en
×

分享给微信好友或者朋友圈

使用微信“扫一扫”功能。
作者简介:

丁陈,男,硕士生,研究方向为永磁电机智能减振降噪.dc1256339592@163.com

通讯作者:

吴兵,男,正高级实验师,研究方向为永磁电机本体及噪声优化设计.396888664@qq.com

中图分类号:TM351

文献标识码:A

DOI:10.13878/j.cnki.jnuist.20230508002

参考文献 1
El-Refaie A M.Fractional-slot concentrated-windings synchronous permanent magnet machines:opportunities and challenges[J].IEEE Transactions on Industrial Electronics,2010,57(1):107-121
参考文献 2
Magnussen F,Lendenmann H.Parasitic effects in PM machines with concentrated windings[J].IEEE Transactions on Industry Applications,2007,43(5):1223-1232
参考文献 3
Cassat A,Espanet C,Coleman R,et al.A practical solution to mitigate vibrations in industrial PM motors having concentric windings[J].IEEE Transactions on Industry Applications,2012,48(5):1526-1538
参考文献 4
Hong J F,Wang S M,Sun Y G,et al.Piecewise stagger poles with continuous skew edge for vibration reduction in surface-mounted PM synchronous machines[J].IEEE Transactions on Industrial Electronics,2021,68(9):8498-8506
参考文献 5
Wang S,Li H F.Effects of rotor skewing on the vibration of permanent magnet synchronous motors with elastic-plastic stator[J].IEEE Transactions on Energy Conversion,2022,37(1):87-96
参考文献 6
Jung J W,Kim D J,Hong J P,et al.Experimental verification and effects of step skewed rotor type IPMSM on vibration and noise[J].IEEE Transactions on Magnetics,2011,47(10):3661-3664
参考文献 7
Peng C,Wang D H,Feng Z K,et al.A new segmented rotor to mitigate torqueripple and electromagnetic vibration of interior permanent magnet machine[J].IEEE Transactions on Industrial Electronics,2022,69(2):1367-1377
参考文献 8
Wang S M,Hong J F,Sun Y G,et al.Effect comparison of zigzag skew PM pole and straight skew slot for vibration mitigation of PM brush DC motors[J].IEEE Transactions on Industrial Electronics,2020,67(6):4752-4761
参考文献 9
Jean L B.Vibroacoustic analysis of radial and tangential air-gap magnetic forces in permanent magnet synchronous machines[J].IEEE Transactions on Magnetics,2015,51(6):1-9
参考文献 10
Zhang J X,Zhang B Y,Feng G H.Influence of pole and slot combination on torque characteristics and radial force of fractional slot permanent magnet machines[J].IEEJ Transactions on Electrical and Electronic Engineering,2021,16(8):1055-1066
参考文献 11
Lin F,Zuo S G,Wu X D.Electromagnetic vibration and noise analysis of permanent magnet synchronous motor with different slot-pole combinations[J].IET Electric Power Applications,2016,10(9):900-908
参考文献 12
Xing Z Z,Wang X H,Zhao W L.Optimization of stator slot parameters for electromagnetic vibration reduction of permanent magnet synchronous motors[J].IEEE Transactions on Transportation Electrification,2022,8(4):4337-4347
参考文献 13
Zhao W X,Zhu S D,Ji J H,et al.Analysis and reduction of electromagnetic vibration in fractional-slot concentrated-windings PM machines[J].IEEE Transactions on Industrial Electronics,2022,69(4):3357-3367
参考文献 14
Yang I J,Lee S H,Lee K B,et al.A process to reduce the electromagnetic vibration by reducing the spatial harmonics of air gap magnetic flux density[J].IEEE Transactions on Magnetics,2021,57(2):1-6
参考文献 15
