Health Evaluation of Lithium Ion Battery Based on Weighted Kalman Filter Algorithm

Abstract

AbstractWith the wide application of lithium ion batteries in various fields, the safety and reliability of lithium ion batteries have been put forward higher requirements, and the health evaluation of lithium ion batteries is very important. In this paper, a new health evaluation method for lithium ion batteries based on weighted kalman filter algorithm is proposed by investigating and analyzing the existing health evaluation methods for lithium ion batteries. Based on the general kalman filter, the weighted kalman filter algorithm was proposed to evaluate the health of lithium ion batteries by constructing the battery SOH double-exponential recession model and the gaussian-type feature correlation mapping model for the health characteristics of lithium ion batteries. Four lithium ion battery data sets provided by NASA were used to simulate and verify the proposed health evaluation method. The verification results show that the health evaluation method of lithium ion battery based on weighted kalman filter proposed in this paper has better evaluation accuracy than the ordinary kalman filter method, with an average percentage error of 0.61%. Moreover, the average absolute percentage error of the health evaluation method for different types of batteries was less than 0.9%, and the method was applicable to all types of lithium ion batteries.KeywordsLithium ion batteryHealth evaluationWeighted kalman filtering