New Energy Vehicle Battery SOH Evaluation Method Based on Charging High Quality Data Set Extraction

Abstract:
In the charging scenario of new energy vehicles, due to the complex and nonlinear electrochemical mechanism of lithium batteries in new energy vehicles, it is difficult to accurately estimate the battery health status by ordinary measurement methods. Aiming at the technical difficulties of poor quality data set and noise sensitive IC curve of data-driven method, an advanced smoothing method based on Adaptive strong tracking untracked Kalman filtering algorithm (ASTUKF) is proposed in this paper. Combined with grey correlation analysis and long-term and short-term neural network, the ASTUKF-GRA-LSTM online SOH estimation model of lithium battery for new energy vehicles is established. The experimental results show that the method is suitable for fast and accurate estimation of SOH of different types of power batteries, improves the calculation accuracy of batteries, and effectively meets the safety analysis requirements of accurate and rapid dynamic evaluation of DC charging piles for lithium battery health status of new energy vehicles.