Lithium-ion batteries (LIBs) for transportation applications have different degradation characteristics depending on the driving and operating conditions. Batteries in electric vehicles cannot avoid high temperatures during a long drive. The uncontrolled temperature rise may cause thermal runaway leading to fire. The temperature rise may be the effect of an abnormal event such as the internal short circuit of the battery. It is highly required to discern the impact of the temperature rise on the state-of-health (SOH) of the battery. However, any abnormality inside the battery is not reflected in the SOH curve. Therefore, the detection of abnormality should be investigated separately. This paper proposes an incremental capacity (IC) based method for SOH prediction and abnormality detection at high-temperature conditions. The peaks of the IC curves are selected as health indicators (HIs). An autoregressive model is trained using IC-based His, and the trained model is used to predict SOH. The variation of peaks in the consecutive cycles is set to a threshold of 3-sigma to detect the abnormality.
