五相混合励磁双凸极电机多目标分层分期优化设计  

Multi-Objective Layered and Phased Optimization Design of Five Phase Hybrid-Excited Double Salient Machine

作　　者：赵耀 徐笠 李东东 张旭飞 林顺富 Zhao Yao;Xu Li;Li Dongdong;Zhang Xufei;Lin Shunfu(College of Electrical Engineering Shanghai University of Electric Power,Shanghai,200090,China)

机构地区：[1]上海电力大学电气工程学院,上海200090

出　　处：《电工技术学报》2024年第22期7045-7058,共14页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金(52377111);上海市青年科技启明星计划(21QC1400200);上海市自然科学基金(21ZR1425400)资助项目。

摘　　要：五相定子槽口永磁型混合励磁双凸极电机定子槽内绕组与永磁体的耦合对优化设计提出了更高的要求,为进一步提升此类电机的功率密度和过载能力,该文提出一种五相20/18极的拓扑结构,对电机的空载特性和功率特性进行分析,确定初始优化参数和目标后进行六参数四目标的分层分期优化。首先,在电机直流饱和前提下,通过槽满率和绕组电流密度的限制提出基本的约束条件并合理分配好定子槽面积;其次,结合正交试验设计法和综合灵敏度分析对电机的本体结构参数进行分层优化,通过有限元分析建立高敏感层参数的数据库并建立数学模型,在此基础上运用改进的非支配排序遗传算法对模型实现分期优化设计,引入佳点集并在算法迭代前期基于线性排名加快收敛速度,在迭代后期施加均值约束避免畸形解,与其他智能算法对比验证了改进算法的优越性;最后,根据优化结构制造实验样机进行实验,验证了所提优化算法的可行性与有效性。The coupling between the winding and the permanent magnet in the stator slot of the five-phase stator slotted permanent magnet hybrid excitation doubly salient machine puts forward high requirements for optimization design.The multi-objective optimization process produces deformed solutions easily,and the optimization results are not comprehensive.Therefore,a five-phase 20/18 pole topology is proposed.According to the limited stator slot area of the machine structure,the no-load and power characteristics are analyzed after the slot area is rationally allocated under the DC saturation state.The slot filling rate and winding current density are taken as constraint conditions to determine the initial optimization parameters and objectives.Multi-objective layered and phased optimization with six parameters and four objectives is carried out.Combined with the orthogonal experimental design method and comprehensive sensitivity analysis,the structure parameters of the machine body are layered,and the low-sensitivity parameters are determined by the single-parameter scanning method.A parameter database of the high-sensitivity layer is established by finite element analysis,and radial basis function neural network modeling is conducted.Secondly,the improved non-dominated sorting genetic algorithm II is used to optimize the model by phases.The best data set is introduced to generate an excellent initial population.In the early phase,the parent selection method based on linear ranking accelerates the convergence speed.In the later phase,the mean value constraint is applied to the merged populations according to the average value of the database to avoid deformed solutions.The optimal design scheme is obtained according to the optimized multi-objective weighting function.Finally,the experiments verify the feasibility and effectiveness of the proposed optimization method.Using the proposed optimization method,the electromagnetic performance of the optimized machine is superior.The output power of the machine is 3649W,increase

关 键 词：定子槽口永磁 混合励磁 槽面积 多目标优化 分层分期 

分 类 号：TM352[电气工程—电机]