Optimization of Commercially Off the Shelf (COTS) Electric Propulsion System for Low Speed Fuel Cell UAV

Abstract

Fuel cell technology offers higher energy density than any other onboard electric power source. However, despite its high energy content per unit mass, the fuel cell is lacking in power density ratio compared to Lithium-Polymer battery which is commonly used for UAV application. Therefore, the design of the unmanned aerial vehicle and its propulsion system must be properly sized and matched. This work addresses this problem through a design-optimization approach. The work focuses on optimizing the selection of the propulsion system components for one specific unmanned aerial vehicle designed in house. The main goal is to reach maximum endurance with the AEROPAK fuel cell system as the power source. Each component of the propulsion system is modeled based on a physical model except for the propeller which uses an artificial neural network. All the models are simultaneously solved with the "fsolve" function in Matlab to evaluate endurance and act as a fitness function for the genetic algorithm. The genetic algorithm is implemented on the outer loop of the optimization as to search the optimal solution according to the fitness function. The optimization process selects the propeller, motor, and the reduction gear ratio that will yield the best cruise endurance for a given unmanned aerial vehicle design. The components used in the optimization are commercially available. It is found out that the low Kv motor and large diameter propeller offer higher endurance since the current requirement to produce the same thrust is lower. The hydrogen consumption is dictated by the amount of current drawn from the fuel cell stack. Through this methodology, the exhaustive tradeoff between these elements can be avoided. The results are then validated against the wind tunnel test data. Results show more than a 13% endurance increase in all cases studied including maximum speed, maximum range, and maximum endurance. Search Terms: genetic algorithm, fuel cell UAV, electric propulsion, design and optimization, multi-disciplinary analysis.