Research On Drive Control Of PMLSM Considering Electromagnetic Parameter Asymmetry

Compared with combining rotating machines and ball/planetary screw structure,Permanent Magnet Linear Synchronous Machines(PMLSMs)realize the conversion between electrical energy and mechanical energy for linear reciprocating motion directly,which have the advantages of high power density,high precision,high speed and simple structure,etc.,and are widely used in applications such as numerical control machine,rail transit,lithography machine,etc.PMLSM has inherent longitudinal end effect,leads to additional detent force generated by primary and secondary besides cogging force,viz.,end force.In addition,the phase winding is no longer symmetrical,resulting in the asymmetry of the permanent magnetic flux linkage and the self-mutual inductance parameters of the stator winding,viz.,Electromagnetic Parameter Asymmetry(EMPA).Both the end force and the EMPA will affect the high-precision motion control of PMLSM.Based on this,this paper studies from three aspects.Firstly,the mathematical model of the motor with single and multiple EMPAs is constructed,and the mathematical model of the electromagnetic thrust of the motor is derived by the virtual displacement method.At the same time,a new salient-pole PMSLM structure is proposed and the optimization design is carried out with the detent force as the optimization target.Based on the optimized motor model,the correctness of the mathematical model is proved by finite element simulation.Also,Design and build a dual-core digital control platform based on TMS320F28379 D and a linear pull mechanical platform including the optimized prototype for subsequent prototype testing and control algorithm verification.Secondly,considering that the traditional Deadbeat Direct Thrust Control(DB-DTC)algorithm is difficult to be directly applied to the PMLSM with EMPA,especially for the salient-pole machine,a simple improved DB-DTC algorithm considering EMPA is realized by constructing the virtual coordinate system without EMPA and its bijective mapping with the real coordinate system.The proposed algorithm is simple in calculation,and the simulation and experimental results prove the effectiveness of the proposed algorithm.Also based on deadbeat principle,considering the even harmonic error problem when the traditional Deadbeat Predictive Current Control(DPCC)algorithm is directly applied to the PMLSM with EMPA,an improved DPCC algorithm considering EMPA is constructed.Furthermore,the parameter robustness of the traditional algorithm in the application of PMLSM with EMPA is analyzed,and then the robustness of the proposed algorithm is analyzed for the parameter mismatch of dq-axis inductance,permanent magnet flux linkage and asymmetric components.The results show that when the motor has EMPA,the parameter robustness range is smaller than that of symmetric motor.Based on this,a simple robust enhancement algorithm is proposed to improve the system robustness while maintaining the high response speed of DPCC.Simulation and experimental results show the effectiveness of the proposed algorithm.Finally,in order to realize the deadbeat operation considering EMPA,online identification of asymmetric components is a must,which can overcome the problems that cannot be realized by traditional identification algorithms.Firstly,the persistent excitation(PE)condition for the convergence of motor parameter identification is analyzed,and the minimum realization of PE condition is given.Considering that there are six parameters to be identified for EMPA motors and their strong nonlinear characteristics,a linear regression equation containing six parameters to be identified is constructed by modifying the original regression equation.However,since six parameters identification will involve the operation of the sixth order matrix,which is difficult to be realized in a single control cycle,a parameter identification algorithm based on High Frequency Injection(HFI)combined with Recursive Least Square(RLS)is further proposed.With the help of HFI,the six parameters of the motor are divided into four inductance-related parameters and two remaining parameters(resistance and permanent magnet flux linkage).The former is identified by the built high-frequency signal response model in real time,while the latter is estimated by RLS,and the problem of model under-ranking and PE condition is solved at the same time.The simulation and experimental results prove the effectiveness of the proposed algorithm.