Search

Lithium-ion batteries experience capacity decline or even deterioration during the working process. Effective estimation of battery health status is a key challenge in the development of battery management systems. This paper proposes a method for estimating the state of health (SOH) of lithium-ion batteries based on the fusion of data-driven models and characteristic parameters. Using differential thermal voltammetry(DTV) to preprocess the experimental data of lithium-ion batteries, this method extracts six useful features, and establishes a SOH estimation model based on two-step Gaussian process regression (GPR) with different kernel functions. The results show that the established model can better approximate the experimental value and shorten the training and prediction time. The average absolute error of SOH estimation is 0.67%—0.97%, which is 20%—30% lower than that of single-step GPR. Therefore, the model has a high robustness and accuracy in estimating the state of health of lithium-ion batteries.