Study Of Fuel Cell Performance Decay Prediction Based On Improved Rain Flow Counting Method

With the increasing demand for clean energy in human society,Proton Exchange Membrane Fuel Cell(PEMFC),which uses the reaction of hydrogen and oxygen to generate electricity,has occupied an increasingly important position.However,pemfcs also face some problems,such as rapid energy decay and short lifetime,which seriously limit their industrialization.In order to predict the fuel cell performance degradation more accurately,a modelling-based approach is used to build a three-dimensional steady-state constant temperature proton exchange membrane fuel cell model in COMSOL and to tune its parameters,so as to study the impact of performance degradation;the number of cycles and the depth of flushing and discharging during the fuel cell operation are the main factors affecting the cell life,and the traditional rain flow counting method is improved for this situation.Based on the improved real-time rainfall counting method,the cycle period is determined,and two models,Support Vector Regression(SVR)and neural network regression,are introduced to investigate the accuracy of life prediction of proton exchange membrane fuel cells under static and dynamic operating conditions.The main contents of document are as follows:The types and application scenarios of five typical hydrogen fuel cells,as well as the structure and working principle of proton exchange membrane fuel cells are analyzed in detail;the modeling software COMSOL and simulation software MATLAB used in this paper are explained.The above content provides a knowledge base for the subsequent study of fuel cell performance decay and life prediction realism assessment in this paper.Based on the model simulation studies of proton exchange membrane fuel cells at home and abroad in the recent decade,a three-dimensional steady-state constant temperature simulation model of the proton exchange membrane fuel cell is constructed by the simulation modeling software COMSOL,combined with mathematical models,and the model data are verified.The polarization curves are proposed to verify the effects of different parameters on the performance decay of the proton exchange membrane fuel cell.The thickness and conductivity of the proton exchange membrane,the thickness of the catalytic layer,the thickness of the gas diffusion layer,the porosity and permeability,and the diffusion coefficient of the reaction gas are used as examples to analyze their effects on the polarization curves of the pemfc and give reasonable suggestions to provide theoretical support for the study of fuel cell performance decay.The traditional rain flow counting method is improved by equating the cell charge/discharge current,depth of discharge and number of charge/discharge cycles to the applied load,load amplitude and number of load reciprocations in the fatigue life of the material according to the basic principle of the rain flow counting method and the characteristics of the fuel cell.A docking algorithm is added to the data processing process to extract the number of sub-cycles in real time,calculate the equivalent load for each sub-cycle,and then use two models,support vector regression and neural network,to portray the decaying trend of pemfc performance.In this paper,the fitted curves for static and dynamic operating conditions are derived from the predicted decay values of voltage,current and power,and compared with the real decay values.The validation results show that the proposed method can accurately predict the fuel cell lifetime,providing a new research idea for fuel cell lifetime prediction.