Nonlinear modeling of PEMFC based on fractional order subspace identification

Abstract

AbstractAiming at the multivariable, nonlinear and fractional‐order characteristics of proton exchange membrane fuel cell (PEMFC), this paper presents a nonlinear state space model based on a novel fractional Hammerstein model subspace identification theory. To reduce the complexity of modeling and choose the suitable input variables, canonical correlation analysis (CCA) method is used to select the most influential factors as the model input variables, and correlation analysis (CA) method is employed to remove the redundant input variables. To guarantee that the input‐output data are derivable at different fractional order, a Poisson moment function (PMF) is employed to construct the fractional order Hammerstein model with a six‐order polynomial as the front static nonlinear unit. To improve computing speed, a fractional differential short memory method (SMM) is proposed to reduce the computation cost of the identification algorithm. Meanwhile, a fuzzy genetic algorithm is adopted to acquire the best fractional order. Simulation results show that the fractional subspace identifying method can avoid fuel cell's internal complexity and PEMFC identification model can describe the working process of PEMFC accurately and quickly, which will provide an ideal control model for some advanced controller.