Detection of Abnormal Clutch Connection Status of Hybrid Vehicle Engine Based on Data Drive
When the generator clutch does not successfully complete the engagement or separation action,the vehicle cannot complete the mode conversion between the electric vehicle mode and the hybrid vehicle mode,and the engine cannot correctly achieve power transmission.Therefore,data-driven technology is used to detect the abnormal connection state of the engine clutch.First,the vehicle power structure and the behavior of the engine clutch are described,and then data difference and pattern extraction are used to preprocess the data.Then,the anomaly detection model training method based on four machine learning algorithms,namely multi-layer perceptron,long short-term memory,convolutional neural network and support vector machine,is presented.Finally,the performance of the model is verified by the actual vehicle test data and the prototype platform test data of Beijing Benz Automobile Co.,LTD.The experimental results show that the performance of the long short-term memory algorithm and convolutional neural network is poor,while multi-layer perceptron and support vector machine can better diagnose the abnormal connection state of the clutch;and the support vector machine has the best performance,with the highest accuracy of 98.7％.
NETL
NSTL
万方数据
