Research On Control Strategy Of Magnetic-geared Compact Motor For Hybrid Electric Vehicle Based On Model Prediction

Due to the structure of mechanical gear in the traditional automobile transmission system,abrasion is inevitable in the long-time operation,which not only increasing the cost of maintenance,but also reducing the control precision.The magnetic-geared compact motor(MGCM),designed as a combination of the permanent magnet synchronous motor(PMSM)and coaxial magnetic gear,could deliver great advantages of avoiding physical contact when transmitting torque and improving the space/utilization of transmission system.In the meanwhile,the model predictive control,a new control algorithm,can effectively inhibit parameters change and uncertainty disturbance.Owing to the lack of research in such control technology on the MGCM,in this thesis,the complementary MGCM is taken as a control object.Beyond the introduction of the structure and mathematical model,the MGCM-based model predictive control strategies are deeply analyzed,and the application of the motor in hybrid electric vehicles is studied and verified.The structure of this thesis is summarized as following:1?Introduce the basic structure and working principle of the MGCM.According to the outer air gap magnetic field orientation of the MGCM,the mathematical model can be represented to a combination of the PMSM and two mechanical ports.2?Introduce the direct torque control system and field-oriented control system of the MGCM and compare the dynamic response and steady response of both control strategies under different simulations.3?Propose the model-prediction-based MGCM control system and compare the dynamic and steady responses of traditional control and model predictive control under different simulations.4?Propose the sensorless-based MGCM model predictive current control.Build the simulation models from the model reference adaptive system and current predictive control.Verify the accuracy of the algorithm according to the relative errors of speed and position under different speeds and loads.5?Experimentally verify the impact of the model predictive control algorithm on motor output performance based on the dSPACE experimental platform.The research of this dissertation was supported by the National Natural Science Foundation of China(51637001,51507002,51607001)and the Natural Science Foundation of Anhui Province(1508085ME87,1708085QE108).