Observer for State and Parameter Estimation in PEM Fuel Cells

Abstract

Water management in polymer electrolyte membrane fuel cells (PEMFCs) is one of the most challenging issues affecting PEMFC efficiency and lifetime. Liquid water active control and supervision techniques are limited by the absence of online sensors that can estimate the liquid water saturation. A promising technique that can be applied in this context are state observers. However, fuel cell models are related with nonlinearities, model uncertainty and sensor noise, which poses as major difficulties in observer design. In this project, a novel nonlinear adaptive observer algorithm will be developed to estimate online the states and unknown model parameters related to the PEMFC water dynamics. The algorithm is shown to provide an accurate estimation even in the presence of sensor noise and model inaccuracies. The convergence of the observer is proven through formal Lyapunov arguments. Moreover, the algorithm is validated through numerical simulations and experimental data.