Reconfigurable Unmanned Aerial Vehicle Design and Control

Transcription

1 Department of Aerospace Engineering IIT Kanpur, India Reconfigurable Unmanned Aerial Vehicle Design and Control Abhishek IIT Kanpur Rama Krishna, Sourav Sinha and Joydeep Bhowmick 1

2 UAV Solutions Developed in MAV Lab at IIT Kanpur Fixed Wing UAV Coaxial rotor Powered Boomerang Quadrotor Convertiplane Conventional Quadrotor IC Engine Helicopter Variable Pitch Quadrotor (IC Engine Powered) 2

3 Innovative VTOL Concepts Under Development Compound Helicopter Quadrotor Tailsitter 3

4 NAAVIK (Autopilot) Navigation for Autonomous Aerial Vehicles by IIT Kanpur (NAAVIK) First multi platform auto-pilot developed in India: implemented and tested for: Fixed wing, Quadrotor, Flapping wing, coaxial helicopter and conventional helicopter Fully developed at IIT Kanpur Offers features of a typical commercial autopilot Loiter / Autonomous Hover: at specified altitude, circle radius and location (lat, lon) Way point navigation Return to home on loss of radio contact Auto takeoff and landing for VTOL vehicles Rugged laptop based ground station for mission planning and telemetry Adaptable to new vehicle platforms 4

5 Outline Background / Motivation Objective Approach Results Summary & Conclusion 5

6 Small UAS Applications Military Reconnaissance and surveillance missions over the hill, around the corner, sensor placement Increase situational awareness (real-time visual and infrared images) Hostage situation Civil Biological/chemical agent detection Power-line inspection Disaster-response management Traffic monitoring (long endurance) Counter-drug operations Payload Delivery Precision agriculture Mining 6

7 Hover Capable Small UAVs Conventional Helicopter Coaxial rotor Cyclocopter Hybrid VTOL Concepts Flapping wing Quadrotor / Multirotor 7

8 Need for Hybrid UAVs Fixed wing Rotary wing Conventional Design Limitations Inability to take-off and Land vertically Poor endurance and stealth. Limited forward flight capability Hybrid UAV designs try to combine best of fixed wing and rotary wing configurations Efficient hover and high speed high endurance forward flight Ease of transition between helicopter and airplane modes 8

9 Hybrid VTOL Concepts Arcturus JUMP ETH Zurich Tailsitter Martin UAV V Bat Bell Eagle Eye Google Project Wing Xcraft X PlusOne 9

10 Hybrid VTOL Concepts Quadrotor Biplane Tailsitter (UMD) Quad Tilt-wing Sabanci University Turkey Quad Tilt-rotor RC Quad Tilt-rotor 10

11 Challenges with Quad Tiltwing and Quad Tiltrotor Higher actuator forces and moments Transition is challenging due to lack of lift during transition due to high AOA on wing (Cetinsoy et al. 2012) Up to 10% loss in thrust due to rotor wake and wing interaction (Radhakrishnan 2006, Vinit et al. 2005) Rotor-rotor and wing-wing interaction in forward flight (Sheng and Naramore 2009) 11

12 Objective Design and development of a novel quadrotor convertiplane VTOL UAV with high efficiency Motor, propeller, wing sizing Estimation of portion of wing to be tilted Quantify effect of tilting of wing Design tandem wing configuration for efficient forward flight Abhishek, Krishna, M. R., Sinha, S., Bhowmik, J., and Das., D, "Quadrotor Convertiplane Air Vehicle", Patent Application No.: , Dated: 14/06/

13 Novel Quadrotor Convertiplane Four tilting proprotors with wing segment in rotor downwash tilting with proprotor Low and high tandem wing configuration Quadrotor type control in hover 13

14 Transition Flight Transition to forward flight is initiated by tilting proprotors to horizontal orientation from initial vertical position Wing segment outboard of rotor downwash starts generating lift as vehicle gains forward speed, aiding in transition process Rotor thrust Wing lift Axis of propeller and wing segment tilted using servo 14

