Design Of RCP Platform For PMSM And Research On DNN Control Strategy

With the rapid development of electric vehicle,the electric motor has become the main traction power of electric vehicle,and the research and system development of motor drive technology have become hot spots.The rapid iteration of intelligent algorithms and power electronics technology have led to the birth of many complete motor drive development platforms and excellent motor intelligent algorithms,which have improved the application level of motor drive technology.In this paper,the control strategy of MTPA based on DNN is studied,and the PMSM control platform based on RCP is designed and built.Firstly,the paper analyzes the application and development of electric motor,introduces the development of electronic control technology and control system,and the working principle of IPMSM.Through the IPMSM mathematical model,the key technology of motor vector control is studied,and the MTPA control principle is introduced in detail.The formula derivation and modeling simulation of MTPA control strategy are carried out.Then,the control platform based on the rapid development control prototype is designed to solve the problems in the research of PMSM drive technology,such as complicated resource allocation,complex algorithm programming and long research and development cycle.According to the working and control principle of IPMSM,a hardware and software platform for docking with d SPACE hardware in the loop real-time simulation system is built,including the hardware parts such as inverter,sensor and protective isolation circuit,and the software part such as algorithm simulation model based on Simulink and online debugging Control Desk interface.Circuit design,PCB drawing and proofing,welding debugging have been carried out successively,the experiment of the motor is carried out.The experimental results show that the design of the experimental platform is feasible,which provides a good development environment for IPMSM dynamic performance optimization,control strategy research and motor servo system pre-development.Finally,through the analysis of various optimization schemes of MTPA control strategy,aiming at the problem of motor torque fluctuation caused by nonlinear changes such as IPMSM parameter change and magnetic saturation,the MTPA control strategy based on DNN algorithm is studied.Because the artificial neural network has achieved remarkable results in solving non-linear function,many researchers have replaced the traditional algorithm with neural network algorithm The algorithm is used to deal with the nonlinear problem of motor parameters.In this paper,an IPMSM modeling technology based on DNN is proposed.Combined with the intelligent algorithm of deep neural network,the motor torque observer is established to analyze and predict the key parameters of the motor,and accurately simulate and evaluate the control performance of the motor drive system.The proposed MSM torque modeling technique does not rely on the modeling of nonlinear or nonlinear systems,thus reducing the difficulty of modeling the nonlinear system.Compared with the traditional MTPA control,the feedback torque data is more accurate and effectively improves the torque ripple problem.