Research On Hybrid Fault Diagnosis Approach Of Hub-driven Electric Vehicle Actuators

Compared with the traditional combustion motor vehicle,in-wheel-drive electric vehicles have the advantages of large available space,high transmission efficiency,and excellent control performance.They are very promising directions in the electric vehicle industry.As a typical over-actuated system,the number of actuators increases,which also increases the possibility of fault.When the actuator fails,the vehicle’s dynamic performance and handling stability will be affected,and even serious traffic accidents will occur.Due to the need for active safety technology for in-wheel drive electric vehicles,actuator fault diagnosis technology has received widespread attention from many automotive manufacturers and academics.Under this background,this paper takes the in-wheel-drive electric vehicle actuator as the object,and implements the hybrid fault diagnosis method of the actuator for the model-based actuator fault diagnosis method in the small slip rate region.Research to improve the accuracy of fault identification..The primary contents of this dissertation are as follows:In this paper,according to the needs of the fusion of model-based and data-driven hybrid fault diagnosis methods,a vehicle dynamics model and a Car Sim/Simulink vehicle model are established.The former is used for the research of model-based fault diagnosis methods,and the latter is used to simulate the real vehicle to obtain the vehicle state data accurately measured by the sensor.The model is verified by simulation tests.The results show that The model is reliable and accurate.Then,the fault diagnosis of the actuator of the hub drive electric vehicle is studied based on the parameter estimation method.By analyzing the fault of the actuator,the fault parameters of the actuator are introduced and the types of faults studied are determined,and the fault diagnosis of the actuator is converted into the estimation of the fault parameters.The key point of the model-based actuator fault diagnosis method is to divide the task of fault parameter estimation into two regions.The small slip region corresponds to the linear tire model and the non small slip region corresponds to the Dugoff tire model.In this way,it is considered that it is difficult to estimate the road adhesion coefficient when the tire is in the area of small slip ratio.The tire force is calculated by the above combined tire model,so as to estimate the fault parameters of each actuator and realize the fault detection,isolation and identification of the actuator.Because model-based actuator fault diagnosis methods have higher requirements for the accuracy of mathematical models,when the mathematical model is not accurate enough,the accuracy of fault diagnosis will also be affected.In this paper,research on the model-based actuator fault diagnosis method shows that the fault parameter estimation accuracy in the small slip rate region is low.therefore,a fault identification method based on support vector regression is proposed,which is an extension of the model-based fault diagnosis method to improve the accuracy of fault identification in the area of small slip rate.Based on the layered fusion approach,a framework for hybrid fault diagnosis of actuators based on model and support vector regression is established.In this paper,aiming at the hybrid fault diagnosis method for the actuator of hub driven electric vehicle,the simulation test is carried out under three different conditions of high,medium and low road adhesion coefficient.The simulation results show that compared with the single model-based fault diagnosis method,the hybrid fault diagnosis method can reduce the mean square error of fault identification from 10^-3 to10^-4,and improve the accuracy of fault identification.