Electric Vehicle Battery System Voltage Fault Diagnosis and Analysis Based on Kernel Density Estimation and Z-Score

Safety of the power battery system is crucial for the reliable operation of electric vehicles. By analyzing the voltage curve characteristics of battery cells, potential faults in the battery system can be effectively identified. This paper proposes a statistical voltage anomaly detection method. First, kernel density estimation is performed within a specified interval using a sliding window to extract anomaly score features, while Z-score features are extracted simultaneously. The combined features are then input into the LOF algorithm to identify whether a fault has occurred within the sliding window. Subsequently, analysis was conducted on real vehicle data, exploring the impact of different sliding window sizes and step lengths on algorithm performance. Finally, various types of faults present in multiple vehicles were analyzed and compared with other Fault diagnosis method, verifying the reliability and effectiveness of this method.