and 70 rev min–1 and the pressure of 600 and 700 mm H20 (Table 2). Research results included in Table 3 show that an increase of the spreading disc rotational speed leads to an increase of correctly filled holes, i.e. filled with single seed; however, to some extent an increase of voids takes place as well. The negative statistical correlation was showed by the impact of the rotational speed of the disc on the number of double fillings and the total number of seeds sown (the sowing standard) (Table 3).
and in the direction of the y axis: 4.6 mm. For the seed outlet point determined in this way, the average outlet angle was determined as α = 29°.
3.3. Seed flight trajectory Analysis of the trajectory of seeds for different circumferential speeds of the spreading disc allowed for obtaining seed flight trajectories that clearly differ from each other (Fig. 4 and 5). The path of the seed in the direction of the x axis increases with an increase of
3.2. Seed outlet point Determined on the basis of the film analysis, the point of outlet of seeds from the disc was characterised by a high stability for all of the sowing speeds; the average standard deviation for all analysed points was, respectively, in the direction of the x axis: 4.1 mm, Croat. j. for. eng. 37(2016)1
Fig. 4 Seed flight trajectory for the disc speed: v1 = 0.166 m s–1
93
J. Walczyk and M. Walczykova
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
the circumferential speed of the spreading disc and, therefore, it should be taken into account during the construction of the coulter and the selection of its placement relative to the sowing set of the seeder. The result of the theoretical calculations of the seed path in a horizontal direction only slightly differs from the result of measurements obtained during the analysis of the film (Table 4). Hence the conclusion that the effect of air resistance and the way of sucking a seed into the spreading disc hole do not significantly alter its flight. For the seeder construction purposes, this path may be determined theoretically.
4. Discussion Based on the results of the laboratory tests, after modification, the disc pneumatic seeder of the Agricola Italiana PK proved to work well as a machine for sowing European beech seeds. When equipped with a mixer, the examined seeder allowed for the sowing of European beech seeds (Fagus sylvatica L.) that meets the requirements of the Polish standard (PN-88/R-36573). The best sowing parameters were obtained for the spreading disc with 5 mm diameter holes, at a vacuum of 800 mm H2O. On the basis of the measurements, it can be stated that the frequency of unfilled spreading disc holes for beech seeds depends more on the value of the applied vacuum than on the rotational speed of the spreading disc. During the sowing in the field, it is very important to ensure proper filling of the spreading disc, which is done by a properly chosen working vacuum and disc holes. Correct operation of the seeder should be checked in each case by making a generally known sowing test. Another important issue is to properly secure the coulter relative to the sowing set of the seeder. For this
Fig. 5 Seed flight trajectory for the disc speed: v4 = 0.664 m s–1 purpose, it is necessary to know the point of seed outlet from the disc and the path travelled by a seed from the time of its outlet from the disc until its fall to the bottom of the furrow. In practice, this path is also influenced by the seeder travelling speed, which is normally directed opposite to the seed flight speed; and ideally they should be equal. Then the seed falls vertically to the bottom of the furrow, which prevents its rolling and improves seeding precision. In general, it may be stated that the coulter wings should extend slightly beyond the point of the seed fall to the bottom of the furrow. This ensures good covering of a seed and prevents its possible rolling in the furrow. The present study provides data on the parameters of the seeder operation settings and it may be helpful for the determination of the coulter attachment point relative to the sowing set. For the purpose of the seeder con-
Table 4 Comparison of the seed flight trajectory in the horizontal direction s (cm) obtained on the basis of film analysis, with the path st (cm) designated on the basis of theoretical calculations, for the disc hole-diameter of 5 mm and vacuum of 800 mm H2O V1 = 0.166, m s–1
V2 = 0.332, m s–1
t, s
x1
xt1
x2
xt2
x3
xt3
x4
xt4
0.000
0.0
0.00
0.0
0.00
0.0
0.00
0.0
0.00
0.025
0.4
0.36
0.8
0.73
1.1
1.03
1.6
1.45
0.050
0.7
0.73
1.6
1.45
2.4
2.05
3.2
2.91
0.075
1.1
1.09
2.5
2.18
3.5
3.08
4.8
4.36
0.100
1.5
1.45
3.3
2.91
4.6
4.10
6.5
5.81
Time of flight
94
V3= 0.469, m s–1
V4 = 0.664, m s–1
Croat. j. for. eng. 37(2016)1
Evaluation of the Possibility of Single-Seed Sowing of Beech Seeds ... (89–95)
struction, it is possible to use the theoretically calculated trajectory of the seed flight, which is close to the trajectory obtained based on the analysis of the film. Application of film technology in this kind of research (Karaye et al. 2006, Walczyk and Tylek 1997, Walczyk 2005) allowed not only for quantitative analysis of the seeder sowing set operation, but also for qualitative analysis. Thanks to using film technique, it was possible to detect the phenomenon of suspending the seeds in the seed chamber, which periodically worsened the seeder operation.
5. Conclusions After modification, the Agricola Italiana PK disc pneumatic seeder turned out to be suitable for sowing European beech seeds. The best sowing results were achieved when the holes of the spreading discs were five millimetres wide, at a vacuum of 600 and 700 mm H2O. The frequency of unfilled spreading disc holes for beech seeds depends more on the value of the applied vacuum than on the rotational speed of the spreading disc. For the purpose of the seeder construction, it is possible to use the theoretically calculated trajectory of the seed flight, which is close to the trajectory obtained based on the analysis of the film.
6. References Anon.: PN-88/R-36573 Maszyny rolnicze. Siewniki punktowe. Ogólne wymagania i badania.
J. Walczyk and M. Walczykova
Anon. 2013: Forests in Poland. Published by The State Forests Information Centre. Warszawa, 47 p. Aleksandrowicz-Trzcińska, M., Żybura, H., Drozdowski, S., 2013: Effect of the substrate type, controlled mycorrhization and application of fungicides in the nursery on the growth of pedunculate oak in the plantation. Sylwan 157(3): 197– 203. Barzdajn, W. 1981: Influence of beech (Fagus sylvatica L.) sowing density in nursery and under a foil tent upon morphological characters of one year old seedlings, success and growth of plantation. Sylwan 125(6): 13–20. Karaye, D., Wiesehoff, M., Özmerzi, A., Müller, J. 2006: Laboratory measurement of seed drill seed spacing and velocity of fall of seeds using high-speed camera system. Computers and Electronics in Agriculture 50(2): 89–96. Wesoły, W., Hauke, M., 2009: Szkółkarstwo leśne od A do Z. Warszawa, 411 p. Walczyk, J., 2005: Operation analysis for Agricola Italiana seeder carried out using film techniques. Inżynieria Rolnicza 10(70): 395–402. Walczyk, J., Tylek, P., 1997: An Analysis of a Point-Method Sowing of Forest Tree Seeds. Sylwan R. 141(3): 57–64. Walczyk, J., Słowiński, K., 2013: Cultivation mechanization of plants in troughs and under covers. In: „ Mobilné energetické prostriedky – hydraulika – životné prostriedie – ergonómia mobilných strojov, Zvolen, 234–244. Walczyk, J., Tylek, P., 2014: Zgłoszenie wzoru użytkowego nr W.123096 z dnia 19.05. 2014 pt. Mieszadło siewnika punktowego typ PK firmy Agricola Italiana. Zhao-qian, W., Ming-gang, L., Jia-yan, S., 2005: A Research on the Precise Drilling Technology of Forest Seed-tapes. Forestry Machinery&Woodworking Equipment 11.
Authors’ address: Prof. Józef Walczyk, PhD. e-mail: rlwalczy@cyf-kr.edu.pl University of Agriculture in Kraków Faculty of Forestry Al. 29 Listopada 46 31-425 Kraków POLAND
Received: October 18, 2014 Accepted: September 12, 2015 Croat. j. for. eng. 37(2016)1
Prof. Maria Walczykova, PhD. * University of Agriculture in Kraków Faculty of Production and Power Engineering ul. Balicka 116 30-149 Kraków POLAND * Corresponding author
95
Original scientific paper
Forest Workers and Steep Terrain Winching: the Impact of Environmental and Anthropometric Parameters on Performance Giovanna Ottaviani Aalmo, Natascia Magagnotti, Raffaele Spinelli Abstract Winching is common in small-scale forest operations, especially on steep slopes, where tractors cannot reach the logs inside the forest. In this case, logs are dragged to the roadside with tractor-mounted winches, for later collection by transportation units. Winching is a heavy task, posing a high physiological stress on winching crew members. The aim of this study was to investigate the relationship between experienced workload, work conditions and operator fitness. The study confirmed the assumption that fit, young operators experience a lower workload than older ones. Workload depends on winching direction, and it is higher when winching downhill than when winching uphill. Results confirmed that gravity is the main factor, and it has a stronger effect than task type and tool weight. Walking uphill with no tools is heavier than walking downhill and carrying a steel cable. As a consequence, tool weight reduction can only palliate the problem, without solving it. Winching crews should be composed of fit, young workers. When the task is assigned to older workers, it is necessary to allow longer rest breaks, accepting a lower productivity. Keywords: steep terrain, winching, workload, heart rate
1. Introduction In many countries, small and fragmented forestland ownership is quite common. Because of the land ownership pattern, small forest owners seek for solutions allowing profit increase with small investments. Farm tractors are a common solution for harvesting small-scale forests (Spinelli and Magagnotti 2012). They can be purchased at a fraction of the cost required for specialised forest machinery, and they are quite versatile, which allows their use in bunching, skidding, forwarding and loading tasks. Winchequipped farm tractors can be used in steep terrain for the extraction of manually felled timber. In steep terrain, limited machine mobility and tight safety restrictions related to environmental concerns, prevent direct access to the loads. Therefore, the tractor stops on the road bank or skid trail and uses its winch to drag the logs to the roadside. This requires someone to pull Croat. j. for. eng. 37(2016)1
the winch cable all the way to the logs, and connect them to the cable using choker chains. The work of pulling the cable up/downhill through the forest is physically demanding, and results in a high energy demand, which often exceeds the endurance limit (Stampfer 1998, Vik 1984). In addition, walking directly up the slope is one of the heaviest and most strenuous activities associated with forestry work. Workload is a measure of the demand experienced by a subject across various means of mental and physical load, resulting from the effects of factors such as task requirements, effort, and performance. Determining the expected levels of overall workload can help evaluate the different tasks and eventually lead to a different system configurations (e.g. pre-determined rest breaks and task assignments) or identify the most efficient team compositions. The efficiency of each harvesting operation in forestry can be increased inter alia
97
G.O. Aalmo et al.
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
by assessing the workers critical workloads for each single task and for each of the task elements. Overall workload gives a general understanding of the demands placed on the workers. Additional understanding regarding the specific types of strains is obtained by examining the individual overall workload channels, e.g., physical workload (Hart 1986). The capacity of performing physical work depends on the ability of the muscle cells to transform food intake into mechanical energy output. This process is affected by parameters such as age (de Zwart et al. 1996), gender, body dimensions and fitness (Rodahl 2003). In motor manual cutting, Hagen et al. (1993) showed that there was no difference between heart rate of younger and older lumberjacks; Lilley et al. (2002) found no influence of age, while Rodgers (1997) found significant influence of age. Body dimensions (height and weight) were found significant in the development of back pain in several researches (Krause et al. 1997, Keyserling et al. 1988, Heuch 2010, Hagen et al. 2010). Afolabi and Akanbi (2013) found that they also significantly affected the aerobic power. Heart rate has been and still is commonly used as the criterion for the evaluation of physical demands of work but also, for example, for determining rest allowances (Rodgers 1997). Because of the linear relationship between heart rate and oxygen uptake, heart rate can, therefore, be used for estimating the workload (Astrand et al. 2003). Fitness of a subject can be evaluated via the maximal oxygen uptake (VO2max), referring to the maximum amount of oxygen that an individual can utilize during intense or maximal exercise. It is measured as millilitres of oxygen used in one minute per kilogram of body weight. It is a factor that can determine the subject’s capacity to perform sustained exercise and is linked to aerobic endurance. This index has been extensively used in forestry studies related to ergonomics (Afolabi and Akanbi 2013, Park et al. 2003, Parker and Kirk 1994, Staal Wästerlund 2001, Trites 1992). The outdoor working environment also affects the physical performance of forest workers (Ovaskainen et al. 2004). The goals of this study are therefore to: Þd etermine the workload experienced during log winching operations, in order to assess how taxing is this specific task, which is especially common in small-scale forest operations, Þ t o investigate the effect of anthropometric and physical parameters such as age, fitness, weight and height in combination with external environmental parameters i.e. site conditions on the
98
forest workers’ heart rate while performing winching in a controlled operation. No previous work has analysed the effect of these factors on the workload experienced specifically during log winching operations. The available evidence for other tasks is confusing, as some papers confirm the strong effect of these factors, while others find no effect at all. Therefore, it is possible that these effects are task-specific, which warrants specific studies for each given task. Table 1 Description of test corridor Test number
Direction of extraction
Slope, %
Length, m
1
Downhill
64
50
2
Uphill
53
48
3
Uphill
66
46
4
Downhill
60
50
5
Downhill
68
47
6
Downhill
64
55
7
Uphill
54
50
8
Uphill
62
47
9
Uphill
60
50
10
Uphill
60
50
11
Downhill
58
51
12
Downhill
53
50
Fig. 1 Farm tractor with Maxwald hydraulic single drum winch Croat. j. for. eng. 37(2016)1
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
G.O. Aalmo et al.
Table 2 Details of the subject’s anthropometric and physiological parameters Subject
Age, years
Experience, years
Weight, kg
Height, cm
BMI
HRr, bpm
Hrmax, bpm
VO2max
1
45
26
85
170
29.4
68
175
38
2
43
24
104
182
31.4
60
177
41
3
52
34
75
165
27.5
61
168
32
4
22
3
73
176
23.6
63
198
60
5
42
13
77
170
26.6
68
178
36
2. Material and Methods The study was undertaken within two different hauling operation sites in a region, where small forest owners use winches mounted on farm tractors for wood extraction. The tractor used in the test was equipped with a Maxwald single drum hydraulic winch (Fig. 1). This machine was equipped with a slack-puller, which eliminated the effect of drum resistance. The test was performed over a total of twelve corridors. The corridors were chosen based on similar conditions in two different harvesting sites: eight on the first site and four on the second. Corridor length varied between 47 and 55 meters. All harvesting residues were removed so that no hindrance was present. At the end of each corridor, a sufficient number of 4 meter logs were prepared, ready to be extracted. The logs average diameter was 20 cm. Corridor slope varied between 53 and 68%. Two settings were defined: (1) downhill extraction and (1) uphill extraction. Each subject was tested in each corridor. They pulled an 11 mm swaged steel cable, weighting 630 g/m to the logs location, set the choker around one log at the time, followed the log until the designed bunching area and released the log (Table 1). The five subjects were selected among several steep terrain crews. They all had significant experience with this type of operations. They all agreed to participate in the test voluntarily. They were entitled to withdraw at any time, or decline to answer specific questions or complete specific tasks, if desired. The performance of physical work is made possible by muscular activity. Muscles, during their movements, use oxygen to release energy. The energy, required for the performance of a given task, is proportional to the amount of oxygen absorbed; the more energy needed, the more oxygen is needed to compensate for the increased blood circulation. ConCroat. j. for. eng. 37(2016)1
sequently, a higher heart rate implies a close relationship between heart rate and oxygen consumption, with the rate increasing in proportion to work intensity (Astrand et al. 2003, Apud et al. 1989). Therefore, the physical workload can be evaluated comparing heart rates measured during resting and working. The task heaviness can be benchmarked by comparing the heart rate attained for each activity with the individual maximal oxygen uptake. The anaerobic threshold is assumed to be below 40% for an 8 hour working shift (Astrand et al. 2003). Heart rate was measured to assess the level of physical stress of each task element, using a Polar RC3 GPS pulse monitor with continuous data logging and storage of the hart rate readings. Resting heart rate was measured and VO2max was predicted through the Polar OwnIndex upon arrival on the work-site, while lying down and resting in silence for 15 minutes. The OwnIndex usually ranges between 20 and 95 and is comparable with the VO2max commonly used to evaluate aerobic fitness (Table 2) (Polar 2015). The task time elements measured in the time study are reported in Table 3. During the test, no delay was recorded. All technical and personal interruptions occurred while the workers were not under the test. Heart rate was recorded for each time element and for each task in accordance to the time study. Rest Table 3 The time study element-break down Work element
Description
Pull out
Pulling out the 630 g/m line until the logs location
Hook
Hooking the logs prepared on the slopes
Walk in
Follow the loads back to the landing
Unhook
Unhook the chokers from the logs
99
G.O. Aalmo et al.
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
Fig. 2 Box plot showing the heart rate response for each time element, direction and subject heart rate was obtained for each subject upon arrival at the work site. Subjects were asked to sit down and rest for 15 minutes. Between each test, the subjects rested for at least 45 minutes. The meteorological conditions for the 5 days of data collection were similar with a mean temperature of 18°C and a mean humidity of 51%. All data were analysed using R statistical software (Team 2008). The heart rate was analysed using mixed effect modes with the nlme package (Pinheiro et al. 2012). The use of mixed effect models allows for modelling the dependent variable giving different inter-
100
cepts and slopes for each subject participating in the experiment (Bates 2005). Parametric statistics response variable is assumed to be linear in time. The analysis was conducted in two phases. The first one aimed at building a mixed effect model for predicting the heart rate for each of the work elements, and the second phase in which the subject specific deviation from the intercept β0 for each model (μ) was tested for each of the anthropometric parameters to visually check any trend explaining how the heart rate is influenced by individual characteristics. The most Croat. j. for. eng. 37(2016)1
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
G.O. Aalmo et al.
parsimonious model was selected out of the various models tested in the building process. The model was chosen on the basis of the Akaike Information Criterion (AIC).
In order to investigate whether the heart rate of each time element is affected by anthropometric parameters typical for each subject in the test, a visual inspection was performed of plots explaining the relationship between µi and different parameters. The first model (mod. 1) predicted the heart rate of The chosen anthropometric parameters are both the subjects during the time elements requiring walkdirectly measured or calculated and they indicate the ing i.e. »pull out« and »walk in« (Table 3) the slope and physical size i.e. weight (kg), the degree of fitness i.e. the best fitting model was found to be: VO2max (mL/(kg×min)) calculated by means of the heart Yij = b0 + b1Slope + b2Speed × Uphill + b3Speed × Downhill + mi rate + eij monitor and the age (years) of the subjects.
lope + b2Speed × Uphill + b3Speed × Downhill + mi + eij
(Model 1)
Where: Yij
rediction of the heart rate at the end of the p time element for each subject (i) and replication (j);
β0
intercept;
β1,β2,β3 unknown parameters to be determined; Slope slope percentage of each test corridor, %; Speed speed at which the subjects were moving on the slopes, m/s; Uphill and downhill dummy variables indicating whether the extraction proceeds uphill or downhill, respectively; µi
random factor;
εij
individual specific error.
For the time elements not requiring walking i.e. hook and unhook the chokers (Table 3), the model selected was:
3. Results A preliminary data analysis showed that there was a significant difference (p<0,05) in heart rate readings for time elements pull-out and unhook for different winching directions, uphill winching to the left, and downhill winching to the right of each plot. For the hook and walk-in elements the difference was less significant (p<0.1) (Fig. 2). The summary statistics for the variables used in the test are presented in Table 4.
Table 4 Summary statistics of the data Variable
Where: Zij
rediction of the heart rate at the end of the p time element for each subject (i) and replication (j); β0 intercept; β1,β2,β3 are the unknown parameters to be determined; Slope slope percentage of each test corridor, %; Time time to perform the task, s; Uphill and downhill dummy variables indicating whether the extraction proceeds uphill or downhill, respectively; µi
random factor;
εij
individual specific error.
Croat. j. for. eng. 37(2016)1
Mean
SD
Count
Heart rate, pull out beats/min
109
173
140.60 21.40
–
Heart rate, hook beats/min
97
160
133.7
15.8
–
Heart rate, walk beats/min
100
164
135.50 17.95
–
Heart rate, unhook beats/min
97
159
134.43 15.36
–
Slope, %
53
68
60.13
4.72
–
0.11
1.09
0.50
0.29
–
Time, hook, s
28
313
85.73
49.91
–
Time, unhook, s
21
93
48.73
14.70
–
Uphill Downhill
–
–
–
–
30 30
23.6
31.4
27.7
2.65
–
VO2max mL/(kg×min)
32
60
41.6
9.41
–
Age, years
22
52
40.8
10.11
–
Mean
SD
Count
Zij = b0 + b1Slope + b2Time × Uphill + b3Time × Downhill + mi + e ij = b0 + b1Slope + b2Time × Uphill + b3Time × Downhill + mi + e ij (Model 2)
Minimum Maximum
Speed, m/s
BMI, kg/m2
Variable
Minimum Maximum
101
G.O. Aalmo et al.
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
Table 5 Estimates and standard errors for the model including movement along the slope Variables
Coefficient
Standard error
t
Coefficient
Standard error
t
P-val
Model »hook« (2a)
Constant
b0
128.07
20.24
6.32
<0.001
Slope
b1
0.28
0.28
1.01
0.31
Speed×uphill
b2
–30.97
7.18
–4.31
Speed×downhill
b3
55.27
20.61
2.01
1.42
Subject variance
Constant
b0
150.77
21.73
6.93
<0.001
Slope
b1
–0.19
0.35
–0.53
0.59
<0.001
Time×uphill
b2
–0.19
0.04
–3.94 <0.001
2.68
<0.001
Time×downhill
b3
0.03
0.03
0.97
0.33
–
–
Subject variance
13.16
3.62
–
–
Model »walk in« (1b)
Model »unhook« (2b)
Constant
b0
125.66
19.49
6.45
<0.001
Slope
b1
0.32
0.26
1.23
Speed×uphill
b2
6.77
7.47
Speed×downhill
b3
–100.19 14.90
Constant
b0
73.13
15.58
4.69
<0.001
0.22
Slope
b1
0.87
0.25
3.49
<0.001
0.96
0.36
Time×uphill
b2
0.37
0.07
4.76
<0.001
21.43
–4.67
<0.001
Time×downhill
b3
–0.06
0.09
–0.69
0.49
3.86
–
–
Subject variance
23.78
4.86
–
–
The parameters of the mixed effect models for the heart rate in the »pull out« element and the »walk in« element (Model 1) are reported in Table 5. As expected, the slope is positive in both models; in the first model (1a) the interaction between speed and direction is negative with the uphill extraction (i.e. pulling the cable downhill), while it is positive correlated to the heart rate when the extraction is downhill (i.e. pulling the cable uphill); the opposite occurs in the model 1b, where the subjects are walking without any loads. Model 1b shows a higher variance in the subjects, meaning that when the subject is not undergoing very heavy workload i.e. pulling the cable uphill, the anthropometric parameters are different in terms of the heart rate. The parameters of the mixed effect models for the heart rate in the elements of »hooking« and »unhooking« (Model 2) are reported in Table 6. In the model 2a, the intercept is quite high and this is in line with the fact that when the subjects perform the hooking task they are under a severe strain in the pull out phase. The time to perform the hooking is not sufficient for recovering (avg=86 sec). The predicted heart rate is lower when the subjects move downhill. The model 2b shows that the HR prediction is lower and this follows the rules of the physiology of recovery. During the walk-in phase, the effort for moving
102
Variables
P-val
Model »pull out« (1a)
Subject variance
Table 6 Estimates and standard errors for the model including the »static« time elements
uphill corresponds to the strain to move the body mass uphill when performing uphill extraction, while in downhill extraction, walking is down-hill. In this way the body has the time to recover and to take the HR back closer to normal activity values. The specific intercept of the HR for each time element and each subject has been extracted from the fitted model and plotted against the anthropometric parameters, age, BMI and VO2max (Fig. 3). The visual inspection of the plots shows a potential relationship between the HR for each time element and both the degree of fitness (VO2max) and the age, while there was no evidence for the effect of BMI. The subject with a higher VO2max i.e. with the best fitness status is the one showing a lower intercept of the model. The same subject is also the youngest giving a lower intercept also in the plots checking the age influence. Also, in this case there is a potential trend in the phases requiring a higher physical effort i.e. pull out and walk in while the trend is not evident in the other time elements.
4. Discussion and Conclusions This study demonstrated that, when implemented on steep terrain, winching is a very heavy task, which becomes increasingly harder as the slope increases. Croat. j. for. eng. 37(2016)1
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97â&#x20AC;&#x201C;105)
G.O. Aalmo et al.
Fig. 3 Plotting the specific intercept for each subject (m) against the subjectsâ&#x20AC;&#x2122; physical parameters
Task demand can be classified according to the severity of workload. Generally, as workload increases, heart rate increases (Miller 2001). The results of our research showed that winching tasks fall in the heavy, very-heavy ranges. Depending on winching direction, the highest effort would occur during cable pull out (if winching downhill), or during walk in (if winching uphill). The effect of cable weight during the pull-out phase is visible in both cases, but it is obviously much stronger when winching downhill (i.e. pulling the cable uphill). Our results show that winching downhill is a much harder task than winching uphill, because the operators have to pull out the cable in an uphill direction in order to reach the logs. Furthermore, downhill winching is heavy on the cable itself, because the load can slide during the drag, causing cable tension to drop and favouring cable nesting into the drum. Croat. j. for. eng. 37(2016)1
Therefore, winching operations should be planned in such a way to avoid downhill winching whenever possible for ergonomic and safety reasons. This study indicates that the effect of slope dominates over the effect of cable weight when the average HR readings for pulling the cable uphill and moving uphill with no load are very close in range, 160.8 and 150.6, respectively. This was confirmed by previous work conducted by Ottaviani et al. (2011) for cable yarder operations, and by Magagnotti and Spinelli (2012) for winches equipped with synthetic cable. Even the effect of speed is mediated by slope, as shown by the fact that increasing pull-out time (i.e. lower speed) will decrease work load when pulling the cable uphill, whereas it will increase work load when pulling the cable downhill. That means that the operator can make gravity work to his own benefit, and use it for faster downhill movements, which will allow re-
103
G.O. Aalmo et al.
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
ducing the time under effort without increasing effort level. Of course, we are not suggesting that workers should rush their downhill walk, as that may involve an increased risk of tripping and may be hazardous. We are just observing a fact and using it to support our conclusion that the slope is the most important factor contributing to the workload experienced by steep terrain workers. Reducing tool weight is not enough, since the operator’s own bodyweight is already difficult to move uphill, on a steep slope. Tool weight reduction may make possible an otherwise impossible task, but it will not turn a heavy job into a light one. As the heart rate is an important indicator for stress levels, adoption of lighter cables can palliate the problem, but not solve it. Other measures aimed at relieving the operators workload could be found in the use of special harness for increasing walking efficiency and decreasing walking effort (i.e exo-skeleton) or dedicated cable recovery devices for pulling out the cable all the way to the hooking site, without an operator walking back and forth all the time. This latter solution is offered by mini-yarders (Spinelli et al. 2010), tong-throwers (Bruce 2009) or the newer cable recovery equipment recently developed in Central Italy. Considering the results of the relationship between the anthropometric parameters and the heart rate predictions for each subject, the study did disclose a significant effect, but it could not solve all the doubts caused by conflicting evidence in the available literature. It was found that the workload was inversely proportional to fitness and directly proportional to age. That could be expected, but had not yet been specifically demonstrated for the case of log winching. Winching is less taxing for younger, fit workers than for older workers, especially if their fitness level (as expressed by the Polar OwnIndex®) is low. A typical winching cycle is constituted by relatively short phases (i.e. time elements): this pattern has great influence on the physiology of recovery which should take place during the less effort-intensive phases. As recovery time is directly proportional to age (Darr et al. 1988), employing crews with older members might impose the need for frequent rest, which will affect productivity. Since winching is a strenuous and intense job, any winching operation would be more efficient and safer if implemented by crews composed mainly by fit young workers, while assigning older or less fit workers to other tasks requiring a lighter workload. In fact, this is common practice in commercial operations, where choker tenders are generally selected among the youngest and fittest subjects. In general, operator
104
selection and careful work planning (i.e. job rotation, crew composition) can palliate the strain placed on the workforce by winching. Further progress in that direction requires adopting radical innovation, going beyond the simple reduction of cable weight and drum friction. Additional research in this field would be necessary, since this study was carried out on a limited number of subjects.
Acknowledgments The authors wish to thank Mr. Sergio Poli and RSAF Lombardia for making available the locations E and the team to perform the study, Maxwald for lending the equipment, and the team members for their patience.
5. References Afolabi, B.O., Akanbi, O.G., 2013: Effects of body mass index on aerobic power (VO2max) and energy expenditure (EE): A case of manual load lifting in agro-processing. IJSER 4(5): 1718–1721. Apud, E., Bostrand, L., Mobbs, I.D., Strehlke, B., 1989: Guidelines on ergonomic study in forestry. ILO, Geneva. Astrand, P.O., Rodahl, K., Dahl, H., Strømme, S.B., 2003: Textbook of work physiology: physiological bases of exercise. New York, McGraw-Hill, 656 p. Bates, D., 2005: Fitting linear mixed models in R. R news 5(1): 27–30. Darr, K.C., Bassett, D.R., Morgan, B.J., Thomas, D.P., 1988: Effects of age and training status on heart rate recovery after peak exercise. Am J Physiol Heart Circ Physiol 254(2): H340– H343. de Zwart, B.C., Frings-Dresen, M.H., van Dijk, F.J., 1996: Physical workload and the ageing worker: a review of the literature. Int Arch Occup Environ Health 68(1): 1–12. Hagen, K., Vik, T., Myhr, N., Opsahl, P., Harms-Ringdahl, K., 1993: Physical workload, perceived exertion, and output of cut wood as related to age in motor-manual cutting. Ergonomics 36(5): 479–488. Hart, S.G., 1986: Theory and measurement of human workload. Praeger, New York. Human Productivity Enhancement Vol 1: Training and Human Factors in Systems Design: 396– 456. Heuch, I., Hagen, K., Heuch, I., Nygaard, Ø., Zwart J.A., 2010: The impact of body mass index on the prevalence of low back pain: the HUNT study. Spine 35(7): 764–768. Keyserling, W.M., Punnett, L., Fine, L.J., 1988: Trunk posture and back pain: identification and control of occupational risk factors. Applied Industrial Hygiene 3(3): 87–92. Krause, N., Ragland, D.R., Greiner, B.A., Fisher, J.M., Holman, B.L., Selvin, S., 1997: Physical workload and ergonomCroat. j. for. eng. 37(2016)1
Forest Workers and Steep Terrain Winching: the Impact of Environmental ... (97–105)
G.O. Aalmo et al.
ic factors associated with prevalence of back and neck pain in urban transit operators. Spine 22(18): 2117–2126.
Polar, 2015: Retrieved 06/03/2015, from http://www.polar. com/en/ support/polar_fitness_test_and_ownindex
Lilley, R., Feyer, A.M., Kirk, P., Gander, P., 2002: A survey of forest workers in New Zealand: Do hours of work, rest, and recovery play a role in accidents and injury? J Safety Res 33(1): 53–71.
Rodahl, K., 2003: Physiology of Work. Taylor and Francis, Inc., Bristol, PA, USA, 266 p.
Magagnotti, N., Spinelli, R., 2012: Replacing Steel Cable with Synthetic Rope to Reduce Operator Workload During Log Winching Operations. Small-scale Forestry 11(2): 223–236. Miller, S., 2001: Workload Measures Document id: N01-006, National Advanced driving Simulator, The University of IOWA, 65 p. Ottaviani, G., Talbot, B., Nitteberg, M., Stampfer, K., 2011: Workload benefits of using a synthetic rope strawline in cable yarder rigging in Norway. Croatian Journal of Forest Engineering 32(2): 561–569. Ovaskainen, H., Uusitalo, J., Väätäinen, K., 2004: Characteristics and significance of a harvester operators’ working technique in thinnings. IJFE 15(2): 67–77. Park, H., Lee, J., Choi, Y., Kim, M., 2003: Evaluation of work intensity by moving work in forest. J Korean For Soc, 92, http:// agris.fao.org/agris-search/search.do?recordID= KR2004005048. Parker, R., Kirk, P., 1994: Physiological workload of forest work. LIRO Report (New Zealand Logging Industry Research Organisation) 19(4): 1–49. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., Team, R.C., 2012: Linear and nonlinear mixed effects models R package version 3.1–104.
Rodgers, S., 1997: Work physiology–fatigue and recovery. Handbook of human factors and ergonomics. 2nd ed. John Wiley and Sons, New York, 268–298. Spinelli, R., Magagnotti, N., 2012: Wood extraction with farm tractor and sulky: estimating productivity, cost and energy consumption. Small-scale Forestry 11(1): 73–85. Staal Wästerlund, D., 2001: Heat stress in forestry work. Acta Universitatis Agriculturae Sueciae, Silvestria 213: 1401–6230. Stampfer, K., 1998: Stress and strain effects of forest work in steep terrain. Proceedings of the IUFRO/FAO Seminar on Forest Operations in Himalayan Forests with Special Consideration of Ergonomic and Socio-Economic Problems in Thimphu/Bhutan. (Heinimann, H.R. and Sessions, J. editors). http://www.upress.de/abstracts/3-933146-12-7.html. Kassel University Press, Kassel: 113–119. Team, R.D.C., 2008: R: A Language and Environment for Statistical Computing. Austria, Vienna, ISBN 3-900051-07-0. Available at http://www. R-project. org, 2008. Trites, D.G., 1992: Ergonomics of tree planting work among British Columbia forest workers. Theses (School of Kinesiology)/Simon Fraser University, 96 p. Vik, T., 1984: Impact of terrain on human effort in forest operations. Paper presented at the Human Resources in Logging. Proceedings of a seminar, Rotorua, New Zealand, 150–155.
Authors’ address: Giovanna Ottaviani Aalmo, PhD.* e-mail: giovanna.ottaviani.aalmo@nibio.no Norwegian Institute of Bioeconomy Research Pb 115, NO-1431 Ås NORWAY
Received: January 1, 2015 Accepted: June 16, 2015 Croat. j. for. eng. 37(2016)1
Natascia Magagnotti, PhD. e-mail: magagnotti@ivalsa.cnr.it Raffaele Spinelli, PhD. e-mail: spinelli@ivalsa.cnr.it CNR Timber and Tree Institute Via Madonna del Piano, Pal. F 50019 Sesto Fiorentino, Firenze ITALY * Corresponding author
105
Original scientific paper
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type Carola Häggström, Mikael Öhman, Lage Burström, Tomas Nordfjell, Ola Lindroos Abstract Exposure to whole body vibration (WBV) is a major concern in mechanized forestry work because its adverse effects may become exacerbated by repetitive hand and arm movements, and nonneutral body postures. Moreover, shock-type vibrations have recently been suggested as a possible agent behind pains in the neck and shoulders of forest machine operators. Shocks have been identified in forwarders during loading, but the effects of crane work in forwarders have, to the best of our knowledge, not been studied. Thus, the aim of this study was to assess contributions of crane work elements, and potential effects of the use of three grapple and brake-link combinations, to vibration exposure levels in a small forwarder. Repeated measurements of cabin WBV were acquired, and work elements timed, as a single experienced operator forwarded wood piles on a standardized track in northern Sweden, using the same forwarder and work procedures with each grapple and brake-link combination. The studied crane equipment was found to have little or no effect on the daily WBV exposure. Furthermore, exposure to shock-type vibrations while loading appears to be due to driving rather than crane work. However, there were fewer collisions with remaining trees while using the tilt grapple with brake link, suggesting its use provides a more relaxed and comfortable work environment for forwarder operators and financial benefits for the forest owner by reducing damage in the remaining stand. Keywords: crane work, forestry, forest machine, seated health, whole body vibration, work elements, work environment
1. Introduction Whole body vibration (WBV) is related to numerous health problems, inter alia various musculoskeletal, digestive and reproductive disorders, low back pain (Seidel and Heide 1986, Bovenzi and Hulshof 1999, Punnett and Wegman 2004, Burström et al. 2014), and more instant effects including motion sickness, sight impairment and fatigue (ISO 1997). In addition to health effects, WBV has been shown to impair performance (Conway et al. 2006), especially in accuracy based tasks, which are typical for crane work during forestry operations. It is important to restrict vibration exposure and monitor its effects on those exposed since a dose response relationship is yet to be established (Pope et al. 2002). Thus, for instance, EU Directive 2002/44/EC restricts daily exposure normalised to an eight-hour reference period, designated A(8), to 1.15 m/s2 or a fourth power vibration dose value (VDV) of 21 m/s1.75, and stipulates that measures should be Croat. j. for. eng. 37(2016)1
taken to reduce the impact of WBV if the A(8) value exceeds 0.5 m/s2 or VDV exceeds 9.1 m/s1.75. A more general guideline is to always minimize occupational vibration exposure (Burström et al. 2014). WBV is a major concern in mechanized forestry work since its adverse effects are exacerbated by repetitive hand and arm movements, non-neutral body postures, and manual lifting (Punnett and Wegman 2004, Okunribido et al. 2006, Lis et al. 2007, Burström et al. 2014). Operators of forest machines have a high prevalence of musculoskeletal symptoms in the lower back, neck and shoulders (Rehn et al. 2002, Jack and Oliver 2008), which may be at least partly linked to their WBV exposure, although the association between WBV exposure and neck and arm pain has not been clearly established (Rehn et al. 2009). However, it is suggested that the high prevalence of neck pain among forest machine operators is associated with exposure to shock-type vibration (Rehn et al. 2009). However,
107
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
shocks may also be more important than sinusoidal vibration with regards to low back pains (Okunribido et al. 2006). Hence, reducing WBV should improve the work environment in forestry. Furthermore, reducing vibrations may also reduce machine wear and damage to the ground (Rieppo et al. 2002). Due in large part to the ergonomic problems, numerous aspects of WBV in forestry work have been intensively researched. These aspects include effects of dampening systems for forestry vehicles (Gellerstedt 1998, Sherwin et al. 2004, Baes 2008) and both chairs and cushioning (Boileau and Rakheja 1990, Sankar and Afonso 1993, Mansfield et al. 2002, Cation et al. 2008, Ji et al. 2015). Vibrations associated with different forest machines and machine systems have also been examined (Rehn et al. 2005b, Gerasimov and Sokolov 2009), and attempts have been made to establish dose response relationships (Rehn et al. 2009), and standardize measurement techniques (Rehn et al. 2005a, Burström et al. 2006). During work studies, forwarding is normally divided into the work elements (WEs) driving (empty or loaded), loading and unloading. Vibration exposure during these WEs has been evaluated, driving has been identified as the major source of WBV, and the operator is exposed to higher vibration levels when driving empty than when driving loaded (Hansson 1990, Rehn et al. 2005a). One of few studies reporting both r.m.s and shock sensitive VDV values found exposure to shock-type vibrations to be common while loading, but the shocks are believed to mainly originate from simultaneous driving between piles in uneven terrain (Rehn et al. 2005a). However, to our knowledge, no previous studies have examined WBV exposure levels in sufficient detail to evaluate exposure during crane WEs. Furthermore, most previous studies have focused on vibrations associated with large forest machines (10 to 20 tonnes), which are almost exclusively used in industrial applications. Thus, there is a lack of information on WBV in small forwarders (lighter than 4 tonnes), which are used by both professionals and self-employed non-industrial private forest owners (cf. Nordfjell et al. 2003, Lindroos et al. 2005). There are serious concerns about both of these groups. Professionals continuously use the machines when working (cf. Passicot and Murphy 2013), so they are highly sensitive to variations in the machine design, while the latter are occasional users who are heavily represented in accident statistics, but difficult to inform about preventive actions (Lindroos and Burström 2010). Thus, for both groups it seems highly important to identify and implement modifications that minimize vibrations.
108
Vibration exposure in vehicles may be affected by not only driver seats and dampening systems, but also working techniques, which are influenced by the operators’ experience and equipment. For example, vibration exposure of professional taxi drivers and train operators reportedly declines as their work experience increases (Chen et al. 2003), and forwarder operators’ working techniques reportedly influence WBV levels while driving loaded (Rehn et al. 2005a). Brake links and tilt grapples are equipment that may alter work techniques during forwarder crane work. Brake-links, placed between the crane tip and grapple (or other tool) are common equipment on large forest machines and help to increase precision by reducing swinging movements of the grapple. Standard brake-links are static, but an active brake-link that can be used to brake when desired has potential capacity to further increase the precision of movements. A tilt grapple provides not only the features of an active brake-link but also the possibility of precisely tilting the grapple and the gripped logs. The use of tilt grapples has been found to increase productivity in forwarding, as well as reducing damage to stands (Fogdestam 2010, Nilsson 2013), but there have been no detailed studies on their effects on vibrations. Thus, there are gaps of information on WBV exposure in small forwarders, the variation between WEs and the effect of crane equipment. Therefore, the aim of this study was to assess contributions of specific crane WEs to the overall vibration exposure in small forwarders and possible effects of three grapple and brake-link combinations on the WBV exposure.
2. Material and Methods 2.1 Experimental design Repeated field measurements of cabin WBV in a forwarder were acquired, while a single operator was forwarding standardized wood piles on a standardized track, using three types of crane equipment. Each monitored work cycle (observation) corresponded to one round on the standardized track, beginning with loading the empty bunk and ending when the last log was unloaded. Through time studies, each work cycle was split into WEs and WBV were analyzed within and over WEs. Thus, the design consisted of two fixed factors (Crane Equipment and WE) within sets of repetitions (blocks). In total there were five blocks. The three types of crane equipment were randomly assigned within blocks to minimize possibilities of order and carry-over effects confounding the results. The field study was conducted during October 7–22, 2013, with one trial (work cycle) per day for the Croat. j. for. eng. 37(2016)1
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
C. Häggström et al.
Fig. 1 Scaled map of the 114 m long standardized track. Positions and sizes (numbers of logs) in the even (e) and uneven (ue) piles are marked by gray lines, trees in the stand by gray dots and the landing area by rectangle first block, and subsequently two trials per day. One designated researcher filmed all the trials and made all the measurements. During the study the temperature was circa 0°C.
2.2 The standardized track The study was conducted in a forest stand in the northern part of Sweden that was selected to represent a typical dense stand that had just been subjected to a first thinning (see e.g. Eriksson and Lindroos (2014) for typical Swedish conditions). The stand contained only Croat. j. for. eng. 37(2016)1
Scots pine (Pinus sylvestris L.) trees, with a basal area weighted mean age of 46 years. The stand density, basal area at breast height and mean tree volume were 1370 stems per ha, 21 m2/ha and 0.1 m3 of solid wood over bark (m3sob), respectively. The ground was extremely flat, sandy and had good carrying capacity (class 1–1–1 according to »Terrain classification for forestry work« Berg 1992). Thus, it was suitable for the tests since risks of confounding the measures of crane equipment induced vibration with vibrations due to terrain structure were minimal despite possibilities of
109
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
Fig. 2 Vimek grapples and brake-links simultaneous crane work and driving. The stand contained a 144 m long roughly circular track (a 2.9 m wide strip road), along which 95 pine pulpwood logs were distributed into 38 piles in a standardized manner for each trial, with a spur leading to a landing (Fig. 1). The volume of the logs was equivalent to one full load for the studied forwarder (3.3 m3sob). The mean length and top diameter of the logs were 4.4 with a standard deviation (SD) of 0.3 and 0.078 (SD=0.014) m, respectively. The mean wood density, based on a sample of five logs, was 997 kg/m3sob. The same logs were used, and the number, positions and sizes (i.e. numbers of logs) of the piles were kept constant during the trials. However, given logs were not always placed in given piles, thus the volumes of the piles may have varied slightly between repetitions. Each pile contained 1–5 logs, and could be handled with one grip of the grapple. The center of each pile was placed at a fixed distance between 1.5 and 5.0 m from the center of the strip road. Twenty-one piles were placed to the outer side of the circular road and slightly fewer (17) to the inner side (due to spatial limitations). The piles were always placed at the same angle with respect to the strip road. In each of the 38 piles, all butt ends were oriented in the same direction. For 20 of the piles, logs were placed so that the
110
butt end surfaces were vertically level with each other (even piles), while for the other 18 piles their vertical positions were varied by up to 0.7–1 m (uneven piles). During each trial, the loading started at the beginning of the track (so there was no driving empty), and the last 36 m of the track was driven with a full load (so there was 108 m of driving while loading). At the landing, logs were unloaded onto a pre-marked area for roadside-piles.
2.3 Base machine and crane equipment A standard 3.5 tonnes Vimek 608.2 BioCombi forwarder was used in the study (Vimek AB, Vindeln, Sweden). The forwarder was equipped with a standard crane with a reach of 5.2 m and a lifting torque of about 20 kNm. The three studied types of crane equipment (grapple and brake-link combinations) were: a Vimek tilt grapple with a Vimek dynamic brake-link (braked tilt grapple, Fig. 2a); a Vimek standard grapple with Vimek dynamic brake-link (brake-link grapple, Fig. 2b); and a Vimek standard grapple with no brakelink (standard grapple, Fig. 2c). The gripping area was the same for all grapples (0.16 m2). The tilting capacity of the braked tilt grapple was 1.3 kNm. The weights of the braked tilt grapple, brake-link grapple and stanCroat. j. for. eng. 37(2016)1
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
dard grapple were 127, 101 and 81 kg, respectively. Thus, the mass of a single pile was not a limiting factor for the crane or any of the studied grapples.
2.4 Operator and work instructions In order to avoid errors between subject variations (Lindroos 2010, Häggström et al. 2015)����������������� , a single operator with previous experience of forestry time studies operated the forwarder throughout the study. The operator was male, 68 years old, familiar with the forwarder used in the study and had 30 years of experience in forwarding. Before the study, he had experience with all the studied types of grapples and brake-links, but little experience with the braked tilt grapple. The operator first had a training session of one work cycle with each of the grapple and brake-link combinations, during which he was instructed to find a preferred working method for all three crane equipment types. He was then instructed to use the selected work patterns throughout the study. The end surfaces of gripped logs were to be aligned before loading only when the operator considered it necessary. When aligning end surfaces, the operator was instructed to do it against the headboard with the standard and brake-link grapples and vertically against the ground with the braked tilt grapple. Between trials, the operator had the chance to get to know the equipment to be used in the following trial while re-arranging the logs along the track. 2.5 Time study A LEGRIA HF S200 high definition video camera (Canon Inc., Tokyo, Japan) was used to record the work in each trial. ProTime Estimation software (Pro-
C. Häggström et al.
planner, Ames, USA) was then used to measure times of the seven defined, non-overlapping work elements (WE) described in Table 1. Collisions between the grapple or lifted logs, and trees or the base machine were also counted.
