State estimation of a lithium-ion battery based on multi-feature indicators of ultrasonic guided waves

Ultrasonic non-destructive testing technology has been applied to battery state estimation applications to ensure the safety of the energy storage system. However, the accuracy and robustness of battery state estimation should be improved. In this paper, the state estimation of a lithium-ion battery based on multi-feature indicators of ultrasonic guided waves is studied. Piezoelectric ceramic ultrasonic probes with a fixed angle are used as the transducers. Eleven feature indicators of ultrasonic signals are analyzed. The appropriate feature indicators for battery state estimation are determined based on sensitivity analysis and correlation analysis. Considering the frequency response characteristics of the probe and the battery, the multi-frequency response characteristics of the battery are analyzed. Finally, seven feature indicators with multi-frequency excitation are selected. Subsequently, an adaptive machine learning model is designed to estimate the battery state. Based on the experimental results, the root mean square error (RMSE) of the battery state of charge (SOC) estimation result is less than 2.36 %. The applicability of the proposed method is verified by battery fully charged and non-fully charged experiments. Meanwhile, the method can quickly diagnose the side reaction process under abuse conditions such as overcharge and overdischarge, which provides a new method for non-destructive battery state evaluation. • Eleven ultrasonic feature parameters are analyzed for battery state estimation. • The multi-frequency ultrasonic guided waves on the battery are analyzed. • An adaptive fusion machine learning model is designed for state estimation. • The method is useful for non-destructive battery state evaluation.