Test Results for a Fuel Cell-Powered Demonstration Aircraft

26-1-392 Test Results for a Fuel Cell-Powered Demonstration Aircraft Copyright 26 SAE International Thomas H. Bradley Georgia Institute of Technology Blake A. Moffitt, Reid W. Thomas, Dimitri Mavris Aerospace Systems Design Laboratory David E. Parekh Georgia Tech Research Institute ABSTRACT A fuel cell powered airplane has been designed and constructed at the Georgia Insitute of Technology to develop an understanding of the design and implementation challenges of fuel cell-powered unmanned aerial vehicles (UAVs). A custom 448W net output proton exchange membrane fuel cell powerplant has been constructed and tested. A demonstrator aircraft was designed and built to accommodate this powerplant and the fuel cell powered aircraft has performed seven test flights to date. Test data show that the aircraft performance validates the models used for design and optimization and that the fuel cell aircraft is capable of longer endurance, higher performance test flights. INTRODUCTION Fuel cells are an attractive technology for implementation as powerplants for aircraft because of their potential for rechargability and high energy density. Where advanced rechargeable batteries can reach energy densities of 15 Wh/kg at the module level, [1] regenerative fuel cells can achieve >8 Wh/kg at the system level [2], and nonregenerative fuel cells can achieve >1Wh/kg [3]. This study is concerned with the application of nonregenerative polymer electrolyte membrane (PEM) fuel cells as the primary powerplant in a medium-scale unmanned aerial vehicle (UAV). Implementation of nonregenerative fuel cells in UAVs is viewed as a first step towards practical fuel cell-powered flight. A few researchers have used small-scale model aircraft to demonstrate fuel cell-powered UAVs [4-6]. Soban and Upton [7] have performed preliminary design studies but without hardware development or system-level validation. AeroVironment has demonstrated a large scale fuel cell UAV [8], but the design and development challenges of fuel cell UAVs are still not well-understood by the general UAV community. In order to begin developing tools and technologies for the design and implementation of fuel cells as powerplants in aeronautical vehicles, a PEM fuel cell UAV demonstration program was started in the Summer of 24 as a collaboration between the Georgia Institute of Technology Aerospace Systems Design Laboratory (ASDL) and the Georgia Tech Research Institute. A portion of the design work is supported by the NASA University Research Engineering Technology Institute (URETI) grant to the Georgia Institute of Technology. The primary research objectives of this program are: the development of validated tools for fuel cell aircraft design, analysis of the design tradeoffs present in fuel cell aircraft, and the demonstration of a series of fuel cell UAVs. This paper presents a summary of the specifications of the aircraft and its fuel cell powerplant. Laboratory and field testing results are analyzed and used to validate the aircraft propulsion model that was used for system design. A summary of the flight tests completed to date is also presented. A photograph of the fuel cell-powered demonstrator aircraft is shown in Figure 1. Figure 1. Fuel cell-powered demonstrator aircraft FUEL CELL POWER PLANT The powerplant for the fuel cell UAV is a 5W 32-cell PEM stack (BCS Fuel Cells, Bryan, TX) with a custom balance of plant. The fuel cell balance of plant was designed for mobility, reliability and minimum weight. The balance of plant (BOP) includes the air management subsystem, hydrogen delivery subsystem, thermal

management subsystem, and system controller. The fuel cell is self-humidified and requires an air manifold pressure of 5 psig and a hydrogen manifold pressure of 3 psig. The fuel cell balance of plant provides a cathode stoichiometry of >2. and 9% peak hydrogen utilization. A schematic of the fuel cell powerplant system is shown in Figure 2. Figure 3. Fuel cell system hydrogen utilization THERMAL MANAGEMENT SUBSYSTEM Figure 2. Schematic of fuel cell powerplant system AIR MANAGEMENT SUBSYSTEM The air management subsystem maintains constant pressure at the cathode of the fuel cell stack and provides variable flow rate control. Variable flow rate control is particularly important