2.6 Vibration measurements Vibrations were measured in three orthogonal axes according to ISO 2631-1 (ISO 1997) using a MTi-G triaxial accelerometer (Xsens, Enschede, The Netherlands) placed on the floor close to the center of the cabin, in front of the chair. The placement ensured that the operator’s weight, height and the chair dampening would not affect the measures. Samples were taken at a frequency of 100 Hz during each, approximately 40 minute long, work cycle, using a XKF Scenario »2.7 Automotive unit« (Xsens, Enschede, The Netherlands). The measuring equipment was checked using a Brüel & Kjær 4294 calibrator after the measurements.
2.7 Data analyses All data processing was performed offline using a commercial software package (MATLAB R2014a 8.3, The MathWorks Inc., Natick, USA) with the »Continuous Sound and Vibration Analysis« program (Zechmann 2013). The acceleration data were converted from the recorded time domain to frequency domain with a frequency range up to 50 Hz, i.e. the maximum frequency range that can be calculated from 100 Hz output. In the analyses, 1/3 octave band values were calculated from 0.1 to 50 Hz. The resulting data were then used to calculate frequency weighted r.m.s. acceleration and VDV values with respect to health effects on a seated driver in accordance with ISO 2631-1
Table 1 Definitions of time study work elements Work element Crane out1
Definition
Priority
Begins when the crane starts moving towards a pile on the ground and stops when grip begins
1
Begins when the grapple is placed against the pile and stops when all logs are gripped and crane in begins
1
Crane in
Begins when the grapple is loaded and the crane starts moving towards the bunk and stops with release
1
Release & reorganise1
The sum of release (which begins when the grapple is inside the supports above the bunk and ends when no log has contact with the grapple) and reorganise (the time the operator spends reorganizing logs on the bunk)
1
Unloading1
Begins when the crane starts moving for unloading on the roadside landing and stops when all logs are unloaded
1
Driving
Begins when the forwarder wheels start to move without the crane being active and stops when the wheels stop or crane movements are initiated, whichever comes first
2
Other working time
All time that is not covered by any of the definitions above, including disruptions
3
1
Grip
1
Note: If multiple work elements were performed simultaneously, time consumption was recorded for the work element with the highest order of priority (lowest number) 1 The WE crane work used in the analysis includes all the crane activities pooled
Croat. j. for. eng. 37(2016)1
111
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
Table 2 Assumptions made during calculation of the time distribution for each work element (WE) during one full day (8 hours) of work with the forwarder Parameter Daily work hours, h
Value 8
Technical utility1, %
Proportion of crane work during loading3, %
Type of work
Work element 1
Contribution, %
Crane work
33–59
0.9
Driving
41–67
50–90
Crane in
7–15
Crane out
3–6
Grip
2–5
88–100
Conversion constant PMh15 to PMh02
Table 3 Ranges of the work elements’ estimated contributions (%) to the total daily WBV dose during the studied forwarding operations, based on the average and maximal measured awz and time distributions presented in Table 2
Work cycle
Work cycle (including driving empty), % Crane activity
Proportion of loading4
45
Proportion of unloading4
15
Release & Reorganise
7–14
Proportion of driving empty4
24
Unloading
15–20
Proportion of driving loaded4
16
Crane Activity (without unloading), % Proportion of Release & Reorganise5 5
Proportion of grip
32 14
Proportion of crane in5
34 5
Proportion of crane out
19
1
Based on Nordfjell et al. (2010) Based on unpublished material, Skogforsk 3 Based on Manner et al. (2013) 4 Based on Rehn et al. (2005a) 5 Based on observed time distribution in the present study 2
(ISO 1997). The weighted r.m.s. values were calculated with respect to all three orthogonal axes, awx (back and forth), awy (lateral) and awz (vertical), and their sum vector (av). Similarly, VDV was calculated for all three orthogonal axes (VDVz, VDVy and VDVz) and the vector value (VDVv) over each measurement period. Furthermore, crest factors were calculated for all orthogonal axes as well as the 8 hour equivalent, A(8), value over each measurement period.
2.8 Statistical analysis Data were analyzed using Minitab 16 (Minitab Ltd, State College, PA, USA). Analysis of Variance (ANOVA) was used to analyze the fixed effects of WE and Crane Equipment type, and the fixed interaction between them, on the vibration measures. The ANOVA models also included the random block effect. A general linear model (GLM) was applied when analyzing the ANOVA models, and Tukey’s Honest Significant Difference (HSD) test of means was used for pairwise comparisons.
112
For WE, two sets of treatment were analyzed. The first set (denoted Work Cycle) included two levels (crane work, i.e. the sum of all crane WEs, versus driving,) and the second set (denoted Crane Activity) included five levels (release & reorganize, grip, crane in, crane out and unloading,). Other working time was excluded from analyses. The same levels of Crane Equipment were used in both analyses. In a third set of analyses, effects of Crane Equipment were analyzed using a single pooled WE (crane work). Alignments of end surfaces, collisions with residual trees and collisions with the machine were included as covariates to investigate the relationships between vibration measures and collisions. In a fourth set of analyses, effects of Crane Equipment type on the number of collisions, and alignments, were analyzed with a GLM including block as a random factor. ANOVA assumptions of independence, homoscedasticity and normality of residuals were not sufficiently violated to require transformation of the data, according to ocular inspection of residual plots. In all analyses, the significance level was set to 5%.
3. Results Each of the 15 observations (five repeated trials with each of the three equipment types) lasted about 40 minutes, providing 9 hours and 22 minutes of recordings in total. Of that time, 5% was classified as other working time with a mean duration per observation of 105 (SD=83) s, which was excluded from further analysis. So, the average duration of work cycles was 2122 (SD=152), 2077 (SD=101) and 2229 (SD=65) s, respectively, for operations with the braked tilt grapple, break-link grapple and standard grapple. Missing Croat. j. for. eng. 37(2016)1
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
C. Häggström et al.
Fig. 3 Mean values of vibrations measured at the floor of the forwarder as a function of frequency (1/3-octave bands) for each indicated work element (WE) during driving and crane work (means of 15 observations, i.e. pooled data for trials with all types of crane equipment)
Fig. 4 Mean values of vibrations measured at the floor of the forwarder as a function of frequency (1/3-octave bands) for the pooled crane work in all three directions (ax, ay, az) and the sum vector (av) (mean of 15 observations, i.e. pooled data for trials with all types of crane equipment)
data for 4, 10 and 4% of the work cycles with the respective grapples were not included in the calculation of the total time.
Based on the assumptions in Table 2, the daily total vibration exposure dose was on average 0.3 m/s2 and the estimated maximum dose was 0.38 m/s2. Gener-
Table 4 Frequency weighted acceleration in the three orthogonal axes (x, y and z), the sum vector (v) and the A(8) value for indicated work elements (based on pooled data for trials with all types of crane equipment) according to »health« in ISO 2631-1. Measurements were taken at the feet Duration Type of work
WE
N
1
Work cycle
Crane activity
Crane work Driving
s
awx
Mean
A(8)
av
VDVv
SD
Mean
SD
Mean
Mean
Mean
Mean
SD
0.14
0.010 3.06
0.37
Crane out
15
262–309
0.11c 0.01 0.16c 0.02 0.28d 0.07 0.34d 0.05 0.028c 0.007 2.20b
0.35
Grip
15
157–313
0.11
a
0.01 0.22
b
0.02 0.19
0.02 0.30
ab
0.01 0.34
c
0.07 0.39
ab
0.07 0.43
0.06 0.042
a
380–638
cd
0.08 0.42
a
0.019 4.24
15
a
0.02 0.32
bc
0.06 0.079
Crane in
c
0.01 0.23
bc
0.08 0.41
A
SD
0.37
a
0.01 0.32
B
SD
0.31A 0.03 0.33A 0.03 0.35A 0.04 0.57A 0.05 0.042B 0.003 4.42A
0.01 0.21
B
SD
0.37
b
B
m/s1,75
A
15 1537–1887 0.15 335–502
awz m/s2
B
15
awy
c
0.06 0.026
a
0.05 0.043
b
0.29
a
0.005 2.21
Release & Reorganize
15
343–628
0.17
0.010 3.01
0.28
Unloading
15
256–345
0.16a 0.02 0.23a 0.03 0.34a 0.09 0.45a 0.08 0.035b 0.009 3.02a
0.44
Note: Mean values within columns and type of work with different superscript letters (A–B for the full work cycle and a–d for crane activities) are significantly different (p<0.05, Turkey’s HSD). WE = Work Element; SD = Standard Deviation 1 Crane Work includes all crane activities pooled
Croat. j. for. eng. 37(2016)1
113
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
Table 5 Frequency weighted acceleration in the three orthogonal axes (x, y and z), the sum vector (v) and the A(8) value for crane work with the indicated crane equipment types according to »health« in ISO 2631-1. Measurements were taken at the feet Duration Crane equipment
N
s
awx
awy
awz
A(8)
av
VDVx
VDVy
10-2m/s2
VDVz
VDVv
10-2m/s1,75
M
SD
M
SD
M
SD
M
SD
M
SD
M
SD
M
SD
M
SD
M
SD
Standard
5
1784–1887
14
1
20
2
32
8
41
7
8.2
2.2
171
15
241
32
301
56
424
54
Brakelink
5
1565–1781
15
1
21
1
33
9
42
6
7.9
2.1
189
16
242
24
313
50
441
22
Braked tilt grapple
5
1537–1846
15
1
21
1
31
8
41
6
7.6
1.6
178
23
222
8
286
50
406
27
ally, driving contributed somewhat more than crane work to the daily dose (Table 3).
release & reorganize had high values in the x-direction (Table 4).
Low-frequency vibrations were more intense during driving than during crane work (Fig. 3). However, there were no visible differences in the frequency spectra of vibrations in the vertical (z) direction between the crane work (Fig. 4) and driving WEs (data not shown). Accelerations in the horizontal directions (x and y) were highest in the frequency range 1.25–4 Hz during crane work (Fig. 4) and 0.25–5 Hz during driving. The frequency distributions for the given WEs were similar when using all Crane Equipment types.
No effect of Crane Equipment type on any vibration measure was found during crane work (Table 5). However, there were significant differences between Crane Equipment types in frequencies of collisions with residual trees. Fewest trees were hit when using the braked tilt grapple and most trees were hit when using the standard grapple (Fig. 5). However, applying collisions and alignments as covariates in the ANOVA did not reveal any significant relationship between collisions or alignments and vibration levels in any direction, nor for the sum vector for any of the vibration measures.
There were significant main effects of Work Cycle on all vibration measures except VDVv. Mean av was significantly higher during driving than during crane work according to the variance analysis. Furthermore, vibration acceleration magnitudes (mean weighted r.m.s) in the predominant vertical z-direction were also highest during driving (Table 4). However, for the time weighted r.m.s. value, A(8), the relationship was reversed (Table 4). An additional set of ANOVAs showed that this relationship between az and A(8) also held for driving versus all the Crane Activity WEs (data not shown). On average, more than four times as much time was spent on crane work than on driving. A high crest factor in the x-direction (mean 12, max 15) indicated occurrences of shocks during crane work. Nevertheless, VDVx, and VDVy were higher during driving than during crane work. In contrast, VDVz was higher during crane work than during driving. Consequently, the overall vector (VDVv) was not affected by WE. Crane Activities significantly affected all vibration measures, but the interaction between Crane Activities and Crane Equipment was non-significant. Crane in and release & reorganize were both the most time consuming Crane Activities and the WEs with the highest average vector vibrations. However, they differed in that crane in had high values in the y-direction while
114
Fig. 5 Average numbers of collisions – with residual trees (Trees), the Machine, or both (T&M) – and alignments of end surfaces of the logs per work cycle with each type of crane equipment. Means within categories with different letters are significantly different (Turkey’s HSD p<0.05) Croat. j. for. eng. 37(2016)1
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
4. Discussion Previous studies (Hansson 1990, Rehn et al. 2005a) have shown that the terrain significantly influences vibration levels and that WBV exposure is highest during driving in forwarder operations. Shock-type vibrations have also been detected during forwarder loading (cf. Rehn et al. 2005a). Nevertheless, effects of forwarder crane work have been seldom addressed, although it accounts for a high proportion of forwarding work: 50–90% of loading and unloading work time depending on extraction distance (Manner et al. 2013), and about 80% of the total monitored time in our study, reflected in higher A(8) values for crane work than for driving (Table 4). However, crane work is often done simultaneously with driving, so vibrations originating from driving confound those from crane work. Therefore, our study was conducted on a very flat, firm and even standardized track to minimize the influence of driving. Unstructured observations by the designated researcher revealed almost no occurrences of simultaneous crane and vehicle movements during the study. This indicates that our attempt to reduce driving vibrations was successful. Nevertheless, despite operating on an even track, the instantaneous vibration levels (ax-z and av) were still higher during driving than during any type of crane work examined in the study. We investigated the effects of six defined crane WEs on WBV, and obtained acceleration values ranging from 0.34 to 0.45 m/s2. Previous analyses of crane WEs during operations of a single-grip harvester have reported generally ca. 0.1 m/s2 lower vibration magnitudes (measured at the cabin floor), ranging from 0.20 to 0.34 m/s2, during both delimbing and felling (Burström et al. 2006). It should be noted that vibration magnitudes are normally lower at the chair, where most vibrations are transmitted to the operator. Indeed. Burström et al. (2006) found that the vibrations transmitted to the seat were lower than 0.04 m/s2 (0–22% of the vibrations at the floor in the x, y and z-directions). Thus, the combined WBVs the operator was exposed to through the seat in the studied small forwarder were probably considerably weaker than the floor-level values presented here. However, it should also be noted that chairs characteristics strongly influence vibration transmissions (Paddan and Griffin 2002), and evaluation or comparison of chairs was beyond the scope of this study. The crane WEs associated with the highest vibrations in our study were associated with handling logs (i.e. crane in and release & reorganize). This implies that the weight and balance at the crane tip influenced Croat. j. for. eng. 37(2016)1
C. Häggström et al.
WBV magnitudes. Nevertheless, despite noticeable shocks caused by impacts that were transmitted as vibrations through the crane to the cabin, no correlation was found between WBV exposure during the pooled crane work and grapple collisions with standing trees or the machine. Thus, these findings, in combination with the non-significant effect of crane equipment type (Table 5) and the predominance of vibrations in the z-direction (Table 4), imply that modifications that increase the stability of the base machine should be considered in attempts to reduce the operators’ exposure to crane work induced WBVs. This recommendation is supported by findings that vibrations are negatively correlated with machine weight (Rehn et al. 2005a). However, other measures, for example improving hydraulics and crane control systems, may also smooth operations and reduce crane work-induced vibrations (Hansson and Servin 2010). As mentioned above, shock-type vibration exposure is common during loading (Rehn et al. 2005a), but we found no association between either vibrations of this type or impacts during crane work. None of the VDV values associated with any Crane Activity were higher than the unloading values either (Table 4), which would also have indicated high frequencies of shocks during those activities (cf. Rehn et al. 2005a). Thus, it is highly likely that high WBV exposure while loading is due to driving between piles. Nevertheless, the differences in collision frequencies between crane equipment types observed in this study would be of interest when selecting thinning equipment to minimize damage to residual trees (Sirén et al. 2013). An experimental setup was used, which is commonly used within forest engineering work studies (Košir et al. 2015) to enable comparison of factors of interest. However, experimental results might be difficult to generalize to other conditions. Since the smoothness of operations also affects vibration levels (Hansson and Servin 2010), these results may not be readily applied to drivers with other experience levels, grip or working technique preferences. Indeed, the rankings of crane equipment types in terms of associated vibrations may differ for other operators under the same conditions (cf. Chen et al. 2003, Purfürst and Erler 2006, Lindroos 2010). Nevertheless, the obtained results regarding crane equipment are consistent with indications of vibration effects from a previous study (Nilsson 2013) and there were no indications of differences in vibration exposure between forwarder operators during crane work (Rehn et al. 2005a). Furthermore, the variation in crane equipment types and associated differences in working techniques did not affect the WBV
115
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
exposure during either crane work overall or the defined crane WEs in this study.
national Archives of Occupational and Environmental Health 72(6): 351–365.
The upper limit of the measured frequency in this study was 100 Hz. Thus the results should be interpreted cautiously. Nevertheless, most vibration was of lower frequency than 50 Hz (Fig. 3 and Fig. 4). Thus, the limitations in measurements should not have had a major impact on the calculated exposure, and the relative levels are fully comparable. Moreover, vibrations during forwarder work depend on numerous factors and measured values are only valid under the prevailing conditions during the study. More research is hence needed to fully generalize forwarder operations with other weight, size and with other dampening systems. Nevertheless, as no significant effect of Crane Equipment was found, it is unlikely that the action value for the daily exposure would be surpassed during forwarder work due to differences in crane equipment or (crane) working technique.
Burström, L., Nilsson, T., Wahlström, J., 2014: Whole-body vibration and the risk of low back pain and sciatica: a systematic review and meta-analysis. International Archives of Occupational and Environmental Health 88(4): 403–418.
5. Conclusions The studied crane working techniques and crane equipment types were found to have little or no effect on the daily WBV exposure with respect to seated health. We found no indication that any crane WE or impacts from making piles should contribute significantly to shock-type vibrations assumed to be associated with neck and arm pains. Thus, the hypothesis that high levels of shock-type vibrations during loading originate from driving in an uneven terrain (cf. Rehn et al. 2005a) seems to hold. However, due to better controllability, there were fewer collisions with trees and the machine when using the braked tilt grapple. Thus its use should make the operator’s work environment more relaxed and comfortable, and provide financial benefits for the land owner by reducing damage to the remaining stand.
6. References Baes, J., 2008: Vibrationsdämpning av skotare. Arbetsrapport. Uppsala: Skogforsk. Berg, S., 1992: Terrain classification system for forestry work. Kista: Forskningsstiftelsen Skogsarbeten. Boileau, P.E., Rakheja. S., 1990: Vibration attenuation performance of suspension seats for off-road forestry vehicles. International Journal of Industrial Ergonomics 5(3): 275–291. Bovenzi, M., Hulshof, C., 1999: An updated review of epidemiologic studies on the relationship between exposure to whole-body vibration and low back pain (1986–1997). Inter-
116
Burström, L., Nordfjell, T., Wästerlund, I., Tabell, L., 2006: Attempts to standardise vibration measurements in a singlegrip harvester. Journal of Low Frequency Noise. Vibration and Active Control 25(1): 11–21. Cation, S., Jack, R., Oliver, M., Dickey, J.P., Lee-Shee, N.K., 2008: Six degree of freedom whole-body vibration during forestry skidder operations. International Journal of Industrial Ergonomics 38(9–10): 739–757. Chen, J., Chang, W., Shih, T., Chen, C., Chang, W., Dennerlein, J., Ryan, L., Christiani, D., 2003: Predictors of wholebody vibration levels among urban taxi drivers. Ergonomics 46(11): 1075–1090. Conway, G., Szalma, J., Saxton, B., Ross, J., Hancock, P., 2006: The effects of whole-body vibration on human performance: A meta-analytic examination. Proceedings from the Human Factors and Ergonomics Society 50th Annual Meeting. October 16–20. Human Factors and Ergonomics Society: 1741– 1745. Eriksson, M., Lindroos, O., 2014: Productivity of harvesters and forwarders in CTL operations in northern Sweden based on large follow-up datasets. International Journal of Forest Engineering 25(3): 179–200. European Parliament and the Council of the European Union, 2002: Directive 2002/44/EC on the minimum health and safety requirements regarding the exposure of workers to the risks arising from physical agents (vibration) (sixteenth individual Directive within the meaning of Article 16(1) of Directive 89/391/EEC) 13 p. Fogdestam, N., 2010: Studier av Biotassu Griptilt S35 i gallring. Arbetsrapport. Uppsala: Skogforsk. Gellerstedt, S., 1998: A Self-leveling and swiveling forestry machine cab. Journal of Forest Engineering 9(1): 7–16. Gerasimov, Y., Sokolov, A., 2009: Ergonomic characterization of harvesting work in Karelia. Croatian Journal of Forest Engineering 30(2): 159–170. Hansson, A., Servin, M., 2010: Semi-autonomous shared control of large-scale manipulator arms. Control Engineering Practice 18(9): 1069–1076. Hansson, J.E., 1990: Ergonomic design of large forestry machines. International Journal of Industrial Ergonomics 5(3): 255–266. Häggström, C., Englund, M., Lindroos, O., 2015: Examining the gaze behaviors of harvester operators: an eye-tracking study. International Journal of Forest Engineering 26(2): 96–113. Croat. j. for. eng. 37(2016)1
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118) International Organisation for Standardization, 1997: ISO 2631-1 Mechanical vibration and shock-Evaluation of human exposure to whole-body vibration. Part 1: General requirements. Jack, R.J., Oliver, M., 2008: A review of factors influencing whole-body vibration injuries in forestry mobile machine operators. International Journal of Forest Engineering 19(1): 51–65. Ji, X., Eger, T.R., Dickey, J.P., 2015: Development of a seat selection algorithm to match industrial seats with specific forestry vibration exposures. International Journal of Forest Engineering 26(1): 48–59. Košir, B., Magagnotti, N., Spinelli, R., 2015: The role of work studies in forest engineering: status and perspectives. International Journal of Forest Engineering 26(3): 160–170. Lindroos, O., 2010: Scrutinizing the theory of comparative time studies with operator as a block effect. International Journal of Forest Engineering 21(1): 20–30. Lindroos, O., Burström, L., 2010: Accident rates and types among self-employed private forest owners. Accident Analysis and Prevention 42(6): 1729–1735. Lindroos, O., Lidestav, G., Nordfjell, T., 2005: Swedish nonindustrial private forest owners: a survey of self-employment and equipment investments. Small-scale Forestry 4(4): 409–425. Lis, A.M., Black, K.M., Korn, H., Nordin, M., 2007: Association between sitting and occupational LBP. European Spine Journal 16(2): 283–298. Manner, J., Nordfjell, T., Lindroos, O., 2013: Effects of the number of assortments and log concentration on time consumption for forwarding. Silva Fennica 47(4): 19p. Mansfield, N.J., Holmlund, P., Lundström, R., Nordfjell, T., Staal-Wästerlund, D., 2002: Vibration exposure in a forestry machine fitted with a saddle type suspension seat. International Journal of Vehicle Design 30(3): 223–237. Nilsson, G., 2013: Griptiltens effekt på skotarens produktivitet. M.A. thesis. Swedish University of Agricultural Sciences, Umeå. Nordfjell, T., Athanassiadis, D., Talbot, B., 2003: Fuel consumption in forwarders. International Journal of Forest Engineering 14(2): 11–20. Nordfjell, T., Björheden, R., Thor, M., Wästerlund, I., 2010: Changes in technical performance, mechanical availability and prices of machines used in forest operations in Sweden from 1985 to 2010. Scandinavian Journal of Forest Research 25(4): 382–389.
C. Häggström et al.
tion, posture and manual materials handling. Journal of Sound and Vibration 298(3): 540–555. Passicot, P., Murphy, G., 2013: Effect of work schedule design on productivity of mechanised harvesting operations in Chile. New Zealand Journal of Forestry Science, 43(1): 1–10. Pope, M., Magnusson, M., Lundström, R., Hulshof, C., Verbeek, J., Bovenzi, M., 2002: Guidelines for whole-body vibration health surveillance. Journal of Sound and Vibration 253(1): 131–167. Punnett, L., Wegman, D.H., 2004: Work-related musculoskeletal disorders: the epidemiologic evidence and the debate. Journal of Electromyography and Kinesiology 14(1): 13–23. Purfürst, T., Erler, J., 2006: The precision of productivity models for the harvester – do we forget the human factor. In: Ackerman, P., Längin, D. & Antonides, M., eds. Precision Forestry in plantations, semi-natural and natural forests: preceedings from the IUFRO Precision Forestry Symposium, 5–10 March. South Africa, Stellenbosch University: 465–474. Rehn, B., Bergdahl, I.A., Ahlgren, C., From, C., Järvholm, B., Lundström, R., Nilsson, T., Sundelin, G., 2002: Musculoskeletal symptoms among drivers of all-terrain vehicles. Journal of Sound and Vibration 253(1): 21–29. Rehn, B., Lundström, R., Nilsson, L., Liljelind, I., Järvholm, B., 2005a: Variation in exposure to whole-body vibration for operators of forwarder vehicles – aspects on measurement strategies and prevention. International Journal of Industrial Ergonomics 35(9): 831–842. Rehn, B., Nilsson, T., Lundström, R., Hagberg, M., Burström, L., 2009: Neck pain combined with arm pain among professional drivers of forest machines and the association with whole-body vibration exposure. Ergonomics 52(10): 1240–1247. Rehn, B., Nilsson, T., Olofsson, B., Lundström, R., 2005b: Whole-body vibration exposure and non-neutral neck postures during occupational use of all-terrain vehicles. Annals of Occupational Hygiene 49(3): 267–275. Rieppo, K., Kariniemi, A., Haarlaa, R., 2002: Possibilities to develop machinery for logging operations on sensitive forest sites. Helsinki: University of Helsinki. Sankar, S., Afonso, M., 1993: Design and testing of lateral seat suspension for off-road vehicles. Journal of Terramechanics 30(5): 371–393. Seidel, H., Heide. R., 1986: Long-term effects of whole-body vibration: a critical survey of the literature. International Archives of Occupational and Environmental Health 58(1): 1–26.
Paddan, G.S., Griffin, M.J., 2002: Effect of seating on exposures to whole-body vibration in vehicles. Journal of Sound and Vibration 253(1): 215–241.
Sherwin, L.M., Owende, P.M.O., Kanali, C.L., Lyons, J., Ward, S.M., 2004: Influence of tyre inflation pressure on whole-body vibrations transmitted to the operator in a cutto-length timber harvester. Applied Ergonomics 35(3): 253–261.
Okunribido, O.O., Magnusson, M., Pope, M.H., 2006: Low back pain in drivers: The relative role of whole-body vibra-
Sirén, M., Ala-Ilomäki, J., Mäkinen, H., Lamminen, S., Mikkola, T., 2013: Harvesting damage caused by thinning of
Croat. j. for. eng. 37(2016)1
117
C. Häggström et al.
Vibration Exposure in Forwarder Work: Effects of Work Element and Grapple Type (107–118)
Norway spruce in unfrozen soil. International Journal of Forest Engineering 24(1): 60–75. Zechmann, E., 2013: Continuous Sound and Vibration Analysis. Matlab Central File Exchange [Online]. Available: http://
www.mathworks.com/matlabcentral/fileexchange/21384continuous-sound-and-vibration-analysis [Accessed April 11 2014].
Authors’ addresses: Carola Häggström, PhD.* e-mail: carola.haggstrom@slu.se Mikael Öhman, MSc. e-mail: mikael.ohman@slu.se Prof. Tomas Nordfjell, PhD. e-mail: tomas.nordfjell@slu.se Assoc. Prof. Ola Lindroos, PhD. e-mail: ola.lindroos@slu.se Swedish University of Agricultural Sciences Department of Forest Biomaterials and Technology SE-901 83 UMEÅ SWEDEN
Received: October 24, 2014 Accepted: December 04, 2014
118
Lage Burström, PhD. e-mail: lage.burstrom@umu.se Umeå University Department of Public Health and Clinical Medicine SE-901 87 UMEÅ SWEDEN *Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Fly Ash in Forest Road Rehabilitation Tomi Kaakkurivaara, Pauli Kolisoja, Jori Uusitalo, Nuutti Vuorimies Abstract Finnish forestry and bioenergy production is seeking novel uses for the fly ash deriving from biomass conversion. There are various possibilities for using fly ash in civil engineering including road construction. The increase in bioenergy production has created more interest for using ash in forest roads. However, no established methods for the rehabilitation of forest roads exist yet. Hence, this research aims to find a suitable construction method that involves using ash that provides adequate bearing capacity. It involved building ten test road sections: two of them were reference sections without fly ash. The study examined the effect of four different rehabilitation methods on the bearing capacity of roads. Measurements were made once before and four times after the rehabilitation. The measuring devices included a light falling weight deflectometer (LFWD), a dynamic cone penetrometer (DCP) and a conventional falling weight deflectometer (FWD). Two of the rehabilitation structures were 50 and 25 cm thick fly ash layers. The other two were 15 and 20 cm thick layers made of fly ash and aggregate in different mixing ratios. In all cases, the constructed layers were paved with aggregate. Statistical comparison showed that the bearing capacity of the rehabilitated road sections had improved compared to the reference sections. The recorded bearing capacities after rehabilitation (during spring thaw in 2012, 2013 and 2014) were about the same as before rehabilitation in summer 2011. Based on this study, fly ash can be recommended as an option for forest road rehabilitation. Keywords: forest road, rehabilitation, fly ash, bearing capacity
1. Introduction Increasing utilisation of bioenergy is increasing the quantity of produced ash. The wood and peat burning processes of the forest industry are producing a significant amount of ash. About 600,000 tonnes of the ash produced in Finland annually is wood- and peatbased fly ash (Emilsson 2006). The total annual production of ash in the country is about 1.5 million tonnes. There are 52 power plants around Finland, which produce at least 1000 tonnes of fly and bed ash per year (Tuhkarakentamisen käsikirja 2012). Ash utilisation is divided roughly evenly between earthworks, landfills and fertilizers or other use (Ojala 2010). The growing ash production necessitates finding sensible ways of utilising ash. Increased utilisation can bring several benefits and there are many potential ways of using fly ash. As a fertilizer, it increases timber growth in forests (Moilanen et al. 2005) and agricultural field crops (Patterson et al. 2004), it can be used as a construction material in road building reducing Croat. j. for. eng. 37(2016)1
the need of natural stone resources (Edil and Benson 2007) or as a filler in concrete (Wang et al. 2008). Ash is no longer considered waste from the energy and forest industries point of view. On the contrary, ash is nowadays a by-product that can be used in various ways. Yet, there are obstacles to expanding ash utilisation. In spite of the above mentioned possibilities of use, a remarkable amount of ash ends up in landfills. From the economic point of view, it is important to minimise the amount of ash going to landfills, because the dumping charge has lately been raised to €55 per tonne. Ash is not as stable a material as mineral soil. Ash can be separated into fly ash and bottom ash based on particle size. The properties of fly ash from wood and peat vary widely depending on the fuel used, the burning process and the combustion gas filtering technique (Korpijärvi et al. 2009). Two features need to be taken into consideration when planning the use of ash: its content of environmentally harmful heavy metals and its technical properties. Ash has been used in the
119
T. Kaakkurivaara et al.
structural layers of public roads because of the flexible licensing policy (Tuhkarakentamisen käsikirja 2012). However, forest roads do not belong to the same class as public roads. Instead, the use of ash in forest roads requires a more specific and precise license granting procedure. Proper use of ash can improve the bearing capacity of forest roads. The forest industry supply chain needs more constant year-round flow of raw materials, which requires adequate trafficability of forest roads. The bearing capacity of forest roads is at its lowest during the spring thaw when their moisture content is high (Salour and Erlingsson 2013). The thawing period either restricts timber haulage or breaks the road structure if haulage continues during the low bearing capacity period. On the other hand, high bearing capacity promotes wood supply management and reduces the need to repair road damage due to rut formation. There are about 135,000 km of forest roads in Finland, of which about 3000–4000 km are rehabilitated annually (Metsätilastollinen vuosikirja 2013). Forestry would benefit greatly if it could combine ash utilisation and forest road rehabilitation. The use of ash in road structures is receiving increasing attention. Related studies have focused mainly on coal ash due to the common use of coal in energy production. There have also been studies on low-volume roads where both coal ash (Edil and Benson 2007) and bio ash (Lahtinen 2001) have been used. The studies indicated that ash is well-suited for road construction. Bio ash has become the focus of studies in recent years. The utilisation of wood-based fluidised and bottom ash in forest roads, where the ash was mixed into the existing road structure (Bohrn and Stampfer 2014), has also been studied. The measurements carried out by a light falling weight deflectometer (LFWD) showed improvement in bearing capacity during the half-year observation period. In another study, wood fly ash and bottom ash were found to have a positive impact on bearing capacity in both a laboratory test and a field survey using the LFWD over a four-week monitoring period (Supancic and Obernberger 2012). In a third study, wood fly ash was mixed into the gravel of an existing road. There, the falling weight deflectometer (FWD) results indicated improvement in bearing capacity during the monitoring period of a year and a half (Vestin et al. 2012). In the three above mentioned studies, ash was mixed into an existing road structure by a road grader. The grader breaks up the old road structure and the ash functions as a binding component. These three positive results were produced by short-term bearing capacity surveys. The aim of this research was to study various rehabilitation techniques and the development of bearing
120
Fly Ash in Forest Road Rehabilitation (119–130)
Fig. 1 Locations of measuring points along the test section. Numbers 1–7 denote measurement points, WP stands for wheel path and CL for centre line capacity when forest roads are renovated with bio fly ash. Four different test structures were made of fly ash. Two of them included only fly ash, while other two contained fly ash mixed with an aggregate. The development of the bearing capacity of the test structures and a reference structure was compared. The goal was to achieve higher bearing capacity especially during spring thaw, when insufficient bearing capacity restricts haulage on forest roads.
2. Material and methods 2.1 Test sections Ten test road sections were established to monitor the effect of ash on bearing capacity. The test sections were located along two forest roads where throughtraffic was possible in Central Finland near Jämsä Municipality (62°1’31’’ N, 24°54’5’’ E). The test sections were prepared before the actual rehabilitations were done. The test sections had seven measuring points (Fig. 1). Four of them were on the wheel path (WP) and three on the centre line (CL). The test sections and measuring points were marked by poles at the road side in summer 2011. That allowed finding the test sections in the following years. A total of ten test sections were established. There were four different test structures and a reference structure on both forest roads. Croat. j. for. eng. 37(2016)1
Fly Ash in Forest Road Rehabilitation (119–130)
T. Kaakkurivaara et al.
Fig. 2 Measuring devices: dynamic cone penetrometer, light falling weight deflectometer and falling weight deflectometer
2.2 Bearing capacity measurements Instead of a single measuring device, this study used three different bearing capacity measuring devices: the conventional falling weight deflectometer (FWD), the dynamic cone penetrometer (DCP) and the light falling weight deflectometer (LFWD). The bearing capacity results can be considered a crucial indicator of successful rehabilitation. The most suitable measure of bearing capacity is elastic modulus (later E-modulus) measured in MPa. More information about bearing capacity and road structure can be acquired by using several measuring devices at the same time. The initial bearing capacity measurements were carried out with LFWD and DCP devices. Post-rehabilitation measurements were carried out with LFWD, DCP and FWD devices. Initial measurements took place in August 2011 and the rehabilitations were done in September 2011. Bearing capacities of the constructed roads were measured four times in 2012–2014. The monitoring period was three years, longer than with Bohrn and Stampfer (2014), Supancic and ObernbergTable 1 Times of measuring rounds and rehabilitation 2011
2011
Summer
Autumn
LFWD
x
DCP
x
Rehabilitation
FWD
Croat. j. for. eng. 37(2016)1
2012
2012
2013
2014
Spring Summer Spring Spring x
x
x
x
x
x
x
x
x
x
x
x
er (2012) and Vestin (et al. 2012). A specific timetable of data collection is presented in Table 1. DCP measurement is based on making a cone tip penetrate into the ground by the impact force of a falling weight. The mass of the falling weight was 8 kg and dropping height about 575 mm. The diameter of the cone was 20 mm and the angle of the tip 60 degrees. The equipment also included a measuring rod, which allowed reading the vertical penetration (mm) after each drop or a certain number of drops. The DCP Penetration Index (DPI; mm per blow) was calculated on that basis. The California Bearing Ratio (CBR) can be estimated from the DPI value with empirical Eq. 1 (Webster et al. 1992). The CBR value can then be used for estimation of elastic modulus (E-modulus) with another empirical Eq. 2 (Powell et al. 1984) that is one of the most well established alternatives among the available equations for making this conversation. More detailed justification and reason for using these equations was presented in a study by Kaakkurivaara et al. (2015). Measurement was continued until the depth of 400 mm was reached. The E-modulus values, derived based on DCP measurements, are hereafter referred to as EDCP.
logCBR = 2.46 − 1.12 × logDPI E = 17.6 × CBR0.64
(1) (2)
Where: CBR California Bearing Ratio, %; DPI DCP Penetration Index, mm/blow; E soil’s elastic modulus, MPa.
121
T. Kaakkurivaara et al.
Fly Ash in Forest Road Rehabilitation (119–130)
The LFWD device used in the study was a Loadman, manufactured by AL-Engineering Ltd of Finland. The operating principle of the LFWD is based on dropping a weight that causes a momentary deflection of ground surface. The mass of the falling weight was 10 kg and dropping height 800 mm. The diameter of the loading plate was 132 mm. The deflection was captured by integrating the acceleration signal twice. The internal electronics of the device calculated and showed maximum deflection (mm) and E-modulus (MPa) immediately after measurement. The Loadman uses Eq. 3 (Pidwerbesky 1997) to convert deflection measurements into E-moduli. The measurement results can later be transferred to computer via a USB connection.
a E = 1.5 p × D
(3)
Where: D deflection under the Loadman loading plate; p vertical pressure on the base plate; a radius of the base plate. The results of the third drop were used to determine the E-modulus in this study. It was done to reduce the influence of the loose surface layer. More detailed information about the use of the LWFD is presented in a study by Kaakkurivaara et al. (2015). The E-modulus yielded by the LFWD was denoted by ELFWD in this study. A conventional falling weight deflectometer was made using the Kuab FWD device. The operating principle of the FWD is also based on a falling weight, which causes a deflection impulse on the road surface. The impulse corresponded to the load of a 50 kN tyre. The E-modulus for the FWD was calculated with the same Eq. 3 as for the LFWD. Later in the article, the E-modulus of the FWD has been denoted as EFWD. At this point it is important to notice that the efficient depth of influence in FWD measurements is far greater than that of LFWD measurements. Thus, the values of E-modulus derived based on LFWD represent basically only the top 200 mm of the road structure, while those based on FWD also include the influence of underlying subgrade up to the depth of 500 mm. More specific information about FWD measurements is available in an article by Kaakkurivaara et al. (2015).
2.3 Rehabilitation operations In this study, the existing road structure was not altered, but a new structural layer was built on top of the old one. Five different test structure types were selected. Four of them contained different amounts of fly ash. Aggregate was also added to the reference
122
Fig. 3 A road grader was used to shape the road surface and mix the ash and aggregate structure in deviation from the others which are normal in rehabilitation and common with forest roads. Two test sections of each type were built. The aggregate used in all test sections was similar. The particle size of the aggregate, crushed from oversize rock material, was between 0–32 mm. The total length of the rehabilitated forest road was 2400 m and a total of 1936 tonnes of fly ash was used. The fly ash was less than one year old. It had been stored outdoors exposed to the elements. Two test sections were chosen as references. They were surfaced with an about 100 mm thick layer of aggregate. Technical implementation of the fly ash test structures involved the following. Firstly, side barriers were formed on both sides of the road to prevent fly ash from escaping into the ditches. The first test structure received a 50 mm layer of fly ash and a 100 mm layer of aggregate. The layers were mixed by a road grader (Fig. 3). The road grader was also used in all other test sections. Hereafter, the first test structure is referred to as #1. The content of fly ash was increased in the second test structure: fly ash and aggregate layers 100 mm thick were mixed. The second test structure is referred to as #2 later. Thus, fly ash was mixed with aggregate in both of the above mentioned test structures on top of the road. Another technique would have been to mix the materials at the storage site before transportation to the construction site. The plan was to have a uniform layer of fly ash on the third and fourth test structures. The thickness of the fly ash layer was 250 mm on the third test structure (#3) and 500 mm on the fourth one (#4). All test road structures were compacted immediately after construction by driving over them with an excavator several times. An environmental permit was required for the operation, Croat. j. for. eng. 37(2016)1
Fly Ash in Forest Road Rehabilitation (119â&#x20AC;&#x201C;130)
T. Kaakkurivaara et al.
oxide content also improves the bearing capacity of the structure due to the chemical bond between calcium oxide and water (Pecqueur et al. 2001). A modified Proctor test (standard SFS-EN 13286-2) was done on the fly ash, which showed that the optimum content of water is between 35 and 40% for compaction. The weather was rainy during the rehabilitation operation, which helped the compaction process.
2.5 Soil sample analyses of existing road
Fig. 4 Test structure profiles. Four different test structures included an ash layer and its provisions required ensuring that the fly ash cannot escape from the road surface after construction. Due to the provisions, all test sections were surfaced with a 100 mm thick layer of aggregate. The thickness of the surface aggregate was hence the same as that of the reference sections. The excavator shaped all road slopes of the test sections. It was especially important in the case of the high profile test structures #3 and #4 to give them horizontal support. Extra soil materials from ditches were added to the road slopes of #4 test structure types. All test structure profiles are presented in Fig. 4.
2.4 Properties of ash The fly ash was obtained from a forest industry power plant fuelled by peat and wood. When the ash used in this study was produced, the burned material was mainly wood-based. The share of peat was higher in winter and lower in summer. The fly ash was collected over one year. The exact mix ratio is unknown, because it varied continuously during the year. Variation in the mix ratio had an impact on fly ash properties. The fluidised bed combustion technique was used where the temperature was about 850°C. The contents of various elements were determined three times from fresh fly ash during the collection period. The content of calcium oxide was most important for the self-hardening process. Three analyses showed that the content of calcium varied between 15 and 19%. Higher calcium Croat. j. for. eng. 37(2016)1
Soil samples were collected from the test sections before the rehabilitations to determine particle size distribution and organic content. Soil samples were taken from the existing road surface structure and subgrade at each test section. The samples describe the properties of the aggregate transported to site from other locations. The subgrade samples describe the properties of the subgrade and the embankment fill layer. Soil material from ditches had been used in the embankment fill layer. The grain size distribution curve was plotted using the wet sieving and pipette methods. Used sieve sizes were 0.63, 2, 6.3 and 20 mm. The organic content of the samples was determined by the burning method for grain sizes below 2 mm. Each plotted distribution curve was compared to the design grading curves of the Finnish Transport Agency (Finra 2005) to estimate the E-modulus according to the Odemark bearing capacity design method (Odemark 1949). Frost susceptibility of the samples was assessed according to the design grading curves of Finnish guidelines (Finra 1993). Specific details of these assessments are presented in a study by Kaakkurivaara et al. (2015). Hereafter, aggregate EGSD refers to the surface layer of the existing road structure and subgrade EGSD to the subgrade and embankment fill layers for E-modu lus estimation based on grain size distributions.
3. Results 3.1 Structure of the road before rehabilitation The test road sections were established on an existing forest road. The E-modulus values estimated based on soil samples taken from the existing road structure before rehabilitation are presented in Table 2. The E-modulus for the surface aggregate material of the old road was generally about 150 MPa. The E-modulus for the subgrade was poor or at best moderate, which is typical for fine sediment soils. The surface aggregate layer was equally very thin and the subgrade material poor at every test section. Thus, the bearing capacity of the whole test section area was poor.
123
T. Kaakkurivaara et al.
Fly Ash in Forest Road Rehabilitation (119–130)
Table 2 Mean elastic modulus (MPa) based on grain size distributions (EGSD) for existing road structure materials before rehabilitation. N stands for numbers of soil samples #1
#2
#3
#4
Reference
Aggregate EGSD
150
150
200
150
125
Subgrade EGSD
23
35
20
18
38
Samples, n
2
2
2
2
2
3.2 Bearing capacities of the rehabilitated road sections Fig. 5 shows the bearing capacities of the test structures in terms of mean MPa. Each value is based on measurements from two test sections. The mean values of EDCP and ELFWD are based on eight measuring points on the wheel path and six on the centre line. The mean values of EFWD are based on four measuring points on the wheel path and two on the centre line. Each measuring round included these measurements on every test structure. The initial situation is represented by the values of summer ‘11, which were measured in summer before rehabilitation in autumn. It is important to keep in mind that after the rehabilitation, one measuring round was undertaken in summer and three rounds in spring. It can be said that, in general, the bearing capacity started to decline after initial positive development. The highest E-modulus values were measured with the DCP for every test structure type. The highest values were measured for test structure #2 (222 MPa) and the second highest for test structure #1 (198 MPa). In the case of test structure #4, the bearing capacity (154 MPa) increased slightly higher than with #3 (135 MPa), but it remained clearly lower than for test structure #1. It is noteworthy that the bearing capacity of #4 test structures was 10–20 MPa lower than for other test sections initially. The positive development of the bearing capacity ended two years after rehabilitation and the results of the fourth measurement round were generally weaker than those of earlier measurement rounds. The bearing capacity of test structure #2 was about 160 MPa and those of test structures #1, #3 and #4 about 120 MPa. The adding of surface aggregate did not increase the bearing capacity of the reference structure. It even decreased from 120 MPa to 100 MPa, as shown by the last measurement round. This indicates that in spring ‘14 measurement conditions may have been different from the other measurement rounds. Nevertheless, the purpose was to do the measurements at the same phase of the frost thaw period.
124
The EDCP results for the centre line were in line with the EDCP results for the wheel path, but the E-modulus values were lower across the board. The ELFWD measurements do not suggest similar clear improvement as the EDCP measurements. The ELFWD results improved only by a few MPa, while the EDCP values increased almost double compared to the initial situation for the #1 and #2 test structures. Initially, the ELFWD measurements showed clear improvement for #1 and #2, when bearing capacity values were over 70 MPa at their highest. In the last measurement round, the bearing capacities were about the same (53–56 MPa) for the #1, #2 and #3 test structures. The most positive observation was made concerning the #4 test structure, whose initial bearing capacity was the lowest (44 MPa), but the highest (59 MPa) in the last measurement round. The lowest bearing capacity was measured for the reference structure, whose bearing capacity values varied widely around 50 MPa during the entire survey. The above mentioned changes in the bearing capacity were measured on the wheel path. The ELFWD results for the centre line were similar to the ELFWD results on the wheel path, but the E-modulus values were lower across the board. FWD measurements were not carried out before rehabilitation. After rehabilitation, the first bearing capacity values (EFWD) were lowest for the reference structures and the #1 test structures (30 MPa). The highest bearing capacity was measured for the #4 test structure and the next highest for the #2 and #3 test structures (47 MPa). Bearing capacity values rose naturally in the summer ‘12 measurement round for all test structure types. The next year (spring ‘13), only #2 and #3 had maintained the achieved bearing capacity (54–55 MPa), while the bearing capacities of the other test structure types had declined. The capacities of all test structure types had deteriorated by the last measurement round, where test structure #3 (41 MPa) fared the best and #4 (38 MPa) next best. On the whole, bearing capacity values were slightly lower in the last measurement round than in the first. On the centre line, EFWD results fell under the above described values varying between 40 MPa (#4, summer ‘12) and 16 MPa (reference, spring ‘13).
