| PEMFC has been applied in many industries,but PEMFC still faces two major bottlenecks,namely limited lifetime and high cost.With great efforts,domestic and international fuel cell technologies have made great progress,and the cost is measured to be down to$2000.00/k W.Meanwhile,the accurate prediction of PEM cell lifetime has been widely studied,as it can both point the way to improve durability and reduce its operation and maintenance cost.In this paper,based on the durability test data of PEMFC with full life cycle up to7016 h and the data of electrochemical characterization every 100 h,a MEA performance degradation model is established by thermodynamics and electrode kinetics theory;based on the model parameters obtained from the test such as electrochemically active area,internal resistance,leakage current,limit current,etc.,the FC voltage value is calculated and the extrapolation method is used to predict the PEMFC life,and then calculate and analyze the main factors affecting the performance degradation;considering the errors caused by random factors,particle filtering is performed on the model parameter data obtained from the actual test,followed by the prediction of the remaining battery life and analysis of the influence of each factor on the performance degradation,and the following conclusions are drawn.（1）FC performance is basically stable or even slightly increasing at 0～1000 h,then decreases at 1000～2000 h,then performance recovers at 2000～3000 h,performance is stable at 3000～6500 h,and FC performance decreases and cannot be recovered after 6500 h until the end of life.the changes of FC performance at different stages are the joint effect of degradation parameters,that is,the degradation characteristics of ECSA,,R,at different stages lead to different stages of FC performance change.（2）Based on the model parameters obtained from the tests such as electrochemical active area,internal resistance,leakage current,limit current,etc.,the calculated voltage values of the cell basically match with the measured voltage values before 4000hours,but after 4000 hours,the calculated voltage values gradually deviate from the measured voltage values,but the change trend of both is the same.The PEMFC voltage data calculated based on the decline mechanism before 3210 h was selected for linear fitting,and the voltage decay rate was 6.45E-6 V·h-1,and the remaining service life of FC was extrapolated to 11007 h with this decay rate.（3）Particle filtering is performed on the model parameter data obtained from the actual measurement to establish a suitable prediction model,and the number of particles is chosen to be 5000 with comprehensive consideration of prediction accuracy and prediction time.With the extension of learning time,the built model can predict the remaining battery life well,so that the total RUL error is basically controlled within10%.（4）At different stages of the whole life cycle,the rate of change of the model parameters themselves and their influence on the FC performance degradation are significantly different.In the whole life cycle,the parameter with the largest contribution to the voltage decay is ECSA,while at the latter stage,the rapid increase,as well as the continuous decrease,their influence on the FC performance degradation is increasingly obvious.In the present experimental conditions,activation loss is the main cause of battery performance degradation.（5）Comparing the calculated results of experimental test data with the PF prediction results,including voltage comparison,RUL comparison,and comparison of the contribution of each parameter change to the voltage loss,the results show that PF can cut the influence of test process noise and make the voltage prediction results closer to the real value,while improving the prediction accuracy of RUL. |
PDF Full Text Request