Study On The Estimation Of SOH And Life Prediction For Li-ion Battery Packs

As the only power source of electric vehicle,power battery has a direct impact on vehicle performance.In order to guarantee the power battery efficient working condition for a long time,it is necessary to monitor the health status of battery,real-time understand battery usage and status,provide the basis for battery monitoring and diagnosis,timely replacement of aging monomer,lengthen the service life of the whole;It will help ensure the range of the vehicle,improve the power and stability,etc.Therefore,it is necessary to accurately estimate the battery state of health,so as to realize RUL prediction.In the battery study,the estimation and prediction of SOH.The cell is the main research object,and using battery pack as the research object is scarce.Therefore,in order to improve the accuracy of miniscule estimation,a large number of complex optimization algorithms and redundant experimental data are used for tracking;Instead,at the expense of The real-time nature of the state estimation greatly,increases the estimated time and brings trouble to practical applications.In order to balance the relationship between real-time performance and accuracy,this article focuses on improving the SOH of battery estimation and prediction,accuracily,while taking real-time performance requirements into account.In this article,the main research contents:The main structure of battery performance is analyzed systematically.The factors both inside and outside that influence the battery SOH are summarized,and the main factors,affecting the battery SOH,are analyzed.By comparing each SOH evaluation indexs,it was determined that the battery resistance was used as the evaluation index.The advantages and disadvantages of the four typical equivalent circuit models of single battery,such as Rint,Thevenin,PNGV and dynamic Massimo Ceraol are compared,Comprehensively.From the point of view of adaptability and accuracy,and based on the Massimo Ceraol dynamic model,an equivalent circuit estimation model for the battery SOH second-order battery is established.At the same time,the parameter identification of the battery model is performed using the recursive least square method.Consider that the internal resistance of the battery has a high degree of nonlinearity and the dynamic changes are not easily tracked,it is not conducive to the tracking and prediction of the classical particle filter algorithm.Through the improvement of the algorithm-risk sensitive particle filter;under the second-order equivalent circuit model,The SOH state estimation of the battery group is carried out.The risk-sensitive particle filter algorithm for battery life prediction is proposed.Based on population growth model,the model of battery capacity recession was established.a battery life prediction framework based on capacity recession model and risk sensitive particle filter is put forward.Finally,the simulation results showed that the proposed battery-based SOH estimation method based on the second-order equivalent circuit model and risk-sensitive filter algorithm,and the battery life prediction framework based on the capacity recession model and the risk-sensitive particle filter algorithm can all meet certain estimates.Accuracy and prediction accuracy also showed a good real-time performance.