Research On Model Predictive Control Strategy Of Five-phase Permanent-magnet Synchronous Motor With Strong Robustness

Multi-phase permanent-magnet synchronous motor(PMSM)has the advantages of high power density,excellent operating efficiency and strong fault-tolerant operation ability.With the development of power electronics technology and the improvement of the control processor manufacturing process,Model Predictive Control(MPC)has been widely used in recent years due to its intuitive and simple structure,fast dynamic response,and easy handling of multi-variable and multi-constraint conditions.It is a research hotspot in the field of control.Subsequently,MPC-driven multi-phase PMSM was applied in the field of motor drive and control.However,in the traditional control strategy,the disturbance caused by the mismatch between the real parameters of the motor and the MPC parameters after the equivalent modeling results in the deterioration of the predicted output performance and the decline of the control accuracy,which affects the reliability and safety of the electric drive control.It poses a huge challenge,so it is crucial to increase the robustness of the MPC system.To this end,this dissertation takes the MPC of five-phase PMSM as the research topic,and conducts research around the robust model predictive control strategy of the motor.The improved parameter online identification method is used to update the MPC in real time,so that the real motor parameters are consistent with the MPC parameters;then,the parameter identification is extended to the faulty motor,and the MPC parameters under fault-tolerant control are updated in real time.The proposed control strategy fundamentally solves the parameter mismatch problem on the basis of ensuring the consistency of the parameters of the control object,thereby improving the robustness of the MPC system.Then,the cost function is reconstructed by using the current error caused by the mismatch of the MPC parameters;the proposed control strategy is further optimized by combining the duty cycle modulation and the virtual voltage vector,so that the parameters can be solved without the aid of external parameter identification or perturbation of the observer.The mismatch problem,thus greatly avoiding the redundancy and uncertainty caused by external algorithms,and increasing the robustness of the MPC system.The main research contents and innovative achievements of the dissertation are as follows:1.Taking a five-phase permanent magnet motor as the research object,the structural characteristics of the motor are briefly explained using the finite element design process.On this basis,the mathematical model of the five-phase motor is established,the coordinate transformation of the control process is deduced,and the MPC principle of the driving motor is expounded.Then use Ansoft Maxwell,LCR measurement method,temperature sensor and other means to calculate the basic parameters of the studied motor offline.The calculated motor parameters are applied in MPC,and the limitations of traditional five-phase motor MPC are analyzed.2.Aiming at the problems that traditional MPC is sensitive to parameter changes and poor system robustness,an improved Model Reference Adaptive System(MRAS)parameter identification strategy is proposed.The method takes advantage of the strong robustness of fuzzy logic control and does not rely on accurate models,and combines the traditional MRAS identification method to design the adaptive rate,forming a high-precision and robust identification system.Then the identified parameters are applied to the MPC and the formula method maximum current control ratio,which can effectively reduce the copper consumption of the motor,enhance the robustness of the control system,and eliminate the negative impact of parameter mismatch on the control system.3.Aiming at the problem that the five-phase motor failure causes the degree of freedom to be reduced,and the parameter changes have a greater impact on the motor fault-tolerant control,a robust model prediction fault-tolerant control strategy is proposed.Firstly,the fault-tolerant control model after motor failure is constructed,and on this basis,the reference adaptive parameter identification of the fault-tolerant model is proposed,and the variation of d-q axis inductance parameters in different states is explored.Then,the identified motor parameters are applied in the fault-tolerant control,and the influence of the permanent magnet flux linkage parameters is eliminated by the incremental model,and the compensation of the resistance parameter changes is considered.Since the proposed control strategy considers the mismatch of all parameters,it can well increase the robustness of the fault-tolerant control system.4.A current error compensation mechanism is proposed to solve the problem of parameter mismatch of MPC system.The strategy reconstructs the cost function by exploiting the predicted current error caused by parameter mismatch,and incorporates optimized duty cycle modulation.The proposed control strategy can improve the vector selection accuracy after parameter mismatch,improve the MPC output performance,and increase the system robustness.5.Aiming at the problem of poor output performance and reduced control accuracy of MPC system caused by parameter mismatch,a robust MPC strategy based on virtual vector modulation is proposed.First,the influence of the parameter mismatch of the permanent magnet flux linkage is eliminated by the incremental model.Then,the virtual vector is used to eliminate the x-y space voltage amplitude,thereby reducing the error caused by the third harmonic current to the reconstruction cost function,and introducing the duty cycle to further optimize the action time of the virtual vector.