High-precision state of charge estimation of electric vehicle lithium-ion battery energy storage system based on multi-scale optimized time-varying bounded smoothing variable structure filtering algorithm

State of charge (SOC) is a crucial parameter in evaluating the remaining power of commonly used lithium-ion battery energy storage systems, and the study of high-precision SOC is widely used in assessing electric vehicle power. This paper proposes a time-varying discount factor recursive least square (TDFRLS) method and multi-scale optimized time-varying bounded layer smoothing variable structure filtering (TSVSF) to obtain a more accurate SOC. Firstly, the TDFRLS algorithm is formed by introducing a time-varying discount factor, which effectively solves the problem of data saturation and simultaneous optimization of speed and accuracy in other improved RLS algorithms and is conducive to realizing high-precision identification of battery model parameters. Then, based on TSVSF, extended Kalman filter (EKF) gain is combined to ensure the stability and accuracy of the predicted system. In addition, the square root algorithm ensures the non-negative quality of error covariance matrix and effectively solves the non-convergence problem in the prediction process. Finally, the professional lithium-ion battery test platform is used to obtain the real-time parameters of the battery under different temperatures and working conditions, and comparative experiments of various SOC estimation algorithms are carried out. Among the four working conditions, the SOC estimation accuracy is the highest at HPPC at 35 °C, and the error is kept within 0136. The proposed algorithm has the highest accuracy and stability compared with four new algorithms. The experimental results show that this method has high accuracy and is significant for accurately estimating electric vehicle electricity.