Prediction method for battery self-discharge voltage drop based on pre-classifier

The self-discharge voltage drop (SDV-drop) is an important indicator in measuring the performance of lithium-ion batteries. Traditional SDV-drop measurement methods are time-consuming and require considerable manpower and material resources. This study proposes a method for predicting battery SDV-drop based on pre-classifier. The features for predicting the SDV-drop are obtained in two ways: direct extraction from the charge–discharge curve and generation based on the classifier. Subsequently, the features are input into the BP neural network model optimized by the Particle Swarm Optimization (PSO) and Levenberg–Marquardt (L–M) algorithms to predict the SDV-drop. Simulation results show that the proposed method can accurately estimate the SDV-drop. • A fast and economic self-discharge voltage drop prediction method is proposed. • The self-discharge voltage drop is predicted by pre-classifier. • Features are extracted from the charge–discharge curve. • The method is verified by experiment with high accuracy.