3.3 Statistical results of paired sample t-test Table 3 presents a paired sample t-test, where the initial bearing capacity result and the result of each measurement round after rehabilitation for every test structure type were compared separately. The initial values were measured in summer before rehabilitation which took place in autumn. Mean values were Croat. j. for. eng. 37(2016)1
Fly Ash in Forest Road Rehabilitation (119â&#x20AC;&#x201C;130)
T. Kaakkurivaara et al.
Fig. 5 Results of bearing capacity measurements for sections are presented in the following order: #1, #2, #3, #4 and reference test structures. The measurements were done with a dynamic cone penetrometer (DCP), a light falling weight deflectometer (LFWD) and a falling weight deflectometer (FWD). CL stands for centre line and WP for wheel path Croat. j. for. eng. 37(2016)1
125
T. Kaakkurivaara et al.
Fly Ash in Forest Road Rehabilitation (119–130)
Table 3 Statistical significance of paired sample t-test results between measurement rounds, when p<0.05. Measurement unit: MPa. Minus indicates declined bearing capacity and plus improved bearing capacity. A blank box indicates insignificant change between values Device Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev.
Class
Test
Summer‘11
Spring‘12
Summer‘12
Spring‘13
Spring‘14
structure
MPa
MPa
MPa
MPa
MPa
+37
+92
+64
–
27
78
41
–
+61
+87
+115
+50
37
17
56
30
–11
–
+13
–
12
–
15
–
––
+58
+36
–
–
64
28
–
–
+12
–
–
–
7
–
–
–
+22
–
–
–
15
–
–
–9
–
–
–
8
–
–
–
–
+19
–
+15
–
14
–
9
+21
–
+62
–
18
–
42
–
–
+38
+40
–
–
28
12
–
+31
–
–
–
24
–
–
–
–
+7
–
+6
–
5
–
4
–
+10
+10
+10
–
6
3
5
+9
+15
+9
+18
6
7
6
8
DCP
WP
#1
106
DCP
WP
#2
107
DCP
WP
#3
112
DCP
WP
#4
97
Loadman
WP
#1
59
Loadman
WP
#2
52
Loadman
WP
#3
54
Loadman
WP
#4
44
DCP
CL
#1
74
DCP
CL
#2
71
DCP
CL
Ref.
66
Loadman
CL
#1
27
Loadman
CL
#2
24
Loadman
CL
Ref.
23
calculated and compared to initial values between measurement rounds. A negative value indicated a decline in bearing capacity and a positive value strengthening of it. Negative values were measured only twice – in the first comparison. Here, the calcula-
126
tion of standard deviation was based on the groups of observations behind both mean values. The number of observations was highest in the column where comparisons were made between measurement rounds of the same season (summer ‘11 and summer Croat. j. for. eng. 37(2016)1
Fly Ash in Forest Road Rehabilitation (119–130)
‘12). Statistically significant observations were almost double in the case of #1 and #2 test structures compared to uniform fly ash structures. The number of statistically significant observations was the same for both measurement classes – WP and CL. If a statistically significant observation was made in the same measurement round concerning a test structure, indicated by both ELFWD and EDCP, the improvement in bearing capacity was higher on the wheel path than on the centre line, as expected. This observation describes well the functioning of the structure since the bearing capacity improved more specifically on the wheel path where compaction of the structure was responsible for the improvement. About the same numbers of statistically significant observations were made with DCP and LWFD devices. Improvements of bearing capacities were generally higher when comparison was based on EDCP values. The ELFWD values of the bearing capacity did not improve as much as EDCP values. Comparison of EFWD results could not be done because initial values were not measured. Development on the centre line measured by ELFWD was positive compared to the reference structure, which showed no improvement on the wheel path measured either by ELFWD or EDCP. In this statistical study, the number and location of the blank boxes of Table 3 must be taken into consideration. It means that statistical improvement was not detected. Some rows, where the combination of the measuring device, measurement class and test structure were not statistically significant, are also missing. They indicate that rehabilitation did not improve the bearing capacity.
3.4 Statistical results of independent sample t-test Table 4 presents a t-test comparison of independent samples, where the results of rehabilitated test structures are compared to results for reference structures of the same measurement round. In Table 4, ELFWD and EDCP values have been divided between wheel path and centre line measurement classes. The table presents the bearing capacity and its standard deviation for each measurement round, if the statistical difference between the reference structure and test structures was significant. EFWD results were also analysed, but statistically significant results were not found. The differences in bearing capacities were quite often statistically significant between the #1 test structure and the reference structure, and between the #2 test structure and the reference structure. The difference between test structure #2 and the reference structure was significant in every measurement round. On the other hand, no statistically significant difference was obCroat. j. for. eng. 37(2016)1
T. Kaakkurivaara et al.
Table 4 Statistically significant results of independent sample t-test based on mean bearing capacity values (MPa) between test structures and reference structure, when p<0.05. The number of observations consists of eight measurements on wheel path (WP) and six measurements on centre line (CL). A blank box indicates an insignificant change between values Device Class Round Mean St. dev. Mean St. dev. Mean St. dev. Mean St. dev. Mean St.Dev. Mean St.dev. Mean St. dev.
DCP
WP
DCP
WP
DCP
WP
DCP
WP
Loadman WP
DCP
CL
Loadman CL
Reference #1 MPa
#2
#3
#4
MPa MPa MPa MPa
Spring
116
–
+52
–
–
2012
36
–
24
–
–
Summer
127
–
+67
–
–
2012
32
–
24
–
–
Spring
122
+47 +100 –
–
2013
26
40
42
–
–
Spring
96
–
+60
–
–
2014
25
–
16
–
–
Summer
52
–
–
2012
7
12
12
–
–
Spring
71
+65
–
–
–
2013
30
55
–
–
–
Spring
28
–
–
–
+13
2014
8
–
–
–
5
+18 +21
served in the case of #3 and #4 on the wheel path. The number of DCP observations was more than double that with the LFWD. Only two statistically significant differences were observed on the centre line. It should be noted that no observations were made during the first measurement round before rehabilitation (summer ‘11).
4. Discussion and conclusions The study presents changes in bearing capacity during a survey covering a period of a little less than three years. In that period, a marked improvement occurred as the initial summer bearing capacity was reached already the next spring after rehabilitation. It meant considerably better trafficability during the most crucial time for haulage. The means of bearing capacity were clearly higher on the wheel path than on the centre line. That indicated that compaction of test structures had taken place as a consequence of light traffic loading after rehabilitation. The weather
127
T. Kaakkurivaara et al.
conditions of spring 2014 can be the reason for the lower bearing capacity since conditions vary from year to year. Weakening of bearing capacity was also observed in the reference structure, whose bearing capacity had not changed before, which supports the above conclusion. A clear difference between the test structure types was indicated by all measuring devices. Similar development of bearing capacity in each measurement round was shown by all measuring devices on all test structures as shown in Fig. 5. In other words, development of bearing capacity was parallel in every measurement round independent of the measuring device. Improvement in bearing capacity can be mainly detected with DCP, to some extent with LFWD, but not at all with FWD. The weather conditions before measurement may be more relevant than the rehabilitation itself in explaining the similar values of EFWD between the test structure types and the reference type. This argument is based on the measuring principle of FWD, where measurement depth is substantially deeper than with the other devices. Therefore, poor existing road EGSD values under the test structure have a negative impact on all measurement results with FWD. It seems that empirical equals of DCP overestimated E-modulus values for these construction materials when compared to measurements of other devices. Time of the year also needs to be taken into consideration; there was high moisture content on road structure during spring thawing season. It may have affected the measurement results. The paired sample t-test concentrated on following the development of the bearing capacity of each test structure type over time. Statistical correlation was observed between the EDCP results when comparisons were made between the initial measurement and three following measurement rounds. Similar correlation did not occur in the case of the fourth measurement round with DCP. The fact that the E-moÂdulus values were calculated by two equations does not matter in the paired sample t-test, because the target of comparison was another DCP measurement, not a measurement made with another device. The aim of this study was not to measure exact ÂE-modulus values, but to observe the changes in bearing capacity by the same measurement device over the time and between different test structures. The empirical equations for DCP are usually working well; if the structures are homogeneous and particle size is not too large. Disadvantage of the DCP is that the measurement operation can be very time consuming on well compacted good quality aggregate. Improvement of the bearing capacity occurred when
128
Fly Ash in Forest Road Rehabilitation (119â&#x20AC;&#x201C;130)
initial ELFWD results were compared to summer measurement round results, whereas comparison to spring measurement rounds did not indicate a clear improvement. There are two reasons for that: the measurement rounds were conducted in different seasons, and spring conditions also varied between years. The ranking of the test structures was not very clear. Statistical significances were observed more often in the case of the #1 and #2 test structures than the #3 and #4 test structures. Moreover, improvement of the bearing capacity seems to be bigger with the #1 and #2 test structures. The lack of observation in the case of the reference structure was a very good finding, because addition of surface aggregate did not improve the bearing capacity in this study. The paired sample t-test confirmed the impact of traffic on improving the bearing capacity along the wheel path, because related observations were made more often on the wheel path than on the centre line. However, the numerous blank boxes (Table 3), i.e. missing observations, challenge the functionality of the test structures. The independent sample t-test excluded seasonal variations and focused on comparing test structure types at a particular moment in time. The independent sample t-test proved that the #1 and #2 test structures had clearly better bearing capacity than the reference structure in each measurement round. Two thirds of the observations were made on the wheel path and one third on the centre line. The tendency of the observations was the same concerning both EDCP and ELFWD. The DCP measurements revealed variation of the bearing capacity on the wheel path between the structures over time. It can be deduced that the used mixture of aggregate and fly ash works better than other structures in this study. The missing statistical observations from the measurement round before actual rehabilitation testify that the test sections had been initially in similar condition. The independent sample t-test does not confirm successful improvement of uniform fly ash structures since only one statistically significant observation was made. The bearing capacity measurements were carried out by three different devices, which meant differences in the results. The bearing capacity results were device-specific and no contradictions were observed. Similar development trends of bearing capacities between the test structures or between measurement rounds guarantee the technical feasibility and reliability of the measurement methods. Bearing capacities showed clear improvement immediately after addition of fly ash. Statistical analyses revealed that Croat. j. for. eng. 37(2016)1
Fly Ash in Forest Road Rehabilitation (119–130)
test structures #1 and #2 had improved. The improvement of test structures #3 and #4 was less pronounced. It should also be noted that it is not possible to exclude eternal factors in field circumstances, as it is in laboratory circumstances. Discussion of results rests on field measurement and full confidence of functionality requires longer survey time. A conclusion can be drawn, although clear improvements in bearing capacities could not be established. The bearing capacity did not improve by increasing fly ash. It seemed to be more a consequence of adding aggregate. The conclusion is not in line with the expectations about the functionality of using fly ash on forest roads. The reason could be inadequate compaction during construction work. Unequal storage times of fly ash and the lack of a better mixing technique may also have affected the results. The results showed, nevertheless, improvement in all test structure types in spring compared to initial measurements since initial summer bearing capacity was reached in spring after rehabilitation. None of the test structures rose above the others based on the received results, but the fly ash and aggregate mix seemed to outperform uniform fly ash layers.
Acknowledgements This paper is part of a doctoral thesis on the suitability and cost-effectiveness of new road rehabilitation methods and techniques for improving trafficability of the Finnish low volume road network. Funding for the doctoral thesis work has been provided by the Marjatta and Eino Kollin Foundation.
5. References Anon., 2012: Tuhkarakentamisen käsikirja, Energiantuotannon tuhkat väylä-, kenttä- ja maarakenteissa. Ramboll Finland Company, Espoo, 65 p. Anon., 2013: Metsätilastollinen vuosikirja, Finnish Statistical Yearbook of Forestry 2013: Finnish Forest Research Institute. 449 p. Bohrn, G., Stampfer, K., 2014: Untreated wood ash as a structural stabilizing material in forest roads. Croatian journal of forest engineering 35 (1):81–88. Edil, T., Benson, H., 2007: Demonstration of ash utilization in low volume roads. Research report 2007–12. Minnesota Department of Transportation, 233 p. Emilsson, S. 2006: International handbook: from extraction of forest fuels to ash recycling. �������������������������� Skogstyrelsen, Swedish Forest Agency, 42 p. Croat. j. for. eng. 37(2016)1
T. Kaakkurivaara et al. Finra., 1993: Yleiset perusteet, Tierakennustöiden yleiset laatuvaatimukset ja työselitykset. Kehittämiskeskus, Tielaitos. Helsinki, 46 p. Finra., 2005: Tietoa tiensuunnitteluun nro 71D. Tie- ja geotekniikka, Tiehallinto, 13 p. Kaakkurivaara, T., Vuorimies, N., Kolisoja, P., Uusitalo, J., 2015: Applicability of portable tools in assessing the bearing capacity of forest roads. Silva Fennica 49(2): 26 p. Korpijärvi, K., Mroueh, U.M., Merta, E., Laine-Ylijoki, J., Kivikoski, H., Järvelä, E., Wahlsröm, M., Mäkelä, E., 2009: Energiantuotannon tuhkien jalostaminen maarakennuskäyttöön�������������������������������������������� . VTT Research notes 2499. VTT Technical Research Centre of Finland, 75 p. Lahtinen, P., 2001: Fly ash mixtures as flexible structural materials for low-volume roads. Finnra Reports 70/2001. Finnish Road Administration, 95 p. Moilanen, M., Silfverberg K., Hökkä, H., Issakainen, J., 2005: Wood ash as a fertilizer on drained mires – growth and foliar nutrients of Scots pine. Canadian Journal of Forest Research 35(11): 2734–2742. Odemark, N., 1949: Undersökning av elasticitetsegenskaperna hos olika jordarter samt teori för beräkning av beläggingar enligt elasticitetsteorin. ������������������������� Stockholm, Statens väginstitut, Meddelande 77. Ojala, E., 2010: Selvitys puu- ja turvetuhkan lannoite- sekä muusta hyötykäytöstä. ��������������������������������� Energiateollisuus ry. Finnish Energy Industries, 46 p. Patterson, S., Acharya, S., Thomas, J., Bertschi, A., Rothwell, R., 2004: Integrated soil and crop management, Barley biomass and grain yield and canola seed yield response to land application of wood ash. Agronomy journal 96(4): 971–977. Pecqueur, G., Crignon, B., Quénée, B., 2001: Behavior of cement-treated MSWI bottom ash. Waste Management 21(3) 229–233. Pidwerbesky, B., 1997: Predicting rutting in unbound granular basecourses from Loadman and other in-situ nondestructive tests. Road & Transport Research 6(3): 16–25. Powell, W., Potter, J., Mayhew, H., Nunn M., 1984: The structural design of bituminous roads. TRRL Report LR 1132, 62 p. Salour, F., Erlingsson, S., 2013: The Influence of Groundwater Level on the Structural Behaviour of a Pavement Structure Using FWD. In BCRRA, The Ninth International Conference on the Bearing Capacity of Roads, Railways, and Airfields, 25–27 June, Trondheim, Norway. Akademika forlag: 485–494. Supancic, K., Obernberger, I., 2012: Wood ash utilization as a binder in soil stabilization for road construction – first results of large-scale tests. ASH 2012, Ashes in a Sustainable Society, January 25–27 Stockholm, Sweden.
129
T. Kaakkurivaara et al. Vestin, J., Arm, M., Nordmark, D., Lagerkvist, A., Hallgren, P., Lind B., 2012: Fly ash as a road construction material. WASCON, Conference proceedings, 8 p. Wang, S., Miller, A., Llamazos, E., Fonseca, F., Baxter, L., 2008: Biomass fly ash in concrete: Mixture proportioning and mechanical properties. Fuel 87(3): 365–371.
Fly Ash in Forest Road Rehabilitation (119–130) Webster, S., Grau, R., Williams P., 1992: Description and application of Dual Mass Dynamic Cone Penetrometer. U.S. Army Engineer Waterways Experiment Station, Instruction Report GL-92-3, 26 p.
Authors’ address: Tomi Kaakkurivaara * e-mail: tomi.kaakkurivaara@gmail.com Jori Uusitalo, PhD. e-mail: jori.uusitalo@luke.fi Natural Resources Institute Finland Green technology Kaironiementie 15, FI-39700 Parkano FINLAND
Received: May 18, 2015 Accepted: August 26, 2015
130
Prof. Pauli Kolisoja, PhD. e-mail: pauli.kolisoja@tut.fi Nuutti Vuorimies e-mail: nuutti.vuorimies@tut.fi Tampere University of Technology Department of Civil Engineering P.O. Box 600, FI-33101Tampere FINLAND * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Timber Truck Payload Management with Different In-Forest Weighing Strategies in Australia Mark Brown, Mohammad Reza Ghaffariyan Abstract A project was carried out to investigate the impact of four different weighing methods on over/ under loading of forestry trucks operating in Forestry Corporation of New South Wales under two types of roads; gazetted (approved for higher legal gross vehicle weight limits) and nongazetted (standard public road gross vehicle weight limits). For all the technologies tested, it was found that there was a substantial under-loading issue ranging from 5.3 to 6.4 tonnes per load on gazetted roads, while the same technology achieved a much better outcome on nongazetted roads with a range of 1.4 tonnes under-loaded to 0.1 tonnes over-loaded on average. There was clearly a large under-loading issue on the gazetted routes. As the same operators with the same technology achieved a much more reasonable outcome on the standard access routes, these results suggest that the GVML available was technically not achievable on the gazetted routes (i.e. not enough volume available to add the weight) or the operators were not aware of or not inclined to load the extra GVML available (i.e. not certain what routes were gazetted or not). As the under load was so consistently close to the extra GVML allowed, the lack of awareness or inclination seems the most likely reasons of under load. The results point also to a more significant role for policy and methods than the technology used for in-forest weighing in achieving effective payload management in forestry haulage. Keywords: forest transport, payload management, on-board scales, loading/transport efficiency
1. Introduction Trucking is often the most expensive phase of a timber-harvesting operation, accounting for as much as 40 percent to 60 percent of the total harvesting cost (Shaffer and Stuart 1998). As a result, all possibilities for reducing the cost of trucking forest products or improving the efficiency of their transport should be examined (Bolding et al. 2009). Several factors such as payload, trip time and fuel efficiency can impact transport efficiency (Acuna et al. 2012, Ghaffariyan et al. 2013). Trucks should be loaded to their maximum legal weight every time as higher payloads will reduce transportation costs per unit which can lead to increased wood demand (Lukason et al. 2011). Load variation can be analysed by measuring load weight to determine any over-loading and underCroat. j. for. eng. 37(2016)1
loading during each transportation cycle. Over-loading may cause considerable safety issues and structural impacts on the roads, which can result in heavy fines. Under-loading will reduce transportation efficiency, which can lead to increased transportation costs. A low cost and simple technique to reduce load variability is for the truck driver to frequently communicate with loader operator to effectively estimate load weight, ideally using an on-board weighing device on the loader or truck. The three basic types of on-board weighing devices are on-board truck scales, portable platform scales, and grapple scales. Both onboard and platform scales can provide single axle and tandem weights as well as net payload weight, while grapple scales record the weight of the wood in the grapple and accumulate grapple loads to calculate net payload weight. These scales can help to increase aver-
131
M. Brown and M.R. Ghaffariyan
Timber Truck Payload Management with Different In-Forest ... (131–138)
age payload while reducing overweight fines or mill penalties (Bolding et al. 2009, Overboe et al. 1998). In a Canadian study (Dayson 2010), a comparison of the delivery point platform scale and on-board truck scale weights of each truck total payload showed the difference varied from –0.1% to 0.9%. McNeel (1990) evaluated the effect of tractor and trailer log truck scales on truck loads. On average, the mean net load weight increased by 2.07 tons when on-board electronic scales were used. Gallagher et al. (2004) analysed the difference in gross vehicle weight (GVW) between trucks that use scales (either in-woods platform scales or electronic on-board scales) and trucks that do not use scales where they found that weighed trucks in wood had higher net payload than other trucks. Variation in payload has negative consequences for both underloading and over-loading. Under-loading increases hauling cost, decreases profit, contributes to mill bottlenecks, and puts more trucks on the highway. Overloading may lead to citations, safety hazards, equipment damage and mill penalties (Bolding 2008). According to a study in the USA, the wood suppliers with the most uniform weights (less weight variations) had a hauling cost savings of 4% to 14% (Hamsley et al. 2007). The moisture content can also impact the payload of trucks and transportation cost, which can be managed to improve truck efficiency (Ghaffariyan et al. 2013). Beardsell (1986) determined using scale house (weighbridge) data and using weighing devices to buffer the problem of GVW variability. The method involved the mill setting a target GVW range, and sending reports on a systematic basis to suppliers indicating their performance relative to the performance of other mill suppliers, which could create suitable base to compare the performances. Brown (2008) studied the wood transport systems in Australia and indicated the current fleets exhibit a wide range of tare weights within each vehicle configuration indicating there is potential for considerable savings in transport costs by equipment selection and management of tare weights. Management of tare weight it primarily done at the time of vehicle specification and purchase where decisions about what components, materials and design can have significant impacts on the tare weight and hence also on the load the vehicle can transport. Maximum payload and allowable axle load can also be impacted by the quality of the roads. Improving road standards (forest roads and highways) can also reduce road user costs in areas line fuel consumption, vehicle maintenance, road maintenance, travel speed and overall productivity (load size), which can contribute to the sustainability of the forest industry and increase the total amount of fibre that can be economi-
132
cally harvested. As an example, the Minister’s Council on Forest Sector Competitiveness recommended subsidies to the forest industry for maintaining primary forest roads in Ontario, Canada (Hajek et al. 2008). Gazetting of a road is a process of assessing designated routes to determine if they can physically and safely carry a higher load than the standard classification. If all water crossings, road geometry, other users, etc. are found to be within defined safety and technical limits, which vary between locations and road types to suit the situation, the route is identified as gazetted. For this study, the exact criteria required for a road to qualify as gazetted were not provided, only what routes were gazetted and their legal GVML. Truck configuration in this study was 7-axle b-double tractor trailers. These heavy vehicles, covered in this study, are allowed an extra 5500 to 6000 kg on their gross vehicle mass limit (GVML) depending on the vehicle configuration and contractor status on gazetted roads. Non-gazetted roads have the standard GVML restrictions of 50,000 to 51,500 kg for 7-axle b-doubles depending on their configuration and contractor status. There is little information available on the effect of the weighing method or road type on the over/under loads of the forestry trucks in Australia. Thus, this project was carried out to investigate the impact of four different weighing methods on over/under load of forestry trucks operating under contract to the Forestry Corporation of New South Wales (FCNSW) on two types of roads; gazetted (approved for higher legal gross vehicle weight limits (GVML)) and non-gazetted (standard public road GVML). While in-forest on-board weight systems are very common in Australian forest transportation, not all operations use them but in the case of this study all loads were weighed upon loading in the forest.
2. Material and methods 2.1 Data collection Data was collected from existing log-haul operations in New South Wales without any prior notice to the operations to help ensure normal operations were observed. Fig. 1 shows the log truck loaded by a grapple loader with pine logs. The trees were felled and processed mechanically by harvesters. Then the logs were extracted to the road side forwarder. The logs were stacked by forwarder into piles along the road side to be loaded later by grapple loader. Using data collected and maintained for commercial purposes by the mills receiving the logs, a 12 month dataset was extracted to ensure a sufficient range of Croat. j. for. eng. 37(2016)1
Timber Truck Payload Management with Different In-Forest ... (131–138)
Fig. 1 Log truck loaded by pine logs at the forest road side being prepared to travel to mill Table 1 Descriptive statistics of recorded parameters for each truck load for gazetted roads N
Mean
Standard deviation
Gross weight, t
13 050
50.37
2.23
Tare, t
13 050
18.60
1.40
Volume, m
13 050
31.88
2.52
Nett weight, t
13 050
31.78
2.57
3
Table 2 Descriptive statistics of recorded parameters for each truck load for non-gazetted roads N
Mean
Standard deviation
Gross weight, t
40 704
49.53
1.45
Tare, t
40 704
18.67
1.45
Volume, m
40 704
30.95
2.02
Nett weight, t
40 704
30.86
2.04
3
data was obtained. The dataset included records for just over 17,700 deliveries by just over 50 individual trucks including 7-axle b-double configurations, operated by 4 contractors. Gross weights, tare weights, load volume and net weights were recorded at each mill using weighbridges certified as legal for trade. The descriptive statistics of these parameters are presented in Table 1 and Table 2 for both types of road. The forest manager also provided: Þ The gross vehicle mass limit (GVML) for each vehicle ID for both gazetted and non-gazetted routes, Croat. j. for. eng. 37(2016)1
M. Brown and M.R. Ghaffariyan
Þ The in-forest weighing system (s) used by each truck ID, Þ Status of the driver for each truck ID (hired driver or owner/operator), Þ Which routes between wood sources and mill destinations included in the data base were gazetted and non-gazetted. If any portion of the route between the logging coupe and the delivery destination was non-gazetted, the load was treated as non-gazetted. As gazetting of the route is typically targeted at operational routes, in most cases the entire route was either gazetted or not, mixed routes were very uncommon. The in-forest weighing methods were then grouped into four categories (scaling methods): Þ Loader scale – weight measured using a load cell system incorporated in the grapple of the loader, Þ Truck scale (driver) – truck based scale using either load cells or air pressure sensors integrated into the truck and trailer suspension and fifth-wheel; operated by a hired driver, Þ Truck scale (owner/operator) – truck based scale using either load cells or air pressure sensors integrated into the truck and trailer suspension and fifth-wheel; operated by the owner of the truck, Þ Loader and truck scale – both loader and truck scales being used. While the technology between air sensors, strain gauges and load cells for truck scale systems is very different, the results achieved from properly used commercial systems are very similar. Past operational observations had indicated that care and attention to use are critical to getting good performance from truck scales and that an owner-operator, being more directly motivated to get the best load performance (maximum load put more profit right in his pocket and overload fines go directly to him), tend to use the scales differently and get significantly different outcomes so the owner-operators using truck scales were examines as a different group. The datasets were then examined and outliers and corrupted entries (weight recorded was less than 60% of the GVML – i.e. half loads, vehicles having made less than 5 deliveries in the 12 month period, missing or unknown vehicle ID, missing or unknown harvest block ID and missing or unknown product ID.) were removed leaving just over 17,700 records for analysis, 13,050 on gazetted roads and 4704 on non-gazetted roads. The collected data for each category on gazetted roads included Loader and truck scale: 2861, Loader
133
M. Brown and M.R. Ghaffariyan
Timber Truck Payload Management with Different In-Forest ... (131–138)
scale: 9289, Truck scale (driver): 475 and Truck scale (owner/operator): 425. The number of observations per each category for non-gazetted roads included Loader and truck scale: 1103, Loader scale: 2593, Truck scale (driver): 467 and Truck scale (owner/operator): 541. The datasets were then statistically analysed to explore the relationship between the in-forest weighing method and over or under loading. The log length and product type were plotted versus load variation but these variables did not have any significant correlation with the load variation.
Tukey is that it shows less significant differences as it applies largest range for the multiple range tests. However, SNK method does not have such a disadvantage (Yazdi Samadi et al. 1998). In this case study, Duncan results were double checked with Tukey and SNK outcome and the statistical significance level of 5% (α=0.05) was applied in the data analysis. The null hypothesis could be expressed as follows: H0: Average under/over load of loader and truck scale = Average under/over load of loader scale = Average under/over load of truck scale (driver) = Average under/over load of truck scale (owner/operator)
2.2 Statistical evaluation The over/under load was calculated by subtracting gross weight from GVML. A frequency histogram of the over/under load data with a fitted normal curve was prepared by SPSS 21 for each road type. The normality of the data was proved by checking the frequency histograms. Then, an analysis of variance (ANOVA) was applied to test the hypothesis of equality of the average of over/under load per each scaling method. As a post-hoc test, Duncan’s multiple range test was applied to derive the homogenous subsets (Zar 1974, Yazdi Samadi et al. 1998) to compare the differences between the pair of treatments. This test could identify what treatment (in this case study means scaling method) was significantly different from the others. There are various post-hoc tests to apply including least significant difference (LSD), Student-Newman-Keuls test (SNK), Tueky, Dunnett and Duncan multiple range test. SNK method seems to be more powerful test than other methods such as least significant difference (LSD). For LSD method, in independent comparison within pairwise comparison of the treatments, for some of them the probability level (α) would be larger than determined probability. With larger number of treatments, the error will be higher. Duncan and Tukey methods do not have this disadvantage of the LSD but the disadvantage of
3. Results 3.1 Over/under load in transportation on gazetted roads The descriptive statistics of the over/under load of each scaling method have been presented in Table 3. The highest standard error (0.09 t) occurred for the over/under load data of truck scales (driver or owner/ operator types), while the lowest standard error belonged to under/over load of loader scale (0.03 t). The frequency histogram is shown is Fig. 2, which indicates the data follows a normal distribution. The skewness and kurtosis values of this data set were 0.095 and 0.438, respectively, while the mean value for over/under load was –5.66 t with a standard deviation of 2.46 t. The null hypothesis was rejected because there were significant differences among the means of over/ under loads for different scaling methods, for data collected on gazetted roads (Table 4). There was no significant difference between the means of the over/ under loads between loader and truck scale vs. truck scale (driver). However, both these groups were significantly different from loader scale and truck scale (owner/operator) in terms of the means of the over/ under loads (Table 5). Application of Tukey and Stu-
Table 3 Descriptive statistics for under/over load (t) for gazetted roads Scaling method
N
Mean
Std. deviation
Std. error
Loader and truck scale
2861
–5.31
2.25
Loader scale
9289
–5.74
Truck scale, driver
475
Truck scale, owner/operator Total
134
95% confidence interval for mean Lower bound
Upper bound
0.04
–5.40
–5.23
2.56
0.03
–5.79
–5.69
–5.48
1.97
0.09
–5.66
–5.31
425
–6.44
1.84
0.09
–6.62
–6.27
13,050
–5.66
2.46
0.02
–5.70
–5.62
Croat. j. for. eng. 37(2016)1
Timber Truck Payload Management with Different In-Forest ... (131–138)
Fig. 2 Frequency histogram for data on gazetted roads
M. Brown and M.R. Ghaffariyan
Fig. 3 Means of under-loads for four types of scaling methods on gazetted roads
Table 4 Analysis of variance for gazetted roads Sum of squares
df
Mean square
F
Sig.
676.85
3
225.62
37.43
0.00
Within groups
78636.48
13046
6.03
–
–
Total
79313.34
13049
–
–
–
Between groups
Table 5 Homogeneous Subsets obtained by Duncan method for gazetted roads N
Subset for alpha = 0.05 1
2
3
Truck scale, owner/operator
425
–6.44
–
–
Loader scale
9289
–
–5.74
–
Truck scale
475
–
–
–5.48
Loader and truck scale
2861
–
–
–5.31
–
1.00
1.00
0.16
Sig.
Fig. 4 Frequency histogram for data on non-gazetted roads
3.2 Over/under load in transportation on nongazetted roads dent-Newman-Keuls approaches offered similar results to the Duncan. Loader and truck scale had the lowest mean under load (–5.31 t), while truck scale (owner/operator) usage resulted in the largest mean under load (–6.44 t) (Fig. 3). Croat. j. for. eng. 37(2016)1
Table 6 includes descriptive statistics of over/under load for each scaling method in transportation on nongazetted roads. Over/under loads from using loader scale and truck scale (owner/operator) had the lowest standard error (0.03 t), while the largest standard error
135
M. Brown and M.R. Ghaffariyan
Timber Truck Payload Management with Different In-Forest ... (131–138)
Table 6 Descriptive statistics for over/under load (t) for non-gazetted roads Scaling method
N
Mean
Std. deviation
Std. error
Loader and truck scale
1103
–1.43
1.68
Loader scale
2593
–0.53
Truck scale, driver
467
Truck scale, owner/operator Total
Lower bound
Upper bound
0.05
–1.53
–1.33
1.50
0.03
–0.59
–0.47
–0.54
1.39
0.06
–0.66
–0.41
541
0.09
0.59
0.03
0.04
0.14
4704
–0.67
1.53
0.02
–0.72
–0.63
Table 7 Analysis of variance for non-gazetted roads Sum of squares
df
Mean square
F
Sig.
1008.73
3
336.24
156.69
0.00
Within groups
10,085.46
4700
2.15
–
–
Total
11,094.19
4703
–
–
–
Between groups
Table 8 Homogeneous subsets obtained by Duncan method for non-gazetted roads N
Subset for alpha = 0.05 1
2
3
Loader and truck scale
1103
–1.43
–
–
Truck scale, driver
467
–
–0.54
–
Loader scale
2593
–
–0.53
–
Truck scale, owner/operator
541
–
–
0.09
–
1.00
0.96
1.00
Sig.
95% confidence interval for mean
tor) in terms of load variation (Table 8) (similar results were achieved by Tukey and Student-Newman-Keuls methods). The loader and truck scale had the largest mean under load (–1.43 t), while the truck scale (owner/ operator) resulted in a small mean over load (+0.09 t) (Fig. 5).
4. Discussion and conclusions The results of this study indicated that the mean under loading varied from 0.5 t to 6.4 t for both types of roads and the mean over loading occurred only in one case (0.1 t). These results contrast with those of an American case study (McNeel 1990), where an increase of 2.1 t mean load weight was achieved through using on-board electronic scales. Although in our case
occurred in the case of truck scale (driver) application, which was about 0.06 t. The frequency histogram of the data is shown in Fig. 4. The skewness and kurtosis values of this data set were –1.64 and 20.58, while the mean value for over/under load was –0.67 t with a standard deviation of 1.53 t. For operations on non-gazetted roads, there was a significant difference between the means of over/under loads for different scaling methods (Table 7). There was no significant difference between means of over/ under load for using loader scale and truck scale (driver). However, both groups were different from loader and truck scale (driver) and truck scale (owner/opera-
136
Fig. 5 Means of under/over loads for four types of scaling methods on non-gazetted roads Croat. j. for. eng. 37(2016)1
Timber Truck Payload Management with Different In-Forest ... (131–138)
study it is not clear whether introduction of on-board scales increased mean load weights, it is clear that maximum payloads have not been reached in most cases. In our case study, if the under loading of the trucks could be eliminated, then the potential saving for the company could be between $3 million and $7 million ($0.60/m3 – $1.20/m3). Significant savings can be achieved through eliminating load variation. Beardsell (1986) found gross annual savings were $153,000 and $431,000 for two different mills, and Deckard et al. 2011 predicted the potential impact on the southern United States wood supply chain at between $44.1 million and $87.1 million. In comparing the study results for the two types of roads, there is clearly a substantial under loading issue on the gazetted roads as compared to the non-gazetted roads. As the same operators with the same technology achieved a much better outcome on the non-gazetted roads, these results suggest that the GVML available was technically not achievable on the gazetted roads (i.e. not enough truck volume capacity available to add the extra weight) or the operators were not aware of or not inclined to load the extra GVML available (i.e. not certain what routes were gazetted or not). The log length and product type did not have any significant correlation with the load variation, which can explain these factors did not influence the under and over loads in this case study. As the data were collected post-operations without any direct observations of the operations, not much insight can be given on the reason, but since the under load is so consistently close to the extra GVML allowed on gazetted roads, the lack of awareness or inclination seems the most likely. Post study discussions with the operations managers revealed that the rate system at the time of the study had no incentive for the operators to load the higher weights on gazetted roads (same 4/t-km rate for both route types so revenue targets are met with non-gazetted loads on both route types), which is an issue that has been addressed with new contract arrangements. A follow-up study is being considered to explore the influence of this new contract arrangement. To remove any load variation, the transport management should consider applying an accurate weighing method at the loading areas before departing the trucks to mills. Distributing the responsibility of weigh measurement over two operators (such as loader and truck operator in this case study) should be avoided, since the study showed the shared responsibility gave poorer results (one assuming the other is looking after it). Where routes have been identified to allow higher loads, operations need to ensure information is readily available to drivers, vehicles asCroat. j. for. eng. 37(2016)1
M. Brown and M.R. Ghaffariyan
signed to the route are able to take advantage of the potential productivity gain of the higher load and contract arrangements are in place so all stakeholders are appropriately motivated to take advantage of the potential productivity gain through increased payloads. Looking at the three results together, it appears the benefit of using the loader scale and truck scale in combination is not necessarily realised in practice with the non-gazetted roads, showing a considerably poorer outcome than when the technologies are used separately. This suggests work methods and techniques can play as great a role, if not greater, than the technology itself, and should be explored in future research. Further research on the performance of the different weighing technologies, under different policy frameworks and methods of usage, need to be explored to better understand how best to achieve efficient payload management in Australian forest haulage operations.
5. References Acuna, M., Mirowski, L., Ghaffariyan, M.R., Brown, M., 2012: Optimising transport efficiency and costs in Australian wood chipping operations. Biomass and Bioenergy 46: 291–300. Beardsell, M.G., 1986: Decreasing the cost of hauling timber through increasing payload. Virginia Polytechnic Institute and State University, PhD dissertation, Blacksburg, VA. 133 p. Bolding, M.C., 2008: Improving trucking efficiency. Sawmill & Woodlot, 26–31. Bolding, M.C., Dowling, T.N., Barrett, S.M., 2009: Safe and Efficient Practices for Trucking Unmanufactured Forest Products. Virginia Cooperative Extension publication, 420– 310. Brown, M., 2008: The impact of tare weight on transportation efficiency in Australian forest operations. CRC for Forestry, Bulletin 3: 1–4. Dayson, P., 2010: Accuracy of log truck onboard weigh scales. Timber measurements society, Portland, Oregon, April 7–9, (http://www.timbermeasure.com). Deckard, D.L., Newbold, R.A., Vidrine, G.G., 2003: Benchmark round wood delivery cycle-times and potential efficiency gains in the southern United States. Forest Products Journal 53(7): 61–69. Gallagher, T., McDonald, T., Smidt, M., Tufts, R., 2004: Let’s talk trucking: weights and loading methods. Technical Paper 05-P-2, Forest Resources Association, Inc. Ghaffariyan, M.R., Acuna, M., Brown, M., 2013: Analysing the effect of five operational factors on forest residue supply chain costs: A case study in Western Australia. Biomass and Bioenergy 59: 486–493.
137
M. Brown and M.R. Ghaffariyan
Timber Truck Payload Management with Different In-Forest ... (131–138)
Hajek, J., Hein, D., Swan, D., 2008: Transportation of raw forest products in northern ontario by trucks. 10th International Symposium on Heavy Vehicle Transport Technology. HV Paris, May 19–22, 377–386. Hamsley, A.K., Greene, W.D., Siry, J.P., Mendell, B.C., 2007: Improving timber trucking performance by reducing variability of log truck weights. Southern Journal of Applied Forestry 31(1): 12–16. Lukason, O., Ukrainski, K., Varblane, U., 2011: Economic benefit of maximum truck weight regulation change for Estonian forest sector. Discussions on Estonian Economic Policy No. 2/2011: 87–100.
Shaffer, R.M., Stuart, W.B., 1998: A checklist for efficient log trucking. Virginia Cooperative Extension publication 420– 094. http://pubs.ext.vt.edu/420-094 (Accessed 7 Nov 08). Overboe, P.D., Shaffer, R.M., Stuart, W.B., 1998: A low-cost program to improve log truck weight control. Forest Products Journal 38(6): 51–54. Yazdi Samadi, B., Rezaei, A., Valyzadeh, M., 1998: Statistical designs in agricultural research. Second edition. Tehran University Publications No. 2346, 764 p. Zar, J.H., 1974: Biostatistical analysis. Eaglewood Cliffs, NJ, Prentice Hall Inc. 620 p.
McNeel, J.F., 1990: Analysis of truck weight modifications for a southern timber hauling operation. Southern Journal of Applied Forestry 14(3): 133–136.
Authors’ addresses:
Received: November 12, 2014 Accepted: February 5, 2015
138
Prof. Mark Brown, PhD. e-mail: mbrown2@usc.edu.au Mohammad Reza Ghaffariyan, PhD. * e-mail: ghafari901@yahoo.com University of the Sunshine Coast Locked Bag 4 4558 Maroochydore, Queensland AUSTRALIA * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope Andreja Đuka, Tibor Pentek, Dubravko Horvat, Tomislav Poršinsky Abstract Skidder mobility during timber extraction is defined by: 1) basic dimensional features of the vehicle, 2) ability to overcome obstacles during movement, 3) traction performance and 4) environmental soundness. Traction performance depends on the ground conditions (soil bearing capacity) and the total effect of all forces on the vehicle. In downhill skidding, the skidder is under great influence of parallel component of forces, adhesion weight and longitudinal terrain slope, which combined result in negative traction force, torque and thrust force. When the horizontal component of rope force is equal to zero i.e. the moment when the weight of the load and resistance to traction are in equilibrium, the slope angle α is a function of load mass distribution factor and skidding resistance factor. This is a »turning point« that can be defined as a critical slope because the load starts to push the vehicle downhill, which results in negative horizontal component of rope force. Depending on skidder Ecotrac 120V dimensional features, centre of gravity, load mass distribution factor, skidding resistance factor of previous research, five different loads were analyzed (1 to 5 tonnes) in order to define the critical slope angle for each of them. Critical slope for downhill skidding of 1 tonne timber is on longitudinal slope of –26%, for 2 tonne timber on –30%, 3 tonne timber on –34%, 4 timber on –38% and for 5 tonne timber on –43% of terrain longitudinal slope. Even though skidding bigger load increases vehicle mobility to even greater slope angles, the most important in downhill skidding, is to avoid blocking of the wheels, which will lead to a complete vehicle slippage and the driver must be constantly aware of that fact. The general recommendation should be that skidding small loads (1 to 3 tonnes) downhill is suitable for smaller longitudinal terrain slopes (up to maximum –34%), while the heavier the load, the further down the slope the skidder can go. The load of 5 tonnes »anchors« the skidder better and therefore it can go on terrain slopes up to –43%, during which less traction force is used (torque is used for braking) and skidder pulls the load by its own weight. It can be concluded that extending the operating range of skidder onto steeper slopes with heavier loads has the potential to decrease harvesting costs and increase productivity. Keywords: skidder, downhill timber extraction, rope force, critical slope
1. Introduction Terrain trafficability is a terrain property that allows vehicle mobility, during which various terrain factors (slope, ground obstacles, soil bearing capacity) show their influence (Janosi and Green 1968, Eichrodt 2003, Suvinen 2006, Lubello 2008). From the standpoint of timber harvesting and forest opening, terrain slope is the most important terrain factor affecting the choice of a harvesting system. Terrain slope affects vehicle stability because all wheels (i.e. tracks) are »in Croat. j. for. eng. 37(2016)1
conflict« with the same macro-topographic conditions. Skidder mobility is its ability to move from point A to point B while achieving its primal goal – timber transport. In timber extraction, vehicle mobility can be considered from two different aspects: 1) extraction on soils of limited bearing capacity (for example lowland forests on gley soils) and 2) extraction in hilly and mountainous forests, where slope and ground obstacles define conditions for application of specialised forestry vehicles. Many parameters define vehicle mobility during timber extraction (Šušnjar 2005, Šušnjar
139
A. Đuka et al.
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
et al. 2010, Poršinsky et al. 2012), of which these four are the most important ones: 1) basic dimensional features of the vehicle (dimensions, turning radius, mass, centre of gravity, longitudinal and lateral angle of stability, clearance, frame and axle oscillation, unloading of the front axle, payload of rear axle, tyres load capacity), 2) the ability to overcome obstacles during movement (ground clearance and lateral vehicle stability), 3) traction performance (dependence of slip, traction power and speed to traction force and soil bearing capacity) and 4) environmental soundness (nominal ground pressure and minimal cone index). Many scientists determined critical terrain slopes for a skidder between 30% and 50%, regardless of extraction direction (MacDonald 1999, Heinimann 1999), while others differentiate between downhill and uphill skidding. So, critical slope in downhill skidding ranges from 23% to 50% (Rowan 1977, Inoue and Tsuji 2003, Lubello 2008) and in uphill skidding from 18% to 30% (Rowan 1977, Inoue and Tsuji 2003, Lubello 2008). Some highlight the importance of load size such as Hippoliti and Piegai (2000), as quoted by Lubello (2008), who reported that an unloaded skidder can overcome the maximum gradient of 40%, but loaded only up to 20% regardless of slope direction. Eger and Kiencke (2003) reported that the effect of dynamic changes in load should be also considered as key factors that affect machine stability. Sarles and Luppold (1986) state that when skidding up the slope, for any increase in the terrain slope of 1% (above the terrain inclination of 10%), the quantity of hooked timber should be reduced by 2.5%. Other scientists emphasise the importance of secondary forest network. According to Heinimann (1999) if skidder is extracting timber on terrain slopes higher than 35%, it should move only on secondary forest road network. Hippoliti and Piegai (2000) note the possibility of skidding timber down the slope of 60%, but only in the case of well-designed and built strip roads. Importance of ground obstacles and soil bearing capacity of forest stand during timber extraction by ground based vehicles is highlighted by Kühmaier and Stampfer (2010). Tendency of anchoring vehicles for timber extraction, and thus moving critical terrain slopes to even higher extents, has become more and more popular in the past couple of years. Sauter et al. (2012) define critical terrain slope as 55% for the skidder with a crane equipped with the additional winch for anchoring the vehicle, and Cavalli (2015) surmised that wheeled machines with chains or bands might have an upper limit of 45%, integral track machines up to 60%, and that tethered machines should be able to operate up to a range of 75 to 85% terrain longitudinal slope.
