A fault detection method of electric vehicle battery through Hausdorff distance and modified Z-score for real-world data

Conventional fault diagnosis methods are tough to detect early faults when the abnormal characteristics of the battery are not obvious. The main purpose of this manuscript is to propose an online fault detection method for lithium-ion battery pack based on the combination of Hausdorff distance and modified Z-score. It enables the detection and location of the internal short circuit fault of the battery pack by detecting the Hausdorff distance between the voltage curve of each cell and the median voltage curve in a moving window. The validity of the algorithm is further verified by utilizing real-world data and compared with threshold method, Pearson correlation coefficient method, and Shannon entropy weighting method. The results show that the proposed method not only possesses higher reliability than these above but also does well without any model. In addition, the proposed method owns good robustness to the battery system with poor consistency and can be applied online.