曹永娟,冯亮亮.基于响应面法的轴向磁场永磁记忆电机多目标优化设计[J].南京信息工程大学学报(自然科学版),2021,13(5):620-627.CAO Yongjuan,FENG Liangliang.Multi-objective optimization design of axial-flux permanent magnet memory motor based on response surface method[J].Journal of Nanjing University of Information Science & Technology(Natural Science Edition),2021,13(5):620-627
参考文献 16
华逸舟,刘奕辰,潘伟,等.基于改进粒子群算法的无轴承永磁同步电机多目标优化设计[J].中国电机工程学报,2023,43(11):4443-4452.HUA Yizhou,LIU Yichen,PAN Wei,et al.Multi-objective optimization design of bearingless permanent magnet synchronous motor using improved particle swarm optimization algorithm[J].Proceedings of the CSEE,2023,43(11):4443-4452
参考文献 17
王群京,郑耀达,刘先增.基于结构参数优化的电机振动噪声的抑制研究[J].电气工程学报,2023,18(2):16-25.WANG Qunjing,ZHENG Yaoda,LIU Xianzeng.Research on suppression of motor vibration and noise based on structural parameter optimization[J].Journal of Electrical Engineering,2023,18(2):16-25
参考文献 18
Wang S,Li H F.Reduction of electromagnetic vibration and noise in permanent magnet motor for EVs by optimizing design of rotor based on GPR-PSO model[J].Journal of Electrical Engineering & Technology,2020,15(3):1231-1243
参考文献 19
Ma C G,Li Q,Liu Q H,et al.Sound quality evaluation of noise of hub permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2016,63(9):5663-5673
参考文献 20
Ma C G,Li Q,Deng L W,et al.A novel sound quality evaluation method of the diagnosis of abnormal noise in interior permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2017,64(5):3883-3891
参考文献 21
Yan B,Yang Y B,Wang X H.Design of a large capacity line-start permanent magnet synchronous motor equipped with hybrid salient rotor[J].IEEE Transactions on Industrial Electronics,2021,68(8):6662-6671
参考文献 22
Aoki S,Hibi T,Ohsugi H.Markov-chain Monte Carlo methods for the Box-Behnken designs and centrally symmetric configurations[J].Journal of Statistical Theory and Practice,2016,10(1):59-72
参考文献 23
Lei G,Chen X M,Zhu J G,et al.Multiobjective sequential optimization method for the design of industrial electromagnetic devices[J].IEEE Transactions on Magnetics,2012,48(11):4538-4541
参考文献 24
Dalbey K R,Karystinos G N.Generating a maximally spaced set of bins to fill for high-dimensional space-filling Latin hypercube sampling[J].International Journal for Uncertainty Quantification,2011,1(3):241-255
参考文献 25
Kailkhura B,Thiagarajan J J,Rastogi C,et al.A spectral approach for the design of experiments:design,analysis and algorithms[J].The Journal of Machine Learning Research,2018,19(1):1214-1259
参考文献 26
Smidova M,Sadilek V,Elias J,et al.Audze-Eglājs criterion for orthogonal and regular triangular grids[C]//1st ECCOMAS Thematic Conference on Uncertainty Quantification in Computational Sciences and Engineering,UNCECOMP 2015.Crete Island,Greece,2015∶664-679
参考文献 27
Cai X W,Gao L,Li X Y.Efficient generalized surrogate-assisted evolutionary algorithm for high-dimensional expensive problems[J].IEEE Transactions on Evolutionary Computation,2020,24(2):365-379
参考文献 28
Beyer H G,Sendhoff B.Robust optimization:a comprehensive survey[J].Computer Methods in Applied Mechanics and Engineering,2007,196(33/34):3190-3218
目录contents

