Research On Fractional Order Subspace Modeling And Predictive Control Of PEMFC

Energy shortage and environmental pollution have become the key issues which will constrain human sustainable development.Increasing the development and utilization of clean,low-carbon and renewable energy has become a main trend of energy development in various countries.As a clean and efficient new energy generation device,proton exchange membrane fuel cell(PEMFC)has many advantages,such as low starting temperature,high reliability,fast response,quiet and convenient for unit modularity.So,PEMFC has a broad application prospect.PEMFC is a complex system with multivariable,strongly coupled and non-linear characteristics.The accurate modeling methods are the basis for research on PEMFC,and the advanced control strategies are the key to improving power generation performance.Recent studies have shown that there are fractional characteristics in the dynamic processes of gas diffusion,heat conduction and electrochemical reaction in PEMFC.Therefore,the fractional order theory is combined with the subspace identification method(SIM)to establish the fractional order state space model of PEMFC in this paper.Based on the obtained model,a model predictive control(MPC)strategy for PEMFC is designed.The main research contents of this paper are shown as follows.(1)An integer order state space model of PEMFC is established.Aiming at the problems that there are lots of iterative optimization and computation in the traditional PEMFC multivariate modeling method,the SIM is introduced to modeling of PEMFC.According to the controllability of the factors which can affect the output characteristics of PEMFC system,the hydrogen flow and stack current are selected as the input variables of the identification model,and the output voltage and power are selected as the output variables.At last,a dual input and dual output integer order state space model is established of PEMFC,and the simulation results show the effectiveness the algorithm.(2)According to the fractional order characteristics existing in the PEMFC power generation process,the fractional calculus theory is introduced into the identification of PEMFC.And the fractional order state space model of PEMFC is established.The Poisson filter is adopted to filter the experiment data,which transforms the fractional differential of input and output data into the differential of Poisson moment function.Thus,the problem of fractional derivability of input and output data can be solved.In order to reduce the computation of the identification algorithm,the short memory method is introduced to transform the iterative calculation of the fractional differential into a product of the matrixes.And,the genetic algorithm is adopted for global optimization of the parameters existing in the fractional order subspace identification algorithm to obtain a better identification result.The simulation results show that the identification algorithm can not only describe the output characteristics of PEMFC well,but also improve the efficiency of identification algorithm.(3)Based on the fractional order state space model of PEMFC,a predictive control strategy for PEMFC is designed.The fractional order state space model of PEMFC is discretized by short memory method,and an augmented discrete state space model for control is obtained.The optimal control law of the closed-loop system is designed based on the augmented state space model.The Hildreth quadratic programming algorithm is introduced to calculate the optimal control law with constraints and meanwhile reduce the computation of the control algorithm.The simulation results show that the proposed control strategy can effectively improve the dynamic response characteristics of the PEMFC output power and reduce the running time of the control algorithm at the same time.