Application For Optimal Designing Of Permanent Magnet Machines By Using A Multi-Objective Hybrid Genetic Algorithm

Permanent magnet synchronous machine(PMSM)is more and more widely used in electric vehicle drive system due to its advantages of high efficiency,wide speed range,high torque density,compact structure and high response.Then the research of optimal design for PMSM is become more and more important.However,the optimal design of machine is a complicated nonlinear optimization task which has multi-parameters,multi-constraints,multiple objective function and multiple local optimum solutions,make its still have many not perfect or unsolved problem,especially the global optimization problem has always been a challenge for motor designer.In this paper,take PMSM for electric vehicle application as research object.The Taguchi method,genetic algorithm(GA),and hybrid genetic algorithm(HGA)which combined Taguchi method with GA,is used in the optimization designing for PMSM.The main work of thesis as follows:Firstly,preliminary design is conducted for stator and rotor structure of PMSM according to the performance requirements for electric vehicle application.Describes the mathematical model of PMSM and the influence of parameters such as PM flux linkage,d/q-axis inductance on performance of machine,and analyzed the measures to improved the flux-weakening capability of PMSM.Secondly,Taguchi method is combined with finite element(FE)software to optimized the structure of PMSM and come true multi-factor and multi-objective optimal design at least number of times of experiments,which greatly improved the efficiency of optimization for machine.The implement process of Taguchi method is introduced in detail.Through suitably selected the optimization-factor,analyzed the effects and its relative importance of each optimization-factor on each performance,then get the optimal combination of optimization-factor for multi-objective optimization design,and FE results validated the effectiveness of Taguchi method for optimization design of PMSM.Finally,the application of GA on the optimal design of PMSM is studied in detail.For the problem of GA on optimization of PMSM,a multi-objective HGA which is combined Taguchi method with GA is proposed.By applied Taguchi method selected the GA parameters to overcome the defect of slow convergent speed.Take advantage of the local searching capability of the Taguchi method by taking the GA optimized design and optimizing it further by the Taguchi method to cover the poor local searching ability of GA which easily trapped in a local extremum during the later evolution stages.The optimization results of the HGA design is compared with the initial design,Taguchi design and GA design,which validated the effectiveness and practicability of proposed muiti-objective HGA.