    摘要

    为了保证永磁无刷直流空心轴电机的输出性能和抑制电磁振动,提出了一种基于非线性多元回归的修正代理模型优化方法.首先,通过AE(Audze-Elglājs)准则确定最优空间填充抽样,并采用核主成分分析(Kernel Principal Components Analysis,KPCA)算法筛选出4个主要变量用于构建代理模型;其次,采用非线性多元回归构建代理模型,决定系数R2值均大于0.9,验证了代理模型的精度;最后,采用鲁棒多目标遗传算法求解代理模型,获得了最优定子槽参数.结果表明,通过优化定子槽参数,电机平均转矩降低了1.3%,不影响输出性能,电机空载、额定负载时最大振动加速度分别降低19%和34.5%,有效地降低了电磁振动,验证了优化方法的有效性和可靠性.

    Abstract

    To ensure the output performance and suppress electromagnetic vibration of a Hollow Shafted Permanent Magnet Brushless DC Motor (HSPMBLDCM),a surrogate-assisted model optimization method based on nonlinear multiple regression is proposed.First,the optimal space-filling sampling was determined by the Audze-Eglājs (AE) criterion,and four main variables were selected by the Kernel Principal Components Analysis (KPCA) algorithm to construct the surrogate model.Next,the surrogate model was constructed using nonlinear multiple regression,and the R2 values of the coefficients of determination were all greater than 0.9,which verified the accuracy of the surrogate model.Finally,the Robust Multiobjective Optimization Evolutionary Algorithms (RMOEA) were used to solve the surrogate model,and the optimal stator slot parameters were obtained.The results show that by optimizing the stator slot parameters,the average torque of the motor is reduced by 1.3%,which does not affect the output performance,and the maximum vibration acceleration of the motor at no load and rated load is reduced by 19% and 34.5%,respectively,which effectively reduces the electromagnetic vibration and proves the effectiveness and reliability of the optimization method.It provides an alternative means for motor vibration damping design optimization with practical significance.

  • 0 引言

  • 分数槽集中绕组(Fractional Slot Concentrated Winding,FSCW)电机具有绕组端部短、铜耗小、效率高、弱磁性能好、齿槽转矩低等优点,近年来得到了广泛的应用[1].然而,与整数槽电机相比,分数槽电机电磁力的空间阶数较低、幅值较大,所引起的振动问题更加突出[2].特别是要求低振动的领域,振动水平是评价电机性能的重要指标.

  • 近年来,国内外学者对电机减振的研究主要集中在电机本体优化上.文献[3]首次采用定子斜槽来降低永磁电机振动,在此基础上,文献[4]通过半解析法研究发现,采用定子斜槽可以完全消除一阶槽频振动,但是电机振动的极频分量削弱效果有限.因此,定子斜槽只能削弱特定频率下的电磁振动,存在局限性.文献[5]考虑到定子材料的弹塑性,采用转子分段斜极抑制电机电磁振动;文献[6]从激振力分析、定子模态分析和振动试验结果等方面,验证了转子斜极能够抑制电磁振动;文献[7]提出一种新的V型内置式永磁电机转子结构,以减小转矩脉动和电磁振动.文献[5-7]结果表明:转子斜极可以有效降低电磁振动,但受转子斜极角度限制.文献[8]设计了锯齿形斜磁极来削弱永磁直流有刷(Permanent Magnet Brush DC,PMBDC)电机的电磁振动;文献[9]在复域中用矩阵的方法归纳了径向电磁力的谐波表达式,指出径向电磁力的最小非零阶为极槽的最大公约数;文献[10]研究了极槽组合对分数槽永磁电机径向电磁力的影响.结果表明:对振动和噪声影响最大的径向电磁力谐波阶次为极槽的最大公约数,随着该阶次的降低,振动会显著增加;槽极组合与电机径向电磁力有关.因此,优化极槽组合可以从提高电机极槽的最大公约数入手,从而降低径向电磁力,以减弱电机的振动水平.

  • 上述电磁振动的削弱措施主要集中在定、转子结构和极槽组合的优化上,而定子槽型参数对电磁振动影响的研究较少.定子槽型结构复杂,通过槽开口、槽深、槽中心宽度等影响电机磁场空间分布[11],从而影响电机气隙磁导.此外,槽开口处不可避免地存在磁路饱和,也会影响电机气隙磁导[12].由于气隙磁导改变而导致径向力发生变化,从而改变永磁电机振动水平.因此,定子槽参数优化是抑制电机振动的一种有效手段.文献[13]为了减小12槽10极分数槽集中绕组永磁电机的电磁振动,设计了不等距齿,但没有给出不等齿结构具体参数的设计方法;文献[14]通过改变定子铁芯的形状来降低气隙磁通密度的空间谐波,从而降低永磁同步电机的电磁振动和噪声,但未考虑频率对气隙磁通密度的影响.由上述分析可知:气隙磁导波形随定子槽型变化而变化;当槽型不佳时,气隙磁导中的谐波将引起振动并会削弱电机输出转矩.到目前为止,尚未有系统的方法研究槽型对电磁振动的影响.而诸多场合都对电机有低振动要求,如汽车、高精度驱动等场合,因此研究槽型对电机振动的影响显得很有必要.

