Research On Parameter Identification Of Permanent Magnet Brushless DC Wheel Hub Motor Of Electric Vehicle

Due to the increasingly prominent of the resources and environmental issues, both each country and automobile manufacturer has to seek substitute for traditional motors that run on fuels oil. With significant advantages of high efficiency, low emissions, the electric vehicle has been regarded as one of the important traffic implements in the future. The performance of vehicle control is directly affected by the effect of the control strategy of motor for the electric machine by in-wheel motor drive. The design of drive motor control strategy is closely related to mechanical parameter(rotational inertia) and electric parameter(resistance,inductance and flux linkage) of electric motor. Therefore, online identification of the parameters has important significance in boosting the whole control performance of the electric machine. Based on the existing algorithm, this thesis has made the further study on parameter identification of permanent magnet brushless DC wheel hub motor and verified by simulation and experiment.Firstly, this thesis analyzes the mathematical model of permanent magnet brushless DC wheel motor and double closed loop control model is constructed under the matlab/simulink environment. In addition, the author has performed sensitivity analyses to the electric machine parameter of the effect of the electric machine control strategy which was influenced by the model.Secondly, the electrical parameters of the permanent magnet brushless DC wheel hub motor are measured offline, which provide parameters criterion for the identifying of online simulation. In view of the disadvantages that identification consequence of forgetting factor recursive least square fluctuates easily, an improved least square identification is put forward by combining forgetting factor recursive least squares algorithm with ordinary least square algorithm and then apply it to parameter identification of electric machine rotational inertia. In allusion to the disadvantages that PSO algorithm are easily fell into the status of local optimum, average best location and Cauchy mutation particle swarm optimizer(AMPSO) are put forward and also apply it to the parameter identification of electric machine.It was improved by simulation that the accuracy of parameter identification can be enhanced by the improved identification algorithm.Lastly, the electrical machine system is designed by the center of dsPIC33FJ64mc804 and it realized identification of electrical machine's parameter by utilizing the two improved algorithms mentioned above on motor control system experimental platform. The experimental result indicates the accuracy of the algorithms.