AIAA DBF '14/'15 Report

School of Engineering Department of Aerospace Sciences

2014/2015 Design Build Fly February 23 2015

Contents 1 Executive Summary

2

2 Management Summary

3

5.4.1 Fuselage . . . . . . . . . . . . . . 36 5.4.2 Internal payload bay . . . . . . . . 37

2.1 Team Organisation . . . . . . . . . . . . .

3

2.2 Milestones Chart . . . . . . . . . . . . . .

4

5.4.3 External ball cage . . . . . . . . . 37 5.4.4 Wing . . . . . . . . . . . . . . . . . 38 5.4.5 Wing attachment . . . . . . . . . . 38

5

5.4.6 Tail . . . . . . . . . . . . . . . . . . 39

3.1 Design Constraints . . . . . . . . . . . . .

5

5.4.7 Tail attachment . . . . . . . . . . . 39

3.2 Scoring Outline . . . . . . . . . . . . . . .

5

5.4.8 Fuselage aft section . . . . . . . . 40

3.2.1 Total Score and Sensitivity . . . .

7

5.4.9 Nose cone . . . . . . . . . . . . . 40

3.3 Configuration Selection . . . . . . . . . .

8

3.3.1 Fuselage and Boom . . . . . . . .

8

Selection and Integration . . . . . . . . . 41

3.3.2 Empennage . . . . . . . . . . . . .

9

5.5.1 Ball drop system . . . . . . . . . . 41

3 Conceptual Design

5.5 Payloads Systems Design, Component

3.3.3 Wing . . . . . . . . . . . . . . . . . 10 3.3.4 Propulsion . . . . . . . . . . . . . 10 3.3.5 Landing Gear . . . . . . . . . . . . 10

6 Manufacturing Plan and Processes

46

6.1 Manufacturing Process Selection . . . . . 46 6.2 Subsystems manufacturing . . . . . . . . 46

4 Preliminary Design

12

6.2.1 Wing manufacturing . . . . . . . . 46

4.1 Design Methodology . . . . . . . . . . . . 12

6.2.2 Fuselage manufacturing . . . . . . 46

4.1.1 Airfoil selection . . . . . . . . . . . 12

6.2.3 Aft fuselage section manufacturing 46

4.1.2 Drag estimation: . . . . . . . . . . 13

6.2.4 Nose cone manufacturing . . . . . 46

4.1.3 Mass estimation: . . . . . . . . . . 15 4.2 Mission constraints . . . . . . . . . . . . . 17

6.3 Landing gear manufacturing . . . . . . . . 47 6.4 Schedule . . . . . . . . . . . . . . . . . . 47

4.2.1 Takeoff . . . . . . . . . . . . . . . 17 4.2.2 Maximum speed . . . . . . . . . . 17

7 Testing Plan

48

4.2.3 Turn performance . . . . . . . . . 17

7.1 Testing Objectives . . . . . . . . . . . . . 48

4.3 Sizing and tradeoffs . . . . . . . . . . . . 19

7.2 Propulsion Test . . . . . . . . . . . . . . . 48

4.4 Estimated aircraft characteristics . . . . . 20

7.3 Electronics Test . . . . . . . . . . . . . . . 48

4.5 Mission performance and Stability analysis 21

7.4 Structural Testing . . . . . . . . . . . . . . 49

4.5.1 Static Stability . . . . . . . . . . . 22

7.4.1 Wing spar and fuselage boom

4.5.2 Tail sizing . . . . . . . . . . . . . . 22

testing . . . . . . . . . . . . . . . . 49

4.5.3 Dynamic stability . . . . . . . . . . 23

7.4.2 Landing gear testing . . . . . . . . 49

4.6 Propulsion . . . . . . . . . . . . . . . . . . 24

7.4.3 Sensor Drop Test . . . . . . . . . . 49

4.6.1 Methodology . . . . . . . . . . . . 24

7.4.4 Flight Test . . . . . . . . . . . . . . 49

4.7 Electronics . . . . . . . . . . . . . . . . . 26

7.4.5 Schedule . . . . . . . . . . . . . . 50

4.7.1 Battery Selection Factors: . . . . . 26

7.5 Flight Checklist . . . . . . . . . . . . . . . 51

4.8 Radio System . . . . . . . . . . . . . . . . 27 5 Detail Design 29 5.1 Aircraft Dimensions and Components . . 29 5.2 Aerodynamic analysis . . . . . . . . . . . 30 5.3 Structural characteristics . . . . . . . . . . 33 5.3.1 Load Paths . . . . . . . . . . . . . 33 5.3.2 Structural Analysis . . . . . . . . . 34 5.4 Aircraft Systems Design, Component Selection and Integration . . . . . . . . . . . 36 U NIVERSITY OF G LASGOW

8 Performance Results

52

8.1 Structural performance . . . . . . . . . . . 52 8.1.1 Wing spar and fuselage boom performance . . . . . . . . . . . . 52 8.1.2 Landing gear performance . . . . 52 8.2 Electronics Test Results . . . . . . . . . . 52 8.3 Propulsion Test Results . . . . . . . . . . 53 8.4 Flight Test Results . . . . . . . . . . . . . 54 Page 1 of 54

1

Executive Summary

The objective of this report is to describe in detail the design, manufacturing and testing conducted by the University of Glasgow AIAA Design, Build and Fly team to create an aircraft that meets the competition requirements. The aim this year is to create a radio controlled aircraft that can perform the following flight missions: fly as many laps as possible in four minutes with no payload, carry 5.0 lb (2.3 kg)1 payload for three laps as fast as possible and fly as many laps as the team desires while dropping a single plastic ball to a specified area each lap. Furthermore, in the ground mission, the team will have to load the specific payload which is three pieces of lumber into the aircraft on the ground, secure the payload and remove it. Then load the amount of plastic balls intended to be carried in the final mission into the aircraft while securing them properly. The ground mission has to be completed as quickly as possible.

The total score for the team is obtained from three components: Total Mission Score, Written Report Score and Rated Aircraft Cost (RAC). Total mission score combines flight and ground mission scores. RAC includes the aircraft empty weight and the number of servos installed. It is also set that the aircraft has to be able to takeoff and land within 60 ft (18 m) and the battery pack cannot weigh more than 2.0 lb (0.9 kg).

From these requirements it was determined through competition score analysis that empty weight and number of servos were the two driving factors in the design, followed by the ability to load and unload the aircraft easily, carry a large number of plastic balls and to have high top speed. For example, if the team goes from having four servos to five servos, the total mission score will go down by 20%. This is a deduction that is virtually impossible to compensate with other improvements. Therefore, the team decided to fly the aircraft without a rudder and to only have a single engine, since ESC would also be counted as a servo.

It was evaluated that the team could not carry a number plastic balls such that their total mass would be higher than the mass of the three pieces of lumber. Hence, flight mission 2 was found to be the most demanding in terms of takeoff and landing requirements due to the highest mass payload required. This dictated the aircraft motor’s power and wing area combination requirements, consequently overpowering it for flight missions 1 and 3. This, however, enabled the team to choose higher pitch propeller to utilize the excess power and increase the top speed for flight mission 1. It also also allowed to operate the motor at maximum efficiency level for mission 3 extending the range and thus increasing the number of laps that can be flown. Greater wing area also made sharper turns possible decreasing the lap time.

The team’s final design was a 1.45 kg empty weight midwing monoplane with maximum flight speed of 23 m s−1 . The main construction materials were carbon fiber and foam. It can carry six plastic balls, takeoff within 18 m carrying mission 2 payload of 2.3 kg. The aircraft is able to complete 6 laps in four minutes for mission 1 and can fly 3 laps for mission 2 in 2 minutes.

1 Imperial Units will be quoted in this report when some competition specification is also quoted in those units. Otherwise metric system is used.

U NIVERSITY OF G LASGOW

Page 2 of 54

2

Management Summary

2.1

Team Organisation

University of Glasgow AIAA Design, Build and Fly team has members from first to fourth year undergraduates.

Overall, the team has 21 members from 9 countries, supervised by faculty advisor. All work was done by the students and on top of regular university curriculum. This was the first time when University of Glasgow has entered this competition and consequently the team management had to start from acquiring equipment, tools, workshop and work their way up step by step. The organisational structure of the team is presented in Figure 1.

Dr Hossein ZareBehtash Faculty Advisor

Arturs Jasjukevics 4 Project Manager

Aki Ruohonen 4 Technical Lead

Vladislav Andrijako 4 Treasurer

Deep Pandya 5 Business Manager

Julius Bartasevicius 4 Chief Designer Thomas Timmons 1 Assistant Treasurer

Aerodynamics

Structures

Electronics

Propulsion

Julius Bartasevicius 4 Sifoyiannos Nevradakis 2 Pavlina Dimitrova 3 Gani Petelov 4

Kiril Boychev 3 Blaga Todorova 3 Junaid Ashraf 1 Stoyan Barbukov 2 Christopher Logie 3

David Thrulbeck 4 Seifallah ElTayeb 2 Thomas Timmons 1 Benjamin Gregg-Smith 1

Angel Zarev 4 Euan McLean 4 Dragomir Kamov 3

Figure 1: Organizational Chart of the team. Number after each name indicated year of study

Team meetings were scheduled every week and utilized online forum and Google Drive as main lines of communication and information sharing. Project manager was responsible for setting up the master schedule, overall project goals and took the final responsibility for all major decisions. Technical lead assisted project manager, was a primary supervisor for the project’s technical side and managed project workshop, tools, equipment etc. The chief designer reported to the technical lead and also to the project manager to provide an extensive checking system for all project critical decisions.

Finance team worked mostly independently from the rest of the team and took the responsibility of financing the project and maintaining public relations. This was especially important since University of Glasgow participates in the competition for the first time and the team had to establish their name and make themselves visible for potential sponsors from scratch.

Furthermore, the team was divided into four sub-teams: Aerodynamics, Propulsion, Electronics and Structures. Each of those had a designated sub-team leader, who would manage the activities of the sub-team and report to the Chief Engineer and through him to Project Manager and Technical Lead. Outline of sub-team tasks is as follows: • Aerodynamics: Sizing of aircraft components, performance and stability estimates • Propulsion: Motor, ESC and Propeller sizing, testing and selection • Electronics: Aircraft battery, servos, receiver and transmitter sizing, testing and selection

U NIVERSITY OF G LASGOW

Page 3 of 54

• Structures: Aircraft fuselage, landing gear, empennage and drop mechanism construction and testing

2.2

Milestones Chart

Project manager created a general schedule for the project that identified all major steps needed to produce a final design. This included everything from early design to flight testing. Milestones chart is shown in Figure 1.

Activity

A/C Design Concept design Preliminary design Prelim. review Detailed design Design freeze Manufacturing Prototype 1 Prototype 2 Final Design Testing Propulsion test Component test Flight test Prototype 1 Flight test Prototype 2 Flight test Final Design Report Draft Edit Due date AIAA DBF competition

October

November

December

January

February

March

W i n t e r

April

Legend Planned Actual Due date Current date

/

E x a m S e s s i o n

Table 1: Master schedule of the project.

Due to the fact that this is the first year the University of Glasgow participates in DBF, the planned and actual schedule disagrees on some occasions. Most delays were caused by administrative issues, such as late receipt of funding, issues ordering tools and equipment, and inability to use certain facilities due to long queuing times. For next year, the team has been accumulating ”lessons learned” style information to preserve experience that will be used in forthcoming competitions.

U NIVERSITY OF G LASGOW

Page 4 of 54

3

Conceptual Design

In the conceptual design phase the mission rules and requirements were analyzed to determine the major design features of the aircraft that would maximize the team score. Sensitivity studies were carried out to quantify the contribution of each section of the competition and to find an optimal compromise for the aircraft configuration.