15 Forward flight mode Once transition is complete, role of rotors changes from lifter to thruster and tandem wings supports weight of the vehicle Conventional A/C controls are employed for flight control 15

16 Summary of Advantages Offered by Current Design Hover losses due to aerodynamic download on wing are minimized Since, only a part of wing tilts with rotor, actuator forces and moments required for tilting are expected to be smaller compared to a full tilt wing Wing segment outboard of rotor downwash helps in transition by contributing to lift generation Low and high tandem wings are expected to reduce wing-wing, propeller-wing and propellerpropeller interactions 16

17 Design Approach Key design parameters are systematically selected through a series of experiments Objective is to design a prototype with 1.6 kg AUW Propeller selection Motor selection Span of tilting wing segment Airfoil selection Wing sizing Integration of prototype 17

18 Hover Test Stand Propeller Motor Nano 43 six component load cell Wattmeter 18

19 Propellers Tested HK 6050 KK 6040 GF 5045 GF 5550 Lumenier 5035 GF Bladed HK 5040 GF 6040 GF

20 Propeller Selection Thrust vs. RPM Aerodynamic Power vs. RPM Thrust (grams) HK 5040 GF 6040 Luminir 5035 KK 6040 GF Bladed GF Bladed GF 5550 GF 6030 HK 6050 Aerodynamic Power (W atts) HK5040 GF6040 Luminir 5035 KK6040 GF Bladed GF Bladed GF 5550 GF 6030 HK ,000 8,000 13,000 18,000 23,000 RPM 0 3,000 8,000 13,000 18,000 23,000 RPM HK 6050 generates maximum thrust for given RPM 20

21 Propeller Selection Power Loading (kg/w) Aerodynamic Power Loading vs. Thrust HK 5040 GF 6040 Luminir 5035 KK 6040 GF 5045 GF blade GF 5550 GF 6030 HK Thrust (in grams) HK 6050 and GF 6040 have best power loading for thrust range of grams 21

22 Motor selection Propeller selected (HK 6050) is paired with eight different motors thrust, electrical power and aerodynamic power for combination is measured using hover stand Tornado (2300 KV), Emax (2300 KV), RCX (2350 KV), Tiger F40 (2500 KV), Cobra (2300 KV), Tiger F40 (2300 KV), Tiger F60 (2450 KV) and Lumenier (2350 KV) 22

23 Motor Selection Electrical Power Loading (g/w) Electric Power Loading vs. Thrust Tornado 2300 KV Emax 2300 KV RCX 2350 KV F KV Cobra 2300 KV F KV F KV Lumenier 2350 KV Efficiency Motor Efficiency vs. Thrust Tornado 2300 KV Emax 2300 KV rcx 2350 KV F KV Cobra 2300 KV F KV F KV Lumenier 2350 KV Thrust (grams) Thrust Luminier 2350 KV motor has highest power loading and consistent efficiency of 70% 23

24 Sizing of Wing Segment Tilting with Rotor Study effect of rotor wing interaction on rotor performance for rotor mounted on wing tip Quantify loss due to download on wing by changing span of wing segment in rotor downwash 24

25 Measurement of Rotor-Wing Interaction Propeller generating downward thrust Wing portion oriented vertically Wing portion oriented horizontally Motor Nano 43 six component load cell 25

26 Sizing of Wing Segment Tilting Thrust (grams) Thrust vs. RPM Isolated rotor Tilt wing span = 0.5R Tilt wing span = 0.75R Tilt wing span = R Tilt wing span = 1.25R Full tilt wing No tilt RPM Electric Power Loading vs. Thrust Electrical Power Loading Isolated rotor Tilt wing span= 0.5R Tilt wing span= 0.75R Tilt wing span= R Tilt wing span= 1.25R Full tilt wing No tilt Thrust (in grams) Pure tiltrotor (only prop tilting) loss in thrust is approximately 30-40% for same RPM Loss in thrust decreases with tilting of wing segment in downwash For wing segment (with 50% of propeller radius) tilting loss in thrust is 18% to 20% For R and 1.25 R case loss in thrust and power loading is negligible 26

