Fault Diagnosis of Permanent Magnet Motor Sensor of New Energy Vehicles in Complex Environment

In the motor driving process of new energy vehicles, it is easy for permanent magnet motor sensor to be in abnormal working state and fault state due to the influence of motor system noise, sensor operating characteristics and poor contact of various components. The complexity of electric vehicle motor operating environment leads to the difficulty of fault diagnosis. Therefore, this thesis puts forward the method of sensor fault diagnosis for permanent magnet motor of new energy vehicle. By analyzing the mechanical movement mode of permanent magnet motor, a mathematical model of permanent magnet motor for new energy vehicles, which is combined with sliding mode observer to capture the sensor fault state of permanent magnet motor at any time, is constructed. Based on this, the feature quantity of the sensor fault state is extracted and input into the support vector machine. According to classification results of the support vector machine, the sensor fault diagnosis of the permanent magnet motor of new energy vehicles can be realized. The experimental results show that the diagnosis structure of the research method is consistent with actual situations for the constant speed form and variable speed form of new energy vehicles.