A voltage correlation correction based robust detection and location method for internal short-circuit faults of lithium-ion batteries

Internal short circuits (ISCs) in lithium-ion batteries (LIBs) typically exhibit continuous, subtle, and unpredictable early characteristics, making them increasingly difficult to monitor under cell inconsistency and low-quality data conditions. To address this challenge, a robust diagnosis scheme based on voltage correlation compensation and kernel independent component analysis (KICA) is proposed. First, a modified correlation coefficient and KICA-based abnormal sample separation strategy is designed to extract normal samples affected by cell inconsistency. Then, the voltage relative slopes between each cell voltage and the reference voltage are updated based on the normal samples. Within each interval, the voltages are corrected using the updated voltage relative slopes, and the residuals between the corrected and reference voltages are further modeled using the KICA model to detect ISC faults. Finally, the faulty cell is identified and located using a cumulative relative partial contribution rate method. Experimental results on a real LIB pack platform demonstrate that the proposed method exhibits strong adaptability under conditions of cell inconsistency and poor data quality.