基于支持向量机和遗传算法的爪极发电机建模及参数优化  

Modeling and Parameter Optimization of the Claw-Pole Alternator Based on Support Vector Machines and Genetic Algorithms

作　　者：王群京[1] 鲍晓华[1] 倪有源[1] 李争[1] 

机构地区：[1]合肥工业大学电气与自动化工程学院,合肥230009

出　　处：《电工技术学报》2006年第4期57-61,共5页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金资助项目(50077005)

摘　　要：爪极发电机广泛应用于汽车、航天等国民经济重要领域,高质量的发电机输出电压是用电设备可靠运行的重要保证。本文从爪极发电机有限元仿真计算出发,通过电机参数样本空间设计,利用支持向量机对其电磁模型进行非线性回归分析,而后利用遗传算法对其结构进行优化。仿真结果表明,所建立的非参数模型是准确可行、并且是高效的。采用支持向量机回归方法和遗传算法相结合对电机的结构参数进行优化,为电机优化设计提供了一种新的思路。The claw-pole alternator is a special type of Synchronous Machine. It is widely used in the areas of automobile, space flight et al. The high quality of the output voltage of the claw-pole alternator is very important for the equipment. This paper mainly aims at nonlinear regression modeling of the claw-pole alternator with support vector machines, which is based on the finite element method simulation, and parameter optimization of the claw-pole alternator is based on genetic algorithms. Simulation results prove feasible and high efficiency of the nonparametric model. The integration of support vector machines and genetic algorithms can be used in the optimum design of the motor and other engineering.

关 键 词：有限元法 支持向量机 遗传算法 爪极发电机 参数优化 

分 类 号：TM301[电气工程—电机]