State-of-health estimation for lithium battery in electric vehicles based on improved unscented particle filter

Abstract

This paper proposes an effective method to estimate the state of health (SOH) of a lithium-ion battery based on the ohm internal resistance R0. Unlike other estimation methods, this work considers the variation of R0 with the state of charge (SOC). The improved unscented particle filter (IUPF) is presented to track and predict R0. That is, an unscented Kalman filter (UKF) is used to generate an importance probability density function in the particle filter, and a method to select the fittest particle in the resampling stage is proposed. Based on the experimental data, a second-order resistance-capacitance equivalent circuit model is set up and the parameters are identified. To verify the accuracy of the proposed method, UKF and IUPF are compared in the prediction of R0 at different SOC points under the same cycle and at the same SOC point of different cycles. The results show that IUPF has certain advantages, and the SOH estimation error is always less than 3% during the charge-discharge stage.