3.1

Design Constraints

The competition rules set several direct limitations on the possible design of the aircraft. Most important of these included: 1. Maximum battery mass is 2 lb (0.9 kg) of NiCd or NiMH type 2. Runway length for takeoff and landing is 60 ft (18 m) 3. Mission has to be conducted without significant damage to the aircraft 4. Complete all the missions: (a) Load and unload payloads of mission 2 and 3 in less than 5 min to avoid heavy penalty (b) Be able to carry flight mission 2 payload (5 lb (2.3 kg)) over 3 laps internally (c) Carry at least one plastic ball externally for mission 3 that can be dropped on command 5. Aircraft has to withstand lifting from the wingtips when fully loaded 6. Payloads must be secured well and ball drop mechanism designed such that the center of gravity (CG) stays within limits 7. Intentional aerobatic maneuvers are not allowed The competition course layout with runway, turns and drop zone is shown in Figure 2. Total length of the course is approximately 3000 ft (1 km).

Figure 2: The flight course layout

3.2

Scoring Outline

Total competition score is given by Score =

U NIVERSITY OF G LASGOW

W RS 路 T M S RAC

Page 5 of 54

where W RS is Written Report Score, T M S is Total Mission Score and RAC is Rated Aircraft Cost. Written Report Score is based on the quality of the technical report submitted by the team in February.

Rated Aircraft Cost RAC is found from RAC = EW · NServo where Empty weight, EW = M ax(M 1, M 2, M 3) hence the highest empty weight for any mission configuration and NServo is the number of servos installed. Definition of a servo for the competition includes, but is not limited to • Conventional radio controlled servo actuator • Speed controller • Electric motor not used for propulsion • Solenoid actuator • Electric relay Total Mission Score Total Mission Score is defined as T M S = GS · F S where GS is Ground Score and F S is Flight Score Ground Score Ground Score is determined by the time taken to perform the ground mission. In the mission the team has to load flight mission 2 payload into the aircraft and secure it. Then the timer will be paused and the payload securing is checked. After this, timer will be resumed while the team loads the number of plastic balls declared in the technical inspection for flight mission 3 into the aircraft and secures them. The time is stopped and final check of the loading is done. The score is then calculated from GS =

Fastest Loading Time Loading TimeU G

where Loading TimeU G is time achieved by the University of Glasgow.

Flight Score Flight score can be calculated from FS = M1 + M2 + M3 where M 1 to M 3 denote individual flight mission scores Mission 1 Ferry Flight

Fly 4 min around the course without a payload to complete as many laps as possible. Score

is given by M1 = 2 ·

Number of Laps FlownU GM 1 Number of Flaps FlownM 1

Mission 2 Sensor Package Transport Mission

Fly three laps around the course with 4.5 00 x 5.5 00 x 10 00 (11 cm x

14 cm x 25 cm) stack of pine boards weighing approximately 5.0 lb (2.3 kg) as fast as possible. Number of points is then M2 = 4 ·

U NIVERSITY OF G LASGOW

Fastest Time TimeU G

Page 6 of 54

Fly as many laps as desired by the team while dropping one plastic ball of 3.759 00

Mission 3 Sensor Drop Mission

(95.48 mm) in diameter, weighing 1 oz (30 g) to a designated area from the aircraft each lap. Mission score then becomes M3 = 6 ·

3.2.1

Number of Laps FlownU GM 3 Number of Flaps FlownM 3

Total Score and Sensitivity

When the terms in the total competition score are now expanded it can be observed that Score =

W RS · =

Fastest Loading Time Loading TimeU G

· 2·

W RS · T M S = RAC

Number of Laps FlownU G M1 Number of Flaps FlownM 1

+4·

Fastest Time TimeU G

+6·

Number of Laps FlownU G M3 Number of Flaps FlownM 3

EW · NServo

Time Therefore the variables for sensitivity analysis are the ground and flight missions performance ratios Fastest etc., TimeU G

empty weight and number of servos as those are the parameters that are determined by aircraft performance Written report score is not affected by aircraft performance and thus will not be considered in sensitivity analysis. Sensitivity plot is shown in Figure 3. It now becomes clear that reduction in empty mass and in the number of servos have the

80 Empty mass Mission 1 score Mission 2 score Mission 3 score Number of servos Ground score

Total score change, %

60

40

20

0

-20

-40 -40

-30

-20

-10

0 10 Parameter change, %

20

30

40

Figure 3: Sensitivity Plot of the mission requirements

largest effect to the total competition score since their product divides the entire competition score expression. For instance 10% increase in empty weight could only be compensated by about 20% increase in mission 3 score or

U NIVERSITY OF G LASGOW

Page 7 of 54

30% increase in mission 2 score. Furthermore ground score plays a vital part in the total score since it multiplies all the flight mission scores. Thus if the team scores 0.2 in ground mission, 80% of the flight mission score will be lost compared to the fastest team. It should be noted that adding a servo causes very sharp drop in the total competition score as the percentage change will be very high when moving from, say, 4 to 5 servos.

Last, the aircraft needs to be designed to perform all missions, thus 60 ft takeoff distance and 3 lap range have to be met at maximum payload, i.e. M2 payload. The requirements are translated to aircraft requirements in Table 2.

Importance Mission Requirement

Aircraft Requirement

1

60 ft Takeoff and Landing with M2 payload

High lift and Thrust-to-Weight

2

Minimize EW

Use of competitive materials (Carbon Fiber etc.), optimized design

3

Minimize NServo

Minimize number of ESC, systems that require servos(flight control, sensor drop), motors not used for propulsion etc.

4

Fast M2 & M3 loading and unloading securely

Easy access to internal payload bay, Simple ball loading mechanism, Straightforward securing systems.

Table 2: Mission requirements translated to aircraft requirements

3.3

Configuration Selection

One of the main design challenges was to design a payload drop mechanism that would not require an extra servo.

Initially it was thought to use a device that would drop a ball based on dynamic pressure. This, however, had to be abandoned because of reliability concerns the changing wind conditions might bring.

The following idea was to perform an aileron roll that would cause a ball to fall. This idea dictated that the balls would need to be dropped from the top of the aircraft and that tolerable rolling characteristics could be achieved. Intentional aerobatics maneuvers were, however, forbidden after rules update and the team had to drop this idea as well. It was decided then to have a more conservative approach and have a dedicated servo to operate the drop system.

3.3.1

Fuselage and Boom Flight direction

M2 Payload

M3 Payload M3 Payload drop direction Figure 4: Fuselage and Boom general configuration

U NIVERSITY OF G LASGOW

Page 8 of 54

Configuration selection started from the most fundamental aspects of aircraft layout. Since empty weight has a very high importance in scoring, a minimalistic approach was taken for the fuselage design. This meant that the fuselage needed to have just enough space for the flight mission 2 payload carried internally and for up to 8 plastic balls (which was defined from endurance limitations) for flight mission 3 to be carried externally. Furthermore the planned usage of composite materials as much as possible shaped the aircraft such that flat composite plates could be used as much as possible. Film covering would be used to minimize time consuming molding, while at the same time maintaining acceptable aerodynamics characteristics.

Those requirements led to a conceptual design with a carbon main boom to which all aircraft components were fixed.

Ball cage that holds the plastic balls was positioned on the bottom side of the boom and the balls would be dropped from the back of the cage. This was done to avoid collisions with the fuselage, empennage and propeller. The cargo compartment for mission 2 payload was mounted above the boom to avoid excessive length and to limit center of gravity movement in mission 3. This configuration is presented in Figure 4.

The conceptual thought that concerned the fuselage layout was aiming for a design that would be as aerodynamically effective as possible. Effectiveness in this case meant the reduction of drag caused by the fuselage, its contribution to the lift coefficient and its optimum functionality in each mission. These requirements by themselves led to a design based on an aircraft’s wing airfoil. Based on experimental data and wind tunnel tests on different types of airfoils and fuselages, and while observing actual designs of transport aircraft at the same time, the fuselage was designed in a way that resembled a semi symmetrical airfoil. Once an initial approach was architectured and pictured, details that would play an important role to its functionality had to be added, with the soothing of the surface from sharp edges being the most important.

3.3.2

Empennage

(a) T-tail

(b) V-tail

(c) Conventional tail

Figure 5: Different tail configurations

The driving requirements for empennage design were lightweight, simplicity of manufacture and effective provision of control. The types of tails considered were conventional, V- and T-tails as seen in Figure 5. T-tail offers good clearance from wing interference by placing the horizontal tail high above the wing thus enabling smaller area. This, however, forces vertical stabiliser to be very rugged so that it can take the loads created by the horizontal tail positioned on top of it. Also, at high angles of attack, tail can be shadowed by the fuselage causing deep stall that can be extremely difficult to recover from.

V-tail can reduce drag of the aircraft by reducing the amount of intersections that cause aerodynamic interference as compared to conventional or T-tail. However V-tail also creates strong coupling of lateral and longitudinal control.

This coupling can create flight behaviour that is difficult to handle when the aircraft does not have a rudder or in high gust situations. Lack of rudder control means that V-tail would lose one of its advantages of enabling pitch and yaw control with high aerodynamic efficiency. Hence it is ineffective and complex solution if only pitch control is required.

Conventional tail is similar to T-tail but a horizontal stabiliser is connected to the fuselage/boom rather than positioned on top of the vertical stabilizer. When the horizontal tail is close to the plane of the wing, downwash of the wing will significantly reduce efficiency of the tail increasing required tail area as compared to an ideal situation. Advantage is that the tail can be connected directly to the fuselage/boom which is a simple solution and does not require strong

U NIVERSITY OF G LASGOW

Page 9 of 54

vertical tail to hold the horizontal tail. Thus conventional tail was found to be the best compromise.

3.3.3

Wing

(a) High wing

(b) Low Wing

(c) Mid Wing

Figure 6: Different wing configurations

The position of the wing was largely dictated by the location of the main boom. The positions considered are illustrated in Figure 6. The boom is the main load carrying structure of the aircraft and it would be inefficient to attach the wing to anything else since this would require heavy strengthening of the parts that would carry the wing. The boom location, on the other hand, was dictated by the requirement to carry plastic ball externally and timber internally in separate compartments and at the same time limit length, height and CG shift to minimum.

Mounting the wing low would have enabled the team to take better advantage of ground effect which is very important due to strict takeoff and landing distance limitations. However, the reasons stated earlier outweighed the advantages that could be obtained from a low wing configuration. Mass especially would have become an issue if the wing had been attached, for example, to the bottom of the ball cage. High wing could have provided extra stability by placing the wing above the CG, however high wing configuration would have required strengthening of the top cargo compartment. Due to these reasons a mid-wing configuration was selected.

3.3.4

Propulsion

Selection of propulsion configuration was made relatively straight forward by the combination of mission 3 sensor drop requirement and the servo count cost which made every ESC to count as a servo. Thus minimal number of ESCs and a good clearance for dropped balls were the two main factors. Two motors could be powered with a single ESC but this would not provide any significant advantages such as differential thrust control and the mass of the system would be most likely higher than with a single motor. Furthermore pusher configuration was quickly discarded due to the fact that a ball could hit a pusher propeller when ejected backwards and propeller ground clearance issues on takeoff and landing. Therefore single motor tractor configuration was chosen.

3.3.5

Landing Gear

(a) Tricycle

(b) Tail Dragger

(c) Bicycle

Figure 7: Different Landing Gear configurations

The main requirements for landing gear design were acceptable ground clearance, light weight, good stability &

U NIVERSITY OF G LASGOW

Page 10 of 54

handling and low drag. Evaluated configurations are presented in Figure 7. Tail dragger configuration is generally the lightest option since the tail gear can be very light due to the small load on it. However ground clearance was an issue because the low position and long length of the ball cage. This would make the main landing gear very tall and cause too high ground angle of attack which in windy conditions causes significant problems. Also the aircraft CG is aft of the main landing gear which makes this arrangement inherently unstable thus the pilot has to use rudder to balance the aircraft.

