Li-ion battery state awareness via nondestructive vibration measurements

As the demand for the Li-ion battery (LIB) surges, ensuring battery safety becomes critical. Traditionally, voltage, current, and other electrical characteristics are monitored to evaluate battery performance. However, relying solely on these signals may not provide sufficient early warning, due to the rapid onset of thermal runaway in LIBs. Battery-state changes are reflected in coupled alterations in the battery electrical, thermal, and mechanical properties. This study investigates vibrational characteristics to monitor the health of LIBs. Vibration signals generated spontaneously by a LIB during normal charging/discharging cycles are collected using sensors placed on the surface of LIB cells, as well as during artificially induced overcharging and overdischarging states. Time series are preprocessed and advanced data processing techniques are used for data classification to identify deviations from normal battery health condition. This work observes vibration signal changes while LIB cells charging at multiple key voltages corresponding to different phase transitions. Vibration signals from LIBs in healthy, previously overdischarged and previously overcharged states demonstrate three distinct clusters, based on extracted features from various algorithms, each corresponding to a specific health state. These findings prove the applicability of employing vibration signals to effectively monitor battery health in a nondestructive way. • The study adopts vibrational features to monitor the health of Li-ion batteries. • This diagnostic method effectively distinguishes LIBs under different conditions. • Vibration signatures reveal priorly misused batteries, even under normal operations. • Diagnostics align well with cathode material phase transitions observed through XRD.