基于改进距离权重双电枢绕组电励磁变磁阻电机多目标优化研究  

Multi-objective optimization of double armature winding wound field variable reluctance machine based on improved distance weighted

作　　者：赵耀 陆传扬 李东东 林顺富 杨帆 ZHAO Yao;LU Chuanyang;LI Dongdong;LIN Shunfu;YANG Fan(College of Electric Power Engineering,Shanghai University of Electric Power,Shanghai 200090,China;State Grid Baoying Power Supply Company,Baoying 225800,China)

机构地区：[1]上海电力大学电气工程学院,上海200090 [2]国网宝应供电公司,江苏宝应225800

出　　处：《电机与控制学报》2023年第11期90-103,共14页Electric Machines and Control

基　　金：教育部“春晖计划”合作科研项目(HZKY20220084);上海市青年科技启明星计划(21QC1400200);上海市自然科学基金(21ZR1425400)。

摘　　要：双电枢绕组电励磁变磁阻电机结构简单且适合在极端环境中运行。针对该电机结构参数变量多,非线性特征明显,常规方法难以快速准确反映结构参数与多优化目标之间的关系等问题,提出一种基于改进距离权重的快速回归建模和多目标遗传算法相融合的结构优化方案。首先,依据磁场调制理论和绕组函数理论推导电机空载反电势和相绕组自、互感模型,并对优化目标进行定性分析;其次,采用缩减样本空间的综合敏感度分析方法,筛选出高敏感度结构参数,并基于改进距离权重的机器学习算法建立电机模型,映射高敏感度参数与多优化目标之间的非线性关系;然后以电机容错性能、电压波动和振动为优化目标,采用多目标遗传算法在设定约束条件下进行结构参数全局寻优。优化结果表明,所选的优化参数能够抑制12%的电机径向力波并在故障时减缓电机25%的相电流;最后,通过样机实验验证了方法的有效性与可行性。Dual armature-winding wound field variable reluctance machine(DAW-WFVRM)has a simple structure and is suitable for operation in extreme environments.In order to address the challenges posed by the high variability of parameters and the pronounced nonlinear characteristics of this motor structure,conventional methods struggle to rapidly and accurately depict the relationship between structural parame-ters and multiple optimization objectives.To tackle these issues,a structural optimization approach was proposed,which integrates improved distance weighting for fast regression modeling and a multi-objective genetic algorithm.Firstly,the no-load back electromotive force and the self-inductance and mutual-in-ductance models of the motor windings were derived based on the magnetic field modulation theory and winding function theory.Qualitative analysis of the optimization objectives was performed.Secondly,a comprehensive sensitivity analysis method was employed to reduce the sample space,identifying highly sensitive structural parameters.A machine learning algorithm based on improved distance weighting was utilized to establish the motor model,mapping the nonlinear relationship between highly sensitive parame-ters and multiple optimization objectives.Subsequently,considering fault tolerance,voltage fluctuation,and vibration as optimization objectives,a multi-objective genetic algorithm was employed to globally op-timize the structural parameters under specified constraints.The optimization results indicate that the se-lected parameters can suppress motor vibration by 12%and reduce phase current by 25%during faults.Finally,through prototype experiments,effectiveness and feasibility of this method are validated.

关 键 词：双电枢绕组 电励磁变磁阻电机 综合敏感度分析 距离权重 可靠性 多目标优化 

分 类 号：TM35[电气工程—电机]