  • 目前,大量的智能算法被应用于电机的分析和优化.文献[15]结合不同的优化方法,对提出的新型轴向磁场永磁记忆电机(Axial Flux Permanent Magnet Memory Motor,AFPMMM)进行多目标优化,通过对试验数据分析对比,得出了低齿槽转矩优化方案;文献[16]基于改进粒子群算法对无轴承永磁同步电机(Bearingless Permanent Magnet Synchronous Motor,BPMSM)输出转矩和悬浮性能进行了优化;文献[17]利用响应面-遗传算法对内置式永磁电机进行了减振优化;文献[18]利用高斯过程回归-粒子群算法对8极48槽永磁电机径向电磁力进行了优化设计,从而降低了电磁振动;文献[19]建立了基于BP神经网络的轮毂永磁同步电机(Hub PMSM,HPMSM)音质评价模型,并采用等级评价法对噪声水平进行主观评价;文献[20]则将配对比较法和等级评价法分别用于HPMSM音质评价,并首次提出两种方法的联合应用,进一步完善了音质评价模型;文献[21]采用遗传算法和神经网络相结合的方法对起线式永磁同步电机(Line-start PMSM,LPMSM)转子结构参数进行优化,确定了最优结构参数.然而,上述研究都仅是将电机优化简化为一个数学模型,然后通过算法迭代对模型进行求解数学最优,而未考虑其他不确定性因素,求得的最优解仅为“理想极点”.定子冲片在生产制造中,难免会受到工艺设计、装配误差、加工误差、弹塑性变形等因素影响,使得电机实际参数超过“理想极点”从而使得电磁性能超过数值阈值.

  • 针对以上永磁电机减振优化方法的不足,本文将代理模型方法应用于电机减振优化中,提出一种基于非线性多元回归修正代理模型的永磁电机鲁棒多目标优化方法.首先介绍电机结构并分析了初始振动水平;其次,阐明所提出优化方法的原理;然后,根据所提出的多目标优化方法,按照最优空间填充抽样(Optimal Space Filling Design Sampling,OSF)、核主成分分析(Kernel Principal Components Analysis,KPCA)、代理模型构建、鲁棒多目标遗传算法(Robust Multiobjective Optimization Evolutionary Algorithm,RMOEA)求解并修正代理模型的步骤对永磁无刷直流空心轴电机进行了定子槽参数优化,获得了最优定子槽参数;最后对比分析电机优化前后的输出性能和振动水平,验证所提方法的有效性.

  • 1 电机基本结构与振动原理

  • 1.1 电机基本结构

  • 目前,激光雷达扫描仪中,其驱动系统在国内外均采用实心轴电机驱动,而采用实心轴电机驱动不但增加了电机中的传动机构,还增加了电机本身的质量、体积,且构造复杂,制造成本高,激光雷达采集精度效果不理想.因此,本文提出一种永磁无刷直流空心轴电机(Hollow Shafted Permanent Magnet Brushless DC Motor,HSPMBLDCM),该电机为8极12槽的分数槽集中绕组永磁电机,其基本结构如图1所示,它是典型的表贴式永磁电机.考虑提高电机气隙磁密平顶波宽度,采用磁环设计,电机高速运行时,转子各部件受到巨大的离心力作用,因此在磁钢外加设护套防止永磁体断裂.空心轴内孔安置采光镜头,对电机有低振动要求.HSPMBLDCM基本参数如表1所示.

