摘要
采用解析法可以得到双绕组永磁容错电机各参量之间的基本表达式,对分析结构参数与电磁量的基本关系有重要意义。确定电感解析式后,利用解析法和有限元法分析槽口参数和最小气隙变化时的电感变化规律,并利用BP神经网络提高电感计算精度。根据设计要求的电感值,结合粒子群寻优算法和高精度电感表达式,可确定槽口参数与最小气隙。通过对电机磁场分析得到了电机反电动势、齿槽转矩脉动等求解的关键为计算电机径向磁密,并得到其主要影响因素为永磁体形状。利用有限元法,对电机永磁体形状进行了优化设计,有效减小了空载反电动势谐波和齿槽转矩脉动。
Basic expressions between various parameters of dual-winding fault-tolerant permanent magnet motor( DFTPMM) were gotten through the analytical method. It was important for analysis of the basic relationships between structural and basic electromagnetic parameters of DFTPMM. After determining the inductance analytical expressions,inductance variations with notch parameters and minimum air gap were analyzed by analytic method and finite element method. BP neural network was used to improve the accuracy of inductance calculation,and combined with particle swarm optimization algorithm and the design requirement of inductance to determine notch parameters and minimum air gap. The calculation of radial flux density was the key to analyze no-load back-EMF,cogging torque ripple and other parameters,and its main factor was the shape of permanent magnet. Using the finite element method,the design of permanent magnet has been optimized to effectively reduce EMF harmonics and cogging torque ripple.
出处
《微特电机》
北大核心
2015年第8期45-48,60,共5页
Small & Special Electrical Machines
基金
国家自然科学基金项目(51077007)
辽宁省科学技术计划项目(2011224004)
关键词
双绕组永磁容错
BP神经网络
粒子群寻优
有限元
dual-winding fault-tolerant permanent magnet motor
BP neural network
particle swarm optimization
finite element