Fuzzy modeling of fuel cell based on mutual information between variables

This paper discusses estimation analysis study of fuel cell using Kalman filter and modeling of the output variables using fuzzy technique based on Gastofan-Kessel (GK) clustering algorithm. The objective of fuzzy model is to represent the dynamics of output variable as a function of most relevant set of input variables. The choice between most relevant and non-redundant input variables is based on mutual information theory. The fuel cell stack system is considered to have important interacting modules. The analytical equations are used to generate data to develop fuzzy model which closely flows the variations of output variables. The dynamic characteristics are predicted taking into account the reactant pressures, their mass flow etc. The performance of identified fuzzy model is tested at different initial load conditions. An analysis of the results based on the computation of the relative errors and on graphical representations is discussed.