Voltage Vector Rapid Screening-Based Improved Predictive Torque Control Of Permanent Magnet Synchronous Machines

Permanent Magnet Synchronous Machines(PMSMs)exhibit the characteristics of small size,high power density,high efficiency and high reliability,and are widely used in the field of AC drives with a bright future.However,the requirements for motor-based drive control are completely different in different situations.In automotive industry,large torque,wide speed and high power density are important indicators to measure the performance of vehicles.However,in the field of intelligent robotics,high precision,high reliablity and fast dynamic response are key criterion of robots' performance.Anyway,as every industry puts forward higher and more demanding demands for motor control,hence,sometimes traditional PI control is unable to meet all the demands simultaneously.Recently,as a relatively new advanced control strategy,Model Predictive Control(MPC)has the merits of simple principle and easy to deal with multivariable constraints.However,so far,MPC has just been successfully applied to the field of processing control,and it is insufficient in the field of motor control.Therefore,it's important and significant to deeply study the application of MPC in motor drive areas.In this thesis,a three-phase PMSM is chosen as a research object and its MPC strategies are deeply studied in this paper.Moreover,some improved strategies are proposed to overcome the inherent defects in MPC.The main content and fruits of this thesis can be divided into the following four parts:(1)The structure and working principle of the PMSM are introduced firstly.The mathematical models based on both the stator and rotor coordinate systems are established respectively,and the simulation model of the three-phase PMSM is built based on Maltab/Simulink.(2)The principle of the MPC strategies is introduced in detail.According to different variables,the MPC can be divided into three different types,namely,model predicted current control(MPCC),model predicted torque control(MPTC),and model predicted flux control(MPFC),respectively.Then,all the three typical MPC control methods are implemented and compared both in simulations and experiments.(3)It is well known that the traditional MPC has significant drawbacks.Aiming at these shortcomings,improved strategies are proposed respectively.1)A new cost function is proposed which can use the voltage vector as the only controlled variable instead of flux amplitude and torque,avoiding the process to tune the weighting factor.2)A rapid screening method is proposed to reduce the candidated vectors,which reduces the calculation burden.3)A duty ratio control method is proposed to alleviate the torque ripples of stator PMSM which can improve the control performance.(4)A PMSM-based drive system of is built including hardware and software,where two hardware systemes are constructed,including d SPACE-and DSP-based platforms.The experimental results confirm the validation and effectiveness of the proposed strategies for improvement of MPC controls.