Bicycle gear would require attaching the wheels to the ball cage which necessitated more mass to strengthen it to carry the weight. Also for ground stability, small wheels would be required at each wingtip, leading to added drag and mass. This solution could be useful when extremely high aspect ratio is used to support the wing upon landing. This design requires very long runways, since the aircraft can not be rolled due to the aft wheel being far behind the CG. Tricycle gear has moderate weight but provides good ground handling qualities. Ground angle of attack can also be set to a more manageable level for high winds than with a tail dragger configuration. The inherent stability of a tricycle gear allows relatively high sideslip angle for landing and thus permits tolerable landing characteristics for a rudderless aircraft. Hence tricycle gear was selected.

U NIVERSITY OF G LASGOW

Page 11 of 54

4 4.1

Preliminary Design Design Methodology

The constraint analysis method was used to design the preliminary configuration of the aircraft. Flying mission 2 was decided to be the most demanding one. This is because it has the biggest payload and also the aircraft has to fly at maximum speed. Mission 3 needs more range, but in terms of power for takeoff, it does not require as much as mission 2. Hence the baseline configuration was designed for mission 2.

The method employs analytical equations of power loading (P/W) as a function of wing loading (W/S) for different mission segments. The mission segment performance is influenced by various constraints, in this case being either competition requirements, environmental conditions or reasonable estimates.

4.1.1

Airfoil selection

With the two main design goals set, the number of servos to be minimum and also mass to be minimum, the airfoil selection was constrained. It was decided not to use any high lift devices, hence a big maximum lift coefficient was the number one criteria for the airfoil selection. Also, it appeared to be one of the biggest influences on minimum mass for takeoff, hence a coefficient in the range of 1.5-1.7 was the first constraint used. Bigger lift coefficients would result in very high trim drag for missions 1 and 3 since the takeoff weight is much smaller than in mission 2.

The second selection criterion was for the airfoil to be optimized for low Reynolds number flow, usually being less than 500,000. Such flow aerodynamics, when it is mostly laminar, require great attention for a couple of reasons.

First, hysteresis is often observed on airfoils operating in such regime. Hysteresis, flow’s dependency on its previous history, is particularly important for stall recoveries, spin flight or high gust conditions. This unsteady phenomenon highly increases the drag and decreases the lift when the angle of attack of an airfoil is reduced after stalling, possibly resulting in loss of the aircraft. Hysteresis is strongly coupled with the second phenomenon of low Reynolds number airfoils - laminar separation bubbles. Laminar separation bubbles form when airflow that is laminar separates because of high adverse pressure gradients and when separated, transitions to turbulent flow. It then curves back and reattaches to the surface, creating a shallow region of reversed flow. Such bubbles increase drag, which can become a few magnitudes bigger than the drag of the airfoil without the bubble, and are to be avoided.

The phenomena of laminar separation bubbles can be controlled with selecting the right transition location during the airfoil design phase with transition ramps or with externally fitted turbulators.2 After analysing the list of low Reynolds number airfoils, three options were chosen. Selig SG6043, which was designed for small wind turbines, Wortmann FX63-137, which originally was designed for human powered airplane use but also adapted by airplane modellers, and Douglas LA203A, another high lift airfoil. The selected airfoils are shown in Figure 8. The 2D polars can be found in Figure 9. For cruise lift coefficient ranging between 0.11 to 0.42 for various missions, the Wortmann FX63-137 had the lowest drag from the three airfoils. Even though it has a very thin trailing edge that might cause problems for manufacturing, it was decided that with careful use of carbon fiber ribs it could be avoided. Hence the selection was concluded with the Wortmann FX63-137 as the airfoil.

2 Michael S. Selig. “Low Reynolds Number Airfoil Design Lecture Notes”. In: the von Karman Institute for Fluid Dynamics, Lecture Series.

2003.

U NIVERSITY OF G LASGOW

Page 12 of 54

Figure 8: Three airfoil profiles considered in the final stage. Red is Wortmann FX63-137, blue is Selig SG6043 and green is Douglas LA203A

(a)

Cl Cd

vs Angle of Attack

(c) Cm vs Angle of Attack

(b) Cl vs Cd

(d) Cl vs Angle of Attack

Figure 9: Aerodynamic graphs of the final three airfoils.

4.1.2

Drag estimation:

Zero-lift drag buildup model from Roskam3 was used to determine drag component by component using 14m s−1 , assumed average flight speed, as a reference.

3 J. Roskam. Airplane Design: Part VI: Layout Design of Cockpit, Fuselage, Wing and Empennage : Cutaways and Inboard Profiles. Airplane Design. DARcorporation.

U NIVERSITY OF G LASGOW

Page 13 of 54

First, wetted areas of the exposed components were calculated, approximating the fuselage as a box a bit bigger than the mission two payload. Then skin friction coefficients were estimated, depending on the component location and the Reynolds number of the flow over the component. Only the tail was assumed to have a laminar boundary layer, since the fuselage is in the wake of the propeller and the wing was also partially covered by the wake.

Form factor, which defines the interference and pressure drag was then included for each component. For the wing and tail it mainly depends on its thickness ratio and for fuselage it depends on its fineness ratio. For the landing gear the drag coefficients suggested by Hoerner4 were used. All the equivalent parasite drag areas and zero-lift drag coefficients are tabulated in Table 3.

Component

Wetted area, m2

Form Factor

Equivalent parasite drag area, m2

Parasite drag coefficient

Wing Tail (horizontal and vertical) Fuselage Landing gear Total Total, including a factor of 1.06

0.7801 0.1659

1.3012 1.1

0.0061 0.0015

0.0143 0.0034

0.2985 -

4.0594 -

0.0065 0.0016 0.0157 0.0166

0.0153 0.0036 0.0367 0.0389

Table 3: Aircraft Drag Buildup

Using XFLR5, average Oswald’s efficiency factor (e = 0.949) for the wing was estimated. Lift-induced drag then was calculated and a drag polar for complete aircraft was created. It can be found in figure 10.

4 S.F.

Hoerner. Fluid-dynamic drag: practical information on aerodynamic drag and hydrodynamic resistance. Hoerner Fluid Dynamics, 1965.

U NIVERSITY OF G LASGOW

Page 14 of 54

Figure 10: Drag Polar of the complete aircraft

4.1.3

Mass estimation:

List of aircraft parts and components was created and mass of each was estimated. Fuselage mass was based on the previous reports which had similar conceptual design. Also, 10 g were added for every ball the aircraft is going to carry, because the ball cage gets heavier. Three component masses were left as functions of different parameters: wing, motor and battery.

Wing density per area was first estimated using previous reports of similar configuration and aspect ratio wings, and later refined after a CAD model was finalized. However, the first estimate proved to be quite close to the actual value.

Some 300 motors were put into a database and a graph of motor mass versus power was created, as shown in Figure 11. Especially in the lower power range, the graph is linear with the deviation increasing as the power increases. After filtering out the overpowered motors according to estimations, the motor mass unknown was substituted by a linear function with required power as a parameter.

U NIVERSITY OF G LASGOW

Page 15 of 54

Figure 11: Motor mass to Power graph

Finally, battery mass was dependent on the required power output and the required capacity. However, the capacity requirement for mission 3, where battery endurance is required, was a smaller constraint than the power requirement in mission 2. After analyzing a few different cells that would fit the performance range, a mathematical model was established, resulting in a battery mass function with battery’s C rating, maximum power required and the energy density of the battery chemical (80 W h kg−1 ) as input parameters.

Battery mass =

Maximum power required C rating · Chemical energy density

The preliminary component masses are tabulated in table 4. With these estimated, mass could be calculated for every point on the power requirement graph.

Component

Mass, kg

Servo Receiver + receiver battery + Fuse ESC Propeller Boom Fuselage Landing gear Empennage Motor Wing Battery

0.03 0.06 0.05 0.02 0.12 0.150 + Nballs · 0.01 0.11 0.11 M2 mass · PW / 3534.6 M2 mass · 0.73945 / WS Maximum power required / 800

Table 4: Initial Mass build up estimations

U NIVERSITY OF G LASGOW

Page 16 of 54

4.2

Mission constraints

The 3 most influential mission segments were considered- takeoff, maximum speed and turn performance. Other two, cruise for maximum range and climb were found to be less demanding.

4.2.1

Takeoff

Takeoff performance initially was the most influential segment. It dictates the minimum wing size with selected maximum lift coefficient. The bigger the Clmax , the lower the stall speed and wing area, decreasing the total mass.

Takeoff also greatly depends on the available field length. A safety factor of 1.25 was used to make sure that it is possible to take off in the given field of 18.3 m.

4.2.2

Maximum speed

The aircraft is supposed to fly at maximum speed during the timed missions. However, since mission 2 score is less important than the empty mass of the aircraft, it was decided not to chase a high maximum speed and only make sure that in windy conditions the aircraft is able to move forward. Research on the environment of the competition location was done for maximum wind speeds. Last year, the winds in Tucson, AZ reached 240 km h−1 . However, flying against such wind for an RC airplane is impossible, so only the wind speeds of three years before were considered, averaging at 70 km h−1 . As a consequence, the maximum achievable speed was chosen to be 82.8 km h−1 (23 m s−1 ).

4.2.3

Turn performance

It was first decided that a high aspect ratio wing is going to be used for smaller induced drag and better cruise performance. This requires higher structural stiffness, especially for the wing root, since bending moments are increased. Most of the previous reports suggested maximum turn load of 5-6 G’s. It was estimated that a load of 5.5G will be achieved during the maximum G turn. However, the actual loads can only be estimated after test flights.

The mission requirements are summarized in a Table 5 and added safety factors can be found in Figure 6. The code operating principle is shown in Figure 12.

Design constraints

Parameters for code input:

F IELD LENGTH MAXIMUM ClMAX

Cl max = 1.63 roll fric = 0.025 t grun = 5

W IND SPEED

cruise h = 15 V max = 23

T URN SPEED

V turn1 = 14

VARIOUS

AR = 8 N balls = 6 h 0 = 680 T 0 = 20

Takeoff Estimated maximum coefficient of lift (WORTMANN FX 63-137) Ground roll friction coefficient Ground run time, s Maximum Speed Desired cruise altitude, m Estimated maximum speed, m s−1 Turns Turn velocity, m s−1 General Estimated aspect ratio Number of balls, an initial guess Ground altitude of Tucson, s Average temperature in Tucson, ◦C Table 5: Mission requirements

U NIVERSITY OF G LASGOW

Page 17 of 54

Geometry inputs: Taper, AR, Volume coefficients, Tail ARs

Typical component masses

Payload inputs: Number of balls carried, mission 2 payload mass

Preliminary drag buildup

Preliminary aircraft mass

Adjust model mass for different power loadings and wing loadings

Adjust drag for every matrix point

Aerodynamic inputs: CL_max

Safety factors

Takeoff Performance

Turn Performance

Wing, motor, and battery densities

Initialize wing loading and power loading matrix

No

Battery mass required the same as predicted?

Design requirements: Takeoff distance, Max speed, Turn speed

Max Speed Performance

Calculate energy required for all the missions

Yes

Empty mass matrix compiled

Minimum mass configuration

Figure 12: Operating principle of the optimization code. Blue represents inputs, grey- outputs. Red shows the iteration loop of the code.

Maximum cruise range performance was also considered, but it didn’t imply any requirements higher than the mission segments mentioned above. Climb angle was chosen so that it does not change the power requirement too, resulting in 8◦ , which seemed to be sufficient. The resulting power loading and wing loading (also wing cubic loading) curves were found and a model and calculated empty masses were incorporated into the graph. The resulting graph can be found in Figure 13.