in a self-humidified fuel cell system because of the risk of under-humidification at low current densities. The 5 psig cathode pressure is regulated with a calibrated ball valve (Microchek 14B14B-5 psi, Lodi, CA). Flow rate is closed-loop controlled by pulse-width modulation of two diaphragm compressors (T-Squared Manuf. T22, Lincoln Park, NJ). These compressors are powered from the fuel cell bus voltage. By using two compressors, and turning one of the compressors off when low flow is required, higher high flow rates and lower low flow rates are achievable than is possible with a single compressor. HYDROGEN DELIVERY SUBSYSTEM Hydrogen is stored on board the aircraft using compressed hydrogen stored within a composite tank (Luxfer Gas Cylinders P7A, Riverside, CA). The hydrogen storage system provides roughly 66 standard liters (SL) of usable hydrogen. Dual regulators regulate the hydrogen pressure to a manifold pressure of 3psi. Hydrogen is purged from the cathode manifold for.2 seconds at roughly 8 second intervals. Hydrogen purge is required to remove water and other contaminants from the anode manifolds. A liquid cooling circuit provides temperature regulation for the fuel cell. A water pump (Laing Thermotech DDC, Chula Vista, CA) circulates water through a custom radiator of aluminum and carbon foam (Poco Graphite, Decatur, TX) construction. Carbon foam is an advanced, lightweight (.56g/cc), thermally conductive material that allowed for the construction of a light weight water-to-air radiator. A number of design iterations were performed to reduce the weight and power consumption of the radiator. In the final design iteration, the fuel cell radiator and fan weigh 578g and rejects 242W over a temperature difference of 2C with 6W of electrical power consumption. Fuel cell product water is sprayed onto the fins of the radiator to allow for evaporative cooling. SYSTEM CONTROLLER The fuel cell system is controlled by an ATMEGA32, 8-bit AVR microcontroller (Atmega, San Jose, CA). The functions of the fuel cell controller are: 1. Thermostatic Fan Control This function turns on the radiator fans when the fuel cell temperature is greater than 45 C. 2. Closed-Loop Compressor Speed Control An inductive current sensor measures the current output of the fuel cell. This function reads the current sensor and calculates the required PWM conditions for the compressors. The compressor speed is closed-loop controlled with a PI controller. 3. Mode Control There are three modes of fuel cell controller operation: Idle, Purge, and Fly. In Idle mode the fuel cell system is not producing power. Purge mode allows the fuel cell controller to purge the cathode manifold of the fuel cell at startup or shutdown. Fly mode is the normal mode of

operation that includes active, closed-loop control of the fuel cell system. 4. Open Loop Compressor Purging Cycle During startup and shutdown, the compressors operate at maximum flow rate to purge the cathode manifold of the fuel cell. 5. Purge Valve Control The fuel cell controller provides a signal to the anode purge valve to purge the fuel cell anode manifold. 6. Data Acquisition The fuel cell controller performs data acquisition, data storage and transmission to a ground station. The fuel cell controller functions as both the fuel cell and aircraft data acquisition system. FUEL CELL SYSTEM PERFORMANCE To validate the design and demonstrate the function of the entire fuel cell system, the system was tested in the laboratory before implementation in the fuel cell aircraft. A total net output power of 448W is available from the 5W rated fuel cell system. The balance of the power between the system net output power and the 523W gross output from the fuel cell is consumed by accessory loads. Figure 4 shows the breakdown of the power consumed by the accessory loads as a function of fuel cell system net output power. The 3W of power consumed intermittently by the cooling system fan is not included in this graph. The compressors make up the largest part of the power consumed by the fuel cell system. The water pump, controller and data acquisition (DAQ) hardware consume a constant 13.3W. 52W/kg and a specific energy of 7.1Wh/kg, both measured at conditions maximum net electrical power. AIRCRAFT DESCRIPTION The medium scale (Sw=188dm2, m=16.4kg) of the aircraft platform is