Parameter estimation algorithm of H-100 PEM fuel cell

Abstract

Polymer electrolyte membrane fuel cells (PEMFCs) have been recognized as one of the most promising eneygy conversion devices for commercial application due to their specific advantages, such as low operation temperature, zero pollutant emission, and high efficiency, etc. Since PEMFC is a highly nonlinear system and some parameters are related to the operation condition, most existing models are difficult to accurately predict the PEMFC characteristics. Thus, it is necessary to exploit parameter estimation methods for PEMFC to online determine the unknown model parameters by using easily measurable data to obtain concrete models. Most of the parameter estimations schemes for PEMFC have been designed based on intelligent optimization techniques. However, optimization methods cannot address the estimation problem online since they focus exclusively on offline searching procedure, which introduces heavy computational costs in the practical implementation and thus cannot be used in the real-time applications. Therefore, this paper aims to exploit real-time adaptive parameter estimation methods for a nonlinear parametric PEMFC system.