Fractional nonlinear subspace modeling for proton exchange membrane fuel cell

Aiming at the fractional and nonlinear characteristics of proton exchange membrane fuel cell (PEMFC), a fractional subspace identification method is proposed to establish a nonlinear state space model of PEMFC. Firstly, in order to reduce the modeling complexity, the model input variables are determined by the typical correlation analysis method and correlation analysis method. Secondly, combined with the fractional differential theory and the Hammerstein model subspace identification method, a Poisson moment function is adopted to preprocess the input and output data to construct the input and output matrix of the subspace identification method. Then, a fractional order short-term memory method is introduced to reduce the computational complexity of the identification algorithm. Finally, the polynomial is selected as the static nonlinear link in the front of the Hammerstein model, and a fuzzy genetic algorithm is adopted to optimize the system of fractional orders and coefficient matrices. The simulation results verified the effectiveness of the proposed identification method which can shorten the computation time, and the obtained PEMFC identification model can accurately describe the dynamic process of PEMFC. © 2019, Editorial Department of Control Theory & Applications South China University of Technology. All right reserved.