140
Besides dimensional characteristics defined in ISO standard 13861 (2000), some authors (Bekker 1969, Janosi and Green 1968, Sever and Horvat 1985, USA Code of Federal regulations 49 CFR 523.2) give additional characteristics that allow bypassing and overriding of macro (slope) and micro (ground obstacles) terrain properties during vehicle off-road movement: 1) approach angle (the smallest angle, in a plane side view of a vehicle, formed by the level surface on which the vehicle is standing and a line tangent to the front tyre static loaded radius arc and touching the underside of the vehicle forward of the front tyre), 2) departure angle (the smallest angle, in a plane side view of a vehicle, formed by the level surface on which the vehicle is standing and a line tangent to the rear tyre static loaded radius arc and touching the underside of the vehicle rearward of the rear tyre), 3) break-over angle (means the supplement of the largest angle, in the plan side view of a vehicle that can be formed by two lines tangent to the front and rear static loaded radii arcs and intersecting at a point on the underside of the vehicle), 4) longitudinal clearance diameter (diameter of a circle that touches the inner side of the tyres from each axle and the lowest hanging point under a vehicle), 5) transverse clearance diameter (diameter of a circle that touches the inner side of the tyres and the lowest hanging point under a vehicle, usually a differential) and 6) centre of gravity position (height from ground, distance from front and rear axles), which is an important constructional parameter that influences load distribution on axles depending on terrain slope during timber extraction. Visser and Berkett (2015) state that, according to Bell (2002), McMahon (2006) and Raymond (2010), extending the operating range of ground-based machinery onto steep slopes has the potential to decrease harvest costs and improve safety. The same authors conclude in their study of 22 machines and effect of terrain steepness during harvesting, that machines exceed slope limits commonly associated with harvesting operations, and exceed them often and for longer periods of time, which is in accordance with Visser and Stampfer (2015), who claim that today there is no guidance on slope limits, based on either science or experience. Authors conclude that many guidelines refer to manufacturer’s specifications, yet few of the major forestry equipment manufacturers provide slope and/or operating limits for their purpose built machinery. The goal of defining limiting terrain slopes for downhill timber extraction of cable skidders should be considered as guidelines for operators and planners, who can then, depending on load size and terrain Croat. j. for. eng. 37(2016)1
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
A. Đuka et al.
Fig. 1 Distribution of forces during timber skidding Croat. j. for. eng. 37(2016)1
141
A. Đuka et al.
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
Table 1 Equations of some parameters of downhill and uphill timber skidding Downhill skidding
Uphill skidding Adhesive weight
Ga = G ⋅ cos a + V
(1)
Vertical component of rope force
V = k ⋅ Q ⋅ cos a
(2)
Horizontal component of rope force
H = Q ⋅ (1 − k ) ⋅ cos a ⋅ m p − Q ⋅ sin a
(3)
H = Q ⋅ (1 − k ) ⋅ cos a ⋅ m p + Q ⋅ sin a
(4)
Front axle load
G1 =
G ⋅ cos a ⋅ a + G ⋅ sin a ⋅ ht − H ⋅ d − V ⋅ c (5) L
G1 =
G ⋅ cos a ⋅ a − G ⋅ sin a ⋅ ht − H ⋅ d − V ⋅ c (6) L
Rear axle load
G2 =
G ⋅ cos a ⋅ b − G ⋅ sin a ⋅ ht + H ⋅ d + V ⋅ ( L + c) (7) L
G2 =
G ⋅ cos a ⋅ b + G ⋅ sin a ⋅ ht + H ⋅ d + V ⋅ ( L + c) (8) L
Drawbar pull
Fv = H − Ga ⋅ sin a
(9)
macro characteristics (slope), define better routes for skidder off-road movement providing better control and manoeuvrability of vehicles.
2. Theoretical Approach During skidding, timber is partially suspended on the vehicle i.e. one part of the load is lifted above ground level and hanged by rope to the rear end of the skidder, while the other part is dragged (trailed) on the ground. Since a part of the load is on ground, only a part of the load weight is actually carried by the skidder rope. While skidding, the force in the rope that carries a part of the timber weight is the so called vertical component of rope force (V), and force that must overcome tractive resistance of timber that is on the ground is called horizontal component of rope force (H). During skidding, the adhesion weight of the skidder is greater than its static weight as the rear axle of the vehicle is under additional influence of the load, while the vertical component of rope force shows its effect. Theoretical approach to distribution of forces during skidding was established by Bennet (1962), who differentiated horizontal, vertical and frictional forces involved in timber skidding of different loads, and since then many scientists used them in their own re-
142
Fv = H + Ga ⋅ sin a
(10)
search (Calvert and Garlicki 1967, Richardson and Cooper 1970, Hassan 1977, Perumpral et al. 1977, Sever 1980, Matthes and Watson 1981, Hassan and Sirois 1983, Hassan and Gustafson 1983, Iff et al. 1984, Horvat 1990, Sever and Horvat 1995, Šušnjar and Horvat 2006, Tomašić et al. 2007, Tomašić et al. 2009, Šušnjar et al. 2010, Poršinsky et al. 2013). Skidding timber on flat terrain begins in the moment when thrust force (brought by transmission system to the wheels) begins to overcome resistance forces (Fig. 1A): 1) skidder rolling, 2) rolling of hooked timber and 3) friction of timber on the ground. During skidding up the slope (Fig. 1C), load distribution becomes more complex and traction begins when thrust force overcomes resistance forces: 1) skidder rolling, 2) terrain slope, 3) rolling of hooked timber, 4) overcoming terrain slope of hooked timber, 5) friction of timber on the ground and 6) overcoming terrain slope of timber on the ground. While skidding timber down the slope (Fig. 1B), thrust force overcomes the same resistance as for skidding timber up the slope, only resultants of the three forces of resistance (terrain slope, overcoming terrain slope of hooked timber, overcoming terrain slope of timber on the ground) are now in the opposite direction, i.e. direction of the vehicle movement. Croat. j. for. eng. 37(2016)1
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
Since skidder movement dynamics is considerably different depending on extraction direction, forces distribution and relating equations are presented in Table 1 for: adhesive weight, vertical component of rope force, horizontal component of rope force, front axle load, rear axle load and drawbar pull (traction force). Load mass distribution factor (k) shows how much load mass is lifted from the ground (hooked on the rope) and how much is pulled on the ground surface (Eq. 11). If the load mass distribution factor is 0.5, this means that the same part of the timber mass is hooked by rope as it is pulled on the ground. Authors (Sever 1980, Hassan and Gustafson 1983, Hassan and Sirois 1983, Iff et al. 1984, Horvat 1987, Šušnjar 2005, Tomašić 2007, Poršinsky et al. 2012) reported that the nature of loading and load mass distribution factor depend on these variables: tree diameter and slenderness ratio, number of trees per load, height of suspended butt above ground, tree form (method of timber processing), timber orientation (thinner or thicker end is above ground). If the load increases, the portion of its weight supported by the ground increases at higher percentage. This increase is also attributed to the butt height above ground, which tends to decrease as the number of trees in the load increases. Tree weight on ground contact length decreases and load mass distribution factor increases with the increase in tree semisuspension height above ground. Load mass distribution factor is unaffected by tree length of up to 20 m.
k=
V (11) Q ⋅ cos a
Skidding resistance occurs due to the effect of load weight pulled on the ground and skidding resistance factor – μp (Hassan 1977, Perumpral et al. 1977, Sever 1980, Hassan and Gustafson 1983, Hassan and Sirois 1983, Samset 1985, Šušnjar 2005, Tomašić 2007). Samset (1975) according to Megille (1954) stated that skidding resistance factor depends on soil type and moisture level, and Samset (1975) according to Dahl (1973) claimed that it also depends on orientation of suspended timber (thinner or thicker end is above ground) and on timber processing method (full-tree, half-tree, etc.). The horizontal component of rope force overcomes the skidding resistance between the load and forest soil and according to known values of force, weight, load mass distribution factor and terrain slope, skidding resistance factor can be determined (Eq. 12).
H ± Q ⋅ sin a mp = Q ⋅ (1 − k ) ⋅ cos a
(12)
In exploring skidder traction features during skidding down the slope, Šušnjar et al. (2010) give some Croat. j. for. eng. 37(2016)1
A. Đuka et al.
limitations identified through two »turning points« of terrain slope. The first »turning point« is determined by the angle of inclination of the terrain in which vehicle no longer achieves positive traction and breaking force i.e. thrust force is equal to zero (Eq. 13).
tg a =
G ⋅ f + Q ⋅ k ⋅ f + (1 − k ) ⋅ Q ⋅ m p G+Q
(13)
The second »turning point« is determined by the angle of terrain inclination in which hooked timber starts to push the skidder down the slope (Eq. 14), which occurs at the time when the horizontal component of the rope force in the rope is equal to zero (H = 0), or when the weight of the load (Q sin α) and traction resistance are in balance.
tg a = (1 − k ) ⋅ m p
(14)
3. Materials and Methods Valid model of skidder–terrain interaction will permit forestry researchers to study and analyse many issues and problems related to skidder performance under a wide range of conditions (different loads, various terrain characteristics, etc.). This way, skidder optimisation and improvement of its operational parameters can be expected. Significance of skidders parameters that affect its off-road performance can be identified without expensive field testing. The results will not only help forestry planners in better forest management, but also practitioners in real-life situations of a skidder off-road locomotion. Analysis was done based on skidder Ecotrac 120V dimensions and centre of gravity (Šušnjar 2005), dependence of skidder Ecotrac 120V load mass distribution factor and skidding resistance factor to affecting parameters (Poršinsky et al. 2012), load distribution during timber extraction on different terrain slopes and five different loads (from 1 to 5 tonnes). Load mass distribution factor (Eq. 15) is a function of (statistically and inversely correlated) load mass, load weight, number of logs per load, load volume. Skidding resistance factor (Eq. 16) is a function (statistically and inversely correlated) of terrain slope and direction of timber extraction i.e. uphill or downhill skidding.
k = 0.62017 − 0.0476 ⋅ Q
(15)
m p = 0.50529 − 0.042 ⋅ a
(16)
Where: Q – Load mass, t α – Longitudinal terrain slope,% (+, – indicate direction of skidding)
143
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
Fig. 2 Technical data of Ecotrac 120V skidder
A. Đuka et al.
144
Croat. j. for. eng. 37(2016)1
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
A. Đuka et al.
Fig. 3 Slope and load influence on skidder adhesion weight and its distribution on both axles Load distribution was determined by calculation and analyses of the following parameters: 1) adhesion weight (Eq. 1), 2) vertical component of rope force (Eq. 2), 3) horizontal component of rope force (Eq. 3), 4) load distribution on front axle (Eq. 5), 5) load distribution on rear axle (Eq. 7), 6) angle of terrain inclination in which hooked timber starts to push the skidder down the slope (Eq. 14), and 7) traction force (Eq. 9). Skidder Ecotrac 120V is a four-wheeled (4×4) articulated forestry vehicle, equipped with a hydraulic Croat. j. for. eng. 37(2016)1
forest winch Hittner 2×80, of the nominal tractive force of 80kN. It is driven by a 6 cylinder diesel DEUTZ engine with the nominal power of 84 kW at 2300 min–1 and maximum torque of 400 Nm at 1500 min–1. Basic technical data of Ecotrac 120V skidder is given in Fig. 2.
4. Results and Discussion Skidder traction performance and force distribution during timber extraction depends on gained forc-
145
A. Đuka et al.
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
es on wheels and forces resisting them, where adhesion weight is a very important parameter. It actually represents the sum of vertical loads on driving wheels during skidding (Fig. 3A). Adhesion weight depends on skidder weight (G), longitudinal terrain slope (α) and the size of the vertical rope force component (V), which is directly influenced by load weight (Q). Adhesion weight is different than empty skidder weight (G) because skidder rear axle is additionally loaded with the full amount of the vertical rope force component (V) that is dispersed to rear wheels through horizontal rollers of the winch. Results of modelling load distribution on skidder axles, on the example of skidder Ecotrac 120V, pointed out that load distribution varies due to the amount (mass) of hooked timber, timber extraction direction (uphill or downhill) and due to longitudinal terrain slope (Fig. 3B and 3C). By increasing longitudinal terrain slope and load mass during uphill timber skidding, there is an increase of load on rear skidder axle due to the growth of the horizontal component of skidder weight (G sin α), which acts against the direction of vehicle movement, and due to the growth of the horizontal component of rope force (H). Axle load distribution of the skidder, during uphill timber extraction, is related to many criteria (limits) derived from previous research: 1) Unloading of the
front axle (Weise and Nick 2003), where at least 10% of the total dynamic load should remain on the front axle (G1 > 0.1 Ga) to retain control; 2) Overloading of the rear axle (Horvat 1990), whereby the load of the skidder rear axle must not exceed the total weight of the skidder (G2 < G); 3) Longitudinal skidder stability (Sever 1980), which is defined as the minimum ratio of load on front and rear axles (G1 : G2> 1 : 3.5), after which longitudinal stability of the vehicle becomes an issue; 4) Permitted tyres load capacity, with regard to the air pressure recommended by the manufacturer (Đuka 2014). In downhill skidding, the load is transferred from the rear to the front axle of a skidder. Increasing terrain slope leads to the growth of load on the skidder front axle due to an increase in the horizontal component of skidder weight (G sin α), which acts in the direction of skidder movement. Increasing the quantity (mass) of hooked timber in downhill skidding will lead to the reduction of the load on the front skidder axle, because of the increase of the vertical component of rope force (V). It is hard to understand the dynamics of load distribution on skidder axles regarding weight (mass) of hooked timber, direction of skidding (uphill/downhill) and slope inclination (Fig. 3B and 3C) without knowing the effect of rope force i.e. its vertical component (V) that carries the hooked load, and its horizontal
Fig. 4 Load and slope influence on horizontal and vertical components of rope force
146
Croat. j. for. eng. 37(2016)1
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
A. Đuka et al.
component of rope force is smaller because less load weight is pulled on the ground.
Fig. 5 Load and slope influence on traction force component (H) which overcomes tractive resistance of the load on the ground. The analysis of horizontal and vertical components of rope force according to longitudinal terrain slope, skidding direction and load mass is shown in Fig. 4. During downhill extraction of timber, the horizontal component of rope force (Fig. 4A) decreases with the increase of terrain slope and load mass, while the vertical component of rope force (Fig. 4B) increases only by increase of load mass i.e. slightly decreases with the increase of terrain slope (due to reduction of the load that is hooked by winch rope and increase of the load of timber on the ground). During downhill skidding, the horizontal component of rope force is greater than the vertical component of rope force for load mass of 1 t and terrain slope higher than 45%, for load mass of 2 t and terrain slope higher than 36%, for load mass of 3 t and terrain slope higher than 27%, for load mass of 4 t and terrain slope higher than 19%, for load mass of 5 t and terrain slope higher than 10%. Throughout downhill skidding, the horizontal component of rope force decreases with the increase of terrain slope and with the reduction of load, by which the vertical component of rope force is always greater than the horizontal component (Fig. 4). The horizontal component of rope force decreases during downhill skidding because load tends to get closer to rear end of the skidder, which makes the vertical component of rope force more important because it holds the load above the ground. Therefore, the horizontal Croat. j. for. eng. 37(2016)1
An important criteria in downhill skidding is terrain slope inclination (α) when the load starts to push the skidder i.e. the moment when the horizontal component of rope force is zero (H=0). When the load pushes the vehicle down the slope, due to the constant thrust of the timber at the back end of a skidder, it can be concluded that, in due time, such performance will result in fatigue of the material and early damage to the vehicle (according to FAO operating hours for wheeled skidder it is between 8,000 and 12,000 depending on operation conditions). It will also have negative influence on psycho-physical state of the driver (as conformed in patent EP2711226 A1 (Eskilsons 2014), in the vehicle-driver interactions, it is essential that the vehicle carries out the driver’s commands in the manner believed to be desired by the driver). The turning point when skidding is no longer recommended for skidding loads up to 1 t is on terrain with longitudinal slope of –26%, for skidding loads up to 2 t on terrain with longitudinal slope of –30%, for skidding loads up to 3 t on terrain with longitudinal slope of –34%, for skidding loads up to 4 t on terrain with longitudinal slope of –38% and for skidding loads up to 5 t on terrain with longitudinal slope of –43%. In uphill skidding, traction force needs to overcome the resistance of the load on the ground (H), but also the resistance of the horizontal component of skidder weight (G sin α), which pulls the vehicle in the opposite direction. With the growth of the inclination angle, traction force grows with the increase of load weight, due to an increase of the horizontal component of rope force (traction resistance) and the weight of the skidder that needs to overcome traction force (Fig. 5). In downhill skidding, the horizontal component of the skidder weight (G sin α) acts in the direction of the skidder and due to its action the skidder overcomes traction resistance of the load on the ground (H), which causes the appearance of negative traction force (Fig. 5) i.e. appearance of braking force. Results of modelling load distribution on skidder front and rear axles, horizontal and vertical components of rope force, based on dimension characteristics of skidder Ecotrac 120V (centre of gravity), knowing load distribution and skidder resistance factors, considering different quantity (mass) of hooked timber, extraction direction (uphill and downhill extraction) and longitudinal terrain slope, are in accordance with previous research that were based on field testing (Šušnjar 2005, Šušnjar and Horvat 2006, Tomašić 2007, Tomašić et al. 2007, Tomašić et al. 2009, Šušnjar et al. 2010).
147
A. Đuka et al.
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
5. Conclusions It can be stated that during downhill skidding no real traction force can be achieved (torque is used for braking), because the skidder pulls the load by its own weight, and also the transfer of power from the motor to the wheels is used for braking due to the large impact of parallel component of the skidder weight. Even though skidding is possible on even greater slope angles than stated above, the most important in downhill skidding is to avoid blocking of the wheels, which will lead to a complete vehicle slippage. When the load pushes the vehicle down the slope, due to the constant thrust of the timber at the back end of a skidder, it can be concluded that, in due time, such performance will result in fatigue of the material and early damage to the vehicle as well as in negative influence on the driver. The general recommendation should be that skidding small loads (1 to 3 tonne) downhill is suitable for smaller longitudinal terrain slopes (up to maximum –34%), while the heavier the load, the further down the slope skidder can go. The load of 5 tonnes »anchors« the skidder better and, therefore, it can go on terrain slopes up to –43%, during which less traction force is used (torque is used for braking) and skidder pulls the load by its own weight. It can be concluded that extending the operating range of skidder onto steeper slopes with heavier loads has the potential to decrease harvesting costs and increase productivity.
6. References 49 CFR 523.2. USA Code of Federal regulations, available at: https://www.law.cornell.edu/cfr/text/49/523.2
Situation and Future Challenges«, March 18–20, 2015, Zagreb, Croatia (retrieved from www.crojfe2015.com/home). Đuka, A., 2014: Development of Terrain trafficability Model for Planning Timber Extraction by Skidder. Dissertation thesis, Faculty of Forestry University of Zagreb, 1–302 (in Croatian). Eger, R., Kiencke, U., 2003: Modeling of rollover sequences. Control Engineering Practice 11: 209–216. Eichrodt, A.W., 2003: Development of a Spatial Trafficability Evaluation System. Dissertation, ETH, Zurich, 1–165. Eskilsons, A., 2014: Method, computer program, control device, system and vehicle with such a system for measuring a physiologic property of a driver and for adaptating control of a clutch. European Patent Application no. EP2711226 A1: 1–22. Hassan, A.E., 1977: Trafficability Study of a Cable Skidder. Transactions of the ASAE 20(1): 26–29. Hassan, A.E., Gustafson, A.L., 1983: Factors Affecting Tree Skidding Forces. Transactions of the ASAE 81–1586: 47–53. Hassan, A.E., Sirois, D.L., 1983: Weight Distribution Characteristics of Semi-Suspended Trees. Transactions of the ASAE 83–2605: 1291–1297. Heinimann, H.R., 1999: Ground-based Harvesting Systems for Steep Slopes. Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium, Corvallis OR, USA, March 28–April 1, 1999. 1–19. Hippoliti, G., Piegai, F., 2000: Technice e sistemi di lavoro per la raccolta del legno. Compagnia delle Foreste, Arezzo, 1–157. Horvat, D., 1987: Skidder Wheel Torque Measuring. Proceedings of the 9th ISTVS International conference, Barcelona, Italy, 2: 531–541. Horvat, D., 1990: Defining Traction Characteristics of a Forestry Tractor – Skidder. Meh. šumar. 15(7–8): 113–118 (in Croatian).
Bekker, M.G., 1969: Introduction to Terrain Vehicle Systems. University of Michigan Press, Ann Arbour, MI, USA, 1–520.
Horvat, D., Spinelli, R., Šušnjar, M., 2005: Resistance Coefficients on Ground-Based Winching of Timber. Croatian journal of Forest Engineering 26(1): 3–11.
Bell, J.L., 2002: Changes in Logging Injury Rates Associated with Use of Feller-Bunchers in West Virginia. J Safety Res. 33(4): 463–471.
Iff, R.H., Koger, J.L., Burt, E.C., Culver, E.W., 1984: C-A-R-TS: Capacity Analysis of Rubber-Tired Skidders. Transactions of the ASAE 82–1594: 660–664.
Bennett, W.D., 1962: Forces Involved in Skidding Full Trees and Tree-length Loads of Pulpwood. Pulp and Manager Magazine of Canada, Woodland Section Index No. 2162: 322–327.
Inoue, M., Tsuji, T., 2003: Management, Technology and System Design of Mechanized Forestry in Japan. Textbook of forestry mechanization technology, Forestry Mechanization Society, Akasaka, Minato-ku, Tokyo, Japan, Forestry Machine Series No. 92, 1–122.
Calvert, W.W., Garlicki, A.M., 1967: Skidding Forces and Trafficability. Bi-Monthly Research Notes, Canada Department of Forestry and Rural Development 23(4): 28–29. Calvert, W.W., Garlicki, A.M., 1968: Tree Length Orientation and Skidding Forces. Pulp and Paper Magazine of Canada 21: 62–64. Cavalli, R., 2015: Forest Operations in Steep Terrain. Presented at Conference CROJFE 2015 »Forest Engineering – Current
148
ISO 13861 (2000) International standard: Machinery for Forestry – Wheeled Skidders – terms, Definitions and Commercial Specifications. First edition, 2000–04–15, 1–9. Janosi, Z., Green, A.J., 1968: Glossary of Terrain-Vehicle Terms. Journal of Terramechanics 5(2): 53–69. Kühmaier, M., Stampfer, K., 2010: Development of a MultiAttribute Spatial Decision Support System in Selecting TimCroat. j. for. eng. 37(2016)1
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
A. Đuka et al.
ber Harvesting Systems. Croatian journal of Forest Engineering 31(2): 75–88.
FORMEC 2012, 8–12 October, Dubrovnik (Cavtat), Croatia. CD ROM.
Lubello, D., 2008: A Rule Based SDSS for Integrated Forest Harvesting Planning. Dissertation thesis, University of Padua, 1–213.
Sever, S., 1980: Research of Exploitation Parameters for Tractors Used for Timber Extraction. Dissertation thesis, Faculty of Forestry University of Zagreb, 1–301 (in Croatian).
MacDonald, A.J., 1999: Harvesting Systems and Equipment in British Columbia. FERIC, Handbook No., HB-12: 1–197.
Sever, S., Horvat, D., 1985: Forestry Tractor with Power of 60 kW. Research study, Forestry Faculty University of Zagreb, 1–187 (in Croatian).
Matthes, R.K., Watson, W.F., 1981: Measurements of Physical Parameters During Skidder Fuel-use Studies. Fourth Annual Workshop, Council on Forest Engineering, Mississippi State University, Mississippi State, USA, 12 p. McMahon, W.S., 2006: Analysis of Fatal Logging Accidents 1988 to 2005. Rotorua (New Zealand): Forest Industry Contractors Association, 8 p. Perumpral, J.V., Baldwin, J.D., Walbridge, T.A., Stuart, W.B., 1977: Skidding Forces on tree Length Logs Predicted by Mathematical Model. Transactions of the ASAE 20(6): 1008– 1012. Poršinsky, T., Šušnjar, M., Đuka, A., 2012: Determination of Load Mass Distribution and Skidding Factors. Nova meh. šumar. 33: 35–44 (in Croatian). Raymond, K., 2010: Innovative Harvesting Solutions: A Step Change Harvesting Research Programme. NZ J Forestry 55(3): 4–9. Richardson, B.Y., Cooper, A.W., 1970: Effects on Articulated Steering of a Rubber-Tired Logging Tractor. Transactions of the ASAE 213(5): 633–635. Rowan, A.A., 1977: Terrain Classification. Forestry Commission, Forestry Record 114. Her Majesty’s Stationery Office (HMSO), Edinburgh, 1–24. Samset, I., 1975: The accessibility of forest terrain and its influence on forestry conditions in Norway. Reports of the Norwegian Forest Research Institute 32.1: 1–92. Sarles, R.L., Luppold, W.G., 1986: Technoeconomic Analysis of Conventional Logging Systems Operating from Stump to Landing. United States Department of Agriculture – Forest Service. Northeastern Research Station. Research Paper NE-577. Sauter, U.H., Sauter, F., Balle F., Lelek, S., Mohrlok, R., 2012: Motor-manual Harvesting System for Large Dimensioned Timber (LTD) on Steep Slopes Supported by Skidders Equipped with Traction Stabilising Winch. Proceedings
Croat. j. for. eng. 37(2016)1
Šušnjar, M., 2005: Interaction Between Soil Characteristics of Skid Trail and Tractive Characteristics of Skidder. Dissertation thesis, Faculty of Forestry University of Zagreb, 1–146 (in Croatian). Šušnjar, M., Bosner, A., Poršinsky, T., 2010: Skidder Traction Performance in Downhill Timber Extraction. Nova meh. šumar. 31: 3–14 (in Croatian). Šušnjar, M., Horvat, D., 2006: Dynamic Wheel Load of a Skidder during Timber Extraction. Glas. šum. pokuse, special issue 5: 601–616 (in Croatian). Suvinen, A., 2006: A GIS-based Simulation Model for Terrain Tractability. Journal of Terramechanics 43(4): 427–449. Tomašić, Ž., 2007: Research of the Technical-working Characteristics of Skidder for Thinnings. Dissertation thesis, Faculty of Forestry University of Zagreb, 1–316 (in Croatian). Tomašić, Ž., Horvat, D., Šušnjar, M., 2007: Wheel Load Distribution of Skidders in Timber Extraction. Nova meh. šumar. 28: 27–36 (in Croatian). Tomašić, Ž., Šušnjar, M., Horvat, D., Pandur, Z., 2009: Forces Affecting Timber Skidding. Croatian journal of Forest Engineering 30(2): 127–139. Visser, R., Berkett, H., 2015: Effect of Terrain Steepness on Machine Slope when Harvesting. International Journal of Forest Engineering 26(1): 1–9. Visser, R., Stampfer, K., 2015: Expanding Ground-based Harvesting onto Steep Terrain: A Review. Croatian journal of Forest Engineering 36(2): 321–331. Weise, G., Nick, L., 2003: Determining the Performance and the Environmental Impact of Forest Machines – Classification Numbers and Performance Diagrams. Proceedings of Austro 2003 – High Tech Forest Operations for Mountainous Terrain, October 5–9, 2003, Schlaegl, Austria, University of Natural Resources and Applied Life Sciences Vienna, CDROM, 1–10.
149
A. Đuka et al.
Modelling of Downhill Timber Skidding: Bigger Load – Bigger Slope (139–150)
Authors’ address:
Received: June 9, 2015 Accepted: June 18, 2015
150
Andreja Đuka, PhD. e-mail: aduka@sumfak.hr Prof. Tibor Pentek, PhD. e-mail: tpentek@sumfak.hr Prof. Dubravko Horvat, PhD. e-mail: dhorvat@sumfak.hr Prof. Tomislav Poršinsky, PhD.* e-mail: tporsinsky@sumfak.hr Department of Forest Engineering, Faculty of Forestry University of Zagreb Svetošimunska 25 10002 Zagreb CROATIA * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Possibilities of Using Small Tractors for Forestry Operations on Private Property Jurij MarenÄ?e, Janez KrÄ? Abstract The difficulty of working conditions and the scope of forestry operations dictate the selection of suitable tractors. In areas with the prevailing private, small and fragmented forest property structure, the small machines may also be applied. Due to their technical characteristics, their scope of operation is often somewhat limited. Nevertheless, these machines can represent a reasonable choice for less demanding working conditions. They may be used for thinning operations or for assembling small loads, mostly in the downhill skidding operations. In regard to the uphill skidding, they may only be applied where the working conditions so allow. The test included AGT 835, a small agricultural tractor, featuring basic forestry upgrades. The test determined its usability and suitability for working operations according to three factors: tractive force, load size and longitudinal skidding incline. The spatial analysis was conducted in regard to the suitability of forests, where the AGT 835 tractor could be used. The share of forest is shown where the tractor could be successfully applied in accordance with its limitations. The study encompassed only private forests, namely the forest areas featuring terrain slope and stand conditions that allow the use of small tractors in the first place. The analysis showed a relatively frequent possibility of using small tractors, which are suitable on a smallscale forest property and which are generally considered as less suitable for forest operations. Keywords: private forests, small tractors, wood skidding, limitations, possibilities of use
1. Introduction Due to difficult working conditions (slope, load size, ground surface, weather conditions) forest operations are regarded as one of the most demanding aspects of forest management. Foresters need to adapt to these situations and apply different machines to transport wood from the point of cutting to the forest road. Apart from skidding, the transport of wood from harsh terrains is most often carried out with cable systems or tractors of varying power and characteristics. This is true for common forest operations executed by professional foresters and qualified and properly equipped contractors. They typically rely on heavy, modern machinery that can operate efficiently in a wide range of often very demanding working conditions. However, the level of equipment used by the majority of private forest owners in Slovenia differs significantly. According to the Slovenian Forest Service Croat. j. for. eng. 37(2016)1
(2012), 75% of Slovenian forests are privately owned. The average forest property is also small and fragmented, averaging just 2.3 ha per owner, with the annual cutting of 3 m3/ha on average (Slovenia Forest Service 2012). This property size and the socio-economic characteristics of forest owners affect forest management intensity substantially (Medved 2000). The majority of owners conduct few, periodic operations in their forests. Usually they use tractors, mainly owned for agricultural use, which are upgraded for forest operations (MarenÄ?e 1997). These tractors differ in size, power, tractive force and other important technical features. Small tractors may also be used in forest operations; however they are less suitable for demanding working conditions. Nevertheless, they prove to be very useful, despite their limitations in less demanding working conditions. Upgraded with additional forestry equipment, however, these tractors can be very useful on small forest tracts (Spinelli and Magagnotti 2012).
151
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
Apart from their technical abilities, the economic efficiency of these machines also contributes to the feasibility of such tractors. With a small capital investment, these small tractors can prove to be very cost effective (Moss and Hedderick 2012). Tractors are not only suitable for skidding, but also in performing other operations. Cable systems can be applied in both directions of skidding (uphill, downhill), but the skidding costs are considerably larger compared to those of tractors for similar operations (Spinelli et al. 2010). Also, small mobile sawing systems are suitable for small-scale forests and can be coupled with tractor systems in a very interesting and usable system for small land owners, especially in terms of increased efficiency and decreased fuel consumption (Lasaux et al. 2009). The search for advantageous technical and economical solutions has now spread to countries where the sustainable principle of forest management has been far less emphasized. The same is true of numerous transition countries that are now becoming aware of the importance of sustainability, professional work, and participation of local communities and small-scale forest owners in forest planning (Nijnik et al. 2009). An entrepreneurial spirit is also frequently crucial in small-scale forestry, in wood and non-wood forest production as well as related services. Many owners do not view their forests only as a source of income attainable through proper machinery and harvesting. Forests often also have a symbolic meaning for them (Niskanen et al. 2007). The concepts of small-scale forests, interrelations, forest dependence, and a close connection to agricultural operations represent important characteristics as well as differences within Europe. Many authors have tried to determine these relations by conducting surveys. Although they established numerous specific similarities, many differences persist among individual countries (Wiersum et al. 2005). The issue of applying small machinery in forestry operations is not emphasized in a majority of countries with different ownership and land property structure, mainly due to different stand conditions, topography, and ownership. Conversely, this issue is very much present in countries characterized by predominantly privately-owned, fragmented forest land, as is also the case presented in this article. Moreover, as more countries face increased urbanization and fragmentation, as well as changing landowner objectives and interests, machinery for small-scale forest operations will become more critical.
152
2. Purpose of the research Large, powerful machinery is often necessary, particularly in case of demanding working conditions. However, in areas with moderate terrain, gradual slopes, and smaller loads, small machinery can be very useful. Because of their technical characteristics, they have numerous use limitations, and thus their use is very limited. They could be used for thinning or small load skidding operations, mostly downhill, while uphill skidding can be accomplished only if conditions allow. The use of these machines obviously calls for motivated forest owners, who manage their land despite its fragmentation and size. The focus was laid mainly on work-related technical specifications of small tractors, thus the issues pertaining to the economy of utilization of a certain type of machine in small-scale forests are not included in this article. For this purpose, the AGT 835 T agricultural tractor was selected for analysis. The tractor is primarily designed for agricultural work, but for the study, it was equipped with modifications for work in light forest working conditions. Its suitability for this work was assessed by an analysis of its technical parameters, mainly the tractive forces it can generate during various skidding loads. Also the load size plays an important role, especially in uphill skidding operations. The study was limited to uphill skidding since it presents most difficulties in this type of operation. Downhill skidding was not addressed in this article. Specific issues examined in the study include: The state of tractors on Slovenian farms – not only their number, but also the types of tractors by technical specifications (especially engine power). The number of machines with lower capabilities that are conditionally applicable for forest work were included; The technical characteristics and performance of small tractors in specific operations – the study deals only with the technical part of the issue (tractive forces, load size in skidding operations); fuel consumption of individual tractor types has not been considered; The extent of forest stands where the tractor with its limitations can be successfully utilized. That is, the study considers only private forests with terrain slopes and stand conditions that allow the use of these machines. Such information is critical, as it indicates the possibility of using small tractors, which are usually considered unsuitable for forest work. The information is also valuable for planning timber harvesting, market analysis, and the development of forest machinery market. Croat. j. for. eng. 37(2016)1
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
3. State of tractors in agricultural holdings The data on the status and condition of tractors was acquired from the periodic registers of the agricultural census conducted by the Statistical Office of the Republic of Slovenia every 10 years. The data consists of information on tractors used mainly for agricultural purposes, but the same machinery when equipped with forestry equipment, is periodically used in forests. According to the periodic data (Poje 2006, 2008, 2010), the power of tractors has increased significantly in the past five decades, (from 19.6 kW in 1952 to 53.5 kW in 2002). The data also reveal that these tractors (some of which are very old) make in average only 280 working hours per year, i.e. less than one hour per day. These facts only further encourage the use of less expensive machinery with minimal technical capabilities for forest operations. It should be noted that this is limited only to cases where landowners conduct operations in their own forests and where the working conditions so allow. The same level of safety needs to be sustained regardless of equipment. Other analyses of working mechanization (Heinimann 1999, Schrottmaier and Handler 2001, Jacke and Drewes 2004, Mago 2007, Hajdu and Mago 2007, Grgić 2009, Savelli 2010) establish tractor and other forestry machines characteristics and utilization potential. Mainly they mention the advanced age of the machinery, insufficient equipment for safe work, whereas in larger lands, where the work is more demanding, they are also equipped with more powerful and efficient machines (Šušnjar 2005, Poršinsky 2005). At the same time, the authors stress the issue of small farms, where there is less and infrequent work. In such cases the use of large and more efficient machines, usually employed in professional forest operations, is irrational mainly due to their costs. For this reason, small and seemingly less suitable machines could also be applied in such situations. This kind of machinery is very important in Slovenia according to the official data. Poje (2012) indicates that almost 100,000 tractors of various powers and types were being used in Slovenia in 2010 according to the Statistical Office of the Republic of Slovenia. The data illustrate the diversity of tractors, farms mostly use tractors of low and mid-power range. There are almost 2/3 of 37 kW tractors, i.e. tractors that are not used in forest operations or are used only where the working conditions allow due to their limited power. The AGT 835 T with its 26.4 kW engine power belonging to the class of most widely used tractors was selected for the analysis. According to its technical characteristics and power, this machine could also be used Croat. j. for. eng. 37(2016)1
J. Marenče and J. Krč
for some forest operations. This tractor was already discussed in previous articles (Jejčič et al. 2003, Košir et al. 2005, Marenče and Košir 2006a, 2006b, 2007). This article aims to emphasize and show as follows: Þ load sizes that the tractor is capable of transporting to forest roads in uphill skidding operations; Þ inclines of skid trails that can still be managed; Þ tractive forces on wheels and the winch required for successful and safe timber harvesting. Previous studies have not addressed the detailed analysis of tractive forces in such machines; Þ facts and information representing a useful guideline to forest owners – especially those managing small forest land who can use tractors that are a favourite machinery among the forest owners. The foregoing parameters can be of help in defining the tractor working range and solve the issue of its suitability and technical limitations in wood skidding operations. An important part of the analysis is also the information on stands and thus the scope of work that could be executed with this tractor in private forests. Thus, in our opinion, the answer to the question about the usability of these tractors, posed in this article, is important for everyday work at the farm. This is particularly true for fragmented and small-scale forests with predominant private ownership structure.
4. Research methods The AGT 835 T tractor, designed for operations in small agricultural lands and vineyards, was additionally furnished with suitable and necessary forestry upgrades (front blade, rear blade with a winch, safety
Fig. 1 AGT 835 T tractor prepared for the test
153
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
Table 1 Technical characteristics, farm tractors AGT 835 T Manufacturer
Agromehanika Kranj, Slovenia
Engine
Lombardini LDW 1503, 4 cylinders diesel
Displacement
1551 cm3
Power
26.4 kW
Transmission
Mechanical
Tyres
750x16
Mass
1085 kg
Additional equipment
Safety frame, front blade, 30 kN winch Krpan, wheel chains
frame, wheel chains, one-drum winch). The machine is equipped with the four-wheel drive and designed for work in lighter working conditions and for skidding small loads (Fig. 1). Some of its technical specifications are displayed in Table 1. The skid trail, on which the test was conducted, was 191 m long, of concave shape, with the longitudinal incline constantly increasing from the forest road to reach its highest level of 27% in the upper portion (Fig. 2). The skid trail was divided into three sections according to its longitudinal incline (up to 10%, from 11 to 20%, over 20%). The results of the measurements (loads and tractive forces on wheels and winch) are presented separately, according to the longitudinal incline of the skid trail. All results and their analysis are shown for the uphill skidding operation, as this represents the most significant skidding problem and limitations when using such tractors. In order to determine the soil humidity at the skid trail, the soil samples were regularly taken during the test days. The humidity changes during the day and for this reason the samples needed for analyses were taken twice a day at the characteristic points of the skid trail. The samples were taken in the morning before the start of skidding operations and also after the works. The values ranged between 36 and 42% during the test. Such values of a momentar soil sample humidity are mainly within the scope of the values usually established at the tractor skid trails (Robek and Medved 1999). The measurements of the required tractive forces were conducted for the unloaded drive and with various load sizes. These were adjusted to the capacity of skidding means according to previous experience. The load always consisted of one, 8 m long, spruce tree
154
Fig. 2 Skidding at the most steep skid trail with varying volumes (0.25 m3, 0.50 m3, 0.75 m3, and 1.00 m3). All loads were weighed in kilograms before the test. The measurement of technical parameters of every individual load size was executed without repetitions in this test. Tests were conducted to establish the skidding limits of the tractor, i.e. the point at which the tractor was incapable of transporting the load due to load size or excessive longitudinal skid trail incline. The point was defined by the load that was too large for the tractor and with tractive forces that the tractor was capable to generate with its engine, but were nevertheless insufficient for skidding due to exceeding longitudinal incline and the resulting slip. Test results were used as indicators and criteria to assess the potential areas for the operation of this kind of machines. The analysis of area suitable for the AGT 835 T tractor was conducted separately for four phytogeographical areas in Slovenia (Alpine, Dinaric, Coastal-Karts, Pannonian) (Zupančič et al. 1987). The analysis included only private forests with the areas (forest stands) suitable for the utilization of the analysed tractor on the basis of two main factors, slope of terrain Croat. j. for. eng. 37(2016)1
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
J. Marenče and J. Krč
and dimensions of trees. The terrain data were acquired from the Surveying and Mapping Authority of the Republic of Slovenia, Digital Terrain Models with 12.5x12.5 m resolution (GURS 2009), while the data on forest stands (ZGS 2013) were gathered from Forest inventory spatial database of forest stands, which encompasses 310,259 units in total, 211,942 of which are privately owned. The attributive part of the stand spatial record also includes data on the growing stock structure according to the average stand diameter at breast height (DBH) and is divided into five classes (I: 10–19 cm; II: 20–29 cm; III: 30–39 cm; IV: 40–49 cm; V: 50 cm or more) enabling a spatial breakdown according to stand diameter structure.
Working conditions (load, longitudinal incline) were selected to reach the tractor limits due to an overload. Skidding efficiency is shown in Table 2.
Suitable areas were determined with the assistance of Geographic Analyses System tools in the IDRISI software package (Eastman 1993) to classify private forests into four classes (suitable according to both criteria discussed above, suitable only from the aspect of stands, suitable only from the aspect of terrain slope, unsuitable according to both criteria).
The tractor with the maximum load did not manage to complete the top section of the skid trail (27%). The test was stopped at the point when the wheel slip brought the tractor to a stop. Thus the limit was reached where the combination of 1 m3 load and the longitudinal skidding incline of 27% did not allow further operation. Downhill skidding operations obviously have different limitations, but this issue is not the subject of this article.
5. Research results 5.1 Technical parameters in skidding with the AGT 834 T tractor 5.1.1 Load and the longitudinal skid trail incline In order to determine the limit values of the tractor capacity, limitations were set in the scope of which skidding of selected loads could be performed. The load mass was gradually increased during the test (from 220 kg to 770 kg), while the skid trail itself with its concave shape and constantly increasing incline (up to 27%) made the work increasingly difficult. Table 2 Skidding efficiency according to the longitudinal skid trail incline and load size Load size
As expected, the tractor did not have any problems when empty (unloaded) or carrying light loads (up to 0.75 m3). The skid trail with the biggest longitudinal incline of 27% did not represent any difficulty in these cases. Despite the increasing wheel slip (Marenče and Košir 2007), the tractor managed to complete the entire skid trail. In such analyses, the load size is mainly expressed by volume; its mass is also stated to additionally explain the level of difficulties related to conditions, especially in relation to the required tractive forces analysed in the second part of this article.
5.1.2 Tractor tractive forces Apart from the load size (in connection to the longitudinal skidding incline), the tractor tractive forces are those that set the limits and define the utilization of a certain machine. This article discusses only the uphill skidding where the level of difficulty is always the highest, especially if the tractor used for forest operations develops correspondingly lower tractive forces due to its lower engine power. The value of the tractor tractive forces was acquired by measuring torques, with the assistance of dynamometers inserted between the axle and individual wheel (Fig. 3). The tangential force transferred to the ground via its rim was calculated, taking the wheel radius into account.
Incline of the skid trail, %
The tangential force was calculated as:
Up to 10%
11 to 20%
Over 20%
No load
ü
ü
ü
3
ü
ü
ü
where:
3
ü
ü
ü
3
0.75 m , (657 kg)
ü
ü
ü
1.00 m3, (770 kg)
ü
ü
STOP at 27%
Ft tangential (circumferential) force; M torque on the wheel axle; rd wheel radius.
0.25 m , (220 kg) 0.50 m , (464 kg)
Where: ü Tractor traversed the entire skid trail STOP Tractor stopped at the specified level of incline due to overload
Croat. j. for. eng. 37(2016)1
Ft =
M rd
(1)
The dynamic wheel radius was considered (Krpan 1962, Sever 1980). It is smaller than the theoretical one and depends on the wheel load, size and form of tire, tire pressure, type of surface, and velocity. The radius
155
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
Fig. 4 Tractive forces on tractor wheels during uphill skidding
Fig. 3 Prior to the measuring test of tractive forces was measured at the skid trail section where the test was conducted. This measurement was performed under conditions that applied to the whole test (Jejčič et al. 2003, Marenče and Košir 2006). The circumferential force on wheels is required to overcome the resistances occurring during the tractor movement and also to skid the load (Košir 1997). This study is mainly aimed at determining the quantity of force the tractor is capable of generating in different working conditions and able to transfer to the ground considering the increasing wheel slip. In other words, the study tries to establish the point where the conditions are too harsh for operation and the utilization of the tractor is not possible due to insufficient tractive forces. At the same time, the aim is also to establish the relationship between the force necessary for tractor movement and the efficient force needed for skidding. The tractive forces on the wheels, which the tractor could generate during the test, are shown in Fig. 4. Expectedly, the lowest load on wheels is recorded during the unloaded uphill movement. The unloaded tractor with its mass of 1.1 t needs a tractive force of approximately 6 kN in the steepest section. By gradually increasing the load (from 0.25 m3 to max 1.00 m3) and the skidding incline, the need for higher tractive
156
Fig. 5 Measurement of pulling forces on the winch of AGT 835 T forces is also increasing. The most demanding part of the test is skidding the largest load (1.00 m3) in the top and steepest part of the skidding trail, where the incline exceeds 20%. Here, the highest necessary tractive force for tractor movement and load skidding (almost 11 kN) was achieved. This is also the point that determines the capacity and thus the limitation and utilization of the tractor. This tractor cannot pass this combination of longitudinal skid trail incline and load. The test also included the measurement of the necessary pulling force on the linkage, i.e. the winch cable. The front part of the load is usually raised from the ground during skidding. The measured force repreCroat. j. for. eng. 37(2016)1
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
sents the actual resultant between its horizontal and vertical component. The vertical component represents the load mass at the rear part of the tractor. The research study focused mainly on the horizontal component, the part of the joint force that is actually needed for skidding. The change of the longitudinal skid trail incline and the load size also alters this force. Higher longitudinal incline and larger load call for the higher pulling force. The remainder is necessary for the tractor movement, i.e. to overcome the rolling resistance and the incline. For this purpose, the rear board was equipped with the frame on which two devices were perpendicularly placed for pulling force measurement. By adjusting the height of the rear board, their horizontal and vertical positions are regulated (Fig. 5). This device was used to obtain the most realistic measurement of the horizontal force component on the winch, while not significantly altering the usual load position on the rear board. The forces required only for load skidding are shown in Fig. 6. The load size was selected according to the expected performance of the tractor. All loads used in the test were weighed – their mass is shown in Fig. 6 above the columns representing pulling forces. The necessary forces do not exceed the value of 2 kN when applying the smallest load. By increasing the load and skidding incline, the forces increase and exceed the value of 4 kN in the most demanding section. This is also the highest value established in the test. The tractor stopped in this part of the skid trail.