chosen for the demonstration aircraft because of: The utility of a larger aircraft platform for potential UAV missions including surveillance and remote sensing. Improved applicability of design methods and validation data. The closer in scale that the demonstration aircraft is to future utilitarian aircraft, the more confidence can be granted to the design and validation data. The suitability in scale and cost for an academic research program. Conceptual design of the aircraft was performed using an integrated model of the powerplant, propulsion and aircraft systems. The characteristics of this design model were explored using a full factorial design of experiments for the 12 design variables: wing area, aspect ratio, tail type, electric motor, fuel cell, propeller, fuselage shape, hydrogen storage type, number of hydrogen storage tanks, and wing configuration. More detail regarding the conceptual design task is available in Moffitt et al. [9]. Fuel Cell Component Power Consumption, Watts 6 5 4 3 2 1 Compressor 1 Compressor 2 Control and DAQ Coolant Pump Electrical Output Power 5 1 15 2 25 3 35 4 45 Net System Output Power, Watts Figure 4. Breakdown of fuel cell powerplant power consumption Overall, the fuel cell system including balance of plant, controls and hydrogen storage has a specific power of Figure 5. Fuel cell demonstrator aircraft configuration The result of the conceptual and lower level design tasks is the aircraft configuration shown in Figure 5. The aircraft features a high aspect ratio, non-swept wing with taper and dihedral in the outer spanwise sections. The aircraft has a streamlined fuselage with an internal volume of approximately 36.7L with a maximum cross sectional area of 438cm 2 and an overall length of 1.12m. The aircraft uses a pusher configuration with a brushless 3-phase motor and a fixed pitch propeller. The aircraft has twin tailbooms supporting an inverted V tail.

Table 1. Specifications for the fuel cell demonstrator aicraft Design Specification Value Wing Area 188 dm 2 Aspect Ratio 23 Span 6.58 m Taper Ratio.67 Taper Change Point.625 Outer Panel Washout 4 deg Tail Area 45.5 dm 2 Length (nose to tail) 2.38 m Mass 16.4 kg Propeller Diameter 55.9 cm Propeller Pitch 5.8 cm Static Thrust / Weight.165 Cruise Airspeed 14.5 m/s Dihedral at Tapered Section 1 deg The lower-level design of the aircraft to refine the aircraft aerodynamics and propulsion system included additional optimization metrics and constraints. When considering stall performance, power limited turning performance, minimum turning climb rate, minimum Reynolds number, minimum bank angle and maximum allowed angle of attack the wing geometry changed slightly from the conceptual design results. The final aircraft design resulted in a slightly higher wing area and aspect ratio than the conceptual aircraft design. Specifications of the aircraft design are given in Table 1. AIRCRAFT CONSTRUCTION Balancing the requirements of light weight and function proved especially difficult for the demonstration aircraft. Analysis performed at the design stage showed that the aircraft performance is highly sensitive to the takeoff gross weight of the aircraft. The large mass of the fuel cell powerplant, necessitated structural analysis and design not usually associated with small scale UAVs. The construction methods and materials were also constrained by budgetary requirements and an aggressive build schedule. The most critical component of the aircraft structural design was the fuselage. The fuselage is designed to support the 8.6 kg fuel cell system, wing loads and tail moments under a minimum 3-g loading. The fuselage consists of a structural internal frame with non-structural fairing. The internal frame is designed to take all major loads during both ground and flight operations. It is made up of a roll-wrapped carbon fiber tubular spar and two roll-wrapped carbon fiber composite tail booms lugged and bonded to a tubular aluminum frame. The fairing provides an aerodynamic shape to the aircraft, and is constructed of a thin layer of fiberglass with honeycomb stiffening in some panels to maintain the desired shape. The main landing gear have a symmetric airfoil cross-section and are machined from 661-T6 aluminum billet. The nose gear is constructed of composite tubing with an aluminum fork. The wings are constructed of a composite spar and solid foam cores sheeted with balsa wood. The empennage surfaces are built from foam and balsa wood without an internal spar. Figure 6. Mass breakdown for the fuel cell demonstrator aircraft

