VOLUMETRIC COMPLEXITY WORKSHOP ONE | 2016 | MOSTAFA ELSAYED | APOSTOLIS DESPOTIDIS TAOLE CHEN | FEI ZHENG | CHERYLENE M. SHANGPLIANG
INDEX PHASE ONE PHASE TWO
PHASE THREE
INTRODUCTION PRELIMINARY FORM-FINDING CAST 01 CAST 02 CAST 03 CENTER POINT/TENSION LINES HYPOTHESIS
PATTERN EXPLORATION & METHODOLOGY CAST 04 CAST 05
METHODOLOGY STITCHING OPERATIONS CAST 06 CAST 07 CAST 08 CAST 09 PARAMETERIZATION
DEFINITION CAST 10 CAST 11 CAST 12 CAST 13 CAST 14 CAST 15
INTRODUCTION Workshop One is a material-based, four-week workshop where students get acquainted with the foundational basics of design research through physical experimentation. This particular workshop limited itself to working with plaster and lycra (also known as spandex). Our team was interested in complex three-dimensional forms from the beginning, as opposed to flat, pattern-dominated projects that are generally seen in this workshop. We turned to patterning later into the process to supplement our initial forays into the three-dimensional world. The logic behind our design was to utilise pattern, as well as modelled fabric form, to achieve more complex, volumetric forms. Each experiment was designed to gain a more fine-grained understanding of how the patterning works in conjunction with the fabric to retain defined, complex forms and asymmetrical/symmetrical curvature that would not deform into formless sacks, as is the general tendency of this particular medium.
PHASE ONE PRELIMINARY FORM-FINDING
With no prior experience in plaster-casting, our priority was to start casting immediately to gain some understanding of material behaviour. In each cast, we combined simple volumetric forms with various stitching techniques to start creating a topographic catalog of techniques based on real-world testing. With cast 03, we formulated a hypothesis that would enable us to control more substantial, three-dimensional forms and preventing the cast from turning into a formless “potato sack�, which is generally the tendency when working with the material.
400 40
0
0
40
600
fabric template and stitching pattern.
CAST 01 The first cast was a solid tetrahedron. Two sides were pinched to each other and one side featured single layer stitch patterns. We made the critical mistake of sewing from the inside, then having to reverse it. It did not work at all. Also, due to a lack of constraints, the cast completely lost its original shape and turned into a sack.
before cast
after cast
600
the form was controlled using tension lines and a center point, that is suspended on the frame.
CAST 02 In the second cast, again a tetrahedron was chosen as the base form. However, this time we scaled down the volume and introduced tension lines along with the concept of a centerpoint. Tension lines refer to lines (usually strings) that are either suspended to something fixed - such as the frame or counter forces. The center point is simply a point where tension lines converge and which is also suspended to the frame, preventing the cast from sagging and dropping to the ground. We were able to retain a defined shape compared to the first cast, however due to the small size it is unclear whether it was the system of tension lines/center point that actually controlled the shape.
before cast
after cast
two center points are added to control deformation.
the center points are fixed to the frame, as well as the corners.
CAST 03 In this cast, we scaled up considerably as well as increased formal complexity by using a cube. We needed to test whether our hypothesis would work on a larger scale. Expansion was controlled somewhat successfully. The cast swallowed a lot of plaster due to it solid core. We ended up using 30kg of plaster.
before cast
after cast
PHASE TWO PATTERN EXPLORATION & METHODOLOGY
After the initial casts, we started exploring patterning in order to achieve fine-grained control of complex forms. The purpose of any pattern that we will eventually develop would be to mitigate uncontrolled expansion when plaster fills the fabric. Our ultimate goal for this workshop is the exploration and understanding of material behaviour on complex, volumetric forms, as most other projects involve flatter designs that emphasize patterning.
3cm
500
3cm
500
CAST 04 & 05 The first attempt at pattern ended in a dead-end. The wave pattern did not let through any plaster during the pour. Also, aesthetically and structurally it did not provide any advantages.
after cast
In the second experiment we tried out a hexagonal pattern. We liked the aesthetic qualities. Also to be noted is its weight in relation to its structural stability. Although it broke after some handling, it was quite stable and we could pinpoint the cause of the breakage to the creases that formed in the ribs due to inbalanced pressure distribution.
after cast
During the process, we have slowly shifted from chaotic experimentation to a more systematized approach. The hexagonal pattern intrigued us and we decided to pursue it further, as we could sense its potential. The pattern is created by a layering of multi-dimensioned pinches and line stitches, which are individually simple operations, but form in the collective intricate pattern that we observed in cast 05. In the following few pages we have outlined the system that was eventually developed through repeated experimentation. This catalogue of methodology will help the reader understand how the pattern was created and the parameters that influenced the final outcome of each cast.
