Model Predictive Control Of Permanent Magnet Synchronous Motor Systems

Permanent magnet synchronous motors(PMSM)are widely used in the national defense industry,manufacturing,traction drive,etc.,because of advantages of simple structure,high power factor,and high efficiency.Model predictive control(MPC)has drawn more attentions due to its excellent dynamic control performance in PMSM drives.However,MPC-PMSM drives still have some problems in steady performance,parameter robustness,disturbance attenuation and so on.Therefore,this dissertation carries out a series of studies that aim at MPC-PMSM drives fed by two-level voltage source inverters.This dissertation first elaborates the basic principle of MPC,including its modeling method,cost function and structure design.On this basis,two kinds of MPC-PMSM strategies--Finite control set(FCS)and continuous control set(CCS)based model predictive control strategies are introduced in this dissertation.Furthermore,main problems and limitations of FCS and CCS-based MPC are discussed from three aspects,i.e.algorithm complexity,parameter robustness and disturbance attenuation ability.The multi-step FCS-MPC could reduce the switching frequency and improve the efficiency of the motor drive system while ensuring good steady and dynamic performance.However,the conventional multi-step FCS-MPC employ an over-complex predictive model and exhaustive optimization method with exponential time complexity,so it is difficult to realize in a short control period.This dissertation proposes a computationally efficient multi-step FCS-MPC for two-level voltage source inverter fed PMSM drives.In the proposed strategy,in order to reduce computation time in multi-step predictive process,a simplified multi-step predictive model is established according to reasonable simplification;a novel optimization method with logarithmic time complexity is proposed to shorten the optimization time for cost function.Combining the above two measures,the proposed strategy can be achieved in a relatively shorter sample period.The performance of CCS-MPC is directly related to the predictive model.If there are unmodelled periodic disturbances in the actual system,MPC will be difficult to suppress the disturbances,thus causing fluctuations of system output.To solve this problem,this dissertation proposes a predictive-integral-resonant control(PIRC)strategy.Compared with CCS-MPC,the proposed PIRC could enhance the suppression ability for disturbances by embedding the internal model composing of the integral and resonant loop.Furthermore,this dissertation applies the proposed PIRC to PMSM drives,and proposes the PMSM control strategy based on the cascaded PIRC,which could suppress periodic disturbances caused by the dead time effects,current sampling errors,and so on.The experimental results show that the PIRC can suppress periodic disturbances in the drive system,thus ensuring good current and speed performance.In the light of limitations of FCS-MPC and CCS-MPC,this dissertation proposes a novel control-set named extended control-set(ECS).In ECS,a new voltage vector synthesis method is deigned,and the corresponding predictive model and cascaded predictive algorithm are set up on the basis of this method.According to ECS,a modified MPC control strategy named ECS-MPC is proposed in this dissertation.The experimental results show that ECS-MPC has excellent dynamic torque performance,and it also has more stable torque control performance and better execution efficiency at the same time.