Safety Factors S takeoff = 60 / 1.25 battery SF = 1.2 eta prop max = 0.7 eta moto max = 0.6 T hot = 30

Takeoff distance (ft), 1.25 divider as a safety factor Battery mass multiplier for safety Estimated propeller efficiency Estimated motor efficiency Hot day temperature at Tucson, AZ, ◦C

Table 6: Safety Factors of the optimization code

U NIVERSITY OF G LASGOW

Page 18 of 54

Empty Mass versus Power Loading versus Wing Loading

Empty Mass versus Power Loading versus Wing Cubic Loading Empty mass, kg Takeoff Climb Performance at 8 degrees Turn speed, 14m/s Max speed, 23m/s Max Range Performance Minimum EM Point

180 2 .5

3

3 2.5

180

2

2

160

160 2

2

1.5 2

3 2.5

100

80 60

20

20 2

3 2.5

40

3 2.5

40

4

1.5

60

1.5

100

2

1.5

80

120

2.5

120

Power Loading, W/kg

140

3

Power Loading, W/kg

140

6 8 10 Wing Loading, kg/m2

12

4

6

8 10 12 14 16 18 20 22 24 Wing Cubic Loading, kg/m3

Figure 13: Power Loading vs Wing Loading and Power Loading vs Wing Cubic Loading. Note that the mass values below the turn constraint line have no physical significance. This is due to optimization code estimating that the airplane stalls when power available is less than that needed for a turn. However, this does not influence the minimum mass point in the design space in any way. Also the mass lines are discrete, because of battery mass increasing not linearly, but as an integer multiplier of battery cell mass.

4.3 Sizing and tradeoffs As can be seen in Figure 13, lowest empty mass point was automatically marked and all the geometry and aerodynamic characteristics for that point were generated. However, the suggested values for wing loading were discussed in a bit more detail.

Wing loading, as explained by Roskam,5 affects many airplane handling qualities. Table 7 presents a summary of those.

Stall Speed Fieldlength (La and To) Max. L/D Ratio: Ride quality in turbulence: Weight:

High W/S

Low W/S

High Long High Good Low

Low Short Low Bad High

Table 7: Wing Loading effects summarized

5 J. Roskam. Airplane Design: Part III: Layout Design of Cockpit, Fuselage, Wing and Empennage : Cutaways and Inboard Profiles. DARcorporation, 2002.

U NIVERSITY OF G LASGOW

Page 19 of 54

As a general rule, loadings up to 3 kg m−2 are for gentle flying, 3 kg m−2 to 6 kg m−2 for trainers and above 6 kg m−2 for aerobatics. However, Reynolds6 suggests that wing cubic loading factor, expressed as W CL =

Weight , Area1.5

would be

more appropriate to estimate the handling qualities for RC airplanes. Including the linear term (dividing by root of area, i.e. some reference length) in the wing loading equation introduces the airplane size as another factor. This makes it easier easier to categorize airplanes according to their controllability. Wing cubic loading for the designed aircraft varies from 5.2 kg m−3 to 13.3 kg m−3 for different missions. According to Myers7 the aircraft would categorise as a sailplane/park flyer for missions 1 and 3, and as an expert sport airplane for mission 2. It was decided that as long as the pilot has enough experience flying the aircraft, handling qualities similar to a sports aircraft are manageable, hence the minimum mass design point was accepted.

Level and Description

Average WCL Factor, kg m−3

1: Includes mostly indoor type models and those that can be flown outside in very light winds, only level with no internal combustion powered planes 2: Includes mostly backyard type models that can be flown indoors in larger venues and outside in low wind conditions, includes a few internal combustion powered planes 3: Includes park flyers, sailplanes, biplanes, 3D planes 4: Includes sport types, biplanes, scale, a few 3D planes, pattern, largest level 5: Includes advanced sport types, sport scale and sport scale warbirds, some twins 6: Includes expert sport types, scale, scale warbirds, twins 7: Includes planes for the expert flier only, twins and multi-motor, true scale, warbirds.

2.39 4.10

5.99 8.51 11.25 14.31 17.52

Table 8: RC Airplane general handling quality levels. http://www.theampeer.org/M1-outrunners/M1-outrunners.htm#CWL

After the preliminary CAD model was created, the number of balls to be carried for mission 3 was discussed. Two constraints limit this. First, battery capacity to fly the laps required. It was calculated that the battery required for mission 2 allows flying another two laps in mission 3. However, the limitations from having the ball cage under the payload bay are bigger. Taking the wing incidence angle into account, roll angle has to be at least 15◦ . Since the ball cage can not be pushed further up front than the nose landing gear, it limits the roll angle unless the landing gear length is increased. And following the main design goal for the mass to be kept at minimum, it was decided to stick with 6 balls.

4.4

Estimated aircraft characteristics

Table 9 summarizes the inputs and outputs resulting from the optimization code.

6 Francis 7 Ken

Reynolds. Wing Cube Loading. Model Builder. Sept. 1989.

Myers. Club Newsletter. Jan. 2014. URL: http://www.theampeer.org/M1-outrunners/M1-outrunners.htm#CWL.

U NIVERSITY OF G LASGOW

Page 20 of 54

Inputs: Used CLmax Used taper Ratio Aspect ratio Number of balls flown in mission 3

1.63 0.6 8 6

Outputs for minimum empty mass configuration: Masses: Total empty mass, kg Total mission 2 mass, kg Wing mass, kg Boom mass, kg Motor mass, kg Battery, kg Other data: Reynolds number of the wing Power loading for mission 2, W kg−1 Wing area, m2 Wing loading for mission 2, kg m−2 Wing cubic loading for mission 2, kg m−3 Wing span, m Wing mean aerodynamic chord, m Root chord, m Tip chord, m Power requirements: Takeoff power requirement for mission 2, W Best L/D power requirement for mission 2, W

1.449 3.717 0.249 0.118 0.112 0.395 223,000 85 0.427 8.7 13.3 1.849 0.236 0.289 0.173 316 109

Table 9: Optimization code parameters

4.5

Mission performance and Stability analysis

Even though the airplane was designed to fly at maximum speed of 23 m s−1 for mission 2, while calculating the of mission 1 the maximum speed was kept the same. The aircraft will experience more trim drag without payload, however this will be compensated by higher dynamic thrust using a higher pitch propeller. An option to raise both ailerons to decrease the lift was discussed, but only after flight tests it could be implied practically. For mission 3 the airplane speed is set to the best L/D speed for maximum cruise performance.

Time, s Energy required, mA h Cl, cruise Stall speed, m s−1 Takeoff speed, m s−1 Best L/D speed, m s−1

Mission 1

Mission 2

Mission 3

230 570 0.114 5.77 6.347 -

121 667 0.291 9.70 10.67 13.6

341 465 0.419 6.54 7.192 7.81

Table 10: Mission Performance Estimations

It was estimated that 6 laps could be flown for mission 1 in 230 s, which, after comparing numbers from previous reports, was a reasonable number. Mission 2 was estimated to take 121 s, a bit more than a half of the mission 1 time. Energy requirement estimation for all the missions was done multiplying the power required for each mission segment by the time it takes to complete the segment. The results are tabulated in Table 10. The aircraft RAC was calculated to be 12.78.

U NIVERSITY OF G LASGOW

Page 21 of 54

4.5.1

Static Stability

Having a big takeoff weight difference in between the missions implies stability issues. After having the preliminary CAD model of the aircraft ready the CGs for different missions could be estimated. The CG change appeared to be quite radical, if it is to be trimmed with the elevator only. However, since possibility to trim with the battery exists, it is more useful to keep the CG within a smaller static margin range, which then results in less trim drag and also helps the pilot by keeping similar control characteristics of the aircraft.

The most important characteristic in context of static stability is the neutral point location. The aircraft is stable as long as its CG is in front of the neutral point, which usually depends purely on the aerodynamics of the aircraft. The distance between the neutral point and the CG location is called the static margin, which as a first estimation was approximated to be 5% MAC. As the static margin grows, the airplane becomes more and more statically stable.

Since it is assumed that the elevator servo and control mechanisms are stiff enough, stick-fixed neutral point is considered.

Usually, it is best to place the wing so that its AC is quite close to the CG of the aircraft, reducing the tail size required.

However, due to structural limitations, it was decided to place the wing in front of the mission 2 payload compartment.

This meant that the CG for mission 2 is behind the wing aerodynamic centre, implying that the tailplane is going to have to provide lift force for stable flight. After assuming the aircraft to be trimmed to the same or more stable configuration with the battery in missions 1 and 3, required neutral point position could be estimated and further tail design could be done.

4.5.2

Tail sizing

0.24

0.22

Tail Mass, kg

0.2

0.18

0.16

0.14

0.12

0.1 0.4

0.6

0.8

1

1.2 Tail Arm Length, m

1.4

1.6

1.8

Figure 14: Tail Arm length vs Tail Mass

For this competition, tail design was a very special and unique case. No rudder meant it is required to make sure that the vertical tail is big enough for sufficient directional stability without any active controls. Also, since the CG is quite far behind the wing’s AC, it meant that the horizontal tail is going to have to be big enough to move the neutral point of the aircraft further aft. An all moving tail for more control authority and a symmetrical NACA 0010 airfoil for optimised trim drag for all the missions were selected. Both surfaces were designed with low aspect ratios to reduce

U NIVERSITY OF G LASGOW

Page 22 of 54

the risk of tail stall.8 To keep the volume coefficients as required, the tail can either be closer to the wing and bigger, or further away, but smaller. Knowing the densities of the tail boom and the foam that is going to be used for the surfaces, an optimisation was done for minimum weight. The resulting graph of tail mass with respect to tail arm can be seen in Figure 14. The tail size was then refined using XFLR5 simulations, which showed a smaller static margin than the optimisation code did. This most likely was the result of the optimization code not including the pitch up moments coming from the very high camber wing airfoil. Final tail characteristics can be found in Table 11.

Selected inputs: Horizontal tail volume coefficient. Most influential parameter for the optimisation.

Vertical tail volume coefficient. Most influential parameter for the optimisation.

Horizontal tail aspect ratio Vertical tail aspect ratio Horizontal tail taper ratio Vertical tail taper ratio CG location for mission 2, m Optimized tail characteristics: Neutral point location, m Static margins for mission 2, %MAC Boom length, m Horizontal tail area, , m2 Horizontal tail span, m Horizontal tail root chord, m Horizontal tail tip chord, m Vertical tail area, , m2 Vertical tail span, m Vertical tail root chord, m Vertical tail tip chord, m

1.09 0.1 5 3.2 0.4 0.5 0.301 0.315 3.8 1.444 0.09 0.672 0.192 0.077 0.07 0.323 0.269 0.135

Table 11: Tail Characteristics

4.5.3

Dynamic stability

The dynamic stability modes were estimated using XFLR5 software while simulating wing and tail only and can be found in Table 12. Without including the fuselage, the simulation resulted in level 1 performance for all modes except the spiral one. Level 2 spiral mode, even dropping lower than level 3 for Mission 2, flying at 10m s−1 , was the biggest predicted handling problem. It was discussed to be a result of a high vertical CG location in relation to the wing. Three solutions were suggested. First was to flight test the airplane with different vertical tail size and dihedral combinations to find the most flyable one. Second one was to redesign the wing box so that the wing AC is above the mission 2 CG. Third one was to introduce a stabilization system sacrificing the mass as a tradeoff.

8 S.

Gudmundsson. General Aviation Aircraft Design: Applied Methods and Procedures. Elsevier Science, 2013.