Fig. 6 Horizontal pulling force on the winch Croat. j. for. eng. 37(2016)1
J. Marenče and J. Krč
Fig. 7 shows the relations between the total tractive force on tractor wheels and the force necessary only for skidding. In certain working conditions, especially in more gradual skid trail sections and with larger load, the necessary force for load skidding is higher than for the sheer movement of the machine. In a similar way, the measured values can also be presented with the efficiency of rolling resistance hf. Where (Sever 1984): Ff = FO − FH
(2)
Where: Ff rolling resistance force; FO circumferential (tangential) force of all wheels; FH horizontal component of the pulling force. And the efficiency of rolling resistance (Sever 1985):
hf =
FH FO − Ff F = = 1− f FO FO FO
(3)
The term actually represents the ratio between the horizontal force on the winch necessary for load skidding and the total tractive force on tractor wheels. In other words: hf represents the share required for skidding, while the rest is available for the tractor movement. Table 3 Efficiency of rolling resistance hf Section
AGT 0.25 m3 AGT 0.50 m3 AGT 0.75 m3 AGT 1.00 m3
To 10%
0.36
0.45
0.58
0.64
11 to 20%
0.23
0.33
0.45
0.49
Over 20%
0.21
0.30
0.40
0.42
Table 3 shows that the tractor requires a higher force for its own movement; the exception is only the skidding of larger loads (0.75 m3 and 1.00 m3), where the tractor needs more than a half of the available force for skidding on the least steep section. The above-stated data do not include wheel slip that always occurs during the tractor movement. The slip even deteriorates the efficiency of the machine in uphill skidding. Its influence is distinctively present especially in the top, the steepest skid trail section where the largest load was skidded. Thus, there is even less available tractive force for the movement of the machine (Table 4) in the most demanding conditions – only 25%. The data on wheel slip were not analysed in detail, since previous research addressed this topic (Marenče 2005, Marenče and Košir 2007).
157
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
Fig. 7 Ratios between the total tractive force and the force on the winch
Hence, Fig. 8 shows only the data on the slip measured in the whole test. Therefore, it is proper to also include wheel slip in the analysis of the actual available tractive forces of the tractor. For this purpose the effects of wheels are expressed with the efficiency on wheel hk. The hk value actually
158
Table 4 Wheel efficiency hk Section
AGT 0.25 m3 AGT 0.50 m3 AGT 0.75 m3 AGT 1.00 m3
To 10%
0.34
0.43
0.53
0.58
11 to 20%
0.22
0.30
0.39
0.42
Over 20%
0.18
0.25
0.32
0.25
Croat. j. for. eng. 37(2016)1
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
J. Marenče and J. Krč
5.2 Selection and analysis of suitable areas
Fig. 8 Slip in relation to the skidding slope and load size conveys the effect of rolling resistance reduced by the wheel slip value. Therefore, the values in the table represent actual efficiencies or shares available for the tractor movement on the skid trail. An additional explanation is needed: in the last section, with the longitudinal incline above 20% and with the load of 1 m3, the average slip of 40% was established. This value represents the skidding average for the last section before the tractor stopped. The value was used to calculate the average wheel efficiency in this section, which is 0.25. At the moment of the stop, the wheels turned without traction, the slip reached the 100% value, and there was no necessary wheel contact between wheels and the ground. This occurred at the 27% incline; the movement of the machine was not possible anymore. The occurrences in the last few meters before the stop are not analysed in detail in this discussion. This analysis of tractive forces and loads can be very useful for assessing the suitability of small tractors for forest operations. The maximum limit of the tractor usability can thus be established: in this case this is represented in skidding up to approximately 20% inclines and with loads of up to 1 m3 in size. The article focuses on the second part of the issue: How many working sites of this type are actually present in private forests? This analysis follows in the next chapter. Both pieces of information are necessary to answer the key question: on how much area are tractors with limited technical capacity and proper forestry equipment able to operate in periodic operations in small-scale forests. Croat. j. for. eng. 37(2016)1
Several criteria and limitations were considered when establishing and selecting the operability areas for these machines. The analysis included only private forests, while the limitations incorporated the values established in the foregoing test. These values represent the limits these tractors cannot exceed due to their technical limitations. The values show the success of operations that include loads smaller than 1 m3 and longitudinal skidding inclines up to 20%. Based on these limits, the analysis of all forest areas excluded those where such machinery cannot be used. The whole territory of Slovenia was divided into four areas that differ by terrain and stand conditions; the division according to phytogeographic regions was applied as the basis (Zupančič et al. 1987). Thus, the results of these analyses within Slovenia differ considerably by terrain difficulty and diverse stand conditions. By disaggregating the results of suitability for the AGT 835 T tractor into phytogeographic regions, the research has greater applicability beyond Slovenia. Two limits were set in determining suitable areas: Þ the difficulty of terrain conditions was defined with the terrain slope limited to 20%; the test revealed that the tractor manages inclines up to this point and has no significant problems in skidding; Þ the stand conditions are defined by the average diameter at breast height; the limit was set at a diameter of 50 cm, as this dimension of properly bucked loads is considered appropriate for skidding operations. The data are shown in Table 5. The analysis indicated that, when both limitations were strictly followed (slope up to 20% and diameter less than 50 cm), an average of 39% of the areas proved suitable. The set limitations exclude all other areas, due to exceeding incline or tree dimensions. Table 5 Shares of suitable areas according to terrain and stand conditions Region
To 20%
20 to 30%
DBH < 50
DBH ≥ 50
DBH < 50
DBH ³ 50
Alpine
25%
5%
10%
5%
Dinaric
35%
15%
12%
8%
Pannonian
60%
13%
9%
4%
Coastal-Karst
43%
11%
11%
7%
Total
39%
8%
11%
5%
159
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
It was also interesting to alter the primary limits, or at least move them, and then compare the share of suitable areas. Here, two possibilities were considered: Þ allowing greater tree diameters (by bucking shorter but still permissible lengths of sorts this alteration is possible; the load does not exceed the specified size of 1 m3), Þ allowing greater terrain slope (20 to 30%), in upslope skid trails skidding is still possible; the inclines do not exceed the value 20%. By altering the limits, the shares of suitable areas also increase. If both limits were changed, the result would be 63% of suitable areas (Table 5). The analysis of the Alpine region expectedly exhibited the most limitations and thus the smallest possibilities of using these machines in comparison to other parts of the country. The terrain slope is the primary reason for limiting the operations in this region. On the other hand, the results for the Pannonian part (NE of the country) are quite the reverse (Fig. 9).
6. Discussion It should be emphasised that the AGT 835 T tractor is not designed for professional operations, but rather
for occasional work in small-scale forests. Due to numerous limitations, presented in the analysis, it does not stand as a reasonable alternative to the machinery used in everyday forest operations. Working conditions are simply too demanding for these tractors to be suitable for the majority of working sites. On the other hand, these tractors are very common on private farms and private forests represent the highest share in Slovenia. Owners of large farms and forest properties operate with larger and more powerful machines, which is necessary mainly due to the extent of work. The machinery can also be used in forest operations in various conditions. The work with these machines, mostly equipped with forestry upgrade, is easier, more efficient and most importantly safer. Furthermore, these machines can manage stands on steeper terrain and transport larger loads. However, the use of smaller machinery for less demanding and infrequent forest operations can be more practical. In such cases, the forestry upgrade is simple and, therefore, also economically feasible. Previous research (Spinelli and Magagnotti 2012) offered similar solutions with additional winch and latching mechanisms. A smaller and properly equipped tractor can also be a good choice for operation in small-scale forests. In this study, the tractor was equipped with a simple single-
Fig. 9 Spatial distribution of all suitable areas (dot represents the selected forest stand location)
160
Croat. j. for. eng. 37(2016)1
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
drum winch, as well as a rear board and a safety cabin. The skidding of smaller loads (up to 1 m3) was analysed considering smaller engine power and uphill skidding. The article does not deal with the option of using cable systems for skidding operations. In addition to small tractors, mini cableways, due to their lower operation costs, represent an additional option (Spinelli et al. 2010). They represent a good solution for skidding on steeper terrains, in terms of productivity as well as safety. Such technology could further increase the share of areas determined in this article as suitable for the use of small tractors. This is especially true for the Alpine region, where the analysed data showed that the terrain slope is the most prominent factor limiting the utilization of this technology.
7. Conclusion The use of small tractors and a simple forestry upgrade is limited. Furthermore, forest operations, where such equipment is practical, are conducted only periodically and the scope of work is also small due to tract size. The economic returns for such instances is frequently questionable. In similar land conditions in other countries, some authors (Moss and Hedderick 2012) stress the importance of this segment of forest operations. That is why the selection of the machine and equipment is important. Simple, affordable and reasonably efficient and safe method of work provides acceptable solutions for small-scale forests also in economic terms. Several studies (Wiersum et al. 2005) dealing with the topic of small-scale forestry within the European region show significant differences between countries. These are indicated in different private ownership structures, land sizes, economic dependence of income from forests. Many of them are not interested in this kind of work. The article shows the private land evaluation in Slovenia in order to assess the possibility of work in such ownership structure, stand and terrain conditions. According to the situation in Slovenia, it has been estimated that the approach analysed in the article can be considered as suitable, reasonable and executable at least for one part of forest owners. Two questions are relevant for them: in what working conditions can they use their machinery and what is the share of areas where the work can be actually executed. The article answers both questions.
8. References Anon., 2011: Poročilo Zavoda za gozdove Slovenije, 133 p. Eastman, J.R., 1993: Idrisi update Manual. Clarc University, Worcester, Massachusetts, USA, 120 p. Croat. j. for. eng. 37(2016)1
J. Marenče and J. Krč
Grgić, I., Levak, V., Matija, R., 2009: Regionalni aspekti opremljenosti obiteljskih poljoprivrednih gospodarstva poljoprivrednim strojevima. Proceedings of the 37th International Symposium on Agricultural Engineering, Opatija, Croatia, 337–345. GURS, 2009: Digitalni model višin DMV 25 in Digitalni model višin DMV 12,5. Geodetska uprava Republike Slovenije. Ljubljana. http://prostor.gov.si/cepp/GURS_izpisvse. jsp?ID={EBE6B040-5373-4821-B19D-525E2BFF8C99} Hajdu, J., Mago, L., 2007: Mechanization of the Hungarian agriculture in present days. Proceedings of the 35th International Symposium on Agricultural Engineering, Opatija, Croatia, 567–575. Heinimann, H., 1999: Ground-based Harvesting Technologies for Steep Slopes. Proceedings of the International Mountain Logging and 10th Pacific Northwest Skyline Symposium, Corvallis, Oregon, 1–19. Jacke, H., Drewes, D., 2004: Kräfte, Schlupf und Neigungen. Ein Beitrag zur Terramechanik forstlicher Arbeitsmaschinen. Forst und Holz 59(6): 259–262. Jejčič, V., Poje, T., Marenče, J., Košir, B., 2003: Development of measuring equipment for forest tractor AGT 835 with mechanical and hydromechanical transmission. Proceedings of the 31st international symposium on agricultural engineering, Opatija, Croatia, 65–74. Jejčič, V., Poje, T., Marenče, J., Košir, B., 2001: Razvoj mjerne opreme za šumarski traktor Woody 110. Aktualni zadaci mehanizacije poljoprivrede. Zbornik radova 29. međunarodnog simpozija iz područja mehanizacije poljoprivrede, Opatija, 111–117. Košir, B., 1997: Pridobivanje lesa (študijsko gradivo), Ljubljana, Biotehniška fakulteta – Oddelek za gozdarstvo in obnovljive gozdne vire, 330 p. Košir, B., Marenče, J., Jejčič, V., Poje, T., 2005: Determining technical parameters in tractor skidding – basis for the choice of tractor. FORMEC: Innovationen in der Forsttechnik durch Wissenschaftliche Kooperation – scientific cooperation for forest technology improvement. Biotechnical Faculty, Ljubljana, Slovenia, 43–55. Krpan, D., 1962: Motorna vozila. Sveučilište u Zagrebu, Fakultet strojarstva i brodogradnje, Zagreb, 134 p. Lasaux, M.J., Spinelli, R., Hartsough, B.R., Magagnotti, N., 2009: Using a Small-Log Mobile Sawmill System to Contain Fuel Reduction Treatment Cost on Small Parcels. Small Scale Forestry 8(3): 367–379. Mago, L., 2007: Survey of the present mechanization of small and medium size plant production farms. Proceedings of the 35th International Symposium on Agricultural Engineering, Opatija, Croatia, 497–506. Marenče, J., 1997: Izbor in gospodarnost prilagojenih tehnologij pridobivanja gozdnih lesnih sortimentov v zasebnih gozdovih. Master thesis. Ljubljana, University of Ljubljana, 282 p.
161
J. Marenče and J. Krč
Possibilities of Using Small Tractors for Forestry Operations on Private Property (151–162)
Marenče, J., 2005: Spreminjanje tehničnih parametrov traktorja pri vlačenju lesa – kriterij pri izbiri delovnega sredstva. Dissertation thesis. Ljubljana, University of Ljubljana, 271 p. Marenče, J., Košir, B., 2006a: Small tractors and small-scale forest property. Formec proceedings, 221–228. Marenče, J., Košir, B., 2006b: Vpliv tehničnih parametrov gozdarskega traktorja ob njegovi izbiri. Gozdarski vestnik 64(4): 213–226. Marenče, J., Košir, B., 2007: Wheelslip in skidding with the AGT 835 T adapted farm tractor. Zbornik Gozdarstva in Lesarstva 82: 25–31. Marenče, J., Košir, B., 2007: Limits of uphill skidding with AGT 835 and WOODY 110 tractors. Formec proceedings, Meeting the needs of tomorrows’ forests: new developments in forest engineering. BOKU, Vienna, 1–5. Medved, M., 2000: Gozdnogospodarske posledice posestne sestave slovenskih zasebnih gozdov. Dissertation thesis. Ljubljana, University of Ljubljana, 228 p. Moss, S.A., Hedderick, D.B., 2012: An Economic Evaluation of a Small-scale timber harvesting operation in western Maryland, USA. Small Scale Forestry 11(1): 101–117. Nijnik, M., Nijnik, A., Bizikova, L., 2009: Analysing the Development of Small-Scale Forestry in Central and Eastern Europe. Small Scale Forestry 8(2): 159–174. Niskanen, A., Pettenella, D., Slee, B., 2007: Barriers and Opportunities for the Development of Small-scale Forest Enterprises in Europe. Small Scale Forestry 6(4): 331–345. Poje, T., Jejčič, V., Cunder, T., 2006: Tehnično stanje traktorjev na slovenskih kmetijah. Acta agriculturae Slovenica 87(2): 343–354.
Poršinsky, T., 2005: Efficiency and Environmental Evaluation of Timberjack 1710B Forwarder on Roundwood Extraction from Croatian Lowland Forests. Dissertation thesis, Zagreb, 170 p. Robek, R., Medved, M., 1999: Some environmental and ergonomic stresses during logging with skidders Woody 110 and Belt 70. International symposium: Emerging harvesting issues in technology transition, Opatija, Croatia, 101 p. Savelli, S., Cavalli, R., Baldini, S., Picchio, R., 2010: Small Scale Mechanization of Thinning in Artificial Coniferous Plantation. Croatian Journal of Forest Engineering 31(1): 11–21. Sever, S., 1980: Istraživanja nekih eksploatacijskih parametara traktora kod privlačenja drva. Doktorska disertacija, Sveučilište u Zagrebu, Šumarski fakultet, 301 p. Sever, S., 1984: Istraživanja nekih eksploatacijskih parametara traktora pri privlačenja drva. Glasnik za šumske pokuse 22: 133–303. Sever, S., 1985: Rezultati istraživanja korisnosti zglobnih traktora. Strojarstvo 27(2): 79–86. Spinelli, R., Magagnoti, N., Lombardini, C., 2010: Performance, Capability and Costs of Small-Scale Cable Yarding Technology. Small Scale Forestry 9(1):123–135. Spinelli, R., Magagnotti, N., 2012: Wood Extraction with Farm Tractor and Sulky: Estimating Productivity, Cost and Energy Consumption. Small Scale Forestry 11(1): 73–85. Šušnjar, M., 2005. Interaction between soil characteristics of skid road and tractive characteristics of skidder. Dissertation thesis. Zagreb, 271 p.
Poje, T., 2008: Analysis of the offer for sale on the Slovene tractor market. Zbornik posveta: Novi izzivi v poljedelstvu 2008, Rogaška Slatina, Slovenia, 401–406.
Zupančič, M., Marinček, L., Seliškar, A., Puncer, I., 1987: Considerations on the phytogeographic division of Slovenia. Biogeographia 13(1): 89–98.
Poje, T., 2010: State of tractor pool in Slovenia. Proceedings of the 38th International Symposium on Agricultural Engineering, Opatija, Croatia, Zagreb, 67–74.
ZGS, 2012: http://www.zgs.si/slo/zavod/informacije-javnega-znacaja/letna-porocila/
Poje, T., 2012: Razvojne tendence traktorskega parka v Sloveniji. 40th Symposium Actual Tasks on Agricultural Engineering, Opatija, Croatia, 23–31.
Wiersum, K.F., Elands, B.H.M., Hoogstra, M.A., 2005: Smallscale forest ownership across Europe: Characteristics and future potential. Small Scale Forestry 4(1): 1–19.
Authors’ address:
Received: September 19, 2014 Accepted: January 26, 2015
162
Assist. prof. Jurij Marenče, PhD.* e-mail: jurij.marence@bf.uni-lj.si Assoc. prof. Janez Krč, PhD. e-mail: janez.krc@bf.uni-lj.si University of Ljubljana Biotechnical Faculty Department of Forestry and Renewable Resources Chair for Forest Techniques and Economics Večna pot 83 1000 Ljubljana SLOVENIA * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Comparison of Two Felling & Processing Methods in Beech Forests Dane Marčeta, Boštjan Košir Abstract In this research, two motor-manual felling & processing methods were compared, assortment and half-tree length, in beech stands. Investigation was done in two compartments in the northern part of Bosnia and Herzegovina (B&H), where four sample plots were chosen that differed by felled tree diameter and harvesting method. On the sample plots A1 and B1 assortment harvesting method was performed and on the sample plots A2 and B2 half-tree length method. In the study, 318 trees were felled in total, of which 163 by the assortment method and 155 by the half-tree length method. With the increase of DBH, productivity was constantly increasing and it was higher when the half-tree length method was applied than the assortment method. The main reason why half-tree length was more productive was the fact that some working operations, like production and stacking of fuel wood, were avoided or minimized. Keywords: motor-manual, assortment, half-tree length, productivity
1. Introduction A harvesting system refers to tools, equipment and machines used to harvest an area, while harvesting method refers to the form in which wood is delivered to the logging access road, and depends on the amount of processing (Pulkki 1997). According to Rebula (1988), working method determines the form and size of assortments transported from the forest. According to him, there are different methods: assortment, halftree length, tree length, full-tree method, part-tree method and chipping method. Pulkki (1997) emphasizes five harvesting methods in use throughout the world: cut-to-length, tree length, full tree, whole tree and complete tree. In the area of Bosnia and Hercegovina (B&H), several studies have been conducted on the introduction of tree length and half-tree length harvesting method (Kulušić et al.1980, Kulušić 1981, Ljubojević 1990, Kulušić and Miodragović 1979). Results of those investigations led to the conclusion that tree length and half-tree length methods are recommended along with better organization of the production process. Some studies proved that long-log methods cause higher damages to the stand, standing tress, seedlings and soil (Doležal 1984, Meyer 1966). Naghdi (2005) comCroat. j. for. eng. 37(2016)1
pared the production rate and costs, as well as damage to the residual stand when using the cut-to-length and tree length method. The productivity of the tree length method was higher than that of the cut-to-length method. Damage to the residual stand in the cut-tolength method was higher than in the tree length method. Adebayo et al. (2007) studied productivity and cost of the whole tree method and cut-to-length method. Their results proved that the whole tree method was more productive than the cut-to-length method, and consequently the production cost was lower. Spinelli et al. (2014) compared motor-manual cut-tolength (CTL) harvesting, motor-manual whole-tree (WT) harvesting, mechanized CTL harvesting and mechanized WT harvesting as applied to the production of energy chips from the second thinning of Mediterranean pine plantations in flat terrain, and concluded that mechanization increased productivity, reduced costs and damages. In Greece, the use of tree length system is introduced mainly in stands with terrain with low inclination, and cutting of stacked wood into length by chainsaw is a typical technical and technological wood harvesting solution (Galis and Spyroglou 2012). In central Sweden, studies of conventional Scandinavian short wood processing vs. a differentiated processing method were performed. The latter
163
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
signifies processing only sawlogs at the logging site. Pulpwood and fuelwood are transported off the site as undelimbed tree sections. Each operation of collecting, processing and transportation of biomass requires some energy consumption at related costs (Vasković et al., 2015). Differentiated processing was found to be recommendable for ergonomic, economical, and efficiency reasons (Bjöerheden 1998). Although, safety hazards increase in motor-manual felling, there are certain advantages because chainsaw felling is not as limited by the ground slope or tree size as mechanized felling. Motor-manual felling is also used to meet management objectives such as pre-commercial thinning, salvage operations and selective harvesting (Behjou et al. 2009). Bojanin et al. (1989) compared harvesting in oak and alder stands. They applied assortment felling & processing method, where technical roundwood and long pulpwood in transport lengths were produced, and emphasized the advantages of producing wood in transport lengths. Application of motor-manual assortment, tree length and half-tree length methods in different working conditions were also investigated by Bojanin and Krpan (1994). They established that instead of classical 1-m length fuelwood, 4-m transport length fuelwood was made. Krpan and Zečić (1996) investigated effective work time in harvesting of poplar by using of group work, where harvesting was done with modified tree length method, after which skidding to forest landing was done, where processing continued. Zečić and Marenče (2005) examined the characteristics of work and efficiency of a work team. Zečić and Krpan (2004) examined group work for felling, processing, skidding and classification in mountainous broadleaf thinning stands. Empirical performance models are generally developed by collecting field data and testing the statistical significance of any relationship with regression analysis (Samset 1990). This technique is used to calculate an equation that can represent the relationship between a dependent variable (typically time consumption or productivity) and one or more independent variables (Costa et al. 2012). Many studies have reported that tree diameter (DBH), ground slope and species of tree influence the felling time in motor-manual felling (Kluender and Stokes 1996, Hartsough et al. 2001, Wang et al. 2004, Ghaffariyan and Sobhany 2007, Ghaffariyan et al. 2013). Notably, the back cut has the highest share of the felling time, and delay times account for about one-fifth of the total working time. Tree diameter (DBH) was found to be the most important factor of time consumption and productivity. In addition to the DBH, distance between trees was also found to influence productivity of felling operation (Behjou et al. 2009). Behjou (2012) established that felling time per
164
tree was mostly affected by DBH, the distance among harvested trees in single-tree selection method and DBH in group selection method. A time study is usually done either as a comparative study, a correlation study or a combination of the two (Acuna et al. 2012). The objective of comparative studies is to compare two or several machines, work methods, etc., while the objective of the correlation or relationship study is to describe the relationship between performance and the factors influencing the work (Nurminen et al. 2006). Time studies can be carried out using continuous time study methods, such as continuous or repetitive timing or indirect work sampling (Samset 1990, Harstela 1991, Spinelli et al. 2013). Wang et al. (2004) developed a productivity model for motor-manual felling, which included variables such as DBH and the distance among harvested trees. Jovanović (1980) conducted time study for two technologies, assortment and tree length method. He used work sampling method for data collection. Poje and Potočnik (2007) studied group work in forestry and concluded that group work demands a highly skilled worker, who is able to perform any work in the group and this requires constant education and employment stability. Technical and economic harvesting of forest biomass depends on various factors related to terrain conditions, transportation networks and harvesting technologies, as well as systems, silviculture and forest operations management (Picchio et al. 2011). Time studies are usually used for the analysis of productivity of various forest biomass harvesting systems (Magagnotti et al. 2012, Picchio et al. 2009, Savelli et al. 2010). Although comparison of cut-to-length method and tree length method provides important information about the effect of log length on the productivity and cost and also damage to the residual stand, it is not sufficiently detailed, because performing cut-to-length method involves large variations in log length that require more detailed studies. Therefore, further comparative studies on the short-log and long-log method are needed to determine various positive and negative aspects of both methods applied under similar conditions (Adebayo et al. 2007). Due to the higher initial costs of mechanized harvesting machines, larger diameters and crowns of hardwoods and the relatively steep terrain in B&H forests, motor-manual felling & processing is still the most commonly used method. Forest practitioners mostly apply motor-manual assortment method. Consequently, processing of wood at the stump has a short log, which could have negative effects on the productivity and costs of skidding. With the production of stacked wood in 1-m length pieces in the forest, the problem arises of increased costs of the cutter in the Croat. j. for. eng. 37(2016)1
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
D. Marčeta and B. Košir
Table 1 Research site description Stand description Method Subcompartment
Felling site A (Sample plots A1 and A2)
Felling site B (Sample plots B1 and B2)
A1 – assortment; A2 – half-tree length
B1 – assortment; B2 – half-tree length
98a, Management Unit »Potoci – Resanovača«
65a, Management Unit »Šiša – Palež«
Altitude, m above sea level
970–1150
690–1230
Inclination, °
15–30
15–30
Exposition
S–SE
W–NW
Limestone, medium or deep rocky land
Limestone and dolomite, medium or deep rocky land
Geologic surface Climate
Mountain, humid
Mountain, humid
Stand
Beech and fir forests with spruce on a series of limestone, predominantly deep soil (Picea-Abieti-Fagetum)
High beech forests on predominantly deep limestone and illimerised soil (Fagetum montanum illyricum)
3
2
Site index Canopy
Dense (0.7)
Dense (0.8)
Management system
Group-selection
Group-selection
Growing stock, m3/ha
513.72
343.74
Cutting intensity, %
14.53
20.94
Average diameter of marked tress
21 cm
35 cm
Medium dense
Medium dense
Regeneration
forest, as well as increased stacked wood transport costs (Halilović 2012). A part of woody biomass with low value or minimal value is often called fuel wood (fire wood or wood fuel) and used as a traditional or classic source of energy (Eker 2014). Wood stacked in the forest applying the assortment method has to be carried out by animal, and due to the lack of animal labour force on the labour market, stacked wood in practice often remains unused in the forest. The problem of stacked fuel wood produced by applying the assortment method can be solved if long fuel wood is produced instead of classical stacked fuel wood. Producing of long wood is a rational solution because productivity increases and human labour decreases (Bajić et al. 2007). The problem of practice is that, in applying the assortment method, cutters often crosscut the stem at the stump and produce definite shape of logs without the supervision of specialists for classification. The aim of this study was to compare two motormanual felling & processing methods, assortment and half-tree length, in beech stands in order to determine the difference in produced wood assortments, productivity and cost competitiveness.
2. Material and Methods Investigation was conducted in the northern part of the B&H in the area of municipality of Ribnik. The Croat. j. for. eng. 37(2016)1
terrain was mountainous, in winter period without or with minor amount of snow. Temperature varied from 0 to 7 °C. Sample plots were placed in two subcompartments (Table 1): subcompartment 98a MU »Potoci – Resanovača« (Felling site A) and subcompartment 65a MU »Šiša – Palež« (Felling site B). When choosing the felling sites, stand conditions and characteristics of forest infrastructure had to be average for beech forests in B&H. In each compartment, two sample plots i.e. work fields were selected, (A1, A2, B1 and B2). Sample plots were selected according to the similarity of their stand and habitat conditions. In this way, as many factors as possible were isolated in order to compare technologies with more reliability. Felling & processing with chainsaw was performed on each sample plot. The width of the work fields was 100 m. The length of each work field was 500 m. So the surface of each sample plot was 5 ha. Assortment motor-manual felling & processing method was performed on the sample plots A1 and B1, where cutters cut the trees with chainsaw and tree processing was done at the felling site. Technical assortments were made and stacked wood (traditional 1-m length fuelwood) was produced and piled. Fuel wood was made from the thinner part of the stem and branches. Half-tree length felling & processing method was performed on the sample plots A2 and B2, where cutting of trees, delimbing and, if necessary, crosscut-
165
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
ting of the stem was done at the site. The stem remained whole or was cut into the transport lengths to allow easier skidding. Stacked wood was made only from branches. Processing continued at the landing site. Workers were in group of two. Both workers were cutters but while one of them worked with chainsaw, the other was an assistant. After half of the working day, they changed roles. During the time when the cutter worked with the chainsaw, the assistant was engaged in several jobs like work place cleaning, accessories carrying, moving the branches away after delimbing, producing and stacking of fuel wood, etc. Productivity was calculated for the working crew. All work was performed by the same working crew so as to avoid the influence of skill and devotion of workers. Working crew was selected on the basis of last tree month productivity sheets. The crew with average productivity was selected. They worked with professional chainsaw Husqvarna 372XP. Productivity was investigated by time and work study method (Björheden et al. 1995, Acuna et al. 2012). Time was divided into time elements, each corresponding to one specific task. Time consumptions for work elements were measured by continuous chronometry method and recorded. The distance between marked trees was measured by measuring tape, slope gradient was measured by clinometers and the produced wood data were collected by measuring the diameter and length of each piece of roundwood and by measuring of pile dimension of 1-m length stacked wood. For conversion of staked volume into solid wood volume, conversion coefficient 0.65 was used. Workers, who worked on processing at the landing site during the research, were paid per shift, not per productivity, because processing at the landing site was part of the study and it is not common for local forestry organizations. Their productivity was roughly determined on the basis of work time and produced wood volume. When data were collected, the influence of different variables on all phases of technological process
Fig. 1 DBH distribution of felled trees was examined on the tree level. Several statistical methods were used (Descriptive statistics, Regression and Multiple regressions with dummy variables, etc.) with the help of software Statistica10. Standard times for both methods were calculated. Cost calculation was done according to official methodology, which is in use in the Public Company »Šume Republike Srpske«, based on Myiata (1980).
3. Results 3.1 Description of sample In the study, 318 trees were felled in total, of which 163 applying the assortment method (sample plots A1 and B1) and 155 applying the half-tree length method (A2 and B2). The average diameter of felled trees on sample plot A1 was 30 cm and varied from 9 to 54 cm,
Table 2 Sample description Sample plot A1 B1
Method
Assortment
N
Vfuelwood, m3
DBH, cm
Vroundwood, m3
Vtotal, m3
Mean
Min.
Max.
Std.Dev.
Mean
Sum
%
Mean
Sum
%
Mean
Sum
%
113
30
9
54
8.39
0.124
14.001
15.53
0.705
76.172
84.47
0.798
90.173
100
50
49
23
78
15.78
0.643
32.149
17.13
3.173
155.500
82.87
3.762
187.649
100
A2
Half-tree
110
27
10
49
7.47
0.072
1.362
1.95
0.629
68.593
98.05
0.636
69.955
100
B2
length
45
50
18
69
13.39
0.288
12.942
8.12
3.256
146.530
91.88
3.544
159.472
100
∑
–
318
–
–
–
–
–
60.454
–
–
446.795
–
–
507.249
100
166
Croat. j. for. eng. 37(2016)1
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
D. Marčeta and B. Košir
Table 3 Characteristics of produced roundwood Sample plot A1 Number of assortments per tree
Assortment
B1 A2
Half-tree length
B2 A1 Average diameter of assortments, cm
Assortment
B1 A2
Half-tree length
B2 A1 Average length of assortments, m
Assortment
B1 A2
Half-tree length
B2 A1 Average volume of assortments, m3
Assortment
B1 A2
Half-tree length
B2
on sample plot B1 it was 49 cm and varied from 23 to 78 cm (Fig. 1). On sample plots A2 and B2, the average tree diameter was 27 cm and 50 cm, respectively, and varied from 10 to 49 cm on A2 and 18 to 69 cm on B2 (Table 2). Total volume of produced wood was 507.25 m3, of which 277.82 m3 in assortments and 229.43 m3 in halftree length. From the total amount, 231.67 m3 of roundwood was in assortments and 215.12 m3 in half-tree length (Table 2). TTK 50S The share of stacked wood on sample plots in assortments (A1 and B1) was 15.53% and 17.13%, respec-
Mean
Min.
Max.
Std. dev.
2.30
1.0
6.0
1.52
1.81
1.0
7.0
0.92
6.23
1.0
12.0
3.03
4.77
1.0
12.0
2.57
25.73
13.0
51.0
7.22
20.01
8.0
46.0
6.89
34.83
12.0
79.0
16.23
26.90
10.0
61.0
13.81
5.16
1.0
9.0
1.72
8.83
3.8
14.5
2.09
4.95
1.0
9.0
1.68
8.66
1.6
18.0
3.23
0.262
0.032
1.079
0.13
0.319
0.035
1.594
0.26
0.501
0.054
2.694
0.45
0.691
0.028
3.505
0.80
tively. On sample plots where half-tree length method was performed (A2 and B2), the share of stacked wood was 1.95% and 8.12%. The number of roundwood assortments per tree and dimension of assortments are presented in Table 3.
3.2 Comparison of samples Multiple regression with dummy variables was performed in order to determine which factors influence effective work time per tree. Results showed that felling & processing method and DBH have significant influence on the level p≤0.05 (Table 4). These results
Table 4 Regression summary for effective time per tree Regression Summary for dependent variable: min/tree R=0.88, R2=0.78, Adjusted R2=0.78 F(3.314)=366.80, p<0.0000, Std. error of estimate: 7.397
N=318 b
Std. err. of b
b
Std. err. of b
t(314)
p-value
Intercept
–
–
–16.0588*
2.137513*
–7.51283*
0.000000*
Felling site
–0.052280
0.036335
–1.7820
1.238457
–1.43885
0.151188
Method
0.232889*
0.026665*
7.2688*
0.832261*
8.73380*
0.000000*
DBH
0.799484*
0.036408*
0.8719*
0.039706*
21.95914*
0.000000*
* Significant at p<0.05
Croat. j. for. eng. 37(2016)1
167
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
Table 5 Regression summary for net volume per tree Regression Summary for dependent variable: m3/tree R=0.94, R2=0.88, Adjusted R2=0.88 F(3.314)=785.87, p<0.0000, Std. error of estimate: 0.649
N=318 b
Std. err. of b
b
Std. err. of b
t(314)
p-value
Intercept
–
–
–1.78630*
0.187694*
–9.51709*
0.000000*
Method
0.004242
0.019402
0.01598
0.073080
0.21865
0.827061
DBH
0.828727*
0.026491*
0.10907*
0.003487*
31.28351*
0.000000*
Felling site
–0.152354*
0.026438*
–0.62669*
0.108748*
–5.76275*
0.000000*
* Significant at p<0.05
indicated that further analysis should be done separately for both examined work methods. Felling site did not show significant influence on the effective time per tree and this result indicated that felling sites with similar work conditions were chosen in accordance with the purpose. Comparison of net volume per tree on each feeling site showed that work method had no significant influence on the net volume per tree but DBH and felling site did show significant influence (Table 5). Trees on the felling site B had larger net volume per tree for the same DBH (Fig. 2). The difference in the net wood volume, as a result of different site index of felling site and different working method, indicated that productivity should be calculated separately for both work method and felling site.
3.3 Analysis of work operations Total studied work time was 4519.44 min on the sample plots where the assortment method was applied and 2502.72 min on the sample plots where halftree length method was applied. From total time, productive work time was 3469.12 min (assortment) and 1913.29 min (half-tree length) with relative share of delays 30.28% and 30.81%, respectively (Table 6). Productive work time was divided into work operations. Most time consuming work operation in applying the assortment method was stacking of fuel wood with 6.29 min/tree, then follows delimbing with 4.44 min/tree and production of fuel wood with 3.78 min/tree. In applying the half-tree length method, most time consuming work operations were delimbing with 4.16 min/tree, stacking of fuel wood with 2.61 min/tree and moving with 1.81 min/tree. The shortest time operation in both methods was preparing of work place. In both methods, most of the allowance work time is related to personal delay, 43% in applying the assortment method and 51% in applying the half-tree length method. Then follow technical delays with the share of 26% and 30%, preparatory-final time with 19% and 11% and organizational delay with the share of 12% and 8%, respectively. 3.4 Analysis of influencing factors on time consumption of work operations
Fig. 2 Net wood volume per tree
168
Research of different factors influencing the time of work operations was done with the regression and correlation analysis. The influence strength was presented with correlation coefficient (R), with the level of significance, p≤0.05. The mathematical models that best show the dependence between variables were presented with regression equations. In work operations, where no significant dependences were evidenced, mean values were used for productivity calculations. Croat. j. for. eng. 37(2016)1
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
D. Marčeta and B. Košir
Table 6 Descriptive analysis of work time Working methods Work operations
Assortment method Sum
Mean
Min.
Half-tree length method
Minimum
Maximum
min/tree
Sum
Mean
Min.
Minimum
Maximum
min/tree
Moving
348.26
2.14±1.53
0.13
8.23
279.72
1.81±1.07
0.35
5.23
Preparing of work place
123.33
0.76±1.04
0.03
7.38
71.59
0.46±0.73
0.08
4.80
Felling
234.23
1.44±1.18
0.12
6.98
187.87
1.21±0.78
0.10
3.37
Delimbing
723.69
4.44±3.37
0.30
18.63
645.38
4.16±3.07
0.67
21.08
Processing
398.42
2.44±4.19
0.12
35.35
148.44
0.96±1.81
0.12
8.85
Production of fuel wood
615.66
3.78±3.45
0.17
19.27
176.47
1.14±2.01
0.20
10.35
Stacking of fuel wood
1025.53
6.29±7.17
0.42
35.00
403.82
2.61±7.11
0.33
40.00
Productive work time, min
3469.12
1913.29
Allowance time, min
1050.32
589.43
Allowance time, % ∑ Total, min
30.28
30.81
4519.44
2502.72
Assortment method
difference from any examined variables and all other work operations depended from DBH (Table 9).
Work operations Moving and Preparing of work place did not show significant difference of any examined variable. Their mean value was used for productivity calculation. All other work operations showed dependence on DBH (Table 8).
3.5 Productivity and costs Productivity was calculated according to the time for each work operation calculated by the regression equation in cases where significant dependence on influencing factors was established or the mean values were used if there was no dependence. The sum of work operations was multiplied by the coefficient of
Half-tree length method In half-tree length method it was similar. Moving and Preparing of work place did not show significant
Table 7 Relative share of work operations in total and productive work time Working methods
Assortment method
Work operations
Half-tree length method
Assortment method
% of productive work time
Half-tree length method
% of total work time
Moving
10.04
14.62
7.71
11.18
Preparing of work place
3.56
3.74
2.73
2.86
Felling
6.75
9.82
5.18
7.51
Delimbing
20.86
33.73
16.01
25.79
Processing
11.48
7.76
8.82
5.93
Production of fuel wood
17.75
9.22
13.62
7.05
Stacking of fuel wood
29.56
21.11
22.69
16.14
100
100
76.76
76.45
30.28
30.81
–
–
∑ Allowance time, % Croat. j. for. eng. 37(2016)1
169
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
Table 8 Time dependence analysis – Assortment method Parameters
N
Independent variable
Equation
Moving
158
Distance, m
–
No significance
Preparing of work place
130
DBH, cm
–
No significance
Felling
161
DBH, cm
Quadratic
0.3810
–0.0075
Delimbing
162
DBH, cm
Linear
–2.1889
Processing
127
DBH, cm
Quadratic
Production of fuel wood
161
DBH, cm
Stacking of fuel wood
156
DBH, cm
Work operation
F
R
p
Standard error
0.04
0.02
0.849
1.54
3.93
0.17
0.049
1.03
0.0009
315.30
0.82
0.000
0.69
0.1794
–
384.04
0.84
0.000
1.83
0.2109
–0.0699
0.0031
234.61
0.81
0.000
2.48
Quadratic
1.2539
–0.0215
0.0021
203.89
0.75
0.000
2.29
Quadratic
1.8507
–0.1583
0.0068
430.71
0.86
0.000
3.69
F test
R
p
Standard error
Intercept
b1
b2
test
Table 9 Time dependence analysis – Half-tree length method Parameters
N
Independent variable
Equation
Moving
155
Distance, m
–
No significance
8.97
0.23
0.003
1.04
Preparing of work place
91
DBH, cm
–
No significance
2.14
0.15
0.147
0.72
Felling
155
DBH, cm
Linear
–0.4833
0.0490
–
565.73
0.89
0.000
0.36
Delimbing
154
DBH, cm
Linear
–1.4454
0.1600
–
293.90
0.81
0.000
1.80
Processing
100
DBH, cm
Quadratic
0.5013
–0.0618
0.0031
260.80
0.85
0.000
0.95
Production of fuel wood
59
DBH, cm
Linear
–1.4280
0.0904
–
55.68
0.70
0.000
1.44
Stacking of fuel wood
51
DBH, cm
Quadratic
0.0264
–0.0916
0.0068
41.65
0.68
0.000
5.28
Work operation
Intercept
b1
b2
Fig. 3 Standard times for felling sites and methods
170
Croat. j. for. eng. 37(2016)1
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
D. Marčeta and B. Košir
Fig. 4 Unit costs for felling sites and methods allowance time and divided by the volume of wood (Fig. 3). The cost of the working day of chainsaw was calculated on the basis of official methodology used by the Public Company »Šume Republike Srpske«. Prices of material and labour, which were valid at the moment of research, were taken as inputs. The cost of one 8-hour-working day was EUR 78.4 (9.80 EUR/hour). In half-tree length method, workers who performed processing at the roadside landing were paid per shift and costs of processing at the landing site were 0.6 EUR/m3 for felling site A and 0.5 EUR/m3 for felling site B (Fig. 4) on the basis of daily productivity of 130.7 m3 for felling site A and 156.8 m3 for felling site B and daily costs of EUR 78.4.
4. Discussion The share of stacked wood was significantly lower in applying the half-tree length method than assortment method, 1.95% (A2 – half-tree length method) vs 15.53% (A1 – assortment method) and 8.12% (B2 – half-tree length method) vs 17.13% (B1 – assortment method). These results were expected in accordance with the working methods. In applying the assortment method, stacked wood was made from branches and from thinner parts of the stem, while in the half-tree length method, it was only made from branches. The decision on the place where producing of roundwood stops and producing of stacked wood starts on the stem was based on recommendations of operation Croat. j. for. eng. 37(2016)1
plans for the specific compartment. The difference in the amount of stacked wood within the same harvesting method was higher on the sample plots with smaller average tree diameter. The reason for that could be that thicker trees have a relatively higher amount of branches above 7 cm diameter from which stacked wood is produced. The average number of assortments per tree was smaller in half-tree length method, 1.81 pieces per tree (A2 – half-tree length method) vs 2.23 pieces per tree (A1 – assortment method) and 4.77 (B2 – half-tree length method) vs 6.23 (B1 – assortment method). The difference was slightly bigger on sample plots with larger DBH, 19% (A2 vs A1) and 23% (B2 vs B1). The diameter of assortments was lower in applying the half-tree length method than the assortment method. The difference was the consequence of the fact that in using the half-tree length method logs were relatively longer and stretched to the thinner parts of the stem, while in using the assortment method these parts were bucked into the stacked wood. The emphasis is on the fact that technical and firewood logs are the same in half-tree length method, which makes skidding of lower value wood cost competitive (Košir 2009, Bajić et al. 2007). The average log length was lower when the assortment method was applied. In applying the assortment method, log length mostly depends on dimension and quality, which are inputs for wood classification. When the half-tree length method is applied, the length mostly depended on the skidding options. The density of remaining trees was the limiting factor in
171
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
most cases. As a consequence of length and diameter, the average volume of assortments was larger in applying the half-tree length method 19% (A2 vs A1) on sample plots with lower DBH and 27% (B2 vs B1) on plots with larger DBH. Relative structure of productive work time showed that work operations Production of fuelwood and Stacking of fuelwood consumed relatively less time in applying the half-tree length method than the assortment method, 9.22% and 21.11% in applying the half-tree length method and 17.75% and 29.56% in applying the assortment method, respectively. Time Delimbing takes relative larger share in half-tree length method, 33.73% vs 20.86% in assortment method. The explanation for that could be the fact that in upper parts of stem in applying the assortment method Delimbing can overlap with the Production of fuelwood and it is hard to define the border between those two operations during time study. Structure of additional times was similar in both methods, coefficient of allowance times are 1.30 for assortment and 1.31 for half-tree length method. Ghaffariyan et al. (2013) determined next-time distribution in chainsaw motor-manual felling: moving to tree 12%, reconnaissance 11%, under cut 27%, backcut 31% and delay 19%. In uneven-aged beech forests, Behjou et al. (2009) established mean delay times of 0.81 min per tree, based on 0.22, 0.44 and 0.15 min per tree for operational, mechanical and personal delays. Regression and correlation analysis showed that in both methods, the operation Moving did not show dependence on distance. The reason could lie in the fact that trees marked for felling were equally distributed in the stand. The reason could also be the discipline of workers. Preparing of work place did not show significant dependence on any examined factor. Other work operations showed more or less strong significant dependence on DBH. Productivity was constantly increasing with the increase of DBH and it was higher in applying the half-tree length method than the assortment method. The reason for higher productivity of the half-tree length method was the fact that some working operations were avoided or minimized in applying the half-tree length method, like production and stacking of fuel wood. Also, bucking was mostly transferred to the landing site, where it could be done in a more productive way.
Unit costs were significantly lower. Large amount of stacked wood remaining after the assortment method is a transport problem for forest managers because it is more and more difficult to find animals on the labour market. Working with animals is very expensive. In some surrounding countries, forest managers sell wood felled by the stump to the local population that produces fuelwood for personal consumption (Vusić 2013), but in B&H local people are, in general, still not interested in it. In applying the half-tree length method, stacked wood also occurred, but in significantly less amount than in applying the assortment method. It only came from the branches, while the other fuel wood remained as roundwood and was skidded together with the other more valuable parts of the stem. This resulted in increased felling & processing productivity and allowed transport of lower value wood.
6. References Adebayo, A.B., Han, H-S., Johnson, L., 2007: Productivity and cost of cut-to-length and whole-tree harvesting in a mixed-conifer stand. Forest Product Journal 57(5): 59–69. Acuna, M., Bigot, M., Guerra, S., Hartsough, B., Kanzian, C., Kärhä, K., Lindroos, O., Magagnotti, N., Roux, S., Spinelli, R., Talbot B., Tolosana, E., Zormaier F., 2012: Good practice guidelines for biomass production studies. COST Action FP0902, WG 2 Operations research and measurement methodologies, CNR Ivalsa – Italy, 1–50. Bajić, V., Danilović, M., Gačić, D., Stefanović, D., Krajnc, N., Dolenšek, M., 2007: Korišćenje nekih resursa u šumama hrasta kitnjaka. Monografija Hrast kitnjak u Srbiji, Šumarski fakultet Univerziteta u Beogradu, 1–498. Behjou, F.K., Majnounian, B., Dvorak, J., Namiranian, M., Saeed, A., Feghhi, J., 2009: Productivity and cost of manual felling with a Chainsaw in Caspian forests. Journal of Forest Science 55(2): 96–100. Behjou, F.K., 2012: Effect of selective cutting type on the chainsaw productivity in Caspian Forests. Journal of Forestry Research 23(4): 699–702. Björheden, R., Apel, K., Shiba, M., Thompson, M.A., 1995: IUFRO Forest work study nomenclature. Swedish University of Agricultural Science, Dept. of Operational Efficiency, Garpenberg, 16 p. Bjöerheden, R., 1998: Differentiated processing in motor manual and mechanized logging. International Journal of Forest Engineering 9(2): 49–59.
