An Indirect State-of-Health Estimation Method for Lithium-Ion Battery Based on Correlation Analysis and Long Short-Term Memory Network

State of Health (SOH) is a key indicator to describe battery's health status, which is important in terms of extending battery life, reducing failures and losses. Battery capacity are commonly used indicators for SOH estimation, but it's difficult to monitor capacity online because of battery working conditions. To overcome this challenge, this paper proposes an indirect method for SOH estimation. The study found a correlation between battery capacity and measured voltage during discharge. This correlation was confirmed using correlation analysis methods. The measured voltage sequence of the battery discharge phase is chosen as a health indicator to measure SOH. An LSTM network was used to create a model predicting battery capacity based on the health indicator. Optimization of LSTM model hyperparameters using whale optimization algorithm. The proposed method achieved a mean absolute percentage error of 0.53%, 11.38% higher than the traditional method.