Motor Over-temperature Fault Estimation and Prediction Method Based on Linear Support Vector Machine Algorithm

Motor over-temperature is a common fault type of new energy battery vehicles. Real-time acquisition and precise prediction of motor temperature have become an important approach to prevent motor temperature failure. Based on the vehicle data collected in the actual working conditions, we constructed multi-dimensional features, used the sliding window model to process the flow data, utilized the Linear Support Vector Machine (LSVM) algorithm in the field of machine learning to realize the classification and early warning of motor temperature fault. To test the accuracy of the model, we also constructed a comparative model for sensitivity analysis. We adopted the commonly used Random Forest and Logistic Regression methods to calculate four evaluation indexes for evaluating the LSVM method. The results show that the LSVM method is superior to the Logistic model and the Random Forest method in training time and training accuracy.