Study On Programmable Optimal Design Of PM-Assisted Synchronous Reluctance Machines

PM-Assisted synchronous reluctance machines(PMASyn RMs)have received a wide attention,due to various inherent merits of high power density,high efficiency,wide speed range,no or little rare earth,etc.In this thesis,the programmable optimal design of PMASynRMs is studied,which is characterized by the integration of automatic modeling,winding distributions,optimization and FEA calculation.Firstly,the theory of symmetric winding distributions based on the star of slots is studied and a general method for winding MMF harmonic analysis is presented.This theory gives the sufficient and necessary conditions of winding feasibility,and analytical results of symmetric winding MMF are obtained by the winding distribution formula.This theory is applied in automatic winding distributions by the computer program.Secondly,the analytical model of PMASyn RMs is presented,which is the theoretical ground of the programmable optimal design.In order to keep the right balance among electromagnetic performance,mechanical performance and manufacturability,the concept of the active saturation for the flux bridge is proposed,and the principle of the active saturation method(ASM)is discussed in detail.ASM can be implemented in such synchronous reluctance machines(Syn RMs)with enhanced flux bridges to compensate the resultant electromagnetic performance loss,which is validated by 2D FEM.Moreover,compared with the End-ASM Syn RM that the permanent magnet placed at the end of the flux barrier,the flux-bridge saturation of the Middle-ASM Syn RM that the permanent magnet placed in the middle of the flux barrier is more severe.Furthermore,the proper position and the optimal shape of the permanent magnets in the rotor are determined based on both the analytical model of PMASyn RMs and ASM.Subsequently,the implementation of the programmable optimal design for PMASynRMs is presented.The optimization for PMASyn RMs can be regarded as multi-parameter and multi-object optimization problems(MOOPs),which can be transformed into two lowdimensional single-object optimization sub-problems(SOOPs).Based on the analytical model of PMASyn RMs,ASM,the proper position and the optimal shape of the permanent magnets in the rotor,the analytical model algorithm based on FEA calculation is studied,which is used in solving the sub-problem I.The Elitist model based Genetic Algorithm(EGA),Particle Swarm Optimization(PSO)and Pattern Search(PS)are introduced,which are used for solving the sub-problem II.When solving the sub-problem II,it is revealed that the optimization with PS is superior,in aspects of optimized machine performance and optimization efficiency,compared with that of E-GA and PSO.Therefore,the analytical model algorithm is used for solving the sub-problem I,and PS is used for solving the sub-problem II.Finally,nine different machines are optimized based on 2D FEM calculation,including different winding types,power levels,slot/pole combination,number of flux barriers and types of permanent magnets.From aspect of optimization convergence process,only a small amount of 2D FEM calculation is needed.From aspect of optimization results,the results of nine optimized machines are satisfied.The torque ripple of the optimized PMASyn RMs with symmetrical rotor structure is also low without the stator chording and rotor skewing,asymmetric flux barrier arrangement and optimal combination of the number of rotor flux barriers and stator slots.7.5k W and 18.5k W prototype machines are manufactured and tested to validate the programmable optimal design of PMASyn RMs.The cost and performance of the induction machines,ferrite and rare earth-assisted Syn RMs with the same power are compared.Moreover,the series products of PMASyn RMs by method of the programmable optimal design are presented.