The mass breakdown of the final aircraft is shown in Figure 6. The weight of the aircraft is dominated by the weight of the powerplant. The fuel cell and balance of plant make up roughly 57% of the total weight of the aircraft. The weight of the hydrogen fuel is included in the hydrogen tank weight. For the test flights described in this paper, the aircraft is fueled with 5g of hydrogen. pilot riding in this vehicle. This arrangement is shown in Figure 7. AIRCRAFT FLIGHT TESTS While the laboratory testing provide a significant amount of data on the fuel cell power system, the ultimate validation of this work comes from integration with the aircraft and flight tests of the total system. The primary goals of this research are to push the boundaries of range and endurance for aircraft by leveraging the high efficiencies of fuel cells. Working towards these goals, the team has developed an incremental flight test plan, focused on achieving low risk aircraft performance milestones, which matches the aircraft s capabilities and the testing facilities available. The goal of this test plan is to collect data regarding the aircraft performance and demonstrate milestones of fuel cell flight, while minimizing the risk to the aircraft. Matching the aircraft s performance capabilities with a suitable flying location proved to be a difficult task. While many sites exist around the country and throughout Georgia for model aircraft flying, the size of this aircraft, as well as the nature of the research provides only a partial overlap with the available field options. Furthermore, most hobby flying sites impose additional performance constraints on the vehicle through their geographic constraints and site obstacles such as trees or power lines. The culmination of these factors does not eliminate hobby sites from contention, but would force the flight test program to take several initial leaps before settling into the incremental schedule. As a potential alternative, the research team investigated the possibility of flying at smaller or private airstrips for full size aircraft. Pursuing this option forced the aircraft up against recent changes imposed by the Federal Aviation Administration (FAA) on the use of unmanned aerial vehicles in the vicinity of airports, which forbids this option, given the timeline of the project. Determined to proceed with the incremental approach, the team looked outside these more traditional aeronautical options for a facility that would allow for the planned initial test flights. Ultimately, the Atlanta Dragway in Commerce, GA was selected as the location for the initial demonstration flights. The dragway provided nearly a thousand meters of unobstructed paved facilities, without any regulatory restrictions. This facility proved to be an excellent location for the initial flight tests. While the use of the Atlanta Dragway alleviated several concerns and challenges, it also posed additional constraints upon the flight test crew. The long straight track, so ideal for the vehicle, made the control of the aircraft much more difficult for the pilot and flight crew. Ultimately it was decided that a chase vehicle would be used to follow the flight tests down the track, with the Figure 7. Fuel cell-powered demonstrator aircraft undergoing flight test with chase vehicle In order to perform longer-range and higher-performance test flights, the aircraft was moved to the flying site of the Georgia Jets radio controlled aircraft club. The Georgia Jets facility, located in Gay, GA, features a 35m paved runway, and a large cleared area. TESTING SUMMARY To date, the team has completed six test flights (Tests #1-6) of the aircraft on hydrogen power at the Atlanta Dragway and a single hydrogen powered circuit test flight (Test #7) at the Georgia Jets facility. Table 2. Test flight summary for fuel cell-powered flights Flight Test Duration Max Altitude Test #1 28 seconds 2.5m Test #2 52 seconds 3.7m Test #3 36 seconds 3.5m Test #4 12 seconds 2.5m Test #5 5 seconds 3.5m Test #6 25 seconds 3m Test #7 12 seconds 25m Test goals Aircraft trim, instrumentation check and pilot familiarization Pilot practice and aircraft trim Initial test of flaps Aircraft trim, further flap testing Final aircraft trim flight Achieve steadylevel flight Achieve steadylevel flight Circuit flight, rate of climb test The purpose of these flights has been trimming of the aircraft controls, pilot familiarization with the aircraft, initial flight testing of the fuel cell system and testing of the flight data acquisition systems. The flight tests also provide the first performance data sets for the fully integrated aircraft. This data allows the team to validate the design models for the vehicle and fuel cell system,