STITCHING OPERATIONS POINT STITCH The most basic operation is the point stitch. It is the most efficient method of restraining the fabric to itself, thus creating tension pockets. symbol
LINE STITCH
The second operation is the line stitch. It is used to close off larger areas of fabric and to create voids.
symbol
PINCHING OPERATIONS The essential operation that forms the distinct hexagonal pattern is the pinch. We have tested several versions of the pinch in an attempt to find the most optimal method. For the last few casts we used exclusively the point-stitch/single layer pinch combination, but we imagine that these different pinches can be used in conjunction to achieve more fine-grained control.
DOUBLE LAYER PINCH The double layer pinch is the fastest to stitch, however it severely limits the plaster from flowing through and the resulting voids are generally too weak to support the whole cast. This could be used in places where weight-savings need to be prioritized.
Cast 06/07_We applied the double-layer pinch to vertical and horizontal surfaces. Interestingly, the resulting forms were almost identical, leading us to the conclusion, that gravity does not have a significant effect on the pattern. For this reason, we think that this pattern is highly suited for complex three-dimensional forms, as it would not be restricted in which direction the surfaces point.
SINGLE LAYER PINCH
The single layer pinch is structurally much more stable. However, the tradeoff is a lot more weight.
Cast 08_When comparing with the double-layer pinch, the thickness is almost the same, but it is a lot heavier.
POINT STITCH + SINGLE LAYER PINCH
The point stitch combined with a single layer pinch offers the best balance between structure and weight. However, the tradeoff is the time-consuming fabrication process as it requires three operations for one pinch. This is the technique that we took into the final cast.
cast 09
PARAMETERS As we got more acquainted with the hexagonal pattern, we realized that aside from the specific method of pinching, there were more variables that affect the outcome of each cast. Put together, it forms a complex system of variables that can potentially adapt to complex formalities.
PARAMETER A_CELL SIZE (S) Altering cell size deforms the entire pattern. This can be exploited to reinforce areas that need stronger constrictions and to achieve complex curvature. Increased cell size expands an area more.
cell size (s) in millimeters.
PARAMETER B_RIB OFFSET (O) The rib offset directly affects the thickness, thus structural stability, of the resulting ribs that form the cell borders. A side effect of increasing offset is the greater shrinkage it causes, as it requires more material to form a larger rib.
rib offset (d) in millimeters, measuring both sides
PARAMETER C_PIN DENSITY Increasing pin density results in a more closed void in the center. Increasing material in the center is desireable, for example, when more structure is required in certain areas.
pin density (d) in amount/cell.
PARAMETER D_NODE PRESSURE (P) As we developed the pattern, we had to introduce an additional parameter to optimize pressure distribution in the ribs. By adding a node the plaster is pushed out from the hyper- expanded nodes into the ribs, preventing the plaster from forming structurally weak creases.
pressure point (d) is a factor. 0.0 is completely open, 1.0 is closed.
PHASE 3 DEFINITION
As we gained a deeper understanding of the pattern, we started thinking about applying it to more complex forms. In the following chapter, we document a series of experiments that test different formal approaches in conjunction with the pattern we have developed.
CAST 10 Cast 10 was mainly a test to explore the effects of all the parameters in synergy. The pattern was morphed assymmetrically, tightening towards the right bottom corner. It was a simple sheet that we then folded into a tube on the bottom. Our hypothesis was that the sheet would naturally develop a curvature towards the denser right side.
I. PROCESS
_the pattern starts out symmetrically on the top and drastically deforms on the right side. _the pattern causes the fabric to shrink to roughly 50% of its original size.
unrolled/unstitched fabric
stitched fabric
_we started experimenting with not stretching the fabric beforehand. _the form grows from an embryonic form into its unfolded form as the plaster enters.
before cast
after cast
II. STITCHING LAYERS
I. cutout with pattern
II. center pins
III. node pressure pts
IV. rib pinches
This diagram shows the layering of stitching operations that make up the pattern.
_incidentally, the hexagonal pattern inherently aids the plaster in distributing evenly, as every node can be abstracted into a Y-shaped intersection that naturally regulates the plaster flow.