27 Performance Measurement in Forward Flight Experimental setup for performance measurement in forward flight 27

28 Thrust Coefficient vs. Advance Ratio 5000 RPM 8000 RPM C T GF 6030 GF 6040 KK 6040 HK 6050 C T GF 6030 GF 6040 KK 6040 HK J (advance ratio) J (advance ratio) 28

29 Thrust Coefficient vs. Advance Ratio RPM CT J (advance ratio) GF 6030 GF 6040 KK 6040 HK 6050 HK 6050 propeller gives high thrust output at all advance ratios and three RPMs tested 29

30 Efficiency in Forward Flight Efficiency ( ) RPM 8000 RPM GF 6030 (Experiment) GF 6030 (Best fit) GF 6040 (Experiment) GF 6040 (Best fit) KK 6040 (Experiment) KK 6040 (Best fit) HK 6050 (Experiment) HK 6050 (Best fit) Efficiency(ηη) GF 6030 (Experiment) GF 6030 (Best fit) GF 6040 (Experiment) GF 6040 (Best fit) KK 6040 (Experiment) KK 6040 (Best fit) HK 6050 (Experiment) HK 6050 (Best fit) J (advance ratio) J (advance ratio) Efficiencies measured are typical for propellers of this size 30

31 Efficiency in Forward Flight Efficiency(ηη) GF 6030 (Experiment) GF 6030 (Best fit) GF 6040 (Experiment) GF 6040 (Best fit) KK 6040 (Experiment) KK 6040 (Best fit) HK 6050 (Experiment) HK 6050 (Best fit) RPM HK 6050 offers highest maximum efficiency of for all three RPMs considered J (advance ratio) 31

32 Wing Sizing Wing sized for 1.6 kg AUW Based on experience with small fixed wing aeromodels wing loading of approximately 5.5 kg/m 2 is chosen. Results in required total wing area of 0.28 m 2 To keep footprint small aspect ratio of 4 is chosen Based on these choices span of each wing is set to 0.7 m and chord length is 0.2 m 32

33 Airfoil selection No horizontal tail to ensure moment equilibrium Airfoil selected should have low pitching moment co-efficient for ease of trimming in airplane mode Four airfoils are compared using XFLR5 1. NACA 0010 (symmetric airfoil) 2. AG36 (low camber) 3. SD7062 (medium camber) 4. S1223 (high camber) 33

34 Airfoil Selection C l Ag36 Sd7062 S1223 NACA Angle of attack (deg) C d Ag36 Sd7062 S1223 NACA Angle of attack (deg) 34

35 Airfoil Selection C l /C d Ag36 Sd7062 S1223 NACA0010 C m Ag36 Sd7062 S1223 NACA Angle of attack (deg) Angle of attack (deg) 35

36 Airfoil Selection Airfoil Cl Stall angle Cd Cm Cl/Cd AG36 Satisfactory Satisfactory Low Very low High SD7062 Satisfactory Satisfactory Low High Low S1223 High Very low Very high Very high High NACA0010 Very low Satisfactory High Very low Low Based on these observations, AG36 is chosen for both wings 36

37 Dimension & Design parameters Gross weight = 1570 grams Wing span = 0.7 m Wing chord = 0.2 m Fuselage length = 0.75 m Fuselage width = 0.15 m Airfoil: AG36 Stall speed = 8.87 m/s Cruise speed = 15 m/s 37

38 Power Required in Hover Total thrust required for hover = 1570 gms Thrust required per motor = gms Power consumed by all motors = 356 W 250 Power consumed by all motors = 356 W Power (watts) Thrust (grams) 38

39 Power in Forward flight For cruising, total thrust required is 328 gms, each motor producing 82 gm by consuming a power of 20 W (each) at RPM of 9800 Total motor power 128 W Power required to overcome vehicle drag 47 W Total forward flight power required: 175 W (50% of hover) Required Thrust (grams) Power (watts) Velocity (m/s) Thrust (grams) 39

40 Stabilization and Control Three different flight modes Hover Transition Forward flight / airplane Requires dedicated control law for each mode to enable smooth transition from hover to forward flight and back Strategy Use RPM for hover control Use conventional aileron, elevator, rudder to control the aircraft during airplane mode Command servos to tilt propellers by 90 degrees to change mode Use weighted combination of hover and airplane controls to ensure stabilization during transition Test in simulation before implementing on real vehicle 40