and to verify the predicted flight performance. This critical feedback will provide the team with significantly greater confidence in the accuracy of future design predictions, and will also aid in flight planning for future test flights. A short summary of the tests completed to date, their duration, maximum altitude, and the goal of the flight is listed in Table 2. FLIGHT TEST RESULTS Figure 8 shows some results from Flight Test #2. The fuel cell voltage and gross current (measured before BOP loads) are plotted on the same axis as the aircraft airspeed. As the aircraft begins takeoff the loading on the fuel cell powerplant increases. The voltage of the fuel cell decreases as the current draw decreases until the maximum power operating condition is reached. At a time of 26 seconds, the pilot rotates the aircraft and the plane becomes airborne. The airspeed plateaus at 32 ft/sec for a few seconds as the pilot gains altitude, but when an altitude of 3.5m is reached, the pilot accelerates at full power until a time of 6 seconds. At that point the pilot scales back the throttle, decreases airspeed, lowers flaps and the aircraft touches the ground at a time of 78 seconds. Voltage (V), Current (A), Airspeed (ft/sec) 8 7 6 5 4 3 2 1 Takeoff Fuel Cell Output Current Fuel Cell Output Voltage Aircraft Airspeed 2 4 6 8 1 Time, sec Landing Figure 8. Sample results from straight-line flight (Flight Test #2) Figure 9 shows data from the same test flight. In this figure, the two compressor speeds are shown on the same plot as the fuel cell temperature. As shown in Figure 9, during low fuel cell power demands, only one compressor is used to provide the cathode air supply. When the fuel cell power demand is increased as the pilot takes off, both compressors work together. When the power demand decreases, as the pilot begins to land, Compressor 2 again shuts off and Compressor 1 provides idle air flow. Figure 9 also shows that for this test the fuel cell does not reach its operating temperature of 5-6C over the course of this test. Fuel Cell Temperature, C 55 5 45 4 35 Fuel Cell Temperature Compressor 1 Speed Compressor 2 Speed 2 4 6 8 1 Time, sec 8 7 6 5 4 3 2 1 Figure 9. Compressor speeds and fuel cell operating temperature from straight-line flight (Flight Test #2) Compressor Speed, rpm The results of this test flight show that the aircraft achieved full fuel cell output power, functional flap deployment, improved aircraft control surface trim, and functional data acquisition. The aircraft took 2m of taxi distance and achieved straight-line flight and landing without incident. VALIDATION OF DESIGN MODEL One of the main goals of both the building and testing the demonstration aircraft is to obtain the data needed to validate and calibrate the models used during the aircraft design. This section describes the propulsion system model and its partial validation using the preliminary test results from the demonstration aircraft. PROPULSION SYSTEM DESIGN MODEL Modeling of the fuel cell system is based on a static polarization curve. During the conceptual design, linear approximations to the polarization curves were used. Although more refined models of the fuel cell can be developed, the computational efficiency of the linear approximation makes it particularly computationally efficient when used in multidisciplinary optimization techniques. Once the stack was purchased and the balance of plant was designed and implemented, refinements to the polarization curve were made using data obtained from laboratory testing. Propeller performance maps for the scale of this aircraft only exist for a few propellers [1] and typical methods