5. Conclusion
Bojanin, S., Krpan, A.P.B., Beber, J., 1989: Komparativno istraživanje sječe i izrade u prorednim sastojinama hrasta lužnjaka i crne johe. Šumarski list 113(9–10): 591–602.
Investigation showed that performing of the halftree length harvesting method was more productive and cost competitive than the assortment method.
Bojanin, S., Krpan, A.P.B., 1994: Eksploatacija šuma pri različitim radnim uvjetima u Hrvatskoj. Šumarski list 118(9– 10): 271–282.
172
Croat. j. for. eng. 37(2016)1
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
D. Marčeta and B. Košir
Costa, C., Menesatti, P., Spinelli, R., 2012: Performance modelling in forest operations through partial least square regression. Silva Fennica 46(2): 241–252.
vanju tehnologije iskorišćavanja bukovih šuma u SR BiH, Naučnoistraživački projekat, Šumarski fakultet Sarajevo, 1–83.
Doležal, B., 1984: Štete u šumi izazvane primjenom mehanizacije. Dokumentacija za tehniku i tehnologiju u šumarstvu 81: 1–47.
Kulušić, B., Jovanović, B., Miodragović, D., Ljubojević, S., Davidović, V., 1980: Prilog poznavanju tehnologije iskorišćavanja bukovih šuma u SR BiH. Šumarski fakultet Sarajevo, 1–82.
Eker, M., 2014: Trends in Woody Biomass Utilization in Turkish Forestry. Croatian Journal of Forest Engineering 35(2): 255–270. Gallis, C., Spyroglou, G., 2012: Productivity Linear Regression Models of Tree-Length Harvesting System in Natural Coastal Aleppo Pine (Pinus halepensis L.) Forests in the Chalkidiki Area of Greece. Croat. j. for. eng. 33(1): 115–123. Ghaffariyan, M.R., Sobhany, H., 2007: Cost production study of motor-manually felling and processing of logs. Forest Science 3: 69–76. Ghaffariyan, M.R., Naghdi, R., Ghajar, I., Nikooy, M., 2013: Time Prediction Models and Cost Evaluation of Cut-ToLength (CTL) Harvesting Method in a Mountainous Forest. Small-Scale Forestry 12(2): 181–192.
Ljubojević, S., 1990: Uticaj nekih faktora na učinke privlačenja bukove i jelove deblovine zglobnim traktorom Timberjack 350A na teškim terenima. Mehanizacija šumarstva 15(5–6): 101–106. Magagnotti, N., Pari, L., Spinelli, R., 2012: Re-engineering firewood extraction in traditional Mediterranean coppice stands. Ecological Engineering 38(1): 45–50. Meyer, G., Ohman, J.H., Oettel, R., 1966: Skidding hardwoods, Articulated rubber-tyred skidders vs. crawler tractors. Journal of Forestry 64(3): 191–196. Miyata, E.S., 1980: Determining fixed and operating costs of logging equipment. General Technical Report NC-55, US Department of Agriculture, Forest Service, St. Paul, 16 p.
Halilović, V., 2012: Komparacija metoda dobivanja šumske biomase kao obnovljivog izvora energije iz hrastovih sastojina. Disertacija, Šumarski fakultet Univerziteta u Sarajevu, 1–154.
Naghdi, R, Rafatnia, N., Jalali, G.H., Hosseini, S.M., 2005: A survey of the efficiency of Timberjack 450C wheeled skidder in Shafaroud forests in Guilan province. Iranian Journal of Natural Resources 57(4): 675–687.
Harstela, P., 1991: Work studies in forestry. Silva Carelica 18: 41 p.
Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analyses of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363.
Hartsough, B.R., Zhang, X., Fight, R.D., 2001: Harvesting cost model for small trees in natural stands in the Interior Northwest. Forest Products Journal 51(4): 54–61.
Rebula, E., 1988: Tehnologija u šumarstvu na početku idućeg tisućuljeća. Mehanizacija šumarstva 13(3–4): 31–40.
Jovanović, B., 1980: Istraživanje utroška vremena za dvije tehnologije rade kod eksploatacije bukovih šuma u SR BiH. Magistarski rad, Šumarski fakultet Sveučilišta u Zagrebu, 1–248.
Picchio, R., Maesano, M., Savelli, S., Marchi, E., 2009: Productivity and Energy Balance in Conversion of a Quercus Cerris L. Coppice Stand into High Forest in Central Italy. Croat. j. for. eng. 30(1): 15–26.
Kluender, R.A., Stokes, B.J., 1996: Felling and skidding productivity and harvesting cost in southern pine forests. In: Proceedings: Certification-Environmental implications for forestry operations, September 9–11, 1996, Quebec City, Quebec, Joint conference Canadian Woodlands Forum, Canadian Pulp and Paper Association, and International Union of Forest Research Organizations, 35–39.
Picchio, R., Spina, R., Maesano, M., Carbone, F., Monaco, A.L., Marchi, E., 2011: Stumpage value in the short wood system for the conversion into high forest of a oak coppice. Forestry Studies in China 13(4): 252–262.
Košir, B, Krajnc, N., Piškur, M., 2009: Uvajanje tehnologije strojne sečnje in izkoriščanja sečnih ostatkov. Konačno poročilo projekta, Biotehniška fakulteta Univerze u Ljubljani, 177p. Krpan, A., Zečić Ž., 1996: Efektivno vrijeme pri sječi i izradi topole primjenom skupnog rada. Savjetovanje »Skrb za hrvatske šume od 1846. do 1996.«, Knjiga 2, 217–226. Kulušić, B., Miodragović, D., 1979: Prilog istraživanja tehnološkog procesa sječe, izrade i privlačenja drveta pri iskorišćavanju borovih i hrastovih šuma u BiH, Radovi Šumarskog fakulteta u Sarajevu 22(5–6): 3–66. Kulušić, B., 1981: Tehnološka tipizacija bukovih šuma i mješovitih šuma bukve, jele i smrče u SR BiH. Prilog poznaCroat. j. for. eng. 37(2016)1
Poje, A., Potočnik, I., 2007: Influence of working conditions on overlapping of cutting and ground skidding in group work. Croat. j. for. eng. 28(2): 157–167. Pulkki, R., 1997: Cut-to-length, tree-length or full tree harvesting? Central Woodlands 1(3): 22–27. Samset, I., 1990: Some observations on time and performance studies in forestry. Meddelelser fra Norskinstitutt for skogforskning 43(5): 80 p. Savelli, S., Cavalli, R., Baldini, S., Picchio, R., 2010: Small scale mechanization of thinning in artificial coniferous plantation. Croat. j. for. eng. 31(1): 11–21. Spinelli, R., Lombardini C., Magagnotti, N., 2014: The effect of mechanization level and harvesting system on the thinning cost of Mediterranean softwood plantations. Silva Fennica 48(1): article id 1003, 15 p.
173
D. Marčeta and B. Košir
Comparison of Two Felling & Processing Methods in Beech Forests (163–174)
Spinelli, R., Laina-Relano, R., Magagnotti, N., Tolosana, E., 2013: Determining Observer and Method Effects on the Accuracy of Elemental Time Studies in Forest Operations. Baltic Forestry 19(2): 301–306.
Wang, J., Long, C., McNeel, J., Baumgras, J., 2004: Productivity and cost of manual felling and cable skidding in central Appalachian hardwood forests. Forest Product Journal 54(12): 45–51.
Vasković, S., Halilović, V., Gvero, P., Medaković, V., Musić, J., 2015: Multi-Criteria Optimization Concept for the Selection of Optimal Solid Fuels Supply Chain from Wooden Biomass. Croatian Journal of Forest Engineering 36(1): 109–123.
Zečić, Ž., Marenče, J., 2005: Mathematical models for optimization of group work in harvesting operations, Croat. j. for. eng. 26(1): 29–37.
Vusić, D., 2013: Pogodnost sustava pridobivanja drvne biomase u smrekovoj šumskoj kulturi. Disertacija, Šumarski faklultet Sveučilišta u Zagrebu, 1–174.
Zečić, Ž., Krpan, A., 2004: Efficiency of group work in harvesting mountainous broadleaf thinning stands. Zbornik gozdarstva in lesarstva 74: 41–57.
Authors’ addresses: Assist. Prof. Dane Marceta, PhD.* e-mail: danemarceta@gmail.com University of Banja Luka, Faculty of Forestry Bulevar Vojvode Stepe Stepanovića 75 78000 Banja Luka BOSNIA AND HERZEGOVINA
Received: August 6, 2014 Accepted: November 18, 2014
174
Prof. Boštjan Košir, PhD. e-mail: bokosir@gmail.com University of Ljubljana, Biotechnical Faculty Jamnikarjeva 101 1000 Ljubljana SLOVENIA * Corresponding author Croat. j. for. eng. 37(2016)1
Original scientific paper
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester in a Hardwood Dominated Stand Eric R. Labelle, Michel Soucy, AndrĂŠ Cyr, Gaetan Pelletier Abstract It is commonly accepted that tree form has an impact on the productivity of single-grip harvesters. However, it remains unclear, which elements of tree form are significant and to what degree they impact harvesting productivity. This is of particular importance in hardwood dominated stands, where hardwood trees often exhibit complex and variable stem and crown architecture that can complicate and prolong the processing phase. With the development of specialized harvesting heads, hardwoods, which were mostly subject to motor-manual operations, are now increasingly being cut and processed with fully mechanized harvesting systems. The goal of this pilot project was to determine the effect of tree form on the productivity of mechanized cut-to-length harvesting. A time and motion study of a single-grip harvester, operating in a hardwood dominated stand, suggests that the presence of a fork or a large branch on the main stem can reduce machine harvesting productivity by 15 to 20%. Keywords: tree characteristics, hardwoods, time and motion study, mechanized harvesting, single-grip harvester, processing, Canada
1. Introduction There is growing interest in using single-grip harvesters in hardwood dominated stands of northeastern North America. However, for certain northern hardwoods, such as sugar maple (Acer saccharum Marsh.), stem form is very complex and wood density is relatively higher compared to most commercial hardwoods (examples of oven dry wood densities: sugar maple, 705 kg/m3; white birch (Betula papyrifera Marsch.), 588 kg/m3; trembling aspen (Populus tremuloides Michx.), 424 kg/m3 (Jessome 1977)). These physical characteristics make the processing phase more challenging and reduce the average harvesting productivity for those species compared to softwoods. Such differences in average harvesting productivity between species have been reported (Huyler and LeDoux 1999, Nurminen et al. 2006, Hiesl 2013) but no studies could be found that determined which tree characteristics actually caused the differences. In the spring of 2013, the Northern Hardwoods Research Institute (NHRI) introduced a tree classificaCroat. j. for. eng. 37(2016)1
tion system that includes a rating of tree form (Pelletier et al. 2013). According to Pelletier et al. (2013), this classification system is seen as an opportunity to improve predictions of harvester productivity and the selection of harvesting systems, amongst other benefits. As such, a pilot study was designed to meet two objectives: i) determine if there is a link between the harvesting productivity of a single-grip harvester and stem form as defined in the new tree classification system; ii) determine if there are other potentially significant stem form characteristics that influence singlegrip harvesting productivity.
2. Methodology The pilot study was limited in extent to a single machine and operator, during day shifts over a one week period. It was designed following an explanatory mixed methods approach (concurrent nested strategy) (Terrell 2012, Creswell 2014). Data were mainly collected via a time and motion study of individual trees harvested. This was complemented by the
175
E. R. Labelle et al.
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
Fig. 1 Location of time and motion study site depicted by star symbol filming of the entire harvesting process to allow a posteriori analysis of the variability in the results and to obtain a broader perspective on the influence of stem form (the qualitative portion of the explanatory mixed methods approach).
2.1 Site, stand, and harvesting system description The study was conducted during regular forest operations on public lands in northwestern New Brunswick (47° 28´ N; 66° 37´ W: Fig. 1). The test area is part of the Atlantic Maritime ecozone and within the climatic zone 4D characterized by having between 1200–1400 annual degree days > 5° C, and 500–550 mm of precipitation from May to September (Rees et al. 2005). The harvest prescription was an overstory removal where all merchantable trees, except sugar maple and yellow birch (Betula alleghaniensis Britton.) with a DBH less than 26 cm are normally harvested. However, for the purpose of this study, sugar maples from 10 to 26 cm DBH were also harvested to allow gaining insight on harvesting productivity in the lower tree DBH
176
categories, thus making the results of the harvest prescription more in line with a clear felling operation. Trees were processed within the harvest area. The target lengths for sawlogs were 265 cm and 287 cm (8’6’’ and 9’4’’), but the mill also accepted short logs of random lengths between 200 cm and 265 cm (6’6’’ to 8’6’’). For pulplogs, the target length was 265 cm (8’6’’) for 90% of the volume with the possibility to produce shorter logs in random lengths down to 112 cm (44’’) to maximize volume recuperation. A test area of approximately 2 hectares with a high proportion (60%) of sugar maple and a slope of less than 10% (southern aspect) was delineated for the study. The stand had an average pre-harvest merchantable basal area of 30 m2/ha, average tree density of 618 stems/ha, average tree quadratic mean diameter of 33 cm, and average tree height of 19 m. Similar stand and site conditions are commonly encountered in mechanized forest operations in New Brunswick. The cut-to-length machine observed was a Landrich single-grip tracked harvester equipped with a Ponsse H8 processing head with topsaw mounted on a 10 m long boom (Table 1 and Fig. 2, Landrich 2012). The Croat. j. for. eng. 37(2016)1
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
Table 1 Manufacturer’s specifications for Landrich single-grip harvester (Landrich 2012) Traction type
Steel rigid tracks – D6 undercarriage
Weight, kg
28,440 with H8 head and fluid levels full
Engine/power, kW
Mercedes Benz – OM906LA Tier 3/205 kW
Width/length/height, m
3.240/8.035/3.675
Boom maximum reach, m 9.965 Head
Ponsse H8 with top-saw
Maximum felling diameter 820 mm Maximum delimbing diameter, mm
740
Delimbing feed rate, m/sec 5 / 30 /force, kN
operator had fifteen years’ experience in mechanized operations. Most importantly, he had been operating the Landrich, equipped with the above-mentioned processing head, for a three year period and was very comfortable and efficient at using its top saw. The har-
E. R. Labelle et al.
vester was equipped with an onboard computer providing optimized bucking solutions. However, because of the high variability in the geometry of sugar maple stems for a given tree diameter, the operator made little use of the bucking solutions proposed. No adjustments of the parameters in the onboard computer were made nor were any guidelines given concerning the use of the proposed optimized solutions. Therefore, the bucking of stems into roundwood products was entirely left to the discretion of the operator.
2.2 Tree form classification Several key characteristics (forks, inclination, etc.) can be assessed to characterize tree form (Larson 1963, Millet 2012). The tree form classification system developed by the NHRI is an assessment of the first 5 meters of each tree on the vertical axis (Pelletier et al. 2013). The classification is derived from the presence and location of large branches and forks, curves, and lean (Table 2). A large branch was defined as having a diameter greater than one third of the main stem (measured below the branch). A fork was also defined as having a diameter greater than one third of the main stem below the fork but where it is impossible to identify the main stem (leader) above the fork (Pelletier et al. 2013).
Fig. 2 Landrich single-grip harvester (courtesy of ALPA Equipment Ltd) Croat. j. for. eng. 37(2016)1
177
E. R. Labelle et al.
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
Table 2 Description of the eight tree form classes (adapted from Pelletier et al. 2013) F1 – Ideal tree form
F2 – Acceptable tree form
– A single stem in the first 5 meters
– A single stem in the first 5 meters
– Without curve or with a curve on one axis
– Light curve on 2 axes or 1 significant curve on the stem
– Inclination of less than 15° from the vertical axis
– Inclination of less than 15° from the vertical axis
F3 – Poor tree form
F4 – Unacceptable tree form
– A main stem and the presence of large branches in the first 5 meters
– Multiple stems or branches in the first 5 meters
– The multiples branches represent potential for roundwood products
– The multiple branches have no potential for roundwood products
F5 – Poor tree form
F6 – Poor tree form
– Multiple stems are present between 0.3 and 1.3 meters from the base of the tree
– A single stem in the first 5 meters
– Single crown
– Significant inclination of more than 15° from the vertical axis F7 – Acceptable tree form
– A principle stem which is divided into a fork between 2.5 and 5 meters from the base of the tree – Inclination of less than 15° from the vertical axis
At the onset of the study, a parallel tree form classification scheme was used to provide additional information. Firstly, form classes F5 and F8 were eliminated since they essentially represent two or more trees from the standpoint of the usual forest inventory. Instead, each stem of the F5 and F8 trees was classified into one of the remaining classes. Secondly, a class »F9« was created for trees that forked at heights between 5 and 10 meters, as work by Plamondon (2010) has shown a significant reduction in harvesting productivity of a single-grip harvester when processing trees with a fork below 10 meters.
2.3 Field procedure and instrumentation Before forest operations commenced, 278 sugar maple trees were pre-assessed as potential study trees. The selection was based on two criteria: i) attempt to cover the full spectrum of form classes and DBH, ii) maintain sufficient spacing between consecutive study trees to allow the research assistant to differentiate logs originating from study trees during the subsequent log measurements for volume estimations. As a result, about one out of every six standing trees was selected as a sample. Study trees were numbered sequentially with paint at breast height. Form and DBH (2 cm classes) of each tree were recorded by a research assistant experienced with the tree form classification system. Continuous time and motion measurements were collected for all trees using the software TS-1000
178
– Light curve F8 – Poor tree form – Multiple stems are present under 0.3 meters from the base of the tree – Can represent a clump of trees from the same species or various tree species
(Gingras 2006). Work cycle was divided in the following elements: felling; processing; moving; brushing; rearranging logs and delays (any complete stop of the machine over a few seconds was recorded as a delay and associated cause was noted). Time per individual study tree was extracted from this data set. The complete time and motion data set was also used to determine a standard duration for work elements that are not common to every tree such as machine movement, brushing of understory vegetation, and for rearranging logs in piles. A short description of each timing element is presented in Table 3. Merchantable volume per tree was obtained by Smalian’s formula using two perpendicular diameter measurements (inside bark) at each end of the logs (mm accuracy) and length measurements to the nearest centimeter. Harvesting productivity per tree was calculated in terms of cubic meter of logs per productive machine hour. To further assess differences in harvesting productivity, tree forms were separated into »acceptable« (F1, F2, F7) and »unacceptable« (F3, F5, F6, F8) categories, as suggested by Pelletier et al. (2013). In order to analyze and explain a posteriori the variations in individual tree harvesting productivity, a digital video camera was mounted in the cab of the harvester and pointed directly towards the harvesting head. The videos were also used to provide a broader perspective on the influence of stem form on specific cycle elements, including non-productive time. Croat. j. for. eng. 37(2016)1
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
Table 3 Description of key machine cycle time elements Generic timing elements
Tasks included
Move
Travel between trees or to separate products
Brush
Cutting non-commercial/merchantable trees and cleaning area to pile logs
Head positioning
Head is moving towards the tree
Fell & drop
Cutting tree
Process
Processing tree (this element indicates the end of a cycle)
Operational delay
Machine stoppage
2.4 Statistical analyses Statistical analyses were performed with the Minitab 17 statistical package. To assess the effect of tree form on machine productivity, one-way analysis of variance (ANOVA) was performed and means were compared using the Tukey pairwise test. Machine productivity and recovered volume were the response variables studied, whereas tree form classes and tree DBH were terms used for comparisons. A significance level of 5% was used throughout all statistical analyses.
E. R. Labelle et al.
3. Results and Discussion Time and motion observations were made during 15.4 hours, which allowed observing the harvest of 645 trees in 14.1 productive machine hours (PMH). The average cycle time required to complete all elements of these 645 trees was 1.31 productive minutes, which is similar to the average delay-free cycle time of 1.29 minute per tree recorded by Huyler and LeDoux (1999) during mechanized CTL operations of a mixed wood stand. On average, 12% of the time was used for positioning the felling head and felling the trees, 71% for processing, and the remaining 17% for movement of the machine, brushing the understory, and rearranging log piles. Out of the 278 trees that were pre-assessed, only 109 were actually recorded during the time and motion studies in the field together with 536 other trees (not pre-assessed trees) for a total of 645 trees. Not all of the 278 initially identified study trees could be monitored due to logistical challenges and from this point forward, all results will pertain to those 109 study trees. Individual tree harvesting productivity ranged from 2.8 to 51.6 m³/PMH with an average of 18.5 m³/PMH, which is similar to the range of 4 to 47 m³/PMH reported by Puttock et al. (2005) for a cut-to-length operation in a mixed-wood stand in Eastern Canada. Again, when not discriminating between DBH classes,
Fig. 3 Harvesting productivity per tree form class along with sample size. Different lower case letters indicate a statistical difference between means at alpha 0.05. Tree schematics are only displayed for a visual representation of different form classes Croat. j. for. eng. 37(2016)1
179
E. R. Labelle et al.
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
mean harvesting productivity ranged from 13.6 m³/PMH for F8 trees to 22.4 m³/PMH for F1 trees with statistical differences between the two means (Fig. 3). Aside from a considerable difference in average DBH (24.6 cm and 35.6 cm) between trees of form F8 and F1, respectively, no further explanation for this statistical difference can be provided. F8 trees are defined by having a fork below 0.3 meters. The limited observations suggest that it is not simply this fork that caused the reduction in productivity but rather the simultaneous combination of other defects and forks on those trees. When separating tree forms into acceptable or unacceptable groups as suggested by the NHRI classification, 54 were acceptable form and 55 unacceptable. In comparison to acceptable trees, the average DBH of the »unacceptable« trees was 5% smaller, the observed average harvesting productivity (m³/PMH) was 18% lower (significantly different p = 0.032 based on a oneway ANOVA) and volume recovered was reduced by 16% (Table 4). This trend is also consistent with results provided by Ramantswana et al. (2013) that indicate a reduced harvesting productivity for Eucalyptus grandis stems of poor form (branches >5 cm maximum diameter at trunk) compared to those of good form when using an excavator-based harvester equipped with a Waratah HTH616 harvester head. The harvesting productivity of unacceptable trees being significantly lower than that of the acceptable trees (Table 4), a separate regression analysis was performed for each form category. When applying a regression with a power function, it is possible to observe that unacceptable trees have a harvesting productivity lower than acceptable trees across all DBH (Fig. 4). However, the models only explain between 25 and 32% of the variability, meaning that other factors play a dominant role in explaining individual tree productivity. In an expanded study offering better representation, data could be analyzed via a single regression using both stem size and form as variables to provide clearer evidence into the influence
of stem form on harvesting productivity. This could result into an improved model of single-grip harvester productivity in hardwood stands. High variability was apparent in both unacceptable and acceptable tree forms and revealed discrepancies in the expected influence of stem forms on machine harvesting productivity. In some instances, trees classified as being of unacceptable form showed the highest harvesting productivity compared to trees of similar size that were classified as acceptable (False negative prediction). In other cases, trees classified as acceptable showed the lowest harvesting productivity compared to trees of similar size classified as unacceptable (False positive prediction). Two approaches were taken to analyze these discrepancies: i) reclassifying which form classes were deemed acceptable on the basis of the presence of forks; ii) a qualitative analysis of the recorded video footage. The reclassification of all trees that had forks in the first 10 meters as being unacceptable resulted in 42% of the trees changing category. However, this reclassification did not reduce the discrepancies in the expected influence of stem form as there were just as many trees of acceptable form with the lowest productivities and trees of unacceptable form showing some of the highest productivity. We then used the harvesting videos to examine ten acceptable and ten unacceptable trees. Of these, half had very high harvesting productivities across tree sizes and the other half had relatively low harvesting productivity. For the trees that showed the lowest harvesting productivity, four main observations were made (no particular order): Presence of rot caused reductions in volume recovered and increased processing time as the operator tried to predict and buck bolts where the rot stopped in the stem. However, this tree characteristic is not considered a »form« characteristic.
Table 4 Average harvesting productivity observed as a function of tree form along with one-way ANOVA results (different lower case letters indicate a statistical difference at alpha=0.05) Tree form
DBH, cm
Harvesting productivity, m3/PMH
Volume recovered, m3/tree
Number of trees
Avg.†
Stand. error‡
Avg.
Stand. error
Avg.
Stand. error
Acceptable
54
33.8a
1.31
20.4a
1.22
0.64a
0.06
Unacceptable
55
32.0a
1.84
16.7b
1.13
0.53a
0.05
Total
109
32.9
1.13
18.5
0.85
0.58
0.04
† average ‡ standard error
180
Croat. j. for. eng. 37(2016)1
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
E. R. Labelle et al.
Fig. 4 Individual tree harvest productivity function of DBH for acceptable and unacceptable form categories. Due to a low frequency in the >60 cm DBH, the two data points identified by a star were omitted from the regression analysis Presence of a crook along the stem often required releasing the tree to relocate the processing head beyond the crook causing delays for the repositioning of the head. It also often meant that the portion of the stem containing the crook needed to be cut off, hence reducing the volume recovered. Presence of large branches (diameter greater than 10 cm; estimated from the video) on a merchantable portion of the stem often required multiple strokes from the processing head for proper delimbing. For some large branches, it was necessary for the operator to release the main stem and reposition the harvesting head to cut the branch. Presence of forks along the stem’s merchantable portion had mixed influences on harvesting productivity. The operator was seen to use five approaches to process forked trees: Þ When a tree forked below 1.3 meters (F5 and F8), each fork was processed as separate trees with only a slight impact on harvesting productivity. Þ When a tree had large forks and clear access to them while the tree was still standing, the operator would cut and process the forks before felling the main stem. This appeared to be an efficient manner to process forked trees. Quantifying this apparent difference in productivity and identifying best practices in harvesting forked trees would warrant a follow-up study. Þ When a tree forked in two relatively small and straight stems, the forks were processed simulCroat. j. for. eng. 37(2016)1
taneously (using the grapple to bend the forks inward), causing no apparent delays. Þ Whenever possible, forks were cut using the top saw while the main stem was processed. Severed forks were subsequently picked up and processed. This method appeared to be less efficient than when the forks were cut and processed from standing trees because picking up a fork on the ground sometimes posed a challenge. Þ When the use of the top saw to cut the forks was impractical due to tree position, the operator would release the main stem and align the head to cut the fork using the main saw. Grabbing the fork and proper positioning of the head to cut the fork appeared to be the cause of significant lost time. After having cut and processed the fork, the operator would pick up again the main stem, reposition the head at the beginning of the stem and continue its processing. When looking at the videos of »unacceptable« trees with an above average harvesting productivity, it was observed that: Þ Trees exhibiting a significant lean (>15°) did not appear to have a lower harvesting productivity. Þ Trees that forked below 1.3 meters were processed as two independent trees and so did not appear to have a significant influence on harvesting productivity when those trees had no other significant defects.
181
E. R. Labelle et al.
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183)
As the goal of the tree classification system used was threefold; predict current and future product distribution, determine harvesting costs, and indicate priorities for tree removal, altering decision criteria aimed at improving the applicability of a single facet, in this case harvester productivity, might negatively impact the general performance of the system. Nevertheless, our observations suggest that for the objective of predicting harvester productivity in hardwood dominated stands, two complementary variables should be collected; presence of a fork or large branch within a height of 10 m instead of only the first 5 m as indicated by the current system, and excessive crooks. The importance of considering forks or large branches up to a height of 10 m because of the almost automatic necessity to release the main stem from the processing head to cut them cannot be understated. The current study was limited in scope but still provided important and relevant findings concerning the influence of stem form on the productivity of a cut-to-length harvester. Through subsequent studies, additional information on the following elements could provide further insight. Improvements to the groupings of form categories specifically targeted for harvester productivity would allow for better and more representative predictions. A more tailored approach could be useful when selecting trees to be removed during pre-commercial thinning and/or commercial thinning treatments in order to reduce harvesting costs of subsequent harvests. Quantifying and observing the influence of rot (frequency, location within the stem, and severity) on volume recovered and machine productivity would provide complementary information, thus giving further insight on the effect of tree form and quality on the productivity of single-grip harvesters. This would allow understanding to what extent the productivity loss recorded with unacceptable trees is due to increased time consumption resulting from unacceptable form, and to what extent it is due to a lower value recovery.
4. Conclusion Hardwood trees can present extra challenges for single-grip harvesters during the processing phase due to their crown architecture. While the extent of this study was limited, it nonetheless suggests that the use of a stem form classification system can improve our understanding of expected single-grip harvester productivity in hardwood stands. On average, trees of unacceptable form showed a 15 to 20% decrease in harvesting productivity compared to trees of acceptable form. At the individual tree level, the stem form classification system used was not able to capture all of the variations. Many trees of unacceptable form showed
182
very high harvest productivity, while trees of acceptable form showed low productivity. Analysis of the videos suggests that simple modifications to the classification system could be made to better account for large branches and forks in the section above the first five meters. This should result in an increased difference in harvesting productivity between trees of poor form versus those of good form. Knowledge of the proportions of trees with different characteristics within a harvest area along with their impact on harvesting productivity could become a key indicator in a context of precision forestry. In addition to helping choose appropriate harvesting equipment, it could also suggest improvements in harvesting techniques and equipment.
Acknowledgements The authors wish to thank Groupe Savoie Inc. and Fornebu Lumber Company Inc. for allowing them to perform this pilot study on their operations. Thanks to Denis Caron, Tommy Caron, and Joël Beaulieu from AGB Foresterie Ltd. who allowed the research team to monitor its Landrich harvester during the study. Technical assistance by ALPA Equipment Ltd. was appreciated. Special thanks to Hector Guy Adégbidi, Ph.D. for his insight on the analysis and to Stephen Wyatt, Ph.D., Diane Beaupré, and Philipp Gloning for internal reviews. Authors also wish to acknowledge the insightful contributions received by external reviewers.
5. References Creswell, J.W., 2014: Research design: qualitative, quantitative, and mixed methods approaches. SAGE Publications, Inc., 273 p. Gingras, J.F., 2006: Time study procedures using Feric’s TS1000 program., 14 p. Hiesl, P., 2013: Productivity standards for whole-tree and cutto-length harvesting systems in Maine. Master of Science thesis, University of Maine, Orono, ME., 150 p. Huyler, N.K., LeDoux, C.B., 1999: Performance of a cut-tolength harvester in a single-tree and group-selection cut. United States Department of Agriculture, Forest Service Research Paper NE-711, 10 p. Jessome, A.P., 1977: Strength and related properties of woods grown in Canada. Environment Canada, Canadian Forest Service, Eastern Forest Products Lab. Ottawa, Ontario, Forest technical report 21 p. Landrich, 2012: Landrich multifunctional track harvester specification document, 4 p. Larson, P.R., 1963: Stem form development of forest trees. Society of American Foresters, Forest Science Monograph 5, 32 p. Millet, J., 2012 : L’architecture des arbres des régions tempérées: son histoire, ses concepts, ses usages. Éditions Multimondes, 432 p. Croat. j. for. eng. 37(2016)1
Effect of Tree Form on the Productivity of a Cut-to-Length Harvester ... (175–183) Nurminen, T., Korpunen, H., Uusitalo, J., 2006: Time consumption analysis of the mechanized cut-to-length harvesting system. Silva Fennica 40(2): 335–363. Pelletier, G., Landry, D., Girouard, M., 2013: A Tree Classification System for New Brunswick. ���������������������� Northern Hardwoods Research Institute, Edmundston, New Brunswick, 53 p. Plamondon, J.A., 2010: Comparaison des volumes et des productivités de deux approches defaçonnage de billes de feuillus: en cour et à la souche. FPInnovations, 12 p. Puttock, D., Spinelli, R., Hartsough, B.R., 2005: Operational trials of cut-to-length harvestingof poplar in a mixed wood stand. International Journal of Forest Engineering 16(1): 39–49.
E. R. Labelle et al.
Ramantswana, M., McEwan, A., Steenkamp, J., 2013: A comparison between excavator-based harvester productivity in coppiced and planted Eucalyptus grandis compartments in KwaZulu-Natal, South Africa. Southern Forests 75(4): 239–246. Rees, H.W., Fahmy, S.H., Wang, C., Wells, R.E., 2005: Soils of Central and Northern New Brunswick. Potato Research Centre, Research Branch, Agriculture and Agri-Food Canada, Fredericton, NB, 137 p. Terrell, S.R., 2012: Mixed-methods research methodologies. Nova Southeastern University, Ft.Lauderdale, Florida, USA, 254–280.
Authors’ addresses: Assist. Prof. Eric R. Labelle, PhD.* e-mail: eric.labelle@tum.de Technische Universität München Assistant Professorship of Forest Operations Hans-Carl-von-Carlowitz-Platz 2 D-85354, Freising GERMANY AND (at time of study) Forest operations research officer Northern Hardwoods Research Institute 165 boul. Hébert Edmundston, NB E3V 2S8 CANADA Assoc. Prof. Michel Soucy, PhD. e-mail: michel.soucy@umoncton.ca Université de Moncton 165 boul. Hébert Edmundston, NB E3V 2S8 CANADA André Cyr, BSc.F. e-mail: andre.cyr@alpaequipment.com ALPA Equipment Ltd. 800 Canada St. Edmundston, NB E3V 3K7 CANADA
Received: June 19, 2015 Accepted: August 7, 2015 Croat. j. for. eng. 37(2016)1
Gaetan Pelletier e-mail: gaetan.pelletier@umoncton.ca Director Northern Hardwoods Research Institute 165 boul. Hébert Edmundston, NB E3V 2S8 CANADA * Corresponding author
183
Preliminary note
Development of Harvesting Machines for Willow Small-Sizes Plantations in East-Central Europe Tomasz Trzepieciński, Feliks Stachowicz, Witold Niemiec, Leszek Kępa, Marek Dziurka Abstract The production of plant biomass in small farms within the Central and Eastern European countries requires the application of agricultural machines adjusted to the scale of production. In the article, new machines for small-sized plantations of energy crops have been presented. Furthermore, the results of strength analysis of three-point linkage mower frame are presented by finite element method. The advantage of the proposed solutions is their simple construction, which is connected with low production cost and simple machine operation. The energy-crop harvesting machines are especially intended for small biomass producers in EastCentral Europe, and the purchase of professional machines is financially inaccessible. The proposed machines are mounted in front or at the back of a typical farm tractor and the chipping units are power-take-off driven. The numerical simulation was carried out using finite element method to study the structural strength of a mower frame. All machine designs proposed by the Rzeszow University of Technology are legally protected by patents and utility models. The presented agricultural solutions have been worked out by authors and a part of them is still being developed in cooperation with biomass producers. Keywords: biomass, energy plants, harvester, renewable energy sources, willow
1. Introduction Woody biomass can come from many sources, including forestry thinning operations logging slash, wood product residues, urban waste wood and woodrelated industries. While material for woody biomass can come from many sources, this profile will focus on the production of trees specifically for bioenergy. Biomass as raw material for European energy production has received raising interest during the last decades from policy makers, energy utilities and farmers (Djomo et al. 2011, Schweier and Becker 2012b). The utilization of natural and renewable energy sources is mainly caused by aspiration to protect environmental and depleted stocks. Among all kinds of renewable energy sources in Poland, the great potential for fast utilization is in biomass especially in biomass of fundamental raw materials i.e. in straw, wood and energy plants. This is also a reaction to the European Croat. j. for. eng. 37(2016)1
Commission (EC) recommendations that established the binding target of 20% share of renewable energy sources in the EU energy consumption by 2020 (COM 2008). In Poland, it has to reach a level of 15% by 2020. Short rotation coppice (SRC) consists of densely planted, high-yielding varieties of either willow or poplar, harvested on a 2–5 year cycle, although commonly every 3 years (DEFRA 2004). Among the various crops for biomass options, SRC is especially regarded as strategic resources of wood products that are fast-growing and high-yielding species (Djomo et al. 2011). Analysis of properties of different types of renewable energy sources (RES) presented by authors (Trzepieciński et al. 2013) leads to the conclusion that woody biomass in East-Central Europe is a RES type characterized by the best synthetic technical and operating parameters. The biomass energy is commonly available and, at the same time, it is characterized by low
185
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
unit cost, high availability and energy comparable with traditional mineral fuels, which allows the common use of biomass in dissipated power engineering based on co-generation systems. Intensive breeding programs of willow production resulted in high production rates for a wide range of climates within the EU. The deficiency of specialized machines adapted to arable areas and their high cost are the following obstacles indicated by potential small producers of biomass, as pointed out by Niemiec et al. (2012b). Increasing of the amount of energy plant plantation causes the necessity of exploration of new technologies allowing efficient harvesting and further processing. Among the various crops, SRC is especially regarded as a strategic resource of wood products, being a fast-growing and high-yielding species that can be managed as a coppice system (Schweier and Becker 2012a). Short rotation coppice is a very extensive form of land management in comparison to conventional agriculture, as crops are harvested in a 2–4 year cycle. The use of the above presented machines in agricultural practice can be useful in solving the most important problems of environment protection, as shown by Spinelli et al. (2008) and Vervijst and Telenius (1999): Þ application of environmentally friendly technology in management system of sewage sludge in production of both energy and industrial plants, Þ utilization of woody biomass for heating purposes or generating of electrical energy not affecting the increase of global warming, Þ elimination of human hazardous substances included in sewage sludge in compliance with the possibility of their environmental control. Different harvesting machines have already been developed. Two main harvesting approaches have been developed for short rotation woody crops, i.e. the harvest-and-storage system (Fig. 1a) and harvest-andchip system (Fig. 1b). In both cases chips are produced from wet stems, collected in an attached trailer and stored as wet chips. In the harvest-and-storage system, wet stems are cut, transported to a storage location to dry, and chipped afterwards to obtain dry chips (Berhongaray et al. 2013). The storage of wet chips implicates a risk of losing dry matter, and possibility of requiring further drying. In the harvest-and-storage system wet stems are cut first and then transported to a storage location. After drying, they are chipped to obtain dry chips. To reach lower moisture content, chips can be dried and stored under roofs or covered chip piles (Balsari and Manzone 2010).
186
Fig. 1 Presentation of the harvest-and-storage (a) and harvest-andchip (b) systems Advantages and disadvantages, costs and harvest capacities of harvesting technologies e.g., Log Lines, Bundle Lines, Chip Lines and Bale Lines were presented and discussed in the latest publications of Abrahamson et al. (2010), Schweier and Becker (2012b, 2012c), Savoie et al. (2012) and Ehlert and Pecenka (2013). In the East-Central European countries, plantations are usually grown on extremely small areas. Furthermore, they are characterized by low technological level. The cost of the harvesters equipped for SRC cropping is approximately 420,000 € (Ehlert and Pecenka 2013). In addition, for the economic operation of highly productive harvest systems, cultivation areas of more than 300 ha are required (Scholz et al. 2009), if the harvester is used for SRC cropping alone. For small biomass producers in East-Central Europe, purchase of professional machines is financially inaccessible and unprofitable. The aim of the article is a presentation of new machines for small-sized plantations of energy crops. The advantage of the proposed solutions is their simple construction, which is connected with low production cost and simple machine operation. The proposed machines are mounted in front or at the back of a typical farm tractor and the chipping units are power-take-off (PTO) driven. All technical solutions of machines proposed by the Rzeszow University of Technology (RUT) are legally protected by patents and utility models.
2. Woody biomass crop 2.1 Biomass harvesters Mechanization in the harvesting of short rotation coppice on farmed land is a prerequisite for the expansion of short rotation coppice cropping (Ehlert and Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Pecenka 2013). On the market, there are many professional high efficiency energy-crop harvesting systems for SRC used for cut and storage. Furthermore, SRC are commonly harvested in a combined cut and chip system with modified forage harvesters used for maize and other crops. They are either self-propelled or tractor mounted. Class Jaguar forage harvesters with different cutting heads dominate in the European agriculture. The productivity of these harvesters was studied by Spinelli et al. (2009). The average machine productivity was 25–35 green tons (gt) per scheduled machine hour (smh–1) and 25gtsmh–1, respectively. Only few studies, e.g. by Becker et al. (2010) and Schweier and Becker (2012c), were identified to analyze other combined cut and chip machines than the Claas Jaguar. The average share of productive working time of New Holland forage harvester FR 9060 equipped with the cutting head FB 130 was 71% (Schweier and Becker 2012a). It took 1.34 pmh–1 to harvest one hectare and the average machine productivity was 31 gtsmh–1. Today, there are few harvesting machines applying cut and storage system and most of them are prototypes (Scholz et al. 2009). Among professional machines, attention should be paid to Empire 2000 with the productivity of 0.75 ha smh–1 and 17.8 gtsmh–1 (Danfors and Nordén 1995), Laughry with the productivity of 0.13 ha smh–1 (Mitchell et al. 1999) and Fröbbesta with the productivity of 0.36 ha smh–1 and 15.7 gtsmh–1 (Danfors and Nordén 1995). The productivity of four machines tested in the study (Forestry Commision 1998) varied between 0.09–0.22 ha smh–1. Resulting costs were two times higher than in the cut and chip system. As stated by Schweier and Becker (2012b), the cut and storage system is still under development but there is great potential for the future.
2.2 Harvesting on small-sized plantations 2.2.1 SRC harvesting methods Today, as reported by Ehlert and Pecenka (2013), few harvesting machines apply this concept and most of them are prototypes. SRC are usually harvested in winter after leaves fall and before leaves set, preferably when soils are frozen (Forestry Commission 1998). In small-sized plantations, there are two different systems to harvest willow, the combined harvest-andchip system and the harvest-and-storage system. Willow is harvested with an agricultural forage harvester (e.g., for maize and sugarcane), either self-propelled or tractor mounted, whose standard header is replaced by a special cutting head. Chips are blown into an accompanying tractor-pulled trailer, which transports the chips to an interim or final storage (Sambra et al. 2008). Croat. j. for. eng. 37(2016)1
T. Trzepieciński et al.
At the RUT establishing technology of energy plant plantation, harvest methods and technology of sludge processing were elaborated. The main aim of machine solutions proposed by the authors is their readjusting to agricultural needs of quite small farms especially in Poland. Due to simple construction of the proposed machines and instruments, they are also cost attractive. One of the premises in performing operations connected with the construction of new machines for planting, logging and processing of plant biomass with ligneous shoots is high cost of high-productivity machines (Niemiec et al. 2011, Niemiec et al. 2012a). So, their application in small farms is not profitable. Meaningful growth of the amount of energy plant plantations, mainly plants with ligneous shoots, requires further exploration of new technologies that would allow efficient harvesting and processing. Over the last years, progress has been recorded in submitting solutions of specialized machines for harvesting and processing of biomass logging from ligneous shoots e.g., mowers, wood chipping machines, chaff cutter for wood, but the problem still exists and requires further research. Machines for small plantations and the adapter for terrain hypsography are particularly investigated. In the case of establishing protective treatments, logging and processing of the biomass in plantations of the energy plants, farm tractors are the basic prime mowers of specialized machines. Moreover, farm tractors usually transport the crop from plantations to further processing or to final management of energy biomass. There are not many commercial offers of machines characterized by low productivity and intended for working in small plantations, and also verified in practice. Characteristic feature of Polish agriculture is the size reduction and low degree of mechanization as well as limited buying power. In this situation, the search for construction solutions adjusted to the requirements of small farms is justifiable. Machine solutions proposed by the RUT are legally protected by patents and utility models (Table 1). To provide the required power and efficiency, the chipping units are mostly PTO driven. Considering the efficiency of application of machines proposed for production, logging and processing of the biomass in small farms, a few factors should be taken into account. First of all, the size of a farm should be considered and in consequence the demand on manpower and degree of accessibility of high power farm tractors. The terrain hypsography and soil structure should be considered from the point of view of possibility of machine utilization. The advantage of
187
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Table 1 Machines developed at the Rzeszow University of Technology Machine
Patent, P / Utility model, UM no.
Year
UM – 116926
2007
P – 386842
2010
Feed mechanism of cut material in wood cutter
UM – 119154
2010
Mobile harvester for biomass logging of plants with ligneous stems
UM – 120576
2011
Harvester for crop and chipping of ligneous stems of energy plants and branches
UM – 119895
2011
Harvester for crop and chipping of ligneous stems of energy plants
UM – 120965
2012
Unit for harvesting and chipping of ligneous stems of plants
UM – 121680
2013
Harvester for crop and chipping of ligneous stems of energy plants and branches
UM – 121863
2013
Wood cutter Mower for tree-like plants
the proposed solutions is their simple construction, which is connected with low production cost and simple machine operation. The whole-tree harvesting is the basic method of the willow harvesting in small-sized farms. In this method there are several ways of harvesting: Þm anually, using a bill hook or similar implement, Þm otor-manually, using a chainsaw or brush cutter, Þu sing tractor-drawn equipment. Obviously the manual and motor-manual approaches are suited for small areas only, where timberl is harvested for household consumption. Kofman (2012) has shown that tractor-drawn equipment can be used on larger areas of coppice, while self-propelled machines are best suited for very large blocks. Some tractor-drawn machines collect shoots on a loading bay. The bundle of the shoots is easy for transporting and subsequent handling. Some machines transport the shoots to the landing, while others unload directly onto the field. 2.2.2 Three-point linkage mower The design process of a new machine started from building a conceptual spatial model of a construction with Autodesk Inventor program. Autodesk Inventor allows to determine kinematic constraints between particular movable parts of the machine, and the program enables to observe possible collisions between parts of the machine. Autodesk Inventor allows kinematic and dynamic analysis of working mechanisms. In order to determine relations between mate components, a wide range of motion constraints and elastic or damping elements, as well as definition of friction coefficient in every constraint, may be used. To under-
188
stand the essence of the kinematic effects, the program demonstrates the simulation in a form of spatial visualization directly on the model of the analyzed mechanism. During the designing of mechanical elements, the question arises whether a specific element with stand load appears in normal operations. In complex constructions, analytical determination of the most efforted places responsible for element destruction is very difficult and often impossible. In order to solve this problem, the optimization of numerical construction of the mower has been considered using ABAQUS program based on finite element method (FEM). ABAQUS allows analyzing physical models of real processes by putting special emphasis on geometrical nonlinearities caused by large deformations, material property, nonlinearities and complex friction conditions. The dimension of the geometrical model of the mower frame corresponded to the conception model. A prerequisite to begin the computation using FEM is the preparation of an accurate numerical model of the selected construction, and then simplification and digitizing of the model in ABAQUS in order to receive equivalent model to the mathematical model of continuous medium. Simplification of the model consists in removing unnecessary details such as small holes, roundings and chamfers, which do not influence meaningfully the accuracy of solutions for the purpose of fast processing of the model. Digitizing consists in the division of the continuous medium on finite numbers of elements of specified shape. In order to obtain the required accuracy of the searching solution, the used elements should be small as much as possible so that they may be approximated inside elements functions using multinomial models. Nevertheless, the reduction of elements number leads to an increase in the number of Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
equal to the thickness of the circular saw tooth used in the machine and it is 4.5 mm. The feed value is equal to the thickness of the chip t during cutting willow stems – about 0.4 mm per tooth (Spychała et al. 2010). Operating speed of the machine is 6 km h–1 and the rotational speed of the circular saw is n=2100 rev. min–1. After inserting the above values into Eq. (1), the value of the force acting on a single tooth of the saw is 5.76 daN.