41 Flight Dynamics Model Modified quadrotor equation of motion Account for different planes of motor-propeller pairs Inertia properties estimated using accurate CAD model with actual component weights Hover I xx (roll) kg m 2 I yy (pitch) kg m 2 I zz (yaw) kg m 2 I xz kg m 2 Forward flight Ixx (roll) kg m 2 Iyy (pitch) kg m 2 Izz (yaw) kg m 2 Ixz kg m 2 Enhanced modeling with provision for propeller thrust vector tilting 41

42 Frame of Reference Ω2 Ω3 Ω1 v y q Aft Fore Ω4 u x p r z w 42

43 Equation of Motion PID Controller implemented to stabilize simulated quadrotor 43

44 Performance of PID Controller Roll Attitude Pitch Attitude (deg) Current state Desired state (deg) Current state Desired state Time (sec) Time (sec) 44

45 Yaw Attitude Tracking (deg) Current state Desired state Time (sec) Yaw control significantly slower than roll and pitch High yaw inertia and small motors and propellers with limited reaction torque How to augment yaw control authority? 45

46 Yaw Control Augmentation Use component of rotor thrust for yaw moment generation Ω2 Aft Ω3 Ω1 Fore Ω4 x 46

47 Effect of Propeller Axis Tilting Desired state (deg) 20 Current state (20 Current state (15 o o tilt) tilt) Current state (10 o tilt) 10 Current state (5 Current state (0 o o tilt) tilt) Time (sec) 47

48 Autopilot PixHawk: Integrated sensor microcontroller board Processor ARM Cortex M4 based STM32f407 (ST Micro electronics) 168 MHz (a big deal for MCU) 32 bit Floating point coprocessor Sensors faster computation 3 axis digital MEMS Gyro ( deg/s) PixHawk 3 axis digital MEMS Accelerometer (+- 16 g) 3 axis digital MEMS Magnetometer (+-12 Gauss) Barometer (+-10 cm) GPS (UBlox) PID controller for attitude stabilization is implemented on Pix Hawk for hovering flight test 48

49 System Integration Wings were made using pink foam Stiffened using Aluminum T-spars Airframe is made using carbon fiber rods Component Weight (gram) Percentage of Total (%) Motors Propellers ESCs Battery Servos Avionics Structure Total

50 Quadrotor Convertiplane Prototypes Prototype II Prototype I 50

51 Flight Test Attitude Tracking Attitude (deg) Roll Roll Roll Setpoint Attitude (deg) Pitch Pitch Pitch Setpoint Time (sec) Time (sec) 100 Attitude (deg) Yaw Yaw Setpoint Time (sec) 51

52 Conclusions For current design propeller operates with nearly same power loading as isolated rotor when span of tilting wing portion is equal to or greater than propeller radius If wing segment with a span of at least 75% of radius of propeller is allowed to tilt and align with propeller downwash, loss in thrust is less than 3% Current novel quadrotor convertiplane design offers significantly higher hovering efficiency in comparison to conventional quad tiltrotor with no wing tilting. For conventional quad tiltrotor Propeller looses up to 40% of its thrust due to download penalty on wing Power loading increases by 50% of that of an isolated rotor 52

53 Conclusions Current convertiplane design is estimated to consume 50% less power during airplane / forward flight mode in comparison to hover mode Making this concept more useful than quadrotors Roll and pitch tracking using conventional quadrotor type control is adequate, yaw stabilization using torque reaction inadequate due to small motor-propeller combination tilting of fore propellers rearward by 10 o and aft propellers forward by 10 o, significantly enhances control authority in yaw with only 1.5% loss in thrust 53

54 Future work Characterize performance in forward flight mode by looking at effect of separation of tandem wings Enhance autopilot for transition flight and forward flight in airplane mode Full flight envelope flight dynamics model has already been developed Carry out structural optimization for more robust yet light weight prototypes 54

55 Thank You! 55