Dynamic Modeling and Parameters Identification of Hydrogen-Oxygen PEM Fuel Cell with Integrated Humidifier Model

Fuel cell is a type of electrochemical energy converter that converts chemical energy stored in fuel into electrical energy. Nonlinear structure, time-varying dynamics and uncertain physical parameters are the challenges of working with polymer electrolyte membrane (PEM) fuel cells. In this paper, grey-box modeling and system identification of flow-through H2/O2 PEM fuel cell with three cells and integrated humidifier is investigated. First, zero-dimensional nonlinear fluidic, thermodynamic and electrochemical modeling of PEM fuel cell is performed. The fuel cell model presented in this research is Multi-Input-Single-Output type. In the following, constant parameters of the studied fuel cell are determined. The system identification process as a multi-input-single-output system is done based on the Prediction-Error minimization, using the method of Trust-Region Reflective Newton. Finally, validation of the obtained model is done with experimental data that not used in modeling. The considered PEM fuel cell was tested under different conditions of temperature, reactant gas inlet pressure and stack current, and 329,420 experimental data were obtained. According to test conditions, data was classified into 189 different modes. The results show that the average voltage error of the identified model compared to the experimental data is equal to 1.03%. Moreover, the correlation between voltage and identified model parameters has been investigated, and the results showed that correlation between voltage and the contact resistance equivalent is higher than coefficients of orifices.