Designing an Intelligent Controller for Improving PEM Fuel Cell Efficiency

Maintaining the optimum performance of a PEMFC over a wide range of operating conditions is one of the greatest challenges in developing efficient and high performing fuel cell systems. This paper presents the effectiveness of neural network based intelligent controllers in regulating the partial pressure of Hydrogen, Oxygen and water under dynamic load conditions. Optimum values for these parameters are obtained using machine learning techniques on the simulation data obtained from the mathematical model of a PEMFC under various operating conditions of pressure, flow rate and humidity at the anode and cathode side of the PEMFC. Windrow Hoff algorithm and Neuro-Fuzzy controllers are used to attain the desired electrical performance of 25 V at 20 A from a 500 W PEMFC system. Genetic algorithm is basically used for validating the performance of the system. Thus an intelligent controller for optimum performance can be designed. The experimental results validates the process.