Modular Linear Permanent-Magnet Vernier Machine And Its Multi-Objective Optimization

Most of the existing city rail vehicle use the linear induction machine,because the linear induction machine has the advantages of simple structure,convenient manufacture and low cost,but the linear induction machine also inevitably has some limitations,such as low efficiency,low force density etc.To solve aforementioned problem,the linear permanent-magnet vernier machine with higher force density appeared,whereas it has the disadvantages of poor fault-tolerant capability and large end detent force.Therefore,this thesis proposes a modular linear permanent-magnet vernier machine with strong fault-tolerant capability and low end detent force.Moreover,this thesis also proposes a multi-objective optimization method with short optimization time and high optimization accuracy,and it is used to optimize the modular linear permanent-magnet vernier machine with multiple optimization objectives to improve the overall performances of the proposed machine.The research contents of this thesis are as follows:(1)The structure and operation principle of linear permanent-magnet vernier machine are introduced,and the aforementioned theoretical analysis is verified by the airgap flux density and its harmonic spectra of linear permanent-magnet vernier machine.(2)A modular linear permanent-magnet vernier machine with strong fault-tolerant capability and low detent force is proposed,and the fault-tolerant capability and detent force of modular and non-modular linear permanent-magnet vernier machines are compared,which indicates that the proposed machine does have strong fault-tolerant capability and low detent force.(3)A multi-objective optimization method with short optimization time and high optimization accuracy is proposed,and it is used to optimize the modular linear permanent-magnet vernier machine with multiple optimization objectives,which indicates that the proposed multi-objective optimization method does have short optimization time and high optimization accuracy.(4)An approximation model named generalized regression neural network model is proposed in the proposed multi-objective optimization method,which has the merit of achieving desired approximation accuracy even with sparse data,and it is used to approximate optimization objectives in sensitive level,while use the response surface model with common approximation accuracy is used to approximate optimization objectives in non-sensitive level,which can provide reasonable models for following multi-objective optimization algorithm.(5)A prototype machine of modular linear permanent-magnet vernier machine is manufactured,and the proposed machine and multi-objective optimization method are verified by the electromagnetic performances of prototype machine.