U NIVERSITY OF G LASGOW

Page 23 of 54

Figure 15: Root Locus plot for the aircraft

Mass, kg Static margin, %MAC CG locations, x, m CG locations, z, m Phugoid damping ratio Phugoid frequency, rad s−1 Phugoid period, s m Short Period damping ratio Short Period frequency, rad s−1 Spiral mode time to double, s Roll mode time to double, s Dutch Roll damping Dutch Roll frequency, rad s−1 Dutch Roll f · d product

Mission 2, 10m s−1

Mission 2, 20m s−1

3.658 5.95 0.301 0.036 0.326 0.516 12.182 1 3.997 1.076 0.126 0.69 21.233 14.66

3.658 5.95 0.301 0.036 0.049 0.325 19.358 1 10.97 6.016 0.036 0.813 31.05 25.25

Table 12: Stability Analysis Results

4.6

Propulsion

In order to ensure a highly optimized propulsion system a multi-step approach is taken as described below.

4.6.1

Methodology

Propulsion selection factors are outlined as follows:

U NIVERSITY OF G LASGOW

Page 24 of 54

1. Motor and propeller pair is able to provide sufficient thrust for preliminary design specified lift-off and cruise flight requirements.

2. Propulsion system is able to provide optimal performance for the mission bearing the most points. Consequently through propeller changes and ESC programming the motor selected will be adapted for best performance in secondary missions.

3. Propeller has to be within ground clearance limitations.

4. Propulsion system can achieve a speed of 23 m s−1 taking into account the overall frame drag generation. This requirement is imposed in order for the aircraft to be able to withstand the high wind profile in Tucson, AZ. This imposes further limitation on propeller size, pitch and motor RPM combinations.

5. The weight of the propulsion system and battery pack combined is to be brought to a minimum, ensuring optimal efficiency of the propulsive system.

Drawing on the above selection the following motor and propeller profiles were established: Propulsion system profile: A low Kv range, 350 Kv to 900 Kv, brushless outrunner type motors were selected due to their: 1. Superior efficiency (a) Lower Kv motors allow higher voltages for a given RPM and power output level. Consequently allowing lower amps for a given input power, and finally lower amperage permit for smaller diameter wiring and physically smaller components at a given efficiency, minimizing waste heat. Furthermore, this allows reduced material costs and weight. Therefore voltage intensive, low kV motors have been found to be more efficient than high kV ones. However, additional tests are required to ensure high grade motor quality and consequently match it to the specific mission requirements for optimal results. It should also be mentioned that the higher voltage requirement impacts negatively on battery cell structure, weight wise, hence a larger range of kV options is used to allow finding the best tradeoff between weight and performance.

2. Allows direct drive - does not require a gearbox and therefore avoids its negative impact on performance.

3. High torque - allows for larger and more efficient propellers, however further testing is to be performed for an accurate estimate of torque roll performance impact produced by the lower range kV motors selected Motor Power Range of 315 W-600 W:

Optimization code predicted a minimum required power of 315 W for the

desired performance, this was consequently compared to theoretical results from optimization software and statistical data from similar motors in order to confirm its consistency. A range of motor maximum output power is presented, adding another dimension to optimization testing , namely optimal efficiency at minimum required thrust complying with the requirements set. The spectrum limits are dictated by brushless motors optimal efficiency loading, usually between 40% and 70% of the maximum load, and the minimal difference in weights between the selected models.

Range of Carbon fibre propellers with main dimensions range 10 00 , to 15 00 , for propeller diameter and 3.7 00 , to 10 00 , for propeller pitch: 1. Chosen to accommodate both speed focused and endurance focused missions. Appropriate propellers are to be paired with motors that have suitable RPM and torque parameters in order to satisfy speed and thrust requirements.

2. Lighter Material - Carbon fiber propellers are significantly lighter than all other commercially available alternatives, providing both weight minimization and lowering torque requirements compared to other propellers.

3. Endurance - Carbon fibre provides excellent elasticity and is therefore significantly more resistant to impacts, allowing to improve time efficiency in flight testing phases.

4. Well balanced - Good quality carbon propellers require very little to none weight balancing, which eliminates the risk of impacting propeller aerodynamic characteristics during optimization.

U NIVERSITY OF G LASGOW

Page 25 of 54

Using the profile generated, various pieces of simulation software including OpenProp, eCalc and MotoCalc, and statistical research on the top contenders of the last four years of DBF, an extensive market research has been done in order to select a range of best suited motors , which is consequently narrowed down to three choices presented in Table 13.

Motor

kV

Max Continuous Power, W

Motor weight, g

Max Continuous Current, A

ESC

T-MOTOR MN4010 Turnigy Multistar 3508

370 640

450 550

112 98

20 30

Turnigy Multistar 4010

485

500

128

26

Afro HV 20A ESC Hobbywing Platinum 40A-OPTO Pro ESC Hobbywing Platinum 30A-OPTO Pro ESC

Table 13: Table of Motors and ESC considered

Electronic Speed Controller

An ESC appropriate for each motor has been chosen with three main parameters in

mind: 1. Is able to support the maximum amperage and voltage that the motors are expected to reach without overheating, therefore an extra ten to twenty percent of current capacity have been added to each one. It should be noted that the afro ESC is selected for voltage intensive tests and is therefore not expected to go over 15 A current in any circumstances.

2. Is programmable in order to accommodate optimization for both battery type and response times.

3. Is as light as possible while complying with first two points.

Fuse selections: A range of fuses will be tested for each motor in order to confirm the characteristics given and ensure safety of the propulsion system. The ESC candidates are shown in Table 13.

4.7 4.7.1

Electronics Battery Selection Factors:

1. The battery must supply adequate power to the propulsion system over the full mission length 2. Meet the takeoff maximum power requirement of 315 W 3. Must have the correct voltage to rotate the propeller at a sufficient angular velocity 4. Mass must be under the 2.0 lb (0.9 kg). maximum 5. As lightweight as possible 6. Battery cell chemistry must be either Nickel Metal hydride (NiMH) or Nickel Cadmium (NiCd) NiMH cell chemistry offers several significant advantages over NiCad mainly a higher maximum continuous discharge rate (10 C) and a larger energy density giving a lower mass for the same energy content.

All performance NiMH cells claim a nominal voltage of 1.2 V and a continuous C rating of 10. As, t = 1/C, this means the minimum flight time will always be 6 minutes, even if the aircraft is on full power for the entire flight. Most NiMH are also capable of supplying a higher ’burst’ current, usually up to 15 C. Selected cells for testing are shown in Table 14.

Using a battery with a higher specific energy means the mass of the battery can be reduced while delivering the same performance.

U NIVERSITY OF G LASGOW

Page 26 of 54

Battery Type/Manufacturer

C rating

Capacity, mA h

Continuous Burst CurCurrent, A rent, A

Specific Energy, W h kg−1

Overlander AA high cap pip Turnigy 2/3A High Power KAN 700 AAA Overlander SubC Sport

10 10 10 10

2000 1500 700 2000

20 15 7 20

88.9 78.3 60.4 61.5

30 22.5 10.5 30

Table 14: Electronics cell selection

For mission 2 the takeoff power is 315 W, the battery specifications can be calculated.

P = IV where I is current, V potential difference and P is power.If the maximum current draw is 30 A, then a voltage of 10.5 V is required, however most battery packs will drop or ’sag’ under load, sometimes as much as 70%. To account for this the nominal pack voltage should be around 10.5 V/0.7 = 15 V. This is equivalent to a 12 to 14 cell pack, where each cell has a nominal voltage of 1.2 V. The lower, 1500 mA h battery would need 20 V, or 16-18 cells.

Mission 1 and 3 follow a similar process and will likely use the same battery and propulsion settings.

4.8

Radio System

Transmitter and Receiver requirements: • At least 4 channel control

• Basic timers

• Failsafe on all channels

• Model memory saves

• Full range

• Dual rates and exponential

• 2.4 GHz or 72 MHz band

• Trim conditions

• Telemetry

• Nickel based batteries used to power both

The FrSky Taranis was selected to meet these requirements as well as being considerably cheaper than competitors.

The X6R will be used, a 6 channel receiver, to allow telemetry usage during testing and to minimise on board weight compared to a 8 or 9 channel receiver.

The channels on the receiver are arranged as follows; 1. Ailerons 2. Elevator 3. Throttle 4. Ball release activation Receiver Battery

The receiver battery must supply sufficient current to give the servos full torque as well as supply

the receiver with a steady voltage. A 4 cell (4.8 V) 1000 mA h Sanyo eneloop NiMH pack was chosen due to its low mass and ability to supply the required power to the receiver and servos for even more than one flight. Testing was to be done on the increase in servo torque and speed from using a 6 V receiver battery, but the 6 V pack may damage other components and was not considered any further.

U NIVERSITY OF G LASGOW

Page 27 of 54

Servos Because the number of servos is a critical component in the aircraft cost factor NSERVO , it is obvious that multiple tasks be carried out by a single servo and to discard any unnecessary servos.

This aircraft requires three servos. A servo to control the elevator providing pitch control, and a larger servo to control both the ailerons and steer the nose wheel. The third will command the ball dropping mechanism for mission 3.

The servos supply the torque and movement to drive the control surfaces. Estimates on the required size, and hence torque and speed, of the servo was taken from experience and the relatively slow speed the aircraft will reach compared to similar sized models.

Obviously a considerably large servo is required if it is to perform multiple functions such as both ailerons and the nose wheel steering. But smaller servos are used for elevator control and release mechanism. A selection of three servos was purchased to fit the required role and are shown in Table 15.

Servo

Mass, g

Torque, kg cm−1

Speed (0-60◦ ), , s

BMS-706 TGY-211DMH HKSCM12-5

26 16 12

4.6 1.9 1.5

0.13 0.12 0.18

Table 15: Servo selection

Problems with single servo ailerons

Normally, a larger and complex model aircraft would use two servos to drive

each aileron individually. This allows for effects such as differential which can help correct adverse yaw during rolls and add the ability to use the ailerons as flaps, a mix known as flaperons. While flaperons are impossible with a single servo aileron system, differential can be added to the ailerons using off 90◦ control horns on the surfaces, meaning the ailerons will deflect more up than down.

Servo Linkages

Pull-pull wires on nose gear are always in tension, meaning there is never any slack in the system.

This can help prevent the aircraft from pulling to one side and help the machine track straight.

Ailerons torque rods are used inside the wing to enable the single servo mounted in the center of the aircraft to control the ailerons near the wing tips.

The elevator to servo connection is done using a single push/pull rod, mounted on the upper surface of the elevator.

Mounting the servo horn on the top side means the rod is in tension when the elevator is in the ’up’ position, reducing slack in the system and allowing for greater control.

Stabilisation The rules allow for some stabilisation so long as there is no automatic element like auto-level or any automatic navigation.

This rules out auto level systems (eg accelerometers), and allows for gyros only which hold orientation. It is thought a gyro placed on the roll axis could prove valuable to correcting the poor spiral mode characteristics. Further testing is definitely required to calibrate the gyro gain and tune the response to prevent flutter yet still augment the flight stability.