III. FAILURE ANALYSIS
Cell Rib
Cell
Before
Cell
Before
90
A
B
Rib
Cell
_the translation from 2d pattern into 3d form changes drastically depending on location, size, and other factors. Essentially, the most important criteria is whether the ribs retained their structural integrity or not. To the left is a diagram showing the annotation system we use to illustrate our findings.
35
Cell
After
Rib
Cell
25 Rib
Cell
BEFORE -
Cell After
AFTER - 3D
Before36
50
5
90
25
Before
A A A
After
Rib
35 90 50
36
After Cell
35
90
50
25
36
Rib
5
Cell
5 35
36
50 160
10
Cell
6 5
15
B B B
5
Rib
160 160
Rib
10
5
Rib
5
10
Cell
Cell
15
Rib
6 6
Cell
15
10
57
6
Rib
15
Cell
85
Rib
10 10
10
57
C C C
5
25
160
15
C
_the ribs are rigid and solid. This is also the most sturdy section of this cast. This rib size even retains it integrity when it collapsed during the casting.
57
38 15
10 10
57
Cell
10 10 10 10
Rib
10
Cell
15
15
10
Rib 10
Rib
s : 160mm o : 85mm d: 5/cell p : 0.5 _this section collapsed entirely. the plaster was not strong enough to withstand pressure and broke when we moved it. Also, it did not have the intended effect of local curvature. s : 20-40mm o : 20-40mm d: 5/cell p : 1.0
85 38
85 85
38
Cell 38
Rib
to be noted is the extreme bulge of the center in the bottom parts. It appears that pin density needs to be increased to account for pressure levels.
_this section held in the beginning. however, it broke in some places after prolonged handling.
Cell Cell
s : 100mm o : 20mm d: 5/cell p : 0.5
CAST 11 Cast 11 was our attempt at combining the centre point hypothesis from Cast 02 with Cast 10. In conclusion, it failed completely. The principles found in Cast 02 could not be transferred directly to freehanging casts.
Centre point
ason to failure: Centre poin does not help much to constrain the shape.
Centre point
I. PATTERN LAYERS
II. TENSION DIAGRAM _the original intent was to limit expansion by counterpulling the sides through a centerpoint. However, the plaster pulled everything downward and caused the entire pattern to collapse.
III. FAILURE ANALYSIS _excessive tensioning has caused the fabric to roll over on itself and create critical creases, completely breaking any structural integrity.
CAST 12 Following cast 11, we tried a different approach to achieve complex forms. This time, two flat panels were to be connected using bridges to achieve a hollow, “light-weight� form. The cast resulted in failure again.
after cast
attempt to flip
after flip
I. PATTERN LAYERS
I. cutout with pattern
II. center pins
III. node pressure pts
IV. rib pinches
II. FAILURE ANALYSIS _the connections between tunnel and panel was extremely flawed, as it created very thin connection points. These broke immediately under pressure.
_without patterning to constrict, the tunnels ballooned in size and swallowed all the plaster. Also, due to the slight asymmetry when hanging the fabric, the plaster flowed to one side, causing uneven deformation.
_the hexagonal pattern is unsuitable to flat panels. We had this suspicion with previous casts, but it was confirmed in this one. The pattern relies on a slight curvature in the ribs to stay solid. When we took it off the frame, it broke instantly in several places.
CAST 13 Picking up on the observations we made in cast 10, we thought of other ways to achieve asymmetrical, complex curvature. In cast 13, we sought to achieve that through a combination of curvature modeled into the fabric itself and the use of our pattern to reinforce the curvature when the weight of plaster takes effect on the fabric.
I. PROCESS
cut out form+pattern
before cast
semi-sucessful flip attempt - the “neck� broke eventually
after cast
II. PATTERN LAYERS
I. cutout with pattern
II. line stitch
III. node pressure pts
IV. rib pinches
III. CURVATURE ANALYSIS
_the pattern was applied by closely comparing the unrolled surfaces with the 3d-Model in order to densify the cells in places of extreme curvature. _we realized that the hexagonal pattern could be simplified to circular cells, resulting in the same effect, but with much higher flexibilty when it comes to morphing the grid and greatly increase simplicity in drawing the pattern.
I V. FA I L U R E A N A LY S I S
_again, the larger cells are much more stable and less prone to breakage. s : 150mm o : 85mm d: 5/cell p : 0.5
_two openings, in conjunction with cells that are too small, likely caused the cast to fracture.