Fig. 2 Components of cutting resistance searched function of node values, which simultaneously results in lengthening the computation time. So, most often, heterogeneous division of the model on elements was used. In places of expected high stress, gradient mesh elements should be concentrated. The following stage in developing the FEM model is to take into consideration boundary conditions and parameters describing material of the mower frame. The mower frame is constructed of S355J2G3 constructional steel. The values of basic mechanical parameters of this steel were determined using tensile test and they are as follows: Young’s modulus E=210,000 MPa, Poisson’s ratio v=0.33, yield stress ReH=353 MPa. The material density ρ is 7865 kg/m3. It was assumed that circular saw is made of 145Cr6 tool steel with the following material properties: E=210,000 MPa, Poisson’s ratio v=0.33, yield stress ReH=380 MPa. The analysis also took into consideration the weight of the mower. The end of the cantilever jib on which the saw is attached is exposed to resultant forces from the cutting process. Values of these forces were determined so as to introduce construction into first plastic strain. The momentary resistance of stem cutting may significantly affect the way and the value of frame loading. To determine the resistance against cutting, a simplified model of the forces distribution (Fig. 2) can be used. Assuming that the circular saw cuts the material parallel to the ground, the cutting resistance R can be divided into two components. The force required to overcome the cutting resistance, which is parallel to the cutting movement of the cutting edge – Fc, is determined from equation:
Fc = rc × w × t
T. Trzepieciński et al.
The value of thrust component of force Ft perpendicular to the cutting force Fc is determined from Eq. (2). Thrust coefficient ct value for the cutting edge with blunt edges is on average 0.5.
Ft = ct × Fc
(2)
Where: Fc Ft ct
cutting force; thrust force; thrust coefficient.
The value of the thrust force acting on a single tooth of the circular saw was 2.88 daN. In the three-year energy willow plantation, the stems diameters are very different. It was assumed that the greatest loading of the mower frame occured when cutting simultaneously two willow stems with a diameter of 70 mm. Under these conditions, the load is transmitted simultaneously through five circular saw teeth (Fig. 3). The arrows present the directions of action of cutting forces and thrust forces acting on the teeth during cutting. The mower frame is composed of 38307 4 node elements called C3D4 in Abaqus terminology. The finite elements mesh of the mower frame is shown in Fig. 4. In order to improve the computational accuracy, the local mesh densification is assumed. Calculations were performed using the implicit finite element code. In the implicit method, the inter-
(1)
Where: rc specific cutting resistance; w chip width; t chip thickness. Specific cutting resistance rc for willow stems for cross-cutting with sharp-edged blade, assumed to be 3.2 daN mm–1 (Spychała et al. 2010). Chip width w is Croat. j. for. eng. 37(2016)1
Fig. 3 Simultaneous cutting of two willow stems with a diameter of 70 mm
189
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Fig. 4 The finite element mesh of the mower frame nal forces are made to balance with the external force through an iterative procedure, from which the deformed state after a time increment can be obtained. One of the merits of this method is that the time increment can be relatively large because of conditional stability of the implicit time integrator, and static solutions can always be obtained by natural characteristics of the method. To solve the Newton-Raphson’s iterations, the time step is controlled by Abaqus automatic incrementation technique. The double precision executable is used in this analysis. The global equation in models based on FEM is represented as:
{}
F (t ) = K {d} + M d
(2)
Where:
boundary non-linearity such as contact, sliding and friction. The solution is obtained as a series of increments with iterations to obtain equilibrium within each increment. Implicit method enables full static solution of the deformation problem with convergence control. The aim of the simulation was to determine the places in construction with potential failure. In order to determine factual strain values in the mower frame, it is necessary to make analytical or experimental evaluation of the real value of load forces of the frame. The value of equivalent von Mises stress localized in the mower frame does not exceed the allowable stress for S355J2G3 steel (Fig. 5). The most loaded place is the welded beam between the mower frame and the cantilever jib (Fig. 6). The analyzed construction is relatively simple, so it was possible to determine approximately the places endangered to fail without the application of advanced programs. However, the present analysis will be used to analyze more complex machines e.g., willow and bush harvester. For economic and trouble-free harvesting of small willow plantations with free arrangement of stems, single-row mower is developed. The structure of a developed and patented mower for cutting down woody plants (Fig. 7) is composed of the frame (1), three-point linkage (2) for attaching to an agricultural tractor and
K global stiffness; M mass; F force matrices; d nodal displacements; d nodal accelerations. The analysis available from ABAQUS, which is used in the model, handles non-linearity from large displacement effects, material non-linearity and
Fig. 6 Distribution of Huber-Mises-Hencky stress (MPa) in the beam between the mower frame and cantilever jib
Fig. 5 Distribution of Huber-Mises-Hencky stress (MPa) in the mower frame
190
Fig. 7 Model of a tractor mower for woody plants Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
T. Trzepieciński et al.
is connected with the frame using articulated joint (5). The remote control hydraulic cylinder is coupled with the extension arm (3). The end of the working arm (2) has rollers mating with the guide (7). The circular saw is driven by the hydraulic engine (10). The drive of the circular saw is transmitted by the reduction gear (8) and the hydraulic pump (9). The mass of the complete tractor-mounted mower was about 450 kg, and it should be mounted in three-point linkage of mediumsized standard tractors (minimum 60 kW).
Fig. 8 Prototype of the patented tractor-mounted mower for woody plants working arm (3) coupled to the frame using articulated joint (4). The ground wheel (5) is mounted to the working arm near its end. At the end of the working arm, a circular saw (6) with a diameter of 600 mm is mounted. To provide the required power and efficiency, the saw is PTO driven. The drive of the circular saw is transmitted from PTO by using a shaft (7), intersecting the axis gear (8) and the belt transmission (9). All movable elements of the machine are protected by safety guards. The other construction solution of the circular saw drive is a long shaft between the saw (6) and the intersecting axis gear (8) instead of the belt transmission (9). The prototype of the tractor-mounted mower for woody plants is presented in Fig. 8. The modification of the presented mower for woody plants is the mower with the retractable arm and hydraulic drive of the circular saw (Fig. 9). The mower structure consists of the frame (1) adapted to three-point linkage of agricultural tractor and the working arm (2) with the extension arm (3). At the end of the extension arm, the circular saw (4) and the ground wheel (6) are mounted. The working arm (2)
Fig. 9 Model of a tractor mower made in Autodesk INVENTOR program Croat. j. for. eng. 37(2016)1
2.2.3 Harvest-and-chip machines In the harvest-and-chip method, there are two main ways of harvesting: Þ tractor-mounted equipment, Þ self-propelled machines. Tractor-mounted equipment is cheaper and has a lower weight than self-propelled machines, but has a lower productivity as shown by Kofman (2012). Biomass may be processed directly in chips by the tractor-mounted harvester (Fig. 10a), especially developed for cutting specified pieces of bevelled energy plants or branches, which are the discards from e.g., cross-cutting of the fruit tree in an orchard. Cutting assembly of the harvester has replaceable heads with three, four or six cutters allowing for the change of cutting length of stems. The machine is intended for harvesting and chipping plants with ligneous stems and branches. The harvester is mounted on the agricultural tractor using three-point linkage (1) and it has a drum cut system. The drive of the harvester is transmitted from power-take-off of the agricultural tractor by belt and gear transmission. The cutting assembly (2) of the harvester has two drums (3) to which circular saws (4) are mounted. The drums (3) have strips (5) that make biomass easy to retract into the throat of the cutting assembly (6). The cutting assembly is composed of feed assembly and knives (9) mounted on two rotational heads (8). An alternative model of harvest-and-chip machine (Fig. 10a) is used for harvesting willow coppice, using a harvest-and-billet system (Fig. 10b). It is capable of producing much coarser chips or even chunks. The normal setting of the machine produces 40–60 mm chips, but by removing some of the knives and lowering the drum speed, 150–200 mm billets can be produced. Tests performed with willows demonstrated the high potential of the new concept as a low cost billets production. The maximal diameter of cut willow stems is 100 mm (Fig. 11). Both cutting and chipping systems are protected by overload coupling. To facilitate the motion of stems to the machine throat, the cutting assembly is equipped
191
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Fig. 10 Model of harvester with discharge chute (a) and belt conveyor (b)
Fig. 11 Billets with a diameter of 100 mm with the guiding arm (10) which bends high stems. The longer chips are transported on the trailer by using the belt conveyor (11) (Fig. 10b). The short chips are transported alternatively by using discharge chute (12) (Fig. 10a). Cutting assembly is equipped with two circular saws with a diameter of 700 mm, which overlap at the joint in a width of about 50 mm. The rotational speed of circular saws should be about 2100 rev/min (Spychala et al. 2010), which allows for a circumferential cutting speed of 100 m/s at the diameter of the saw of about 700 mm. This speed value is necessary to obtain a smooth cut surface and reduces the risk of splitting phloem under the bark, which could contribute to the snags putrefaction. Therefore, machines for willow harvesting use circumferential speeds of cutting saws
192
ranging between 63 and 118 m/s (Lechasseur and Savoie 2005). The chipping assembly (Fig. 12a) is a constructional solution adapted from a wood cutter (Table 1) and is composed of two main units: feeder (1) and cutting assembly (2). The feeder comprises two shafts (3) on whose circumference toothed blades (15) are located. In the chipping unit, two heads (4) have blades (5) arranged uniformly at the heads circumference. The stem (6) is cut to billets (7) whose length dependents on the ratio of the rotational speed of both feed shafts and cutting heads. As shown in Fig. 12b, the feeder of stems comprises two feed shafts, the driving (8) and the clamping (9), arranged one above the other. The lower driving shaft (8) is coupled to the feeder frame (10) by units of bearing and unidirectional clutches (11). In turn, the upper feed shaft (9) is connected to the feeder frame (10) and the lower drive shaft (8) by compression springs (12). The springs are slidably coupled with guides (13). The lower and upper shafts are covered by a suitably shaped cover (14). The shafts (8) and (9) have toothed blades (15), which protect hard wood from slipping, thus ensuring accuracy of cut to a predetermined length. The cutting head (Fig. 13) has a body (1) with a shaft hole (2). A knife seat (3) is placed In the body. In the bottom of the knife seats, adjusting screws (4) are placed. Knives (5) are secured with wedges (6) bolted to the body with screws (7). Depending on the required length of chips, the chipping assembly has three, four or even six knives located uniformly on the head circumference. We cooperate with the following companies: R&D Centre Inventor Sp. z o.o. and SIPMA S.A. – one of the leading manufacturers of agriCroat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185â&#x20AC;&#x201C;199)
T. TrzepieciĹ&#x201E;ski et al.
Fig. 12 Chipping (a) and feed (b) assemblies
Fig. 13 Cutting head
cultural machines in Poland. The results of our collaboration are prototypes of chipping machines manufactured as electric motor driven (Fig. 14a) and PTO driven (Fig. 14b). These machines were manufactured by modification of the above mentioned harvesters. The field testing of machines indicates full usefulness of chipping devices. In the alternative solution, the shafts of circular saws can be driven by hydraulic engines. As the speed of tractor/machine arrangement can be changed depending on field conditions, the use of a hydraulic engine, controlled by electric-hydraulic distributor, will allow for optimum selection of the saws angular velocity so that the circumferential speed is higher than the speed of the farm tractor. The speed of the tractor will depend on field conditions, but the maximum speed cannot exceed the recommended
Fig. 14 Prototypes of chipping devices: electric motor-driven (a) and PTO-driven (b) Croat. j. for. eng. 37(2016)1
193
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Table 2 A qualitative and quantitative comparison of the harvestand-chip and harvest-and-billet methods for harvesting SRC-willow, prepared on the basis of own experience and Kofman’s study (2012) Factor
Area
Harvesting productivity tonnes/ha
Drying possibility
Markets, ha
Area
Cut-and-chip(1)
Cut-and-billet(1)
≤1 ha
–
–
1¸5 ha
+
+
³5 ha
++
++
≤1 ha
<10
<12
1¸5 ha
<25
<20
≥5 ha
<25
<20
≤1 ha
+
++
1¸5 ha
–
+
³5 ha
––
+
all
>5
1–5
8 km h–1 (Lechasseur and Savoir 2005). Cutting height of stems of 100 mm above the ground is in the range usual in similar solutions i.e., 50–100 mm.
Characteristic features of the presented harvester are: Þ possibility of application of each typical agricultural tractor with a power above 60 kW, Þ multifunctionality: chipping on plantation or stationary chipping, Þ by the operating parameters, the harvester is adapted for especially small plantations of energy plants. Both methods of SRC harvest are highly profitable in the case of larger plantations (Table 2). The harvestand-billet method can be used because billets dry a lot easier than chips (Table 2). However, billets generally have to be re-chipped before being used as fuel or may be used as a whole in family households that use biomass for heating. Limited resources of energy, risks connected with emission of thermal gas and low efficiency systems of thermal energy transmission lines require exploration of other methods of generation, transmission and consumption of electric and thermal energy. One of the solutions of this problem is the idea of generating electric and thermal energy directly in a flat. Single-family homes especially use the system of combined heat and power for home (CHPH) in association with Stirling Engine supplied with billets or chips (Szewczyk and Trzepieciński 2012). It has been
Fig. 15 Machine for harvesting and chipping of energy plants – the top (a), isometric (b) and tractor mounted (c) view
194
Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
T. Trzepieciński et al.
Fig. 16 Back (a) and front (b) views of prototype of one-row side-operating SIPMA SR 1010 HERO harvester
Fig. 17 Prototype of two-row side-operating SIPMA SR 2010 HERO harvester (a) and cutting system (b)
shown (Kofman 2012) that for a large-scale use, the harvest-and chip is the cheapest way of harvesting willow SRC.
3. Industrial application of harvesting technology The result of cooperation of Rzeszow University of Technology and SIPMA joint-stock company, one of the Croat. j. for. eng. 37(2016)1
leading manufacturers of agricultural machines in Poland, are a series of harvesters for short rotation willow coppice. Both one-row (Fig. 16) and two-row (Fig. 17a) harvesters are side-operating and tractor-pulled singlepass harvest-and-chip machines. The harvester should be accompanied by an additional, separate tractor–trailer combination to collect the chips. The machines are intended to work on plantations with 0.75 m between rows. The header attachment for the »HERO« series
195
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Fig. 18 View of the cut shoots
The two front rear feed rollers in the header grab the crop and draw it back toward the feed roller module of the harvester. The lower front feed roller in the header has deep paddles to help flip the base of each stem up into the two feed rollers behind it that are spring loaded. The harvesters are equipped with a PTO driven drum type chipper. Four knives are mounted on the rotor. Many elements of both single and double-row machine are hydraulically driven, namely: Þ discharge pipe tilt mechanism, Þ lid tilt mechanism, Þ system of angular adjustment of bar position, Þ control system of cutting height, Þ circular saw drive.
Fig. 19 Drive mechanism (a) and woody crop feeding system (b) of SIPMA SR 1010 HERO harvester harvesters uses circular saws mounted in a horizontal plane to cut short rotation crops. The circular saws (Fig. 17b) with diameter of 0.4 m are driven by the HPL MA2 hydraulic engines controlled by electric-hydraulic distributor. The hydraulic engine drive of saws is much simpler than the structure of mechanical gears that transmitted the power from power-take-off (PTO) to circular saw. Hydraulic drive is characterized by high flexibility that fully eliminates failures due to overloading of the cutting head. Operating speed of the machine is 6 km h–1 and the rotational speed of the circular saw is n=2100 rev. min–1. The use of cutting system with carbide-tipped circular saw blades does not lead to negative tearing of fibrous tissue and stem bark (Fig. 18). However, the carbide blades were vulnerable to rocks and had to be replaced after approximately 10 ha. Feeder drums located vertically over the saw blades rotate at approximately 160 rev. min–1 to actively collect and feed the crop back into the harvester throat. The header has four horizontally mounted feed rollers (Fig. 19) that help feed cut stems back into the chipping unit.
196
All the mentioned mechanisms are controlled by waterproof mobile control panel located in the cab. Chip size distribution is, therefore, the best indicator of how a chipper performs in terms of product quality. Dimensions of fuel wood chips are limited to 50 mm, however, at least 90% of the chips should not exceed 40 mm in size. In the case of storage, too exces-
Fig. 20 Example of willow chips Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
T. Trzepieciński et al.
sive fragmentation of chips in combination with high humidity is disadvantageous because of the possibility of the digestion of stored chips. The presence of oversize particles may prevent the use in small and medium-size plants, those that generally offer the best price opportunity. However, very fine chips are known to store poorly. The size distribution of willow chips (Fig. 20) meets the geometric requirements of fuel wood chips. The major fraction of the crop are chips with length of about 35–45 mm. This confirmed the proper synchronization of feeding speed and rotational speed of the drum chipper.
can be used directly for heating in the single-family housings. Considering the efficiency of application of the proposed machines for harvesting and chipping the biomass in small farms, a few factors should be taken into account. First of all, the size of a farm should be considered and in consequence the demand on manpower and the grade of accessibility of high power farm tractors. Considering the possibility of using the proposed machines, terrain hypsography and soil structure should also be taken into account.
Field tests confirmed that prototypes of willow harvesters meet the main design guidelines, i.e.:
5. References
Þ machines provide handling of at least 3 year old willow with maximal stem diameter of shrub of 70 mm, Þ tractor power required to drive harvesters does not exceed approximately 120 kW, Þ drum type chipper can produce chips that are 35–50 mm long, Þ effective material capacity is 30 Mg per hour, Þ prototypes should meet the requirements of Road Traffic Act, Þ objectives of harvesting include the ability to harvest willow coppices containing stems of up to 120 mm in diameter, Þ harvesters are equipped with centrifugal discharge pipe allowing direct transport of chips into a trailer at the distance of 5 m.
4. Conclusions The proposed agrotechnical machines have original construction and take into account the needs of small producers of ligneous biomass. Machines intended for operation in small and medium-sized plantations of energy plants are characterized by simple construction and they do not require highly skilled workers. In the case of establishing protective treatments, logging and processing of biomass in plantations of energy plants, the farm tractors are the basic prime mowers of specialized machines. The majority of developed machines are mounted on a farm tractor using levers of toolbar assembly. For a small-scale use, the whole-shoot cutting method is probably the best. Alternatively, for small-scale use, the harvest-and-chip method can be used but a special dryer is required for drying chips and it is very expensive. So, the harvestand-billet method is more economic, because billets dry much easier than chips. Furthermore, the billets Croat. j. for. eng. 37(2016)1
Abrahamson, L.P., Volk, T.A., Castellano, P., Foster, C., Posselius, J., 2010: Development of a harvesting system for short rotation willow & hybrid poplar biomass crops. In: SRWCOWG Meetting, Syracuse NY, USA, 165–183. Balsari, P., Manzone, M., 2010: Evaluation of different wood chips storage techniques. In: Proceedings of the Symposium on Forestry Mechanisation (FORMEC), Padova, Italy, 10 p. Berhongaray, G., El Kasmioui, O., Ceulemans, R., 2013: Comparative analysis of harvesting machines on an operational high-density short rotation woody crop (SRWC) culture: One-process versus two-process harvest operation. Biomass and Bioenergy 58: 333–342. Becker, R., Scholz, V., Wegener, J., 2010: Maschinen und Verfahrenfür die Ernte von Kurzumtriebsplantagen. In: Bemmann A., Knust C. (eds.). AGROWOOD. Kurzumtriebsplantagen in Deutschland und europäische Perspektiven. Weissensee-Verlag, Berlin, 88–102. COM (Commission of the European Community) 2008: 19 final: Proposal for a Directive of the European Parliament and of the Council on the promotion of the use of energy from renewable sources. Brussels 23. January, 61 p. Danfors, B., Nordén, B., 1995: Sammanfattande utvärdering av teknikoch logistic vid salixskörd. Slutrapport over analys av hanterings – och transportsystem vid skördav Salix. JTI Rapport. Jordbrukstekniska Institutet, Uppsala. DEFRA, 2004: Best practice guidelines for applicants to Defra‘s energy crops scheme. http://www.naturalengland.org. uk/Images/short-rotation-coppice_tcm6-4262.pdf. [Cited 10 Aug 2014]. Djomo, S.N., Kasmioui, O.E., Ceulemans, R., 2011: Energy and greenhouse gas balance of bioenergy production from poplar and willow: a review. GCB Bioenergy 3(3): 181–197. Ehlert, D., Pecenka, R., 2013: Harvesters for short rotation coppice: current status and new solutions. International Journal of Forest Engineering 24(3): 170–182. Forestry Commission, 1998: Harvesting and comminution of short rotation coppice. Harvesting machine trials. Technical Development Branch. Technical Note 8/98. Forestry Commission, Dumfries.
197
T. Trzepieciński et al.
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
Kofman, P.D., 2012: Harvesting short rotation coppice willow. Harvesting/Transport 29: 21–24. Lechasseur, G., Savoie, P., 2005: Cutting, bundling and chipping short rotation willow. The Canadian Society for Engineering in Agricultural, Food, and Biological Systems. Paper 05-080, 11p. Mitchell, C.P., Stevens, E.A., Watters, M.P., 1999: Short-rotation forestry – operations, productivity and costs based on experience gained in the UK. Forest Ecology and Management 121(1): 123–136. Niemiec, W., Stachowicz, F., Szewczyk, M., Trzepieciński, T., 2011: Technological progress in production, logging and processing of the biomass. SSP – Journal of Civil Engineering 6(2): 85–92. Niemiec, W., Stachowicz, F., Trzepieciński, T., 2012a: Maszyny przeznaczone do zbioru wysokołodygowych roślin energetycznych na małych plantacjach. Technika Rolnicza Ogrodnicza Leśna 4: 10–12. Niemiec, W., Stachowicz, F., Szewczyk, M., Trzepieciński, T., 2012b: Production technology and management of energetic plants with lignified shoots. ECONTECHMOD: An Inter-national Quarterly Journal on Economics of Technology and Modelling Processes 1: 31–34. Sambra, A., Sřrensen, C.A.G., Kristensen, E.F., 2008: Optimized harvest and logistics for biomass supply chain. Proceedings of European Biomass Conference and Exhibition, 2–6 June, Valencia, Spain. http://pure.agrsci.dk:8080/fbspretrieve/1350892/OC4.1.pdf. [Cited 10 Aug. 2014].
Scholz, V., Lorbacher, F.R., Spikermann, H., 2009b: Technologien der Ernte und Rodung von Kurzumtriebsplantagen. In: Reeg T., Bemmann A., Konold W., Murach D., Spiecker H. (eds.). Anbau und Nutzung von Bäumen auf landwirtschaftlichen Flächen. WILEYVCH Verlag GmbH & Co., Weinheim, 99–112. Schweier, J., Becker, G., 2012a: Evaluation of the New Holland FR 9060 forage harvester equipped with the coppice header 130 FB to harvest short rotation coppice: Results from field studies in Germany. Manuscript. Schweier, J., Becker, G., 2012c: New Holland forage harvester’s productivity in short rotation coppice: Evaluation of field studies from a German perspective. International Journal of Forest Engineering 23(2): 82–88. Spinelli, R., Nati, C., Magagnotti, N., 2008: Harvesting shortrotation poplar plantations for biomass production. Croatian Journal of Forest Engineering 29(2): 129–139. Spinelli, R., Nati, C., Magagnotti, N., 2009: Using modified foragers to harvest short rotation poplar plantations. Biomass and Bioenergy 33(5): 817–821. Spychała, W., Łowiński, Ł., Pawłowski, T., Zbytek, Z., 2010: Simulation research of impact of cutting resistance on machine frame. Journal of Research and Applications in Agricultural Engineering 55(2): 134–137. Szewczyk, M., Trzepieciński, T., 2012: Application of biomass-powered stirlingengines in cogenerative systems. An International Quarterly Journal on Economics of Technology and Modelling Processes 1(2): 53–56.
Savoie, P., Current, D., Robert, F.S., Hébert, P.L., 2012: Harvest of natural shrubs with a biobaler in various environments in Québec, Ontario and Minnesota. Applied Engineering in Agriculture 28(6): 795–801.
Trzepieciński, T., Niemiec, W., Stachowicz, F., 2013: Wybrane problemy projektowania kosiarek do ścinania drzewiastych roślin i pielęgnacji terenów zieleni. Technika Rolnicza Ogrod nicza Leśna 1: 13–15.
Scholz, V., Eckel, H., Hartmann, S., 2009a: Processes and costs of SRC cropping on agricultural land. ���������������� In Die Landwirtschaftals Energieerzeuger, KTBL-Schrift 476: 67–80.
Verwijst, T., Telenius, B., 1999: Biomass estimation procedures in short rotation forestry. Forest Ecology and Management 121(1): 137–146.
198
Croat. j. for. eng. 37(2016)1
Development of Harvesting Machines for Willow Small-Sizes Plantations ... (185–199)
T. Trzepieciński et al.
Authors’ addresses: Tomasz Trzepieciński, DSc., PhD.* e-mail: tomtrz@prz.edu.pl Prof. Feliks Stachowicz, DSc., PhD. e-mail: stafel@prz.edu.pl Rzeszów University of Technology Faculty of Mechanical Engineering and Aeronautics 8 Powstańców Warszawy Ave. 35-959 Rzeszów POLAND Witold Niemiec, DSc., PhD. e-mail: wniemiec@prz.edu.pl Rzeszów University of Technology Faculty of Civil and Environmental Engineering and Architecture 6 Powstańców Warszawy Ave. 35-959 Rzeszów POLAND Leszek Kępa, MSc. e-mail: info@sipma.pl SIPMA S.A. 26 Budowlana St. 20-469 Lublin POLAND
Received: October 03, 2014 Accepted: April 09, 2015 Croat. j. for. eng. 37(2016)1
Marek Dziurka, MSc. e-mail: mdziurka@rndcentreinventor.pl R&D Centre INVENTOR sp. z o.o. 4 Ciepłownicza St. 20-469 Lublin POLAND * Corresponding author
199
Subject review
Carbon Footprint of Forest Operations under Different Management Regimes Giulio Cosola, Stefano Grigolato, Pierre Ackerman, Sergio Monterotti, Raffaele Cavalli Abstract Different forest management regimes have different carbon footprints due to alternative operational strategies and options. Data concerning CO2 emissions (kg m–3) in felling, extraction, comminution and transport operations, performed under two different forest management regime (close-to-nature and plantation), were collected through a systematic literature review involving 162 scientific papers and compiled into a database. Results show that, within limits, forest operations in plantations produce lower emissions due to easier operational conditions, while transportation in both close-to-nature and plantation based forest operations reported the highest levels of emissions. Literature came from a variety of sources and often differed in context due to factors such as technology, work technique, operator skill and environmental conditions. These factors have been shown to highly affect the results obtained from the studies. Nevertheless, it has been possible to summarize most of the information gathered and to highlight the most representative driving factors in CO2 emissions throughout different forest management regimes. Keywords: CO2 emissions, forest management regimes, forest operation, carbon footprint, wood harvesting, wood extraction, wood transport
1. Introduction Forest ecosystems store more than 80% of all terrestrial aboveground C and more than 70% of all soil organic C (Jandl et al. 2007). As a consequence, the ever increasing concentrations of carbon dioxide (CO2) in the atmosphere have highlighted the importance of forests as a mitigation agent (IPCC 1993, Routa et al. 2012). Globally, forest vegetation and soils contain about 1146x1012 kg of C, 37% of which is found in low-latitude forests, 14% in mid-latitude forests and 49% at high-latitude forest (Dixon et al. 1994). The annual CO2 exchange between forests and the atmosphere via photosynthesis and respiration accounts for about 50×1012 kg of C per year, approximately 7 times that caused by anthropogenic sources (Jandl et al. 2007). Accordingly, forests are considered important in limiting the increase in atmospheric carbon dioxide since trees sequestrate substantial amounts of carbon from the atmosphere to be stored in both above and below ground biomass through the process of photosynthesis. Croat. j. for. eng. 37(2016)1
However, during forest harvesting operations, carbon is released to varying degrees depending on the product being harvested and on emissions from the machines used in the process (Liski et al. 2001). Principal sources of CO2 in forest operations result from direct core emissions related to fuel used by machines (Knechtle 1997, Schwaiger and Zimmer 2001, Klvač et al. 2003, Gonzalez-Garcia et al. 2009a, Gonzalez-Garcia et al. 2009b, Valente et al. 2011, González-García et al. 2012, Klvač et al. 2012, Picchio et al. 2012, Vusić et al. 2013). However, these emissions are rather low when considered in a global context (Berg and Lindholm 2005). For example, in countries with high forest coverage, such as Sweden, these specific emissions amount to about 1% of total national emissions (Athanassiadis 2000). CO2 emissions in forest harvesting operations are also influenced by stand and terrain conditions, wood species, management methods, operator performance and machinery limitation or design (Van Belle 2006, González-García et al. 2009a, Gonzalez-Garcia et al. 2009b, Kärhä 2011, Vusić et al. 2013, Alam et al. 2014).
201
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
Therefore, with increasing mechanization of forest operations, it can be expected that emissions could increase (Berg 1997, Athanassiadis 2000) even though forestry activities do not tend to emit vast amounts of greenhouse gases (GHG). The necessity for a low carbon emission system still exists, bearing in mind that GHG emissions in the European Union must be reduced by 40% by 2030 (with 1990 as base-line). This proposed reduction must, however, be cost effective and sustainable in the long run (EU 2014). The aim of the present work is to investigate, by using the data retrieved by a systematic review on the scientific literature published over a period of 20 years, the different carbon footprints due to various operational strategies and options occurring under specific forest management regimes and highlight the general principles that can be recognized in order to enforce more environmentally friendly harvesting practices.
2. Materials and Methods 2.1 Systematic review To synthesize and discuss issues and findings of CO2 emission from forest operations, a systematic review was conducted to retrieve relevant scientific publications over the last 20 years (1994–2014). A systematic review consists in a process of identifying and evaluating multiple studies on a topic using a clearly defined methodology (Wolf et al. 2015). In our case, literature search was located by Scopus and Google Scholarsearch engine using English search terms and their various combinations applying Boolean operators (AND OR), wild-cards (for any group of characters (*) or for a single character (?)) (the search strings were combined as follow: 1. AND 2. AND 3.). 1. Search string »forest*« OR »stand« OR »*wood*« OR »*timber« OR for forest and »spruce« OR »beech« OR »pine« OR »poplar« OR forest products: »eucalyptus« OR »plantation« OR »close-to-nature« 2. Search string »operation« OR »logging« OR »harvest*« OR for forest »forward*« OR »extraction« OR »skid*« OR operations: »*haulage« OR »transport*« OR »machin*« OR »*mechaniz*« 3. Search string »emission?« OR »CO2« OR »GHG« OR »greenhouse*« for emissions: OR »fuel consumption« OR »productivity« OR »rate« OR »time« OR »LCA« OR »life cycle«
The search was carried out in the context of »���� ����� Forestry, Agriculture and Environmental Science��������� «�������� or ���� »��� En-
202
gineering« and only articles reporting results about CO2 emissions in forest operations were analyzed. In addition, information not specific to CO2, but that could be worked back through to calculated emissions, was used. To assist the search, an »Evidence-Based Approach to Scoping Reviews�������������������������������� «������������������������������� was followed based on the following methodology (Landa et al. 2011): 1) Define and refine research search terms, 2) Identify databases and search engines and query using the search terms, 3) Create and apply the inclusion and exclusion criteria filters, 4) Verify that the sub-selections are representative.
2.2 Database All the identified literature was re-organized into a database built in Microsoft Access®. The framework of the database (Fig. 1) is taken up in the following tables: Þ Literature, Þ Emissions, Þ Survey, Þ Machine technical data. In the Literature table, all the principal features of analyzed the papers were reported, such as the Title, Year, Author/s and Country. A link to the relative Portable Document Format (PDF) file was also provided. The Emissions table, in which all the most relevant data and values were collected, was connected with the Literature table through a »one-to-many« relation between the ID field, where a unique ID identified each paper. Another ������������������������������� »������������������������������ one-to-many������������������� «������������������ relation connected the Survey table to the Emission table through the field »ID_S« (survey). In the former, specific data of the field survey areas were reported when available. The database also included specific tables containing technical data on the relative categories of machines (e.g., harvester, forwarder, slash bundler, skidder, tractor, cable yarder, excavator, chipper and truck) according to the way in which information was provided by each study. They were then simply connected to the Emission table through the field »ID_M« (machine).
2.3 Boundaries and Functional Unit The boundary of the study was fixed to activities performed under two different forest management approaches [close-to-nature silviculture (CTN) and plantation (P)] and related to the harvesting site and the transport of forest products. Harvesting operations were considered as carried out under semi-mechaCroat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
G. Cosola et al.
Fig. 1 Framework of the CO2 Database showing the relationships between different tables nized (SM) and fully mechanized (FM) levels. Hence, only data on emissions from the functional phases of felling, extraction (primary transport) and transportation (secondary transport) were collected. Other work stages typical of forestry operations in plantation, such as site preparation and tending, were not considered. The functional unit (FU) was expressed as kilograms of CO2 directly emitted for every cubic meter of fresh (moisture content of 50%) wood processed and then expressed in kg CO2 m–3. Even if, at times, it was Croat. j. for. eng. 37(2016)1
possible to distinguish between over bark (o.b.) and under bark (u.b.) diameter, this distinction was eventually not used. »Directly emitted« means that only core direct emitted CO2 (EPA 2008) was considered. The choice of considering only CO2 gas lies in the fact that it represents the main air pollutant factor in terms of Global Warming Potential (GWP) (González-García et al. 2009a, Gonzalez-Garcia et al. 2009b), whereas all other gases represent only a minimal percentage, with a total amount less than 1% (Gillenwater 2005).
203
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
Table 1 Simple emission factors reported in some studies Fuel
CO2
Emission unit
Diesel
2.65
kg l–1
Diesel Diesel Diesel
2.672 2.6569 2.6
References
Used for
(Holzleitner et al. 2011)
Trucks
–1
(Devlin et al. 2013)
Trucks
–1
(Zorić et al. 2014)
Trucks
–1
(Korpilahti 1998)
Forwarders, chippers and trucks (Diesel density of 840 g/l is assumed)
–1
(Van Belle 2006)
Chippers [Coefficient derived from (CWAPE 2002)]
kg l kg l kg l
Diesel
3
kg l
Diesel
3.188
kg kg–1
(Karjalainen and Asikainen 1996)
Heavy-duty diesel vehicles and farm equipment [Coefficient derived from (IPCC 1993)]
Diesel
73.3
g MJ–1
(Karjalainen and Asikainen 1996)
Heavy duty diesel vehicles and farm equipment [Coefficient derived from (IPCC 1993)]
Diesel
982
g km–1
(Karjalainen and Asikainen 1996)
Heavy duty diesel vehicles and farm equipment [Coefficient derived from (IPCC 1993) and assuming moderate control and fuel economy of 2.8 km l–1]
Diesel
74.06
g MJ–1
(Dias et al. 2007)
Different machine and equipment [Coefficient derived from (Perry and Green 1997, Normand and Treil 1985)]
Diesel
74.10
g MJ–1
(Valente et al. 2011)
Different machine and equipment [Coefficient derived from (IPCC 2006)]
78.15
–1
(Engel et al. 2012)
Different machine and equipment [Coefficient derived from (Öko-Institut 2008)]
–1
Diesel
g MJ
Diesel
78.20
g MJ
(Engel et al. 2012)
Different machine and equipment [Coefficient derived from (Öko-Institut 2008)]
Diesel*
260
g MJ–1
(Athanassiadis 2000)
Different machine and equipment [Coefficient derived from engine test and must be assuming a 40% thermal efficiency of the engines]
Gasoline
78.31
g MJ–1
(Engel et al. 2012)
Chainsaw [Coefficient derived from (Öko-Institut 2008)]
Gasoline
14.36
g MJ–1
(Engel et al. 2012)
Horse-drawn trailer equipped with an engine powered crane [Coefficient derived from (Öko-Institut 2008)]
Gasoline
69.30
g MJ–1
(Dias et al. 2007)
Different machine and equipment [Coefficient derived from (Öko-Institut 2008)]
* Environmental class 3; Environmental class 1; Rapeseed Methyl Ester
In fact, CO2 equivalent (CO2eq) considers the sum of the three principal gases such as CO2, N2O and CH4 weighted for their GPW for a time horizon of 100 years with reference to CO2 GWP (IPCC 2006). Methane and N2O emissions depend not only upon fuel characteristics, but also on the engine combustion technology type, conditions within the engine combustion chamber, usage of pollution control equipment, and local environmental conditions (Lloyd and Cackette 2011). When weighted by their Global Warming Potentials (GWPs), CO2 typically represents over 99% of the GHG emissions from the stationary combustion of fossil fuels (Gillenwater 2005). So, in those cases conversion to normal values of CO2 was achieved using a conversion factor of 0.99 (Gillenwater 2005). All retrieved papers were divided in three groups according to the origin of the emission values: Þ Emission: papers in which CO2 emission values are stated;
204
Þ Fuel consumption: papers in which CO2 emission values are not stated, but they can be extracted through direct or indirect measurement of fuel consumption; Life Cycle Assessment (LCA): papers in which emission CO2 and GHG emissions are provided in the measuring and assessing procedures of environmental performance of forest operations by isolating, when possible, base values of CO2 from the rest of the information provided.
2.4 Emissions Emissions analyses are mostly done in an indirect manner. Only one study reports the amount of exhaust emissions calculated by a portable emission measurement system (Lijewski et al. 2013). In all the other studies, more attention was paid to variables and coefficients used according to machine and physical fuel characteristics. Different analysis Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
G. Cosola et al.
Table 2 Calorific value of fuels Fuel type
MJ l–1
References
Used by
Diesel
36.14
(Alt1n et al. 2001)
Self-determined
Diesel
36.55
(McDonnell et al. 1999)
(Klvač and Skoupý 2009)
Diesel
37.00
(Bailey et al. 2003)
(Spinelli and Magagnotti 2013)
Diesel
38.60
Supposed
(Routa et al. 2011)
Diesel
35.87
(AGQM 2009)
(Engel et al. 2012)
Diesel
36.29
(IPCC 2006)
(Valente et al. 2011)
Diesel
38.65
(EPA 2008)
Self-determined
Diesel
36.83
(Normand and Treil 1985) and (IPCC 2007)
(Dias et al. 2007)
Diesel
35.30
(SEA 2008)
(Lindholm et al. 2010)
Swedish environmental class 3
36.00
(Furuholt 1995; Grägg 1999)
(Athanassiadis 2000)
Swedish environmental class 1
35.30
(Grägg 1999)
(Athanassiadis 2000)
Rapeseed methyl ester
33.10
(Grägg 1999)
(Athanassiadis 2000)
Blend of semi-refined rapeseed oil (25%) and Diesel fuel (75%)
35.67
(McDonnell et al. 1999)
Self-determined
Gasoline
34.48
(Perry and Green 1997) and (IPCC 2007)
(Dias et al. 2007)
Gasoline
34.63
(EPA 2008)
Self-determined
Gasoline
32.48
(AGQM 2009)
(Engel et al. 2012)
methods have been developed during the past decades by both researchers and national agency teams to address this problem. Important values have been collected, analyzed and, if necessary, used in this study in order to process other similar available data. The studies, whose results have been computed through the use of coefficients or equations from other studies, were marked in the database with the ID code of the respective study. In the simplest cases, a coefficient for liter or MJ of fuel consumed (Table 1) was applied to obtain a value for CO2 emitted. Energy content or calorific value (CV) of fuels also varied between studies. Data concerning the leading references of the values, the references and in which paper they were used are reported in Table 2. Some studies offered more complex equations, which were used if they could be fitted to the specific case (Table 3).
2.5 Fuel consumption Fuel consumption (FC) is defined as the amount of fuel in liters (l) consumed by a machine during one working hour (h), and its measurement unit is expressed as l h–1. In emission analysis, fuel consumption is indeed an important value when CO2 emissions are computed indirectly. Croat. j. for. eng. 37(2016)1
From 1994 to 2014, a total of 118 studies including direct or indirect fuel consumption values were identified. During the considered time period (1994–2004), some equations were proposed and used to compute fuel consumption indirectly for different forest harvesting systems (Table 4). In order to measure indirectly the fuel consumption, it is also significant to mention the generic formula proposed in the FAO Forestry Paper 99 (FAO 1992).
2.6 LCA LCA studies in forestry have a wider context than the ones dealing with machine emission and fuel consumption and report values of CO2 and other GHG emissions relative to energy consumed. From 1994 to 2014, a total of 27 studies were identified. In most cases, LCA calculations relate to machine construction, repairs and other materials involved. The more complex part is the interpretation of the results to isolate base values of CO2 from the rest of the information provided. Moreover, and despite efforts by ISO 14400, the major drawback of LCA studies is the lack of uniformity in variables and scope (Berg 1995, Schwaiger and Zimmer 2001, Heinimann 2012).
205
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
Table 3 Paper whose CO2 emission equations have been used in the database calculations Equations and References Dias et al. (2007)
SEij = EWTi ⋅ Cij ⋅ VWj ⋅ CEFj ⋅ NCVj ⋅ FCOj ⋅
44 −3 ⋅ 10 12
Where: SEij is CO2 specific emissions associated with operation i due to the consumption of fuel j, g CO2 ha–1 or g CO2 m–3ub EWTi is effective work time of operation i, h ha–1 or h m–3ub Cij is consumption of fuel j in operation i, l h–1 VWj is volumetric weight of fuel j, kg m–3 CEFj is carbon emission factor of fuel j, kg C GJ–1 NCVj is net calorific value of fuel j, MJ kg–1 FCOj is fraction of carbon oxidized of fuel j Density and carbon content approach (EPA 2008) n CO Eec = ∑ Fi ⋅ FDi ⋅ Ci ⋅ FOi ⋅ 2(m.w.) C(m.w.) i=1 Where: Eec is Emission from Energy (fuel) Consumption, g FU–1 Fi is volume of Fuel Type i combusted FDi is density of Fuel Type i, mass / volume Ci is Carbon Content Fraction of Fuel Type i, mass C / mass fuel FOi is Fraction Oxidized or Fuel Type i
Klvač and Skoupý (2009) Eec = FC ⋅ EF ⋅ CV ⋅ TE
Where: Eec is Emission from Energy (fuel) Consumption, g FU–1 FC is Fuel Consumption, l FU–1 EF is Emission Factor, g MJ–1 CV is Calorific Value, MJ l–1 TE is Thermal Efficiency, %
Table 4 Fuel consumption equations specifically developed for different forest harvesting systems Equations and References (Nordfjell et al. 2003)
(Nordfjell et al. 2003)
y = −0.110 + 0.00047 ⋅ x Note: Specific for forwarder VALMET 890 (130 kW), in Pine stand for loads of saw-logs already stacked
0.00638 ⋅d y = 0.288 + L
y = −0.026 + 0.001⋅ x Note: Specific for forwarder VALMET 890 (130 kW), in Pine stand for loads of saw-logs already stacked Where: y is fuel consumption, l m–3ub x is average extraction distance, m
Where: y is fuel consumption, l m–3ub L is load size, m3ub d is average extraction distance, m Note: Generic for forwarder in a Pine stand with easy condition terrain for final felling, with loads of both sawlog and pulpwood already stacked
Freitas (2004)
Klvač and Skoupý (2009)
y = 0.134 ⋅ x
y = 8.203 ⋅ e3.655⋅ x⋅10 Where:
Where: y is Fuel Consumption, l PMH0–1 x is Power of machine, kW Note: Generic, in clear-cutting system stand conditions (coppice stem selection and pre-commercial thinning are excluded)
However, LCA provides a systematic way to measure and assess the environmental properties of products and processes (Athanassiadis et al. 2002) along
206
−3
y is Fuel Consumption, l PMH0–1 x is Power of machine, kW Note: Generic, for harvesters and forwarders in clear felling
with the methodology on how to conduct these studies. Unfortunately, the technical parameters required for analysis, such as fuel consumption or CO2 emisCroat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
sions, are not usually measured. These parameters are rather collected from interviews, technical specifications, agency reports or scientific articles. Hence, in many cases their results are meaning less from an operative point of view, as they do not represent anything outside of pure coefficients used for estimations (even if they are properly chosen by Authors). Beyond these aspects, in many cases, parameters used in LCA studies came from findings already analyzed and recorded in the database, so their retrieval has been avoided to minimize redundancy in the results.
3. Results
G. Cosola et al.
Table 6 Comparison of CO2 emissions in felling operations, according to different management approaches (P – Plantation or CTN – Close-to-nature) and mechanization levels (FM – Fully mechanized or SM – Semi-mechanized) Observations
Average
Max.
Min.
kg CO2 m–3
N SM (CTN)
23
0.63
2.48
0.10
FM (CTN)
158
3.66
70.17
0.87
SM (P)
4
0.60
1.01
0.02
FM (P)
39
3.94
16.35
0.55
3.1 Forest operation CO2 emissions Values of CO2 emissions related to 523 different forestry operating conditions were thus recorded in the database and were analyzed at various levels of detail. A general analysis follows in the subsequent three sections referred to as the three phases of »harvesting« (felling and extraction), »chipping« and »transport«. 3.1.1 Felling and primary transport In the CTN management system, values of 6.69 kg CO2 m–3 and 3.94 kg CO2 m–3 are, respectively, found for FM and SM harvesting systems. In P management system, values of 5.80 kg CO2 m–3 and 3.52 kg CO2 m–3 were recorded for the same harvesting systems. FM harvesting system seems to show a higher impact (Table 5) than SM system in terms of CO2 emissions, both in P and CTN management approaches. This is mainly due to the use of chainsaws instead of self-propelling machines (Berg 1997), which can also be seen in Table 6. Lower average emissions in the P approach are likely due to easier and more productive working contexts.