U NIVERSITY OF G LASGOW

Page 28 of 54

5 5.1

Detail Design Aircraft Dimensions and Components Wing Aerofoil Span MAC Wing Area (S) Aspect Ratio Incidence Angle Static Margin for mission 2

Wortmann FX63-137 1.85 0.24 0.43 8 -3.7 3.8

Fuselage Length Width Height

0.996 0.16 0.18 Tail Surfaces

Aerofoil Span MAC Wing Area (S) Aspect Ratio Incidence Angle Tail Arm

Horizontal Vertical NACA 0010 0.672 0.323 0.143 0.209 0.09 0.07 5 3.2 0 0 1.252 1.242

Table 16: Final Design Dimensions and Aerodynamic Parameters

Controls ESC Receiver Servos

Afro HV 20A ESC FrSky F6R BMS-706 (Aileron and Gear Steering) TGY-211DMH (Elevator) HKSCM12-5 (Sensor Drop)�

Propulsion

Motor Mission 1 Mission 2 Mission 3

T-MOTOR MN4010 14x10 15x8 15x8

Table 17: Aircraft Electronics and Propulsion components

U NIVERSITY OF G LASGOW

Page 29 of 54

Aircraft Component

Wings Fuselage Propeller Tail Landing Gear Ball Drop Mechanism Motor Front Electronics (Receiver, aileron servo and low voltage alarm) Ball Cage Mission 1: Empty configuration Battery, position 1 Static margin = 11.5% Mission 2: Sensor Package Transport Sensor Package Battery, position 2 Static margin = 4.68% Mission 3: Sensor Drop 6 x Payload Balls Battery, position 1 Static margin = 7.65%

Weight

CG Location

(g) 245 342 55 137 99 29 120 52

X (mm) 47 0 0 -1 0 2 0 14

Y (mm) -12 19 20 28 -147 -11 20 15

Z (mm) -224 -298 37 -1192 -327 -523 10 -157

26 1585 470

0 8 0

-99 7 41

-409 -288 -154

3853

3

41

-304

2268 470

0 0

64 41

-333 -71

1993 408 470

6 0 0

-13 90 41

-297 -336 -154

Table 18: Balance Table

5.2

Aerodynamic analysis

After having the final CAD model freezed, further aerodynamic analysis could be done. STAR-CCM+ and XFLR5 were used for that.

For all the CFD tests, a �bullet-shaped� domain extending 8 wing spans (or body lengths) from the actual geometry was created to simulate half-body with a symmetry plane. Trimmed prism layer mesher was used, resulting in 3-4 million cell volume mesh. The mesh had to be refined around all the edges, wing-tip vortices, airfoil leading edges and smaller details like the ball cage strips.

First, the wing lift predictions were compared for all the available sources. Wind tunnel test results of a 2D airfoil,9 2D XFLR5 test results, 3D finite wing XFLR5 and 3D STAR-CCM+ simulation results were compared and plotted in Figure 16. 3D simulation had a smaller lift curve slope than the 2D, as expected. Also the lift coefficient predicted with STAR-CCM+ reached a value of 1.66, just a bit higher than what was used in the preliminary design. This confirmed estimations from the preliminary phase to be very close to the actual values.

9 Michael

S. Selig and Bryan D. McGranahan. In:

U NIVERSITY OF G LASGOW

Page 30 of 54

Figure 16: Comparison of finite wing performance using different engineering tools

Fuselage, tail, main landing gear and ball cage combination was simulated separately at 0◦ angle of attack over velocities range of 5 m s−1 - 35 m s−1 . The CFD model for the configuration at reference speed of 14 m s−1 calculated the parasitic drag coefficient to be 0.0239. Interestingly, the value predicted in the preliminary design stage was just 2% different, 0.2364 without the wing. This proved the validity of using empirical drag estimation methods for preliminary design stage. The refined mesh and the streamlines can be seen in Figure 17.

Figure 17: Refined mesh and streamlines visualized

The drag coefficients of configuration without the wing and the wing separately were added together to get a complete U NIVERSITY OF G LASGOW

Page 31 of 54

drag polar. Interference drag between the wing and the fuselage was omitted, since, according to Hoerner,10 drag increase due to wing placement at the bottom of the fuselage and fuselage incidence angle cancel each other out.

Dynamic thrust estimation using an online tool11 for 14x10 propeller was done and is plotted in Figure 18 together with the drag polars.

As can be seen, the actual drag at low speeds for mission 2 is higher than for mission 1. This arises because of the lift induced drag at higher lift coefficients. For higher speeds the trim drag for mission 1 increases and becomes higher than mission 2 drag. The preliminary drag polar was higher than the actual one, probably being a result of wrongly estimated planform efficiency. The maximum speeds for both, CFD and preliminary drag models were smaller than expected for the selected propeller. Achievable speed of 17.2 m s−1 for mission 1 and 14.4 m s−1 for mission 2 do not cover the design target of 23 m s−1 . To cope with the winds, a more powerful motor and battery might have to be considered after proving the drag models with flight tests. It also has to be noted that the online tool for dynamic thrust estimation is said to underestimate the propeller performance, hence the actual maximum speeds will be higher.

Figure 18: Drag vs Thrust

Wake structures behind the model were analysed as shown in Figure 19. Filtering the mesh by vorticity, high turbulence zones could be identified. The strongest zones appeared behind the first ball cage rib, fuselage end and landing gear wheel. Weaker zones could be found behind second and last cage ribs, landing gear mount and intersection between the fuselage and the landing gear. The most problematic spot, the end of the fuselage is going to be blended with the main boom for the final model. The cage ribs are going to be sanded and streamlined as long as the structural strength is conserved and landing gear fairing might be considered if drag reduction will compensate for the mass addition. One more ”red” zone can be seen at the joint of the main fuselage compartment and the front fairing. Separation occurs there and sanding the frontal fairing to a more rounded shape is suggested to decrease 10 Hoerner,

Fluid-dynamic drag: practical information on aerodynamic drag and hydrodynamic resistance.

Staples. Article. Apr. 2014. URL: http://electricrcaircraftguy.blogspot.co.uk/2013/09/propeller- static- dynamicthrust-equation.html.

11 Gabriel

U NIVERSITY OF G LASGOW

Page 32 of 54

the adverse pressure gradient. It is to be noted that the propeller wake effects were not included in the simulation because of the complexity of the model required.

Figure 19: Wake structure behind the fuselage, landing gear, ball cage and the tail

5.3

Structural characteristics

The structure sub team’s design goals were to minimize the structural weight of the aircraft, ensuring that the structure will provide adequate strength and rigidity for high wing loading during turns.

5.3.1

Load Paths

The main structural components of the aircraft are several carbon fiber tubes. A single 12 mm outer diameter tube runs along the span of each of the wing pieces at 25% chord. To ensure torsional rigidity of the wing, a 4 mm outer diameter tube runs along the span of each of the wing pieces at 75% chord. A 12 mm outer diameter carbon fiber tube runs through the foam nose cone, the carbon fiber bulkheads and the foam aft part of fuselage. Those carbon fiber tubes and the connections between them form the main load paths. The load paths are presented in Figure 20.

U NIVERSITY OF G LASGOW

Page 33 of 54

Figure 20: Load Paths

5.3.2

Structural Analysis

The main structural components of the aircraft were analyzed. An assumption was made that the carbon fiber tubes along the wing pieces and the fuselage carry all of the structural loading. The empty aircraft undergoes a load factor of five during turns. The loading applied to the 25% chord spars is shown in Figure 21. An assumption was made that the 25% chord spar of the wing features distributed loading due to the lift and point loading due to the weight of the aircraft.

Figure 21: 25% chord spar loading

Based on the loads, shear and moment diagrams were created. This analysis does not account for spar deflection, which may cause other structural components to fail well before the spar fails. A diagram is shown in Figure 22.

U NIVERSITY OF G LASGOW

Page 34 of 54

Figure 22: Loading, shear and moment for wing 25% chord spar

The main load carrying structure of the fuselage is a carbon boom. As with the rest of the components, the intention was to identify the lightest, yet sufficiently stiff and strong boom. The two available suitable options were 12 mm pultruded and rolled carbon fibre tubes. Maximum load during 5.6G turn has been modelled using ABAQUS FEA. Since the supplier provided a range of values for tensile moduli, it was decided to use a safety factor of 1.5 rather than 1.2, that was used during the actual testing, in order to make sure that the selected boom satisfies the performance requirements. Hence 8.4G loading was used for modelling. The pultruded boom was predicted to deflect significantly, and tensile modulus exceeds the maximum strength of the material, which immediately ruled it out as an option.

Rolled boom, however, while being about 10% heavier, would experience a much smaller deflection ,as can be seen in Figure 23. The stress is also within maximum tensile strength. It was thus decided to acquire a rolled tube and conduct live tests in similar conditions.

U NIVERSITY OF G LASGOW

Page 35 of 54

Figure 23: Main boom deflection under 8.4G MTOW load factor

5.4 5.4.1

Aircraft Systems Design, Component Selection and Integration Fuselage

The fuselage was designed around the internal payload bay. It had to have a small frontal area and to be as light as possible. It consists of a carbon fiber tube of 12mm outer diameter running along the entire length of the fuselage, carbon fiber motor mount, carbon fiber bulkheads and flat carbon fiber spars connecting the bulkheads. A carbon fiber plate secured between the motor mount and the first bulkhead is used as a mounting plate for electrical components.

The plate is also secured to the fuselage spar. As an effort to minimise weight the nose cone and the empennage were not made by placing additional bulkheads and spars but were made entirely from foam. The foam empennage and nose cone are lighter but still contribute to the structural rigidity of the fuselage. The nose cone is secured to the motor mount, the first fuselage bulkhead, the mounting plate for electrical components and the fuselage spar. The empennage is secured to the third fuselage bulkhead and the fuselage spar. A 12 mm ID spar which is perpendicular to the fuselage spar is located between the first and the second fuselage bulkheads. The 25% chord spars from each wing piece slide into this spar. The remaining space between the bulkheads above the spar is filled with foam.

Another carbon fiber plate between the second and third bulkhead is used for mounting payloads. The third bulkhead features the mount for the main landing gear. Renderings of the fuselage are shown in Figure 24.

U NIVERSITY OF G LASGOW

Page 36 of 54

Figure 24: Fuselage

5.4.2

Internal payload bay

The internal payload bay is situated between the fuselage second and third bulkhead. A rectangular carbon fiber plate is placed between the bulkheads and onto the fuselage spar. The plate serves as a mounting point for the payload.

The payload bay is accessible by a foam cover. Two carbon fiber plates are glued to the foam cover that are inserted into openings in the second bulkhead. A simple locking mechanism at the back of the foam cover prevents movement of the cover during flight. It is doing this by interacting with the third bulkhead. This mechanism is comprised of a bent rod which latches onto an opening in the third bulkhead. Internal payload bay is depicted in Figure 25. The payload and battery are secured with velcro straps.

Figure 25: Internal payload bay

5.4.3

External ball cage

The external ball cage is situated below the fuselage and consists of three flat carbon fiber brackets and two flat carbon fiber spars as seen in Figure 26. The brackets are designed in such a way that the external payload bay can be detached from the fuselage. The payload bay is specifically designed to carry a payload in the form of balls which can be dropped with the help of the ball drop system.

U NIVERSITY OF G LASGOW

Page 37 of 54

Figure 26: External Payload Bay

5.4.4

Wing

The wing is made up of two pieces. Each piece consists of carbon fiber ribs, two carbon fiber spars and a foam leading edge, trailing edge and aileron. A 12 mm outer diameter spar is present at 25% chord and a 4mm outer diameter spar at 75% chord. A 10 mm external diameter spar of 300 mm length is inserted 150 mm into one of the 25% chord spars of the wing pieces. The 25% chord spar of each wing piece is inserted in the wing attachment tube. The wing can be seen in Figure 27.

Figure 27: Wing

5.4.5

Wing attachment

The wing attachment consists of a carbon fiber tube featuring 12 mm internal diameter placed below the fuselage spar and between the first and second fuselage bulkheads as depicted in Figure 28. The wing attachment spar features a small hole. As mentioned before a carbon fiber tube of 10 mm external diameter and length of 300 mmis inserted 150 mm into the 25% chord spar of one of the wing pieces. This spar is secured to the 25% chord spar and features a small hole. The 25% chord of the other wing piece also features a small hole. When the two 25% chord spars of each wing piece are inserted into the wing attachment a three spar connection is formed. With the holes aligned a pin is placed, which secures the two wing pieces. With this pin the wing can easily be attached or detached from the fuselage.