based on classical momentum theory are not accurate enough for designing a power limited aircraft. Thus a custom propeller analysis code was developed based on a mixture of momentum and blade element theory using Goldstein s vortex model [11,12]. Three different propeller geometries were accurately measured for use with the propeller code. In addition, scaling of the propeller profiles and pitch distributions was used to help predict the performance of other fixed pitch propellers similar to the measured propellers. The electric motor was modeled using a lumped parameter equivalent circuit model of the motor, as shown in Figure 1. This model uses a no-load current (I ), motor internal resistance (R m ), controller resistance (R cont ), and motor voltage constant (K v = ω m /V m ). Values of the lumped parameters were obtained from motor manufacturers during conceptual design and were then measured from motors actually purchased for the final detailed design of the propulsive system. By combining the measured fuel cell stack polarization curve with the propeller and electric motor models, the predicted propulsive performance of the aircraft could be calculated for any number of flight conditions. For example, Figure 11 shows both the maximum predicted thrust and the maximum available propulsive power generated by the optimized propulsion system. MODEL VALIDATION The first data used for validation were obtained by bench testing the full propulsion system. Current, voltage, and motor speed were measured for all combinations of two motors and two propellers. Figure 12 shows both the predicted and measured data for the motor and propeller selected for the final aircraft. Note that the propulsion model offers a good estimate of the fuel cell current and voltage up to about 23 amps. At currents above 23 amps, the voltage of the fuel cell begins to rapidly drop causing the power output to plateau and then decrease. As a result, the motor speed would also begin to decrease slightly after 23 amps. For flight purposes, the throttle was limited so that the maximum allowable current was 23 amps, thus predicting performance at higher currents was unnecessary. The model used to calculate the values in Figure 12 is the linear fuel cell polarization model used for conceptual design. More details regarding the conceptual design are present in Moffitt et al. [9]. Figure 1. Electric motor model 3 3 The most important data for validation in Figure 12 are the motor speed data. Motor speed is dependent on all components in the propulsion system. Note that the predicted motor speed agrees with the measured data. In order for this to happen, the code must correctly predict the torque vs. angular speed relationship for the propeller, fuel cell and motor in combination. 25 25 Thrust, N 2 15 1 2 15 1 Power (T*V), W 5 5 1 15 2 25 3 Airspeed, m/s Thrust Power 5 Figure 11. Predicted maximum available propulsive thrust and power Figure 12. Propulsion system bench test results The data in Figure 12 suggest that the propulsion model is adequately predicting performance under static

conditions. However, to be useful, the model must also be able to predict performance as airspeed is varied. To verify that the model could predict performance over a range of airspeeds, a test flight of the aircraft was performed. A well-characterized nickel-cadmium battery pack was used to power the flight test. Use of the battery pack for initial flight testing allows for the characterization of the aircraft independent of the fuel cell powerplant and also allows the aircraft to be flown without risk to the fuel cell system. Data from the battery flight test is shown in Figure 13. The flight lasted nearly 1 seconds and consisted of a simple circuit around the field. The airspeed varied from to 25 m/s. Using the measured flight airspeed in conjunction with a linear battery model, the battery current and propeller thrust was predicted and compared to measured data. The predicted current is very close in magnitude to measured data and closely predicts the trends measured throughout the flight. The predicted thrust also closely matches the measured data. This comparison provides validation for the dependence of the propulsion system model on airspeed. Current, A; Airspeed, m/s 5 45 4 35 3 25 2 15 1 5 25 5 75 1 125 15 Time, sec Measured Current Predicted Current Measured Airspeed Measured Thrust