When splitting harvesting into its two components, felling (Table 6) and extraction (Table 7), the latter presents higher average values of emissions at the semimechanized level. This is in accordance with its generally less productive context (selective cutting, steep terrain, vulnerable sites). Reasons behind CO2 emissions patterns can be found among fuel consumption and productivity factors, such as machine type, logistic organization, stand characteristics, type of treatment, site conditions and of course, operator skill and attitude (Nordfjell et al. 2003, Kärhä et al. 2004, González-García et al. 2009a, Gonzalez-Garcia et al. 2009b, Alam et al. 2014).
Table 7 Comparison of CO2 emissions in extraction operations, according to different management approaches (P – Plantation or CTN – Close-to-nature) and mechanization levels (FM – Fully mechanized or SM – Semi-mechanized) Observations
Average
Observations
Average
Max.
Min.
Min.
kg CO2 m–3
N
Table 5 Comparison of CO2 emissions in harvesting operations (felling and extraction), according to different management approaches (P – Plantation or CTN – Close-to-nature) and mechanization levels (FM – Fully mechanized or SM – Semi-mechanized)
Max.
SM (CTN)
36
4.26
11.06
0.24
FM (CTN)
67
3.04
6.77
0.97
SM (P)
6
2.92
5.9
1.14
FM (P)
11
2.25
5.18
0.42
kg CO2 m–3
N SM (CTN)
9
0.62
2.48
0.10
FM (CTN)
16
6.64
41.83
1.17
SM (P)
3
0.85
1.01
0.69
FM (P)
8
4.23
10.02
1.75
Croat. j. for. eng. 37(2016)1
Focusing on SM extraction, differences in emissions do not seem to depend on the type of extraction system used, i.e. Ground Based System (GBS) or Cable Based System (CBS). Table 8 reports that there are minimal differences between these systems (4.01 kg CO2 m–3 for GBS and 4.42 kg CO2 m–3 for CBS).
207
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
Table 8 Comparison of CO2 emissions in extraction operations, according to different extraction systems (GBS – Ground based system and CBS – Cable based system) in CTN (Close-to-nature) management approach Observations
Average
Max.
Min.
kg CO2 m–3
N GBS (CTN)
14
4.01
9.33
0.24
CBS (CTN)
22
4.42
11.06
1.15
Table 10 Comparison of CO2 emissions in harvesting operations (felling and extraction), according to different management approaches (P – Plantation or CTN – Close-to-nature), mechanization levels (FM – Fully mechanized or SM – Semi-mechanized) and silvicultural treatments, such as CC (Clear cutting), SHW (Shelterwood cutting) and SC (Selective cutting). SC and SHW cuttings belong only to the CTN management approach, because no data considering the P management approach were available Observations Average
Observations
Average
Max.
Min. –3
kg CO2 m
N SM (CTN)
14
4.01
9.33
0.24
FM (CTN)
66
3.04
6.77
0.97
SM (P)
6
2.92
5.90
1.14
FM (P)
11
2.25
5.18
0.42
Greater differences can be found within the GBS itself than between the various levels of mechanization or management approaches (Table 9). The organization at landings depends on machines available and site conditions. Table 10 presents emissions for harvesting according to three main silvicultural treatments of CC (clear cutting), SHW (ShelterWood Cutting) and SC (selective cutting). SHW and SC treatments show higher values of emissions in the FM context. This is in accordance with the higher percentage of wood harvested by thinning operations (where productivity of mechanized felling can be severely affected due to tree size and working conditions) when compared to CC treatment. SHW treatment implies a natural regeneration, which is why no data were available for the P management approach (Table 10). In the case of CC in P, preparatory thinning practices such as early thinning (ETH) or thinning (TH) made before a final cutting (FC) were also considered. Data about thinning considerably affected the results, as shown in Table 11. In both P and CTN management approaches, fully mechanized thinning operations (ETH and TH) emit
208
Min.
kg CO2 m–3
N
Table 9 Comparison of CO2 emissions in extraction operations with GBS (Ground based system), according to different management approaches (P – Plantation or CTN – Close-to-nature) and mechanization levels (FM – Fully mechanized or SM – Semi-mechanized)
Max.
CC
SM (CTN)
18
4.72
11.83
2.32
CC
FM (CTN)
138
4.68
11.76
1.94
CC
SM (P)
10
3.53
6.92
1.14
CC
FM (P)
49
5.73
21.53
0.97
SHW
SM (CTN)
40
3.67
11.81
1.25
SHW
FM (CTN)
94
9.44
75.28
2.82
SHW
SM (P)
N/A
N/A
N/A
N/A
SHW
FM (P)
N/A
N/A
N/A
N/A
SC
SM (CTN)
9
2.27
6.72
0.10
SC
FM (CTN)
6
7.45
9.20
5.86
SC
SM (P)
N/A
N/A
N/A
N/A
SC
FM (P)
N/A
N/A
N/A
N/A
more CO2 per cubic meter than final cutting (Berg 1997). The same pattern occurs for semi-mechanized operations in P management approaches, whereas an opposite effect is seen in CTN. In this case, the results are mostly influenced by the emissions during extraction operations that, as mentioned above, are associated with a more difficult working environment. 3.1.2 Chipping In many circumstances, timber harvesting includes the chipping of the residues or of the whole trees. Chipping can be done directly in the stand using small machines. With easy terrain accessibility, it can be done at the roadside using more powerful machinery or at a terminal with either mobile and/or stationary machinery (Liška et al. 2010). Average values found for CTN and P approach are reported in Table 12. Even in chipping, it appears that the P context has less of an impact with 5.36 kg CO2 m–3 emissions compared to 9.70 kg CO2 m–3 in the CTN management approach. Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
G. Cosola et al.
ETH
SM (CTN)
2
2.69
2.84
2.55
ETH
FM (CTN)
14
6.95
9.10
5.20
ETH
SM (P)
1
5.90
–
–
ETH
FM (P)
13
9.15
19.88
6.12
TH
SM (CTN)
2
2.69
2.84
2.55
TH
FM (CTN)
55
12.07
76.94
3.14
TH
SM (P)
22
3.13
9.33
0.00
TH
FM (P)
N/A
N/A
N/A
N/A
However chipping is usually a highly energy demanding operation (Pan et al. 2008, Valente et al. 2011), and fuel consumption depends on the size and type of the material to be chipped as well as on the wood density (Van Belle 2006, Röser et al. 2012, Spinelli et al. 2013, Spinelli and Magagnotti 2013). Regarding different types of chippers, it is widely recognized that disc chippers produce more uniform woodchips than drum chippers, especially if fed with good quality raw material. In contrast, flexible small branches may pass through the disc slots uncomminuted, resulting in low chip quality (Spinelli and Hartsough 2001b). Also, dealing with smaller chippers, the disc chipper has higher energy efficiency, using less fuel per unit of product. This may be due to its simpler design, which integrates comminuting and discharge systems into one synergic device. In contrast, the drum chipper is more productive, since it cuts with the same energy along the length of its knives. They, however, produce finer particles (Spinelli et al. 2013).
FC
SM (CTN)
23
4.81
11.91
2.32
3.1.3 Secondary transport
FC
FM (CTN)
156
4.87
11.76
1.94
FC
SM (P)
7
2.73
4.23
1.16
FC
FM (P)
20
5.20
13.44
0.97
Table 11 Comparison of CO2 emissions in harvesting operations (felling and extraction), according to different management approaches (P – Plantation or CTN – Close-to-nature), mechanization levels (FM – Fully mechanized or SM – Semi-mechanized) and silvicultural treatments such as ETH (Early thinning), TH (Thinning) and FC (Final cutting) Observations Average
Max.
Min.
kg CO2 m–3
N
Table 12 Comparison of CO2 emissions in chipping operations, according to different management approaches (P – Plantation, CTN – Close-to-nature) Observations
Average
Max.
Min. –3
kg CO2 m
N CTN
6
9.70
14.20
5.24
P
13
5.36
11.94
1.66
Table 13 Comparison of CO2 emissions in secondary haulage, according to different transported material (woodchips or timber). Values derived from unitary values of kg CO2 m3 km–1 referred to a transportation distance of 100 km (50 km load, 50 km unload) Observations
Average
Max.
Min.
kg CO2 m–3
N Woodchips
30
11.50
47.62
3.17
Timber
8
7.04
17.40
3.03
Croat. j. for. eng. 37(2016)1
As mentioned above, transport generally represents the highest degree of emissions both in timber and energy wood chains (Karjalainen and Asikainen 1996, Schwaiger and Schlamadinger 1998, Schwaiger and Zimmer 2001, Berg and Karjalainen 2003, Berg and Lindholm 2005, Pan et al. 2008, Picchio et al. 2009, England et al. 2013). The only exception was found in Spain (Dias et al. 2007, Gonzalez-Garcia et al. 2009a, Gonzalez-Garcia et al. 2009b), where harvesting emissions are higher than those of the secondary transport. The main factor affecting fuel consumption in transport is the distance travelled (Schwaiger and Zimmer 2001, Holzleitner et al. 2011, Devlin et al. 2013). Besides distance, the amount of uphill road travel and road condition as well as road design can influence levels of emissions (Pan et al. 2008, Holzleitner et al. 2011). In fact, higher fuel consumption for driving empty was observed because empty trucks usually run uphill. Moreover, lower values for fuel consumption were related to a reduction of travelling on forest roads (Holzleitner et al. 2011). Chip transportation causes higher fuel consumption compared to round wood because of its lower bulk density (Whittaker et al. 2011). Table 13 reports the values of CO2 emission in timber and woodchips transportation. Average values of 11.50 kg CO2 m–3 and 7.04 kg CO2 m–3 were calculated, respectively, for woodchips and timber transport over a distance of 100 km (50 km load + 50 km unload).
209
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
4. Discussion Regarding exhaust emissions, it is accepted that substituting conventional fuels with biofuels can reduce gas pollution on a large scale. For example, by using Rapeseed Methyl Ester (RME), as much as 6.8 kg CO2 m–3 of CO2 emissions are reduced compared to using Diesel fuel (Gonzalez-Garcia et al. 2009a, Gonzalez-Garcia et al. 2009b, Klvač and Skoupý 2009). In addition, CO2 emissions per cubic meter depend on the level of machine maintenance. In fact, the maintenance shortage affects negatively the productivity by increasing the time lost on repairs (Senturk et al. 2007, Gerasimov et al. 2012, Röser et al. 2012, Spinelli and Magagnotti 2013). In fact, when maintenance is neglected, a lower machine performance and a higher number of delays due to repair time can be expected. As a consequence, productivity decreases and fuel consumption increases. Senturk et al. (2007) suggests that an adequate number of spare parts should be maintained in order to prevent any loss of time in case of urgent maintenance or repair works. Moreover, operator’s training, expertise and attitude play a fundamental role in reducing fuel consumption and thus emission as reported by Nordfjell et al. (2003), Kärhä et al. (2004), Kärhä and Vartiamäki (2006), Mederski (2006) and Alam et al. (2014). Finally, a rational harvest planning is essential in maintaining high productivity levels, meaning a higher efficiency in terms of fuel consumption and emissions. More specific features can be considered regarding harvesting, especially relating to the felling and extraction phases. In either CTN or P approaches with a FM or SM, harvesting operations can be achieved by FT (Full tree) or CTL (Cut-to-Length) work systems. CTL is common where trees are motor-manually or mechanically felled, delimbed as well as crosscut at the felling site, and then extracted. Contrarily, in FT systems, trees are felled and extracted to the landing area where they are delimbed and crosscut. In order to understand the advantages of each system, the strength sand weaknesses of the main unique machines in various working conditions are discussed.
4.1 Felling 4.1.1 Chain saw Knechtle (1997) and Berg (1997) have observed that motor-manual felling gives rise to lower emissions per unit of wood than mechanized felling. For Swedish forestry, the magnitude of the difference between felling methods is so great that even the deployment of resources for transport personnel between home and
210
work sites or between work sites is not sufficient to balance this difference (Berg 1997). Nevertheless, harvesters are fourfold more efficient than a chainsaw, producing less exhaust emissions per kW. This results in a better ecological performance (Lijewski et al. 2013). On the other hand, it has also been recognized that SM harvesting systems produce higher emissions in the extraction phase, since forwarding productivity is influenced by the number and the size of the loads (Laitila et al. 2007, Laina et al. 2013). Beside these aspects, manual felling remains a lower cost solution (Laina et al. 2013), but ergonomically marginal (Laina et al. 2013, Lijewski et al. 2013). 4.1.2 Harvester In many studies carried out on different site conditions, tree volume is the most important factor affecting harvester productivity (hence consumption and emissions) (Sirén and Aaltio 2003, Kärhä et al. 2004, Jiroušek et al. 2007, Laina et al. 2013). CO2 emissions end up as a tradeoff between the power of machines and the size of trees to be harvested. For example, larger harvesters use more energy, but when processing large trees the energy used is lower than with smaller machines processing small trees (Berg and Lindholm 2005, Klvač and Skoupý 2009). It would appear that smaller harvesters (up to 80 kW), including a tractor with a processor, can operate with the same productivity level as medium-sized harvesters (80–120 kW) in the thinning process. Consequently, they can be run at a fuel consumption and cutting cost lower than those of medium-sized harvesters (Kärhä et al. 2004). With regard to the work method, in P management approach, the use of a cutting-area between two striproads was the most efficient working method in thinning using harvesters with short booms (<8 m), even if more damages might occur. Correspondingly, the strip-road method (without cutting-strips) was most efficient when working with harvesters with long booms, although the distribution of the remaining trees is not so even (Kärhä et al. 2004). In thinning operations in stands of Scots pine managed by CTN management approach, the midfield operation technique (i.e. harvester is combined with the chain saw in areas that cannot be reached by the harvester between the two skid-roads) was always more productive and much less fuel costly than skid-road one (Mederski 2006). 4.1.3 Harwarder In an effort to make thinning operations affordable for mechanized processes, the harwarder has appeared as a possible solution. In this case, tree size, Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
removals per hectare and number of timber assortments are the factors affecting productivity when the forwarding distance is limited to 250 m (Wester and Eliasson 2003). Harwarders are most competitive when two timber assortments are applied in small stands (stem volume <0.1 m3) with short forwarding distance (<250 m) (Sirén and Aaltio 2003). Another possible solution is to fit a feller-buncher head to the forwarder, but studies have found that tree volumes should not be less than 0.05 m3 (Gingras 2004, Rottensteiner et al. 2008).
4.2 Primary transport 4.2.1 Forwarder Haulage distance and payload are the ariables affecting forwarder productivity and fuel consumption (Nordfjell et al. 2003, Tiernan et al. 2004, Jiroušek et al. 2007, Laitila et al. 2007). Considering payload, there are no significant differences in fuel consumption when driving loaded or unloaded. However, consumption per unit volume of wood is greater in transporting pulpwood than sawlogs. This is due to longer loading times and smaller volumes of pulpwood, which decreases productivity and increases consumption (Nordfjell et al. 2003). Moreover, when extracting both pulpwood and sawlogs, a two-pass forwarding technique is more productive than a mixed-load forwarding technique (Kellogg and Bettinger 1994). With regards to site conditions, slope appears to be dramatically significant. Uphill extraction can reduce productivity by 1–5 m3 PMH0–1 with an obvious increase in fuel consumption. In easy site conditions (gentle slope <10% and even roughness), forwarder productivity is significantly higher in clear-felling sites when compared to thinning sites (2.0 m3 PMH0–1). In addition, productivity of forwarders with a 10 m boom is up to 9 m3 PMH0–1 greater than that of forwarders with a 7 m boom (Tiernan et al. 2004). The main advantages of forwarders over skidders include: less soil disturbance and damages; enhanced work safety and ergonomics; longer extraction distances (hence, reduced road density requirements); less labor; and finally, reduced landing area requirements for handling short wood (Kellogg and Bettinger 1994, Tiernan et al. 2004). The same advantages can be seen when comparing forwarders to tractors with trailers (Spinelli et al. 2012b) notwithstanding larger volumes extracted per cycle. Nevertheless, forwarding with tractors with trailers can offer a technical benefit in terms of higher travel speeds in many small scale forestry harvesting operations (Magagnotti et al. 2013b). Croat. j. for. eng. 37(2016)1
G. Cosola et al.
4.2.2 Skidder Grapple and cable skidders (in steeper terrain) are usually used in clear cutting operations. In fact, their relatively large size makes them more effective in working in the open. Field-measured productivity results are significantly different between cable (43.9±7.5 m3 PMH0–1) and grapple skidders (123.9±3.9 m3 PMH0–1), but no difference has been recorded between unloaded and loaded travel speeds (Ackerman et al. 2014). In similar conditions (Poplar or Eucalyptus plantations), substituting a skidder with an articulated frontend loader could offer a better solution for flail chipping. Although a loader working in conjunction with a skidder takes more time for essentially every extraction element, during extraction and chipping at landing in fast-growing tree species plantations, it can perform 60% more than the skidder because of its larger payload. The capacities of both the skidder and the loader exceeded the productivities of the flail-chippers, so they had excess time. The loader had enough time to handle the landing work. The skidder grapple and decking blade, however, were less suited to moving residues at the landing, and the skidder did not have much excess time, so a second machine was required for landing duties (Spinelli and Hartsough 2001a). Instead, in the context of small-scale forestry, other more versatile machines are usually used, such as a tractor with a winch. New mini skidders, when compared with common agricultural tractors or forestryfitted tractors, are more environmentally friendly in terms of energy inputs and GHG emissions during wood extraction operations, both in thinning and final cutting (Vusić et al. 2013). This is the case even when compared to crawler tractors in mountainous conditions (Spinelli et al. 2012a). The analysis of working time indicated that equipping a mini-skidder with a double drum winch is important in high forest thinning, due to the smaller size of trees, while for regeneration cuts in high forests, the double drum winch becomes almost redundant. In terms of productivity, energy inputs and emissions, skidding is negatively influenced by slope (uphill over 15%) (Vusić et al. 2013). Finally, in forests with protective rather than productive purpose, All Terrain Vehicle (ATV) are an optimal solution for full-tree bunching and skidding operations (<200 m) in first thinning and coppice harvesting since they cause negligible impacts on the ground (i.e. no effect on the roots), and they are able to work on slopes up to 50% (Savelli et al. 2010). 4.2.3 Cable Crane Aerial cable crane extraction systems are applied in mountainous regions, contributing to better quality
211
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
of the logs extracted, lower damages to the harvesting site and a reduced necessity for forest roads (Ozturk and Demir 2007, Senturk et al. 2007, Valente et al. 2011). Aerial systems can reduce the energy required per functional unit considerably when compared with fully mechanized systems (Klvač et al. 2012) or with other traditional systems, such as the short-woodsystem (SWS) in alpine regions (Dias et al. 2007). The productivity of cable cranes is effectively influenced by lateral outhaul, inhaul and in lateral phases (Senturk et al. 2007). In particular, productivity decreases with an increasing number of bundles and with higher extraction distances (Zimbalatti and Proto 2009). 4.2.4 Animal Animal (horses, mules or oxen) log extraction systems, apart from being the most environmentally friendly solution, show the lowest emissions (Engel et al. 2012, Cerutti et al. 2014). In specific conditions, they can also be competitive in terms of productivity and costs (Magagnotti and Spinelli 2011, Cerutti et al. 2014). Draught horses represent an efficient log extraction tool in steep terrain and in low-intensity cuts, as generally offered by closed canopy forests. Horse skidding incurs lower unit costs than tractor skidding when the extraction distance is short or when skid trails are not available. The cost-efficiency of horse skidding increases significantly when two horses are paired per driver (Magagnotti and Spinelli 2011).
4.3 Integrated harvesting Integrated harvesting has been developed during last decades with the growing importance of forest residues for energy use in heating and combined heat and power (CHP) plants (Friso et al. 2011). Comparisons between conventional product and integrated fuel wood production harvesting in Canada suggests that production costs are highly variable depending on the harvesting system used and the ratio of conventional products to fuel wood (Puttock 1995). The same was seen in a poplar plantation in Italy where a more integrated pulp and chip strategy generally created higher revenues than the exclusive production of woodchips (Spinelli and Magagnotti 2011). However, the additional woodchips produced are generally not the result of potential (i.e. harvesting residues, wood from thinning, coppice stands and short rotation forests), but more a matter of economic feasibility. Harvesting conditions, roadside landing capacities, road transportation distances, operating volumes and storage capacities of heating and CHP plants, availability of production machinery, type of forest woodchips produced and, notably, the total supply
212
chain costs all influence the selection of the forest chip supply chain (Kärhä 2011). Most likely, chipping will move from roadside locations closer to the heating and power plants since the closer plant chipping is performed to their processing destination, the more costefficient is the process (Kärhä 2011). Increasing fuel demand will result in a larger supply area for the energy producer and lead to increasing transportation costs. The analysis of different chip production systems resulted in the identification of two major challenges: firstly, the design of the chipping and transport interface, and secondly, the need to reduce transportation costs. Through drying the material, compressing harvesting residues as well as increasing payloads an improved utilization of load volumes can be achieved. For example, drying wood in storage areas near the forest enhances the transportation productivity by 50%. Similarly, bundling harvesting residues pays off, especially for longer transportation distances (Stampfer and Kanzian 2006). Still, in harvesting agro-forestry plantations, the removal of stumps can be accomplished by two different approaches: grinding or extraction. Results of a study suggested that the use of a stump grinder instead of a stump extractor or backhoe excavator is particularly advisable in terms of productivity, costs and energy inputs (Lindholm et al. 2010). Particularly in the latter case, direct inputs are much higher than indirect inputs. Moreover, stumps extraction negatively contributes to the accumulation of carbon in the agro-forestry soils (Picchio et al. 2012). Also, slash bundlers are capable of increasing the productivity of both transport and chipping of forest residues as they collect branches in pressed and tight bundles. The amount of residue available on the unit surface, its average size and its distribution on the field are the three main parameters that most affect bundling productivity and fuel consumption (Cuchet et al. 2004). Productivity is considerable if the piles are stacked on both sides of the strip road (Kärhä and Vartiamäki 2006).
4.4 Secondary transport There are several ways of decreasing energy demands in wood secondary road transport, such as reducing transport distance, adjusting load factors, designing better route-planning systems, improving roads (curve geometry and surfaces), adopting more fuel-efficient driving techniques and using the best available transport carriers (Berg and Lindholm 2005, González-García et al. 2009a, Holzleitner et al. 2011, Pierobon et al. 2015). Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
5. Conclusion Despite the small number of papers related to the direct measurement of forest operation emissions, the study provides an effective approach for data and information collection in the field. In particular, it defines a first overview of the carbon emissions from forest operations that encompasses a variety of operative contexts from different countries. Even if the collected information are limited, they are useful in providing some concepts of the level of emissions that can be expected under certain conditions with certain machines in selected harvesting systems. It should be noted that the more detail is required, the greater the risk that specific variables from one study may bias the results. This is particularly true when dealing with harvesting operations, where many variables are involved. However, despite specific cases, there are some important general principles that can be recognized in order to enforce more environmentally friendly harvesting practices that would reduce CO2 emissions. Results comparing forestry harvesting, primary transport and chipping show a higher efficiency in P management approaches compared to CTN management approaches. Secondary transport CO2 emissions are more affected by the type of product transported, with a higher efficiency for logs compared to woodchips. Impacts from fully mechanized or semi-mechanized operations can have different patterns according to specific site conditions. CO2 emissions in forestry can be reduced at different operative levels, starting by: Þ Using the most environmentally friendly technologies (e.g. Tier 4 Diesel engines); Þ Substituting pure Diesel fuel with Diesel-biofuel blends as far as possible; Þ Maximizing machine productivity and reducing maintenance delays for the proper application of machine maintenance; Applying the best harvesting plan according to site conditions, forest management, machines features and drawbacks. With particular regard to this last point, and keeping in mind the high dependence of emissions for both primary and secondary transport on the type and slope of planned routes, it is necessary to underline the importance of using GIS tools to improve the environmental aspects of management logistics. This is particularly important in terms of road networks, which play a fundamental role in operation plan decision making (Cavalli and Grigolato 2010). Croat. j. for. eng. 37(2016)1
G. Cosola et al.
Even if the P management approach results in lower emissions for felling and chipping operations, the present study does not consider the phases of site preparation and stand tending, which indeed, have a high environmental impact. This may result in a higher global impact for P management approaches when compared to CTN management approaches in terms of kg CO2 m–3. Nevertheless, restricted access to natural forests is making plantation forestry increasingly important as a source of wood (Spinelli and Hartsough 2001a). While GHG emissions are evidently a pressing issue, the research conducted here was subject to many shortcomings, due to a high diversity in the coefficients and methodology used by researchers and the technological evolution of engines during the analyzed period. Besides LCA studies, for which drawbacks, weak points and proposals for improvement have already been highlighted by Heinimann et al. (2012), in all other cases a more standardized manner for collecting productivity, fuel consumption and GHG emissions data should be developed. This would ensure the comparability of results and the repeatability of experiments, both fundamental elements of the scientific method. Proposals for the development and harmonization of new operational research and assessment procedures for GHG emissions should be promoted following an approach similar to procedures adopted by other forestry frameworks/organizations, e.g. in sustainable forest biomass supply in the frame of Cost Action FP-0902 (Magagnotti et al. 2013a). With reference to the difficulties encountered here, some suggestions to enforce the effectiveness of the scientific forestry literature include: Þ Defining a homogeneous silviculture terminology for reference purposes; Þ Reporting values of coefficients necessary to switch from PMH0 (or SMH0) to PMH15 (or SMH15), or from SMH to PMH, or vice-versa; Þ Reporting values of coefficients to allow the reader to switch from values for the weight of wood to values for the volume of wood, taking into account its own bulk density and the moisture content.
Acknowledgements The study was carried out within the framework of the Programme EU FP7 PEOPLE 2011 IRSES, Theme Marie Curie Actions, Project »Climate-fit forests« (GA 295136). We thank Philipp Gloning for the useful comments and suggestions.
213
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
6. References Ackerman, P., Pulkki, R., Gleasure, E., 2014: Modelling of wander ratios, travel speeds and productivity of cable and grapple skidders in softwood sawtimber operations in South Africa. Southern Forests: a Journal of Forest Science 76(2): 101–110. AGQM, 2009: Information zum Gesetz zur Einfuerung einer Biokraftstoffquote durch Aenderung des Bundesimmissionsschutzgesetzes und zur Aenderung energie- und stromsteurrechtilche Vorschriften [Information for law introducing a biofuel quota by amending the Federal Pollution Control Act and amending energy and current tax regulations]. Available at: from http//:www.agqm-biodiesel.de. [Accessed April 5th 2009] Alam, M., Walsh, D., Strandgard, M., Brown, M., 2014: A log-by-log productivity analysis of two Valmet 475EX harvesters. International Journal of Forest Engineering 25(1): 14–22. Altın, R., Cetinkaya, S., Yücesu, H.S., 2001: The potential of using vegetable oil fuels as fuel for diesel engines. Energy Conversion and Management 42(5): 529–538. Athanassiadis, D., 2000: Energy consumption and exhaust emissions in mechanized timber harvesting operations in Sweden. Science of the Total Environment 255(1–3): 135–143. Athanassiadis, D., Lidestav, G., Nordfjell, T., 2002: Energy use and emissions due to the manufacture of a forwarder. Resources, Conservation and Recycling 34(3): 149–160. Bailey, A.P., Basford, W.D., Penlington, N., Park, J.R., Keatinge, J.D.H., Rehman, T., Tranter, R.B., Yates, C.M., 2003: A comparison of energy use in conventional and integrated arable farming systems in the UK. Agriculture, Ecosystems & Environment 97(1–3): 241–253. Berg, S., 1995: The environmental load of fossil fuels in Swedish forestry-an inventory for a LCA. Proceeding of the LifeCycle Analysis – A challenge for forestry and forestry industry. European Forest Institute, Hamburg, Germany, 3–5 May, 57–65. Berg, S., 1997: Some aspects of LCA in the analysis of forestry operations. Journal of Cleaner Production 5(3): 211–217. Berg, S., Karjalainen, T., 2003: Comparison of greenhouse gas emissions from forest operations in Finland and Sweden. Forestry 76(3): 271–284. Berg, S., Lindholm, E.L., 2005: Energy use and environmental impacts of forest operations in Sweden. Journal of Cleaner Production 13(1): 33–42. Cavalli, R., Grigolato, S., 2010: Influence of characteristics and extension of a forest road network on the supply cost of forest woodchips. Journal of Forest Research 15(3): 202–209. Cerutti, A.K., Calvo, A., Bruun, S., 2014: Comparison of the environmental performance of light mechanization and animal traction using a modular LCA approach. Journal of Cleaner Production 64: 396–403. Cuchet, E., Roux, P., Spinelli, R., 2004: Performance of a logging residue bundler in the temperate forests of France. ���� Biomass and Bioenergy 27(1): 31–39.
214
CWAPE, 2002: Coefficients d’émission de CO2 de différentes sources d’énergie primarie [CO2 emission coefficients for different primary energy resources]. Comité de Direction: 03 Décembre 2002. Commision Wallonne pour l’Energie, Namure (Belgrade), 9 p. Devlin, G., Klvac, R., McDonnell, K., 2013: Fuel efficiency and CO2 emissions of biomass based haulage in Ireland – A case study. Energy 54: 55–62. Dias, A. C., Arroja, L., Capela, I., 2007: Carbon dioxide emissions from forest operations in Portuguese eucalypt and maritime pine stands. Scandinavian Journal of Forest Research 22(5): 422–432. Dixon, R.K., Solomon, A.M., Brown, S., Houghton, R.A., Trexier, M.C., Wisniewski, J., 1994: Carbon Pools and Flux of Global Forest Ecosystems. Science 263(5144): 185–190. Engel, A.M., Wegener, J., Lange, M., 2012: Greenhouse gas emissions of two mechanised wood harvesting methods in comparison with the use of draft horses for logging. European Journal of Forest Research 131(4): 1139–1149. England, J.R., May, B., Raison, R.J., Paul, K.I., 2013: Cradleto-gate inventory of wood production from Australian softwood plantations and native hardwood forests: Carbon sequestration and greenhouse gas emissions. Forest Ecology and Management 302: 295–307. EPA, 2008: Direct Emissions from Mobile Combustion Sources. EPA – United States Environmental Protection Agency Climate Leaders, Washington DC, 31 p. EU, 2014: A policy framework for climate and energy in the period from 2020 to 2030. Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Region: COM/2014/0015. European Union, Brussels, 18 p. FAO, 1992: Cost Control in Forest Harvesting and Road Construction. FAO Forestry Paper: 99. Food and Agriculture Organization of the United Nations, Rome, 106 p. Freitas, C., 2004: Análise dos encargos horários com a utilização das máquinas pesadas na agricultura e floresta. IHERA – Gabinete de Gestão do Parque de Máquinas, Direcção de Serviços de Projectos e Obras, Instituto de Desenvolvimento Rural e Hidráulica, Lisboa, 13 p. Friso, D., Grigolato, S., Cavalli, R., 2011: Energetic and exergetic analysis of steam production for the extraction of coniferous essential oils. Biomass and Bioenergy 35(9): 4045–4056. Furuholt, E., 1995: Life cycle assessment of gasoline and diesel. Resources, Conservation and Recycling 14(3): 251–263. Gerasimov, Y., Seliverstov, A., Syunev, V., 2012: Industrial round-wood damage and operational efficiency losses associated with the maintenance of a single-grip harvester head model: a case study in Russia. Forests 3(4): 864–880. Gillenwater, M., 2005: Calculation Tool for Direct Emissions from Stationary Combustion. World Resources Institute, Washington DC, 94 p. Gingras, J.F., 2004: Early studies of multi-tree handling in Eastern Canada. International Journal of Forest Engineering 15(2): 18–22. Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
G. Cosola et al.
González-García, S., Bacenetti, J., Murphy, R.J., Fiala, M., 2012: Present and future environmental impact of poplar cultivation in the Po Valley (Italy) under different crop management systems. Journal of Cleaner Production 26: 56–66.
Kellogg, L.D., Bettinger, P., 1994: Thinning Productivity and Cost for a Mechanized Cut-to-Length System in the Northwest Pacific Coast Region of the USA. Journal of Forest Engineering 5(2): 43–54.
González-García, S., Berg, S., Feijoo, G., Moreira, M. T., 2009a: Environmental impacts of forest production and supply of pulpwood: Spanish and Swedish case studies. The International Journal of Life Cycle Assessment 14(4): 340–353.
Klvač, R., Fischer, R., Skoupý, A., 2012: Energy Use of and Emissions from the Operation Phase of a Medium Distance Cableway System. Croatian Journal of Forest Engineering 33(1): 79–88.
Gonzalez-Garcia, S., Berg, S., Moreira, M.T., Feijoo, G., 2009b: Evaluation of forest operations in Spanish eucalypt plantations under a life cycle assessment perspective. Scandinavian Journal of Forest Research 24(2): 160–172.
Klvač, R., Skoupý, A., 2009: Characteristic fuel consumption and exhaust emissions in fully mechanized logging operations. Journal of Forest Research 14(6): 328–334.
Grägg, K., 1999: Emissions from two truck engines and two off-road engines. MTC: 6806. AB Svensk Bilprovning Motortestcenter, Haninge, 14 p. Heinimann, H. R., 2012: Life cycle assessment (LCA) in forestry – state and perspectives. Croatian Journal of Forest Engineering 33(2): 357–372. Holzleitner, F., Kanzian, C., Stampfer, K., 2011: Analyzing time and fuel consumption in road transport of round wood with an onboard fleet manager. European Journal of Forest Research 130(2): 293––301. IPCC, 1993: Greenhouse Gas Inventory Reference Manual. First Draft. US Enviromental Protection Agency, Washington, 360 p. IPCC, 2006: 2006 IPCC Guidelines for National Greenhouse Gas Inventories. Volume 4: Agriculture, Forestry and other Land Use. National Greenhouse Gas Inventories Programme, Institute for Global Environmental Strategies (IGES), Hayama, Kanagawa. IPCC, 2007: Climate Change 2007: The Physical Science Basis. Cambridge (First ed.). Cambridge University Press, 996 p. Jandl, R., Lindner, M., Vesterdal, L., Bauwens, B., Baritz, R., Hagedorn, F., Johnson, D.W., Minkkinen, K., Byrne, K.A., 2007: How strongly can forest management influence soil carbon sequestration? Geoderma 137(3–4): 253–268. Jiroušek, R., Klvač, R., Skoupý, A., 2007: Productivity and costs of the mechanised cut-to-length wood harvesting system in clear-felling operations. Journal of Forest Science 53(10): 476–482. Kärhä, K., 2011: Industrial supply chains and production machinery of forest chips in Finland. Biomass and Bioenergy 35(8): 3404–3413. Kärhä, K., Rönkkö, E., Gumse, S.I., 2004: Productivity and Cutting Costs of Thinning Harvesters. International Journal of Forest Engineering 15(2): 43–56.
Klvač, R., Ward, S., Owende, P.M.O., Lyons, J., 2003: Energy audit of wood harvesting systems. Scandinavian Journal of Forest Research 18(2): 176–183. Knechtle, N., 1997: Materialprofile von Holzerntesystemen– Analyse ausgewählter Beispiele als Grundlage für ein forsttechnisches Ökoinventar [Material profiles from forest harvest systems. Analyses of chosen examples as a base for forest LCI]. Departement Wald- und Holzforschung, ETH, Zürich, 65 p. Korpilahti, A., 1998: Finnish forest energy systems and CO2 consequences. Biomass and Bioenergy 15(4–5): 293–297. Laina, R., Tolosana, E., Ambrosio, Y., 2013: Productivity and cost of biomass harvesting for energy production in coppice natural stands of Quercus pyrenaica Willd. in central Spain. Biomass and Bioenergy 56: 221–229. Laitila, J., Asikainen, A., Nuutinen, Y., 2007: Forwarding of Whole Trees After Manual and Mechanized Felling Bunching in Pre-Commercial Thinnings. International Journal of Forest Engineering 18(2): 29–39. Landa, A.H., Szabo, I., Le Brun, L., Owen, I., Fletcher, G., Hill, M., 2011: An evidence-based approach to scoping reviews. The Electronic Journal of Information Systems Evaluation 10(4): 173–175. Lloyd, A.C, Cackette, T.A., 2001: Diesel Engines: Environmental Impact and Control. Journal of the Air & Waste Management Association 51(6): 809–847. Lijewski, P., Merkisz, J., Fuć, P., 2013: Research of exhaust emissions from a harvester diesel engine with the use of portable emission measurement system. Croatian ������������������������ Journal of Forest Engineering 34(1): 113–122. Lindholm, E.L., Berg, S., Hansson, P.A., 2010: Energy efficiency and the environmental impact of harvesting stumps and logging residues. European Journal of Forest Research 129(6): 1223–1235.
Kärhä, K., Vartiamäki, T., 2006: Productivity and costs of slash bundling in Nordic conditions. Biomass and Bioenergy 30(12): 1043–1052.
Liška, S., Safarik, D., Klvač, R., Kupcak, V., Jirousek, R., 2010: General assessment of Jenz Hem 420 D chipper operating in the Czech Republic conditions. Proceeding of the FORMEC 2010 – Forest Engineering: Meeting the Needs of the Society and the Environment. CLEUP, Padova, July 11–14, 8 p.
Karjalainen, T., Asikainen, A., 1996: Greenhouse gas emissions from the use of primary energy in forest operations and long-distance transportation of timber in Finland. Forestry 69(3): 215–228.
Liski, J., Pussinen, A., Pingoud, K., Mäkipää, R., Karjalainen, T., 2001: Which rotation length is favourable to carbon sequestration? Canadian Journal of Forest Research 31(11): 2004–2013.
Croat. j. for. eng. 37(2016)1
215
G. Cosola et al.
Carbon Footprint of Forest Operations under Different Management Regimes (201–217)
Magagnotti, N., Kanzian, C., Schulmeyer, F., Spinelli, R., 2013a: A new guide for work studies in forestry. International Journal of Forest Engineering 24(3): 249–253. Magagnotti, N., Picchi, G., Spinelli, R., 2013b: A versatile machine system for salvaging small-scale forest windthrow. Biosystems Engineering 115(4): 381–388. Magagnotti, N., Spinelli, R., 2011: Financial and energy cost of low-impact wood extraction in environmentally sensitive areas. Ecological Engineering 37(4): 601–606. McDonnell, K., Ward, S., Leahy, J. J., McNulty, P., 1999: Properties of rapeseed oil for use as a diesel fuel extender. Journal of the American Oil Chemists’ Society 76(5): 539–543.
Routa, J., Kellomäki, S., Kilpeläinen, A., Peltola, H., Strandman, H., 2011: Effects of forest management on the carbon dioxide emissions of wood energy in integrated production of timber and energy biomass. GCB Bioenergy 3(6): 483–497. Routa, J., Kellomäki, S., Peltola, H., 2012: Impacts of Intensive Management and Landscape Structure on Timber and Energy Wood Production and net CO2 Emissions from Energy Wood Use of Norway Spruce. BioEnergy Research 5(1): 106–123. Savelli, S., Cavalli, R., Baldini, S., Picchio, R., 2010: Small scale mechanization of thinning in artificial coniferous plantation. Croatian Journal of Forest Engineering 31(1): 11–21.
Mederski, P. S., 2006: A comparison of harvesting productivity and costs in thinning operations with and without midfield. Forest Ecology and Management 224(3): 286–296.
Schwaiger, H., Schlamadinger, B., 1998: The potential of fuelwood to reduce greenhouse gas emissions in Europe. Biomass and Bioenergy 15(4): 369–377.
Nordfjell, T., Athanassiadis, D., Talbot, B., 2003: Fuel consumption in forwarders. ����������������������������������� International Journal of Forest Engineering 14(2): 11–20.
Schwaiger, H., Zimmer, B., 2001: A comparison of fuel consumption and greenhouse gas emissions from forest operations in Europe. Energy, Carbon and Other Material Flows in the Life Cycle Assessment of Forestry and Forest Products. Joensuu, European Forest Institute, 33–53.
Normand, X., Treil, A., 1985: L’Industrie Du Raffinage Du Petrole: Lecons sommaires. Institut Français du Pétrole, Paris, 363 p. Öko-Institut, 2008: Global Emission Model for Integrated Systems (GEMIS) – Version 4.5. Darmstadt. Ozturk, T., Demir, M., 2007: Transporting of spruce timbers by URUS MIII cable system from selective forests of Artvin region. Building and Environment 42(3): 1278–1282. Pan, F., Han, H.S., Johnson, L.R., Elliot, W.J., 2008: Net energy output from harvesting small-diameter trees using a mechanized system. Forest products journal 58(1–2): 25–30. Perry, R. H., Green, D. W., 1997: Perry’s Chemical Engineers’ Handbook. New York (7th ed.). McGraw-Hill., 2640 p. Picchio, R., Maesano, M., Savelli, S., Marchi, E., 2009: Productivity and energy balance in conversion of a Quercus cerris L. coppice stand into high forest in Central Italy. Croatian �������������� Journal of Forest Engineering 30(1): 15–26. Picchio, R., Verani, S., Sperandio, G., Spina, R., Marchi, E., 2012: Stump grinding on a poplar plantation: Working time, productivity, and economic and energetic inputs. Ecological Engineering 40: 117–120. Pierobon, F., Zanetti, M., Grigolato, S., Sgarbossa, A., Anfodillo, T., Cavalli, R., 2015. Life cycle environmental impact of firewood production – A case study in Italy. Applied Energy 150: 185–195. Puttock, G.D., 1995: Estimating cost for integrated harvesting and related forest management activities. Biomass and Bioenergy 8(2): 73–79. Röser, D., Mola-Yudego, B., Prinz, R., Emer, B., Sikanen, L., 2012: Chipping operations and efficiency in different operational environments. Silva Fennica 46(2): 275–286. Rottensteiner, C., Affenzeller, G., Stampfer, K., 2008: Evaluation of the Feller-Buncher Moipu 400E for Energy Wood Harvesting. Croatian Journal of Forest Engineering 29(2): 117–128.
216
SEA, 2008: Energy indicators 2008 Theme: Renewable Energy. Kungsgatan, Swedish Energy Agency (SEA), 21 p. Senturk, N., Ozturk, T., Demir, M., 2007: Productivity and costs in the course of timber transportation with the Koller K300 cable system in Turkey. Building and Environment 42(5): 2107–2113. Sirén, M., Aaltio, H., 2003: Productivity and Costs of Thinning Harvesters and Harvester-Forwarders. International Journal of Forest Engineering 14(1): 39–48. Spinelli, R., Cavallo, E., Eliasson, L., Facello, A., 2013: Comparing the efficiency of drum and disc chippers. Silva Fennica 47(2): 1–11. Spinelli, R., Hartsough, B.R., 2001a: Extracting whole short rotation trees with a skidder and a front-end loader. Biomass and Bioenergy 21(6): 425–431. Spinelli, R., Hartsough, B.R., 2001b: A survey of Italian chipping operations. Biomass and Bioenergy 21(6): 433–444. Spinelli, R., Magagnotti, N., 2011: Strategies for the processing of tree tops from hybrid poplar plantations. Baltic Forestry 17(1): 50–57. Spinelli, R., Magagnotti, N., 2013: Performance of a smallscale chipper for professional rural contractors. Forest Science and Practice 15(3): 206–213. Spinelli, R., Magagnotti, N., Laina Relaño, R., 2012a: An alternative skidding technology to the current use of crawler tractors in Alpine logging operations. Journal of Cleaner Production 31: 73–79. Spinelli, R., Schweier, J., De Francesco, F., 2012b: Harvesting techniques for non-industrial biomass plantations. Biosystems Engineering 113(4): 319–324. Stampfer, K., Kanzian, C., 2006: Current state and development possibilities of wood chip supply chains in Austria. Croatian Journal of Forest Engineering 27(2): 135–145. Croat. j. for. eng. 37(2016)1
Carbon Footprint of Forest Operations under Different Management Regimes (201–217) Tiernan, D., Zeleke, G., Owende, P.M.O., Kanali, C.L., Lyons, J., Ward, S.M., 2004: Effect of Working Conditions on Forwarder Productivity in Cut-to-length Timber Harvesting on Sensitive Forest Sites in Ireland. Biosystems Engineering 87(2): 167–177. Valente, C., Spinelli, R., Hillring, B.G., 2011: LCA of environmental and socio-economic impacts related to wood energy production in alpine conditions: Valle di Fiemme (Italy). Journal of Cleaner Production 19(17–18): 1931–1938. Van Belle, J.F., 2006: A model to estimate fossil CO2 emissions during the harvesting of forest residues for energy-with an application on the case of chipping. Biomass and Bioenergy 30(12): 1067–1075. Vusić, D., Šušnjar, M., Marchi, E., Spina, R., Zečić, T., Picchio, R., 2013: Skidding operations in thinning and shelterwood cut of mixed stands – Work productivity, energy inputs and emissions. Ecological Engineering 61: 216–223.
G. Cosola et al.
Wester, F., Eliasson, L., 2013: Productivity in final felling and thinning for a combined harvester-forwarder (Harwarder). International Journal of Forest Engineering 14(2): 45–51. Whittaker, C., Mortimer, N., Murphy, R., Matthews, R., 2011: Energy and greenhouse gas balance of the use of forest residues for bioenergy production in the UK. Biomass and Bioenergy 35(11): 4581–4594. Wolf,C., Klein, D., Weber-Blaschke, G., Richter, K., 2015: Systematic Review and Meta-Analysis of Life Cycle Assessments for Wood Energy Services. Journal of Industrial Ecology. Article in press, DOI: 10.1111/jiec.12321 Zimbalatti, G., Proto, A.R., 2009: Cable logging opportunities for firewood in Calabrian forests. Biosystems Engineering 102(1): 63–68. Zorić, M., Šušnjar, M., Pandur, Z., Mihaljević, K., 2014: Fuel Consumption and Greenhouse Gas Emission in Timber Haulage in Croatian Forestry. Nova mehanizacija šumarstva 35(1): 89–97.
Authors’ addresses: Giulio Cosola, MSc. e-mail: giulio.cosola@gmail.com Prof. Stefano Grigolato, PhD. e-mail: stefano.grigolato@unipd.it Sergio Monterotti, MSc. e-mail: sergio.monterotti@gmail.com Prof. Raffaele Cavalli, PhD.* e-mail: raffaele.cavalli@unipd.it Dept TESAF – University of Padova Viale dell’Università 16 35020 Legnaro ITALY
Received: February 10, 2015 Accepted: October 13, 2015 Croat. j. for. eng. 37(2016)1
Prof. Pierre Ackerman, PhD. e-mail: packer@sun.ac.za Dept of Forest and Wood Science Stellenbosch University Private Bag X1 Matieland, 7602 Stellenbosch SOUTH AFRICA * Corresponding author
217
ISSN 1845-5719
9 771845 571000