U NIVERSITY OF G LASGOW

Page 38 of 54

Figure 28: Wing attachment

5.4.6

Tail

The horizontal and vertical tail are made out of foam, and they feature a carbon fiber spar of 10 mm outer diameter which makes them more stiff and resistant to damage. Illustration is provided in Figure 29.

Figure 29: Tail

5.4.7

Tail attachment

The tail attachment consists of two carbon fiber tubes with 12 mm outer diameter and carbon fiber rod with 10 mm outer diameter. The horizontal carbon fiber tube of 10 mm outer diameter is perpendicular to the carbon fiber rod. The vertical carbon fiber tube of 12 mm outer diameter is perpendicular to both the horizontal carbon fiber tube and the carbon fiber rod. The tail spar of 10 mm outer diameter passes through the horizontal carbon fiber tube of 12 mm outer diameter. This allows the horizontal tail to pivot about its spar acting as a stabilator. The vertical tail spar of 10 mm outer diameter is attached to the vertical carbon fiber tube of 12 mm outer diameter. Tail attachment is rendered in Figure 30.

U NIVERSITY OF G LASGOW

Page 39 of 54

Figure 30: Tail Attachment

5.4.8

Fuselage aft section

This component provides the aft aerodynamic profile of the fuselage. It is designed to be lightweight. Instead of making the shape of the empennage with carbon fiber bulkheads, foam was used. The foam empennage features lightening holes and the fuselage spar runs through it. The empennage is secured to the fuselage spar and the third fuselage bulkhead. The servo for the horizontal tail is located at the very end of the foam empennage. The servo is attached to mounts made out out of plastic which are attached to the fuselage boom. This part also houses the sensor drop mechanism. This part of the fuselage is shown in Figure 31.

Figure 31: Aft part of the fuselage

5.4.9

Nose cone

The nose cone as presented in Figure 32 provides the forward aerodynamic profile of the fuselage. It is designed to be lightweight, so it features lightening holes just like the empennage. The nose cone is secured to the fuselage spar, the motor mount, the first fuselage bulkhead and the carbon fiber plate used for mounting electrical components. The top aft section of the nose cone is removable offering quick access to the electrical components located inside.

U NIVERSITY OF G LASGOW

Page 40 of 54

Figure 32: Nose Cone

5.5 5.5.1

Payloads Systems Design, Component Selection and Integration Ball drop system

The ball drop system is made of one carbon fiber tube of 14 mm external diameter and two carbon fiber plates (paddles) of 4 mm thickness attached to it. There is an angle between the paddles. The carbon fiber tube features a control horn which is connected via a push-pull rod to the control horn of the servo driving the ball drop system.

As the control horn of the servo moves left to right the carbon fiber tube rotates and so do the the carbon paddles attached to the tube. As the paddles rotate the aft paddle gets out of the way of the last ball in the external payload bay and the forward paddle gets in the way of the next ball. This prevents additional balls from falling. After the last ball falls the paddles are rotated which allows the next ball to move in place. Airflow pushes the balls backwards. Ball drop system is depicted in Figure 33.

Figure 33: Ball Drop system

U NIVERSITY OF G LASGOW

Page 41 of 54

SolidWorks Student Edition.

For Academic Use Only.

166.67 6.56 241.68 9.52

380 14.96

566 22.28

SCALE 1:14 1848 72.76

339 13.35

709 27.91

999 39.33 199.09 7.84

89.88 3.54

320 12.60

180 7.09

374 14.72

270 10.63

429 16.89 575 22.64

172.92 6.81

287.26 11.31

1532 60.31

Cessna/Raytheon Missile Systems Student Design/Build/Fly competition TITLE:

AIRCRAFT 3-VIEW DRAWING THE REST OF THE DISTANCES FROM PROFILE TO PROFILE UNTIL THE END OF THE WING Is 120mm

SIZE

120 116 4.72 4.57

160 6.30

B

UNLESS OTHERWISE SPECIFIED DIMENSIONS ARE GIVEN IN MM AND THEN IN INCHES

SCALE: 1:10 (Unless specified)

Drawing number №1

SHEET 1 OF 4

19

23 8

2

17

7

4 18 6 9

5 16

3

1

Quantity

1

Main Spar

1

2

Fuselage

1

3

Crafting Foam

1

Crafting Foam Carbon Fibre

2 2

Carbon Fibre

2

7 8 9

Fuselage Forward Firing Wing Foam Main Wing Spar WingSpar at 75% Chord Aileron Torque Rod Wing Rib Propeller

Carbon Fibre Foam and Carbon Fibre

2 16 1

10

Propeller Spinner

11

Carbon Fibre

1

12 13

Electrical Mounting Plate Nose Wheel Main Wheel

Steel Carbon Fibre Carbon Fibre 2014 Aluminium Alloy

1 2

14

Front Landing Gear

Plastic and Rubber Plastic and Rubber Steel

15

Main Landing Gear

Carbon Fibre

1

16

Payload Holder and Release Mechanism

Carbon Fibre

1

Steel

1

Crafting Foam Carbon Fibre

1 1

Plastic

1

21

Fuselage Locking Mechanism for Main Payload Fuselage Aft Firing Tail Horizontal Spar Main Spar to Tail Connector Vertical Tail Spar

Carbon Fibre

1

22

Vertical Stabiliser

Crafting Foam

1

23

Horizontal Stabiliser Profile Horizontal Stabiliser

Carbon Fibre

2

Crafting Foam

2

6

20

24

Material

4 5

22

21

Part

17 18 19 20

10

11

24

14

15 12

SolidWorks Student Edition.

For Academic Use Only.

13

1

1

Cessna/Raytheon Missile Systems Student Design/Build/Fly competition

TITLE:

AIRCRAFT Structural Arrangement SIZE

B

UNLESS OTHERWISE SPECIFIED DIMENSIONS ARE GIVEN IN MM AND THEN IN INCHES

SCALE: 1:8 (Unless specified)

Drawing number â„–2

SHEET 2 OF 4

257 10.12 Payload 1

G E

NOTE: Top Foam Hatch has been ommited for clarity

K Payload 2 housing

Release paddle 1

Payload 2

Release paddle 2

Payload 2

Fuselage Locking Mechanism for Main Payload

Cessna/Raytheon Missile Systems Student Design/Build/Fly competition

DETAIL E SCALE 1 : 2

TITLE:

Location of motor battery for mission 1 and 3

SolidWorks Student Edition.

For Academic Use Only.

DETAIL K(Payload 1 Housing) SCALE 1 : 6

SIZE

DETAIL G SCALE 1 : 1

B

Payload Layout UNLESS OTHERWISE SPECIFIED DIMENSIONS ARE GIVEN IN MM AND THEN IN INCHES

SCALE: 1:10 (Unless specified)

Drawing number â„–3

SHEET 3 OF 4

SCALE 1:15 TOP VIEW

SolidWorks Student Edition.

For Academic Use Only.

DETAIL E SCALE 1 : 2

G

LOW VOLTAGE ALARM

E RELEASE MECHANISM SERVO

RECEIVER

FRONT LANDING GEAR & AILERONS SERVO

SERVOS BATTERY PACK

ELEVATOR SERVO MOTOR

MOTOR BATTERY

H SERVO MOUNT FRONT LANDING GEAR

MAIN LANDING GEAR

DETAIL G SCALE 1 : 1

PROPELLER PROP SPINNER

Cessna/Raytheon Missile Systems Student Design/Build/Fly competition TITLE:

ELEVATOR SERVO

SYSTEMS LAYOUT SIZE

B

DETAIL H SCALE 1 : 1

RELEASE MECHANISM SERVO

UNLESS OTHERWISE SPECIFIED DIMENSIONS ARE GIVEN IN MM AND THEN IN INCHES

SCALE: 1:10 (unless specified)

Drawing number №4

SHEET 4 OF 4

6 6.1

Manufacturing Plan and Processes Manufacturing Process Selection

The manufacturing process selected for fabrication plays an important role in aircraft weight and overall flight score.

There are a lot of manufacturing methods for radio controlled aircraft. The most common are outlined here.

Balsa build-up -the most widely used method, parts from balsa are easy to manufacture by hand, semi-complex shapes can be made with thin sheets of balsa.

Foam - often heavier than a balsa build up, foam parts can be manufactured very easily with a hotwire machine.

Foam parts are excellent for making complex shapes such as fuselage lofted sections or wing leading or trailing edges.

Carbon fiber - very rigid, carbon fiber parts can be manufactured relatively easily out of pre cured carbon fiber sheets using a CNC machine. Other carbon fiber products such as tubes and rods are readily available.

6.2

Subsystems manufacturing

In the following sections, the manufacturing processes used for the wing, fuselage, empennage, nose cone and other components are documented.

6.2.1

Wing manufacturing

The wing is composed of CNC cut carbon ribs, two carbon fiber spars, and foam. The ribs are aligned along the spars at their respective locations. Epoxy is used to glue the carbon fiber ribs to the carbon fiber spars. Foam blocks are inserted at the leading edge between each pair of ribs. A manual hotwire cutting tool is then used to shape the foam blocks. The carbon fiber ribs are used as a hotwire cutting templates. The same is done for the trailing edge.

The foam provides surface to which the covering film is attached, as well as extra rigidity at low mass penalty.

6.2.2

Fuselage manufacturing

The fuselage is composed of CNC cut carbon fiber parts, carbon fiber boom and foam. CNC-cut carbon fiber bulkheads are aligned along the main boom at their respective locations. Epoxy is used to glue the carbon fiber parts.

The wing attachment carbon fiber spar is placed between the first two bulkheads below the fuselage boom. Payload and electrical carbon fiber mounting plates are placed directly on the main boom. The aft fuselage section and nose cone are aligned with the bulkheads and glued to them and the fuselage boom. Foam provides surface to which the film is attached, as well as stiffens the structure.

6.2.3

Aft fuselage section manufacturing

Profiles which form the shape of the empennage are cut from thick cardboard which are then used as hotwire cutting templates. A manual hotwire cutting tool is then used to cut the lofted shape between the two profiles. The lofted cut foam parts are then aligned along the fuselage boom and glued with epoxy to it and to the bulkheads.

6.2.4

Nose cone manufacturing

Same technique when manufacturing the aft section is used to manufacture the nose cone. A section of the nose cone is cut with a manual hotwire tool which acts as an access hatch to the electrical equipment. After all fuselage sections are combined, they are covered with film.

U NIVERSITY OF G LASGOW

Page 46 of 54

6.3

Landing gear manufacturing

The landing gear shape is cut from thick cardboard which is then used as a profile for hotwire cutting. A manual hotwire cutting tool is used to cut the landing gear mold from foam. Carbon cloth is layered on mold multiple times as shown in Figure 34 and covered with epoxy, which is then left to dry for at least 24 hours. It is then sanded down to remove imperfections and acquire the desired shape and size.

Figure 34: Landing Gear manufacturing

6.4

Schedule January Activity

19

26

February 02

09

16

March 23

02

09

16

April 23

30

Prototype 1 Carbon fibre & Foam cutting Wing assembly Main Section assembly Landing Gear assembly Insert Motor, Servos & Wiring Prototype 2 Carbon fibre & Foam cutting Wing assembly Main Section assembly Landing Gear assembly Insert Motor, Servos & Wiring Final Design Carbon fibre & Foam cutting Wing assembly Main Section assembly Landing Gear assembly Insert Motor, Servos & Wiring

06

Legend Planned Actual Due date Current

/

Table 19: Manufacturing Schedule for the project

Every effort has been made to follow the manufacturing plan, however for the reasons described above there was only a limited amount the team could achieve. This, however, was sufficient to prove the ability to manufacture all components required and confirm that further plans are possible to meet.

