External Short Circuit Fault Diagnosis Based on Supervised Statistical Learning

This paper proposes a fault diagnosis method for external short circuit detection based on the supervised statistical learning. The maximum likelihood estimator is used to capture the statistical properties of the fault and non-fault datasets, and the Gaussian classifier is applied to distinguish the two states. Validation experiments demonstrate the good performance of the proposed method in dynamic conditions. Compared to the prevailing fault detection methods, this method does not require extensive modeling work, determines the fault completely based on the data in existing fault occurrence, and can be adopted easily with the trend of big data and connected vehicles.