_the smaller cells do not provide enough material to create structural form. s : 30mm o : 5mm d: 5/cell p : 0.5
a band of larger cells runs through the less curved section of the cast, theoretically providing structure.
CAST 14 Given our current catalog of knowledge, there was one limitation we have yet to test - Large scale casts. Cast 14 takes our pattern and applies it at a much larger scale. A new frame was built for this purpose and we simplified the overall form to focus more on the pattern. The pattern was set up systematically to test a range of things. Ultimately, this cast completely failed as we underestimated the effects of gravity on the pattern.
unstitched fabric
before cast
stitched fabric
after cast
I. PATTERN LAYERS
offset: 90 cell: 320 offset: 70 cell: 240 offset: 40 cell: 200 offset: 30 cell: 125 offset: 25 cell: 50 offset: 30 cell: 125
cell: 165 offset: 60 cell: 210 offset: 40 cell: 260 offset: 25
I. cutout with pattern
II. center pins
III. node pressure pts
IV. rib pinches
_large cells on top to decrease weight. offset: 90 cell: 320 offset: 70 cell: 240 offset: 40 cell: 200 offset: 30 cell: 125 offset: 25 cell: 50 offset: 30 cell: 125
_gradually dense cells in the center to retain curvature in the center.
cell: 165 offset: 60 cell: 210 offset: 40 cell: 260 offset: 25
_large cells with narrow ribs to adjust for pressure.
I I . FA I L U R E A N A LY S I S
cast after removing the frame
_gravity completely overwhelmed any considerations we have thought through resulting to plaster filling up the bottom cells. After pouring about 25kg, we decided to stop the cast. It became obvious that our pinning strategy would not work for the bottom segment.
Even though it failed, the aesthetic quality is quite pleasing.
CAST 15
FINAL CAST Learning from the last cast, we adjusted our system to include a binary option between completely closed cells and partially open cells. Using this method, we hope to prevent the excessive ballooning of the bottom section. Going further, we also realized that this binary system could be used to strategically add structure, or save weight. Also, all nodes had a pressure of 1.0 in the interest of efficiency, as half pinches take an immense amount of time to stitch.
unstitched fabric
before cast
stitched fabric
after cast
I. PATTERN LAYERS
1880
2230
2230
1880
1880
2230
I. cutout with pattern
II. center pins
III. node pressure pts
IV. rib pinches
3d stitching
1800mm
2230
1880
1880
2230
2200mm
1800mm
II. STRATEGY structural column
structural column
pouring
structural column
pouring
pouring hollow hollow
hollow
cell: 350 offset: 20
cell: 320 offset: 30
cell: 300 offset: 50
cell: 280 offset: 80 cell: 260 offset: 100 closed off cells to control expansion
cell: 240 offset: 50
cell: 220 offset: 50
_cells decrease in size going downwards to account for expansion.
_bottom two rows completely closed off to avoid excessive deformation.
_three solid columns going upwards to provide structure. The rest is hollow to save weight.
_large ribs in the center to achieve deformation.
I I I . PA R A M E T E R T R A N S L AT I O N A N A LY S I S
_vertical columns using open cells worked. the cells expanded as expected.
_rib 3d-expansion after cast with nodes distributing pressure to ribs.
_small cells with thick ribs tend to collapse. There seems to be an optimal range for rib thickness.
_closed cell bottom worked as hypothesized. expansion was limited to an acceptable range.
I V. FA I L U R E A N A LY S I S
Cell (closed)
Rib
Cell
Before
A
30 30
After
320
35
0
60 246
Cell
B
C
50 50
300
240
198
Rib
Cell
110 6 135
140
5 120
Cell
Rib
Cell
55
2 50 50
27
205
Cell
15
3 194
Rib
Cell
_generally, the cast was a success. A majority of the ribs turned out solid with a well-controlled form. s : 320mm o : 60mm d: p : 1.0
_at some sections the nodes/ ribs ballooned without regularity. We were not able to find out how to mitigate this issue. It may be something as simple as a slightly imperfect stitch. s : 300mm o : 100mm d: p : 1.0 _this section appears crumpled. But the viewer will find that the material has gone towards the back where it caused the whole form to “bend� forward. s : 240mm o : 100mm d: p : 1.0
right view
front view
V. M O D E L D O C U M E N TAT I O N
left view
back view