Predicted Thrust Figure 13. Propulsion system flight test data for a battery powered test flight The final validation of the propulsion model was performed by using the data from the fuel cell powered flights. A non-linear polarization curve derived from bench-top testing of the fuel cell system was used to model the fuel cell. Validation was performed by comparing the predicted and measured motor speed and thrust. Figure 14 shows this comparison for the fuel cellpowered circuit flight, Flight #7. Note that the propulsion model accurately predicts the magnitude of the motor speed and correctly predicts the increase in motor speed near the time of 75 seconds. The thrust data for this flight shows a dip in magnitude during climb out (1s- 3s) that may be an artifact of the strain-gauge measurement setup. There is good correlation between 8 7 6 5 4 3 2 1 Thrust, N the measured and predicted thrust for the other portions of the flight. Motorspeed, rad/s 4 35 3 25 2 15 1 5 Motorspeed Measured Motorspeed Predicted Thrust Measured Thrust Predicted 2 4 6 8 1 12 Time, sec 5 45 4 35 3 25 2 15 1 5 Thrust, N Figure 14. Predicted vs. measured data for the circuit fuel cell test flight (Flight #7) Together, these comparisons show that the behavior of the fuel cell aircraft propulsion system can be predicted to the accuracy required for design. With refinement and further testing the accuracy of the model will be improved. CONCLUSIONS Theoretical design models for fuel cell aircraft are present in the field, but the modeling, optimization, and implementation challenges of fuel cell aircraft were not well understood. The design, construction and testing of a fuel cell-powered demonstration aircraft is an important step towards realization of fuel cells as a powerplant for aviation. The design of the fuel cell system has resulted in a feasible fuel cell powerplant with >45W net output and 9% hydrogen utilization. The fuel cell demonstrator aircraft is optimized around the use of a fuel cell powerplant. The result is a highly effective platform for extended testing and evaluation of the fuel cell-powered aircraft concept. Testing of the fuel cell demonstrator aircraft has followed a conservative, incremental approach. To date, seven flight tests have been performed to gather data regarding the performance of both the fuel cell system and the aircraft. These flight tests have included low-altitude straight-line and higher-altitude circuit flights. Data gathered from flight testing validates the conceptual design models and the more refined system models. Future tests will focus on evaluating the rate of climb,

endurance, maneuverability, and power requirements of the aircraft with the goal of further refinement of design techniques and modeling tools. REFERENCES 1. Anderman, M. Brief Assessment of Improvements in EV Battery Technology since the BTAP June 2 Report, California Air Resources Board, 23. 2. Burke, K. A., High Energy Density Regenerative Fuel Cell Systems for Terrestrial Applications, NASA/TM-1999-29429. 3. Bradley, T. H., Moffitt, B. A., Parekh, D. E., Mavris, D., 26. Validated Modeling and Synthesis of Medium-scale PEM Fuel Cell Aircraft, 4th International ASME Conference on Fuel Cell Science, Engineering and Technology, June 18-21 26, Irvine, California. 4. Scheppat, B. Betriebsanleitung für das brennstoffzellenbetriebene Modellflugzeug, Fachhochschule Wiesbaden, 24. 5. Kellogg, J., Fuel Cells for Micro Air Vehicles, Joint Service Power Expo, Tampa, Florida, May 2-5, 25. 6. Ofoma, U. C., and Wu, C. C. Design of a Fuel Cell Powered UAV for Environmental Research, AIAA- 24-6384. 7. Soban, D. S., and Upton E., Design of a UAV to Optimize Use of Fuel Cell Propulsion Technology, AIAA 25-7135. 8. AeroVironment Press Release, AeroVironment Flies World s First Liquid Hydrogen Powered UAV, June 28, 25. 9. Moffitt, B. A., Bradley, T. H., Parekh, D. E., Mavris, D., 26. Design and Performance Validation of a Fuel Cell Unmanned Aerial Vehicle, AIAA 26-823. 1. Merchant, M. P., and Miller, L. S. Propeller Performance Measurement for Low Reynolds Number UAV Applications, AIAA 26-1127. 11. Phillips, W. F., Mechanics of Flight, John Wiley and Sons, Inc., Hoboken, New Jersey, 24. 12. Adkins, C. N., Liebeck, R. H., Design of Optimum Propellers, AIAA 21st Aerospace Sciences Meeting, Reno, NV, January 1-13, 1983.