In Operando Health Monitoring for Lithium-Ion Batteries in Electric Propulsion Using Deep Learning

Battery management systems (BMSs) play a vital role in understanding battery performance under extreme conditions such as high C-rate testing, where rapid charge or discharge is applied to batteries. This study presents a novel BMS tailored for continuous monitoring, transmission, and storage of essential parameters such as voltage, current, and temperature in an NCA 18650 4S lithium-ion battery (LIB) pack during high C-rate testing. By incorporating deep learning, our BMS monitors external battery parameters and predicts LIB’s health in terms of discharge capacity. Two experiments were conducted: a static experiment to validate the functionality of BMS, and an in operando experiment on an electrically propelled vehicle to assess real-world performance under high C-rate abuse testing with vibration. It was found that the external surface temperatures peaked at 55 °C during in operando flight, which was higher than that during static testing. During testing, the deep learning capacity estimation algorithm detected a mean capacity deviation of 0.04 Ah, showing an accurate state of health (SOH) by predicting the capacity of the battery. Our BMS demonstrated effective data collection and predictive capabilities, mirroring real-world conditions during abuse testing.