Research On Parameter Robustness And Prediction Structure Optimization Of Model Predictive Control For Permanent Magnet Synchronous Motor

Nowadays,with the development of power electronic technology,model predictive control has been widely concerned as a new control strategy.Model predictive control is directly based on the physical model of the controlled object,and realizes the control purpose through the cost function containing a variety of constraints.It has obvious advantages such as intuitive concept,simple implementation method and excellent response performance,and has great development potential in the application of motor drive field.However,the control performance of model predictive control is extremely dependent on the accuracy of motor mathematical model and reasonable combination of weighting factors,which seriously restricts the development of model predictive control.Especially in the application field of electric vehicle drive system,the motor is in the operating environment of intense vibration,closed and large temperature variation,and the problem of parameter mismatch and external disturbance becomes more and more significant.The model predictive torque control and model predictive direct speed control are selected as the research objects in this manuscript.Combined with the control strategy of parameter updating mechanism,parallel predictive control structure and ultra-local model,the system can eliminate the weight factor and enhance the robustness of the system against parameter mismatch and external disturbance,which effectively solves the two main problems faced by model predictive control.The optimized control algorithm can be applied to more industrial scenes,which has a great promotion effect on the popularization of model predictive control algorithm,and has high research value and significance.The main research contents of this paper are as follows:(1)In order to reduce the sensitivity of the model prediction torque control performance to the motor parameters and improve the ability of the system to resist external disturbances.By analyzing the error between the predicted value and the measured value,the motor parameters can be updated in real time according to the variation of the error,and the mismatched motor parameters can be controlled in a small range.Then the model reference adaptive control method of the speed control loop was used to enhance the robustness of the system against the parameter mismatches and external disturbances.(2)In order to enhance the torque dynamic response ability of the model predictive direct speed control,the control objectives in the conventional cost function of the model predictive direct speed control are optimized,and the torque and flux error control terms are introduced into the cost function.Meanwhile,in order to eliminate the increased weight factors in the optimized cost function,the composite prediction error cost function is decomposed into independent forms of velocity error,torque error and flux error,and the purpose of eliminating the weight factors was realized by the optimized multiobjective parallel prediction structure.Combined with the extended state observer,lumped disturbances including the un-modeled part of the traditional model and external disturbances were introduced in the control process to improve the anti-disturbance ability of the system.(3)In order to further enhance the robustness of the system against parameter mismatch and external disturbance on the basis of eliminating the weight factor,the ultralocal model is introduced into predictive control,and the ultra-local models of model-free parallel predictive control and model-free hybrid parallel predictive speed control were established.Moreover,the expansion state observer is used to improve the observation speed of disturbance,so as to eliminate the weighting factor and improve the robustness of the system.(4)A motor experiment platform based on Micro Lab Box toolbox was built,and the selection of components and the principle design of the experiment platform were introduced.The proposed control algorithm was experimentally analyzed on the motor experimental platform,and the effectiveness and superiority of the proposed control strategy were illustrated.