Multi-vector Model Predictive Flux Linkage Control Strategy For Permanent Magnet Synchronous Motor Drive System

Permanent magnet synchronous motors(PMSMs)have been widely used in many applications due to their advantages e.g.,small size and high efficiency.Generally,the conventional strategies of the PMSM drives(e.g.,vector control and direct torque control)may suffer from the issues of low system dynamics and unsatisfactory steady-state performance,which is not preferable for the applications that require high precision and fast dynamics.As an emerging control strategy,features like fast dynamics and easy incorporation of multiple input and multiple output(MIMO)systems and multivariable constraints makes model predictive control(MPC)attract much popularity in motor drives.However,several challenges of the existing MPC methods,e.g.,heavy computational burden and large torque ripples,should be addressed properly.In this paper,the surface-mounted PMSM(SPMSM)is regarded as the research object.And,with the goal of control performance enhancement and computational burden reduction,the MPC methods applied in PMSM drives are intensively studied.The works of this paper is organized as follows:In Section II,according to the working principle of PMSM,the mathematical models of PMSM in different reference frame are established.The voltage values corresponding to the switching state of a two-level inverter are given.Then,a brief introduction of vector-controlled SPMSM drives is provided,which is accepted as the cornerstone of the whole research in this paper.In Section III,the working principle,the basic characteristics and the cost function design of MPC method are detailed.Inspired from vector control strategy,the MPC methods applied in SPMSM drives can be classified three categories,i.e.,single-vector model predictive current control(MPCC),single-vector model predictive torque control(MPTC)and singlevector model predictive flux control(MPFC).The performance of the above MPC methods are investigated through simulations,and a comparison in terms of steady-state performance and system dynamics is made.In Section IV,considering the time delay in the digital controller may adversely affect the performance of the conventional MPC strategies,the performance of the two-step MPC strategy and the time-delay compensation unit is elaborated.Then,to deal with the issue of heavy computational burden,a cost function in the MPC methods is proposed,which has the ability of accurately tracking the voltage vector errors.Moreover,another issue of ripples in transit processes may challenge the performance of the MPC methods.The solution toward this is to increase voltage vector in every control period.With this,the steady-state performance improvement and the current ripple elimination are achieved.Simulations are performed to verify the effectiveness of the proposed MPC strategies.In Section V,according to the hardware-in-the-loop(HIL)test platform that consists of a real-time simulator(OPRT 5600)and a digital signal processor(DSP)controller(TMS320F28335),the performance of the above-mentioned MPC methods is experimentally examined.