Performance Degradation And Life Prediction Study Of Proton Exchange Membrane Fuel Cell

In order to solve the global energy shortage problem,save energy and reduce emission,proton exchange membrane fuel cell（PEMFC）using hydrogen as energy source,which have the advantages of high power density,high efficiency and no pollution,have been widely applied in the field of portable,mobile power and fixed power generation.However,fast performance degradation and short life have severely hindered the large-scale commercialization of PEMFC.In this regard,this paper uses a model-based and data-driven approach to establish a two-dimensional steady-state proton exchange membrane fuel cell simulation model in COMSOL to analyze the impact of fuel cell parameters on its performance.Designed human-computer interaction interface to make PEMFC simulation performance analysis easier to operate.Based on artificial intelligence-related algorithms,a model was established to predict the life of the PEMFC stack.The following is the research content of this paper:Proceeding from the overview of proton exchange membrane fuel cell,the categories of PEMFC and their application scenarios are explained.The basic components of PEMFC and their related functions are introduced.The working principle of PEMFC is explained.The indicators describing the performance of PEMFC are proposed.The COMSOL,MATLAB and COMSOL co-simulation and GUI modeling simulation tools used in this paper are introduced.The above work laid the foundation for the performance degradation analysis and life prediction work below.Based on the actual fuel cell related parameters and polarization curve test data,a two-dimensional steady-state PEMFC simulation model was established in COMSOL,combined with mass,momentum,composition,energy conservation equations and current distribution,water transfer models in proton exchange membranes.The fuel cell（stack）model was subjected to grid-independent division and experiment data verification,etc.Finally,a PEMFC model was obtained in accordance with actual working conditions.Based on the Live Link ? for MATLAB interface in COMSOL,the MATLAB GUIDE toolbox is used to design a human-computer interaction interface for PEMFC model performance analysis,which makes the analysis of battery performance degradation more convenient.Aiming at the analysis of the influencing factors of PEMFC performance degradation,this paper reviews the degradation mechanism of proton exchange membranes,catalysts,and gas diffusion layers.The thickness and conductivity of the proton exchange membrane,the electrochemical specific surface area,thickness,and porosity of the catalyst,the thickness,porosity,and permeability of the gas diffusion layer,and the diffusion coefficient and transmission coefficient of the reactive gas were analyzed.Then the impact of four driving conditions on the performanc of PEMFC is analyzed,which provides theoretical support for the improvement of battery performance and optimization design.The least squares and back-propagation neural network algorithm was used to build the reference model of PEMFC stack voltage decay,and the PEMFC stack voltage prediction model established by the autoregressive moving average time series algorithm was compared with them.A PEMFC stack power prediction model is established and the predicted value is compared with the value calculated by the PEMFC simulation model.The comparison results meet the accuracy requirements and the models can be used for the lifetime prediction of PEMFC stack.