Lithium battery charge state estimation based on improved Unscented Kalman filtering

Abstract

The state of charge (SOC) of lithium batteries is one of the key parameters to ensure their safe operation. In response to the traditional untraceable Kalman filtering (UKF) algorithm in the process of estimating the state of charge (SOC), the covariance matrix is non-positively determined, which leads to the termination of the algorithm. In this paper, an improved trace-free Kalman filtering method with singular value decomposition Unscented Kalman filtering (SVD-UKF) is proposed to estimate the SOC. singular value decomposition is used instead of Cholesky decomposition to improve the accuracy and stability of SOC estimation. First, a second-order RC equivalent circuit model is established, a lithium battery experimental platform is built to obtain charge/discharge data, and the parameters of the second-order RC model are identified by combining the hybrid pulse charge/discharge test and the 1stopt software, and then the battery SOC estimation is carried out by using the improved untraceable Kalman filtering algorithm, and the SVD-UKF estimation results have a smaller error with the actual value compared with the traditional UKF through experimental analysis, the estimation accuracy is high, and the real value can be converged quickly when the initial value is inaccurate, and the average absolute error is reduced by about 14.5% compared with the traditional UKF, which has good accuracy and robustness.