U NIVERSITY OF G LASGOW

Page 47 of 54

7

Testing Plan

Components testing is an essential part of aircraft design. It allows the verification of predicted performance and provides data on actual performance. Based on tests performance can either be increased or decreased to optimize the design according to project specification.

7.1

Testing Objectives

The objectives of aircraft testing were to verify that the design can meet all the rules and requirements of the competition and at the same time accumulate the highest possible competition score. The general goals were as follows • Determine the most suitable propeller and motor combination – Carried out on a test rig to identify propeller-motor combinations that provide greatest static and dynamic thrusts • Evaluate individual cell and complete battery performance – Maximal discharge rate and actual performance under load such as voltage drop and capacity reduction from the rated value. Verify sufficient cooling in flight by stress testing at worst expected • Ensure sufficient structural integrity in all expected flight conditions – Wing and main boom bending test at simulated 6G load and landing gear at 3G • Test handling characteristics and landing and takeoff performance – Test flight schedule at different payloads and flight conditions will be carried out to verify and adjust sizing and configuration if required. Ground handling and stability needs to be verified as well.

• Ensure the reliability of the sensor drop mechanism in various flight situations

7.2

Propulsion Test

The next step in the selection process is to test the motors static thrust while utilizing different propellers in order to determine the best combination.This will be done using a carefully calibrated load cells and a power supply unit.

The voltage and power output will be kept constant in order to satisfy battery limitations. Optimal ESC and fuse selection for the given motor will be implemented in the corresponding test circuit, along with insulation tape, and bullet connectors for safety.

As stated above main selection will be done based on optimal characteristics for the mission bearing the largest amount of points , therefore a static thrust data table will be created for each appropriate motor and propeller set.

A secondary propeller will also be chosen for the selected motor with respect to other missions optimization.

The selected pairs will consequently be tested in a circuit with batteries instead of PSU, in order to confirm that the results are viable and apply fine tuning where needed .

Finally ESCs will be reprogrammed in order to achieve best results with the battery circuits.

Practical Testing Sequence:

After the lab testing sequence is complete the propulsion system will be used in a

series of flight tests to further confirm the selection and fine tune it where required.

7.3

Electronics Test

Battery testing will be carried out with the use of a custom built, computer controlled, constant current discharger that uses modular load boards consisting of regulators that provide a set amount of power to banks of resistors. Each load board carries a pair of thermistors for monitoring board temperatures to provide a layer of safety during extended

U NIVERSITY OF G LASGOW

Page 48 of 54

and strenuous testing, as well as 2 relays, each of which connects one load circuit to the battery, allowing for finer control of the load.

These ’load boards’ are connected to both the battery and control board, which detects the load boards currently connected, controls the relays to control the load as well as monitoring the load board temperatures. In addition, the control board also monitors cell voltage and temperature for logging. The system allows for battery testing to mirror real world flight profiles, as it is possible to upload a test profile to the control board processor which can vary the load during testing to simulate take-off power, turning, climb and landing loads on the battery.

7.4 7.4.1

Structural Testing Wing spar and fuselage boom testing

For the wing spar testing, the objective is to simulate a load during 5G turn with a 1.2 safety factor which results in 6G total load. The loads are at MTOW. The lift distribution will be represented as a point load acting at the mean aerodynamic chord. Water bottles will be used as a load. The deflection of the spar at the wing tip at different loads will be measured. Similar approach will be used with the fuselage boom. The boom will be suspended from two points, wing and tail aerodynamic centers, and load applied at aircraft center of gravity. Deflections will be measured at the maximum deflection point of the boom.

7.4.2

Landing gear testing

For the landing gear testing the objective is to simulate a 3G load representing impact during landing. The landing gear is secured in such way that lateral motion is restricted. The load is applied to the landing gear at the place where the landing gear is attached to the fuselage. The deflection of the landing gear is then measured.

7.4.3

Sensor Drop Test

Sensor drop mechanism will be tested as a standalone unit to investigate the operating limits of the system. The ball cage will be tilted at different angles from horizontal to model the force created by the air flow in a variety of airspeeds. Also the cage is to be tilted from vertical to test sufficient performance even at high bank angles that could be encountered.

7.4.4

Flight Test

Flight testing is to be conducted following a carefully planned schedule to slowly expand the flight envelope of the aircraft. Testing was divided into following parts 1. Taxi Testing • The aircraft will be taxied at gradually increasing velocities at different payloads and the aircraft behaviour will be observed. This is to be continued until 80% of the planned takeoff velocity is achieved. This will ensure tolerable ground handling qualities such as turning and stability.

2. Empty Flight Tests • The aircraft will be flown in ferry condition. First flight was planned to only include a small hop to at a low altitude to verify general flight behaviour and controllability. The flight envelope will be slowly expanded until maximum speed and load factor are reached.

3. Flight with variable payloads • After successful empty flight test programme the aircraft will be loaded with variable payload weights and CG positions to verify flight handling characteristics in different payload configurations before MTOW test can be carried out.

U NIVERSITY OF G LASGOW

Page 49 of 54

4. Maximum Payload Test • Aircraft to be loaded to MTOW to measure landing and takeoff distances as well as tolerance to maximum planned load factor in flight.

5. Sensor Drop Testing • At the last stage of the flight test programme the sensor drop mechanism will be tested and verified to work reliably in all expected drop conditions.

7.4.5

Schedule

November Activity

03

Propulsion test Component test Wing Spar test Fuselage Boom test Battery cell test Landing gear test Drop sensor test Prototype 1 taxi&flight Prototype 2 taxi&flight Final Design taxi&flight

10 17 24

December 1

8

January

February

March

April

15 22 29 5 12 19 26 02 09 16 23 02 09 16 23 30 06

W i n t e r

Legend

Planned Actual Due date Current Date

/

E x a m

Figure 35: Testing Schedule

Despite having to shift the actual to the right, all major components were successfully tested before the report submission.

U NIVERSITY OF G LASGOW

Page 50 of 54

7.5

Flight Checklist Inspection Item

Task Initial Checks

Aircraft

Visually verify that the aircraft in general is in normal condition without

noticeable cracks, missing parts or unsecured components Wing

No fractures present and wing is firmly in place. Control surfaces move

easily.

Empennage

All control surfaces move easily. No cracks or loose components.

Payload

Payload is properly secured and the payload is correctly located. If

balls are loaded verify smooth action of the drop mechanism.

Batteries

Make sure battery contacts and leads are in good condition for propul-

sion, receiver and transmitter batteries.

Landing Gear

No visible cracks and wrapping is intact.

Batteries are fully charged.

Battery positions are appropriate for the payload carried.

Firmly in place, no fractures. Tyres spin easily and front gear turns in

normal limits.

Motor

Mount is free of fractures. Securely in place.

Propeller

Careful visual inspection of the propeller. No cracks or fractures.

Mounting is appropriate and firm.

Center of Gravity

Ensure CoG matches with the loaded payload. Max aft position allowed is 0.3m from the front tip of the main boom.

Pre Flight Checks Flight Log

Fill in Flight Log details

Batteries

Connect batteries and then connect fuse. Make sure the connections

are good.

Radio controls

Transmitter on first, Receiver on second.

Control surfaces

Verify that maximum deflections are nominal and front gear turns with

and front landing

aileron inputs (ailerons full left-right, elevator full up-down), response

gear

to control inputs is consistent. Small slack. Range check.

Motor

While holding the aircraft from behind, throttle up to full power swiftly.

Note response. No excessive vibrations or unusual noise.

Post Flight Checks Fuse and Batter-

Unplug the fuse and batteries

Radio controls

Receiver off first, transmitter second

Aircraft

Visually verify that the aircraft is in normal condition without noticeable

ies

cracks, missing parts or unsecured components.

Flight Log

U NIVERSITY OF G LASGOW

Ensure flight is properly documented. Report any discrepancies.

Page 51 of 54

8 8.1 8.1.1

Performance Results Structural performance Wing spar and fuselage boom performance

(a) Wing Spar

(b) Main Boom

Figure 36: Structural Testing

The maximum deflection at the tip was measured and the general behaviour of the spar was carefully observed.

Significant amount of bending resulted at maximum load factor but the spar remained intact. For the prototype, it was desired to have a larger safety factor. For the final production model, destructive testing will be performed to more aggressively optimise the weight of the aircraft.

Deflection measurement

1, mm

2, mm

Average, mm

2G 4G 6G

24.6 35.4 50.1

24.4 35.3 50.3

24.5 35.35 50.2

Table 20: Fuselage boom deflections under different loads

Deflection measurement

1, mm

2, mm

Average, mm

2G 3G 4G 5G 6G

24.2 50.6 74.5 105.3 141.7

24.5 50.2 74.9 104.7 140.1

24.35 50.4 74.7 105 140.9

Table 21: Wing spar deflections at wingtip under different loads

8.1.2

Landing gear performance

The landing gear successfully withstood the 3G load. The deflection at the centre of the gear was just about 1.2cm.

There was no permanent twist or bending after unloading, and it was concluded that a lighter landing gear will be manufactured for the next prototype in order to optimize mass further.

8.2

Electronics Test Results

The battery cells tested so far are the Turnigy 2/3A 1500 mA h and the KAN 700 mA h. Both of these cells were tested at 5 C and 10 C, to compare the performance at high current draws. For the purposes of easy comparison C rating U NIVERSITY OF G LASGOW

Page 52 of 54

and % Capacity were used to give nondimensionalized results. Results are plotted in Figure 37.

Figure 37: Battery Discharge test results

From initial constant current tests it is apparent that a higher discharge rate causes the battery voltage to sag and drop below the stated value for all batteries. The KAN700 cells perform well initially by holding a higher voltage than the Turnigy cells, but they drop off steadily. At 10 C, the KAN cell fails to deliver its full stated capacity and drops below the minimum cell voltage. The Turnigy 2/3A 1500 mA h cells perform well, delivering the full stated capacity at both C levels and holding the voltage in a relatively flat level for most of the tests. There was only a significant drop off after 80% capacity, but this is not important as the aircraft has short flight times and will likely only use a portion of the battery capacity. Because of the ability to supply a high current for a longer time period, the Turnigy 2/3AA 1500 mA h cells were selected to be used in the battery.

8.3

Propulsion Test Results

The static test was conducted in the wind tunnel as can be seen in Figure 38. The wind tunnel was not running as the diameter of the propellers tested exceeded the size of the wind tunnel outlet.

U NIVERSITY OF G LASGOW

Page 53 of 54

Figure 38: Propulsion wind tunnel test setup

After extensive testing on a large array of motor-propeller pairs, a database of set specifications was created. The results in Table 22 were used to estimate mission efficiency as a function of minimum thrust required depending on mission specification, estimated maximum speed , and optimal battery weight. Drawing on these the T-motor MN4010 was found to be the most suitable option complemented by a 15x8 carbon propeller for sufficient thrust capabilities required for missions 2 and 3, and a 14x10 propeller providing additional speed at the expense of static thrust for Mission 1. The preselected ESC for this motor (Afro HV 20A) was also confirmed to work well in this configuration with no signs of overheating. Results for the configuration can be seen in the following table: Power Actual, W

Prop

Voltage, V

thrust, g

RPM

current, A

317.2 316.4

15x8 14x10

25.9 25.9

1903 1825

5178 5784

12.25 12.22

Table 22: Static Thrust test data

The thrust data obtained agrees well with power requirements produced by the mass optimization code, meaning that the propulsion matches the design and the aircraft is expected to perform as required.

8.4

Flight Test Results

To team’s great disappointment and despite its best effort, it was impossible to produce a flying prototype on time for various reasons described above. To compensate for that at least partially, an additional analysis was made to estimate the aircraft performance using CFD for drag estimation and FEA for structures.

U NIVERSITY OF G LASGOW

Page 54 of 54