Incremental Expansion Analysis and State-of-Health Estimation for Lithium-Ion Batteries

Accurate analysis of internal mechanisms and estimation of the state of health (SOH) are critical for safe and reliable lithium-ion batteries (LIBs) operation. Conventional health monitoring techniques such as the incremental capacity analysis (ICA) predominantly rely on electrical parameters, ignoring other sources of characteristics. These parameters are often sensitive to external noise and insufficiently reflect aging processes. Mechanical signals, due to their high sensitivity and easy accessibility, hold significant potential for revealing the aging state of LIBs. Hence, a novel method for analyzing the battery mechanical properties and estimating SOH based on expansion is proposed. The incremental expansion analysis (IEA) method is developed to explore the internal degradation mechanism under different working conditions, and the analysis reveals that the peaks on the IE curve exhibit strong correlations with SOH. Furthermore, a convolutional neural network- long short-term memory-attention mechanism (CNN-LSTM-Attention) model is built to estimate SOH by utilizing peak features as input, which aims to comprehensively consider the local feature information, temporal degradation trends, and feature relevance of battery aging. Experimental results demonstrate that the method achieves high estimation accuracy and robustness, with the root mean square error (RMSE), and mean absolute error (MAE) are within 0.18% and 0.12%, respectively.