Research And Experiment Based On Model Predictive Control Of The Permanent Magnet Synchronous Motor

Permanent magnet synchronous motor has been widely used in the field of AC drives, because of its excellent performances. For example, simple structure, reliable operation, high power factor and low loss. With society development, many motor occasions are in great demand for higher motor control performance. In the present vector control system, the ordinary PI control strategy can’t satisfy the need of high-performance control motor precise mathematical model. Because the ordinary PI method has many defeats, such as, easily be affected by uncertainties, and difficult when adjusting parameters.Model Predictive Control (MPC) as an intelligent control algorithm has features:rolling optimization and feedback correction. Therefore, the model predictive control can achieve very high control performance and has been successfully applied to the field of process control, for MPC’s wonderful ability to inhibit parameters change and uncertainty disturbance. The research based on MPC in the motor field is vastly insufficient, however, it is a very important and significant research direction that’s worth trying in motor areas. On the basis of PMSM’s state space equations and vector control algorithm, this thesis bring in MPC to overcome the disadvantages of PI method.Firstly, it brings MPC algorithm in current loop that is traditionally used in speed loop and designs model predictive controller respectively in the speed loop and current loop, and call it as "double-loop MPC". The advantages of double-loop MPC is that, the control parameters and control periods can be designed differently. For that speed changes slower than current, so the current control period is set generally smaller than speed control period. Secondly, the difference between the speed and current control period is becoming smaller and smaller, as a result, two control loops can be replaced by a single loop which is called "single-loop MPC". The nonlinear model of PMSM is made linear, and the constant term of the equations is eliminated in the thesis. To some extent, single-loop MPC can simplify the system and improve the dynamic and static performances. In addition, in order to explore how the parameters affect the performances of the system, this thesis conduct theoretical and simulation analysis and sum up a set of detailed parameter tuning rules. And so, it becomes more quickly and easily to get better parameters that is suitable for the system. Finally, the thesis conduct MATLAB/Simulink, design Model predictive control algorithm programs for PMSM and Labview programs used to watch the experimental waveforms in real time, and set up experiment platform based on TMS320F2812 chip to verify whether the MPC algorithm is feasible and effective, and then compare the results of PI and double-loop MPC. The simulation and experiment results prove the effectiveness and superiority of Model predictive control, get ideal control performances of PMSM and achieve the desired consequence of the subject.