Identification of Nonlinear Multi Input Multi Output Model of PEM Fuel Cell Stack System

Fuel cells provide clean and efficient energy and recently there have been intense development by many companies in order to use in various applications (stationary or mobile). The performance, stability and robustness of electricity generation in these power sources are dependent on understanding and estimating the transient and steady state behavior of the fuel cell system. The goal of this project is to find a low order model and proper estimator for the multi input multi output (MIMO) nonlinear fuel cell stack system which can be used in control purposes in future studies. After analyzing transient state of fuel cell, the system is linearized about the operating range. Then a MIMO bilinear model is chosen among several ones and Recursive Least Squares (RLS) method is used to estimate MIMO bilinear model parameters. At the final stage, the measurement noise is added to the system, and instrumental variable (IV) method adopted for bilinear model is used to estimate the parameters and system simulation. This research is based on the data obtained by Matlab Simulation model by Dr. Pukrushpan [1]. In future studies data derived by the actual fuel cell stack system can be used.