State of charge estimation of lithium-ion batteries based on ultrasonic guided waves by chirped signal excitation

Data -driven methods for the state of charge (SOC) estimation of lithium-ion batteries have emerged as a prominent research topic, but currently they are limited by the reliance on electrical features in certain scenarios. As a result, ultrasonic detection technology has gained popularity in battery diagnostics in recent years due to its nondestructive, efficient, and high -precision characteristics. Nevertheless, the selection of an optimal excitation frequency for ultrasonic signal remains an ongoing challenge. To obtain the optimal frequency quickly and easily, this paper presents a SOC estimation method for lithium-ion batteries using ultrasonic guided waves excited by a chirped signal to avoid time-consuming and tedious single-frequency excitation tests at different center frequencies. Four center frequencies with high modal purity and large amplitude are selected through time-domain analysis and the Hilbert-Huang transform. The signal local amplitude (SLA) and time of flight (TOF) serve as two ultrasonic features for capturing the internal mechanical dynamics of the battery at different SOCs. SOC estimation is performed using a long short -term memory (LSTM) neural network based on the extracted ultrasonic features. Experimental results on two pouch lithium-ion battery cells in 14 cycles showed that the proposed method achieves the highest accuracy at 60 kHz on Cell 2, with a root mean square error of 3.07 %.