  • 1.2 电磁振动分析

  • 根据麦克斯韦张量法,作用在定子铁芯内表面的径向电磁力密度[18]可表示为

  • 图1 HSPMBLDCM电机基本结构

  • Fig.1 Structure of the proposed HSPMBLDCM

  • 表1 HSPMBLDCM基本参数

  • Table1 Parameters of the proposed HSPMBLDCM

  • fr(θ,t)=Br2(θ,t)2μ0.
    (1)
  • 式中,θ为周向坐标,Br为气隙磁通密度的径向分量,μ0为真空磁导率.

  • 忽略磁路饱和,永磁电机径向气隙磁场可表示为

  • Br(θ,t)=Br1+Br2.
    (2)
  • 式中,Br为径向气隙磁通密度,Br1为永磁磁场气隙磁通密度,Br2为电枢反应磁场气隙磁通密度.

  • Br1Br2,可以表示为磁动势Fθt)与气隙磁导λθ)的乘积:

  • Br1=F(θ,t)1λ(θ),
    (3)
  • Br2=F(θ,t)2λ(θ).
    (4)
  • 式中,Fθt1为永磁磁场气隙磁动势,Fθt2为电枢反应磁场气隙磁动势,λθ)为等效气隙磁导.

  • Fθt1Fθt2λθ)可以表示为

  • F(θ,t)1=μ=1,3,5, Fμcosμ(ωt-pθ),
    (5)
  • F(θ,t)2=v=3k±1,k=1,2,3, Fvcos(ωt-vpθ+φ),
    (6)
  • λ(θ)=λ0+m=1,2,3, λmcos(mZθ).
    (7)
  • 式中,μ为永磁磁场谐波阶数,v为电枢反应磁场谐波阶数,m为齿谐波阶数,Fμμ次永磁谐波磁动势幅值,Fvv次电枢反应磁场谐波磁动势幅值,λm为气隙m阶齿谐波磁导幅值,Z为定子齿数,φ为磁动势初始相位.

  • 将式(5)—(7)带入式(1)可得:

  • fr(θ,t=Br2(θ,t)2μ0=F(θ,t)1+F(θ,t)2λθ22μ0=[a+b]22μ0
    (8)
  • 式中,a为永磁磁场气隙磁通密度,b为电枢反应磁场气隙磁通密度.

  • 在本研究中仅为了说明径向电磁力幅值与磁导、磁动势的关系,因此仅在表2中给出了由永磁磁场各阶次谐波相互作用产生的力波分量具体阶次和频率特性.

  • 从表2中电磁力振幅的表达式可以看出,径向电磁力的幅值与电机磁势Fμ和等效气隙磁导λθ成正比.因此,在保证电机电磁性能变化不大的前提下,可以仅通过优化定子槽形、减小等效气隙磁导,来降低径向电磁力的幅值、削弱电磁振动,而不需要修改电机的其他参数.

  • 1.3 初始振动水平

  • 由图1所示电机结构,在ANSYS Maxwell中建立8极12槽HSPMBLDCM电磁模型,得到作用在定子铁芯内表面的节点电磁力.假设电磁力沿着轴向均匀分布,通过传递二维电磁模型中的节点电磁力,并将其作为电磁振动激励加载到三维定子模型中.最后,可以准确地获得电机的振动加速度如图2所示.

  • 额定负载时,气隙磁通密度由ab共同作用,因此电机在额定负载下的振动加速度大于空载时的振动加速度.振动在(μ1±μ1f1多频点突出,符合表2分析结果.电机的最大振动加速度出现在3 200 Hz,即8f1处振动最大.由于忽略了漏磁影响,因此电机振动在非(μ1±μ1f1处也有较小振动,仿真结果符合理论推导.因为电磁振动和气隙磁场密切相关,电磁振动的减小通常会导致输出转矩的减小,因此,本文为了保证良好的电机输出性能和低振动水平,选取平均径向电磁力密度Fa、平均输出转矩Ta、转矩脉动Tr作为优化目标.

  • 表2 径向电磁力特定阶次和频率

  • Table2 Specific order and frequency of radial electromagnetic force

  • 图2 电机在空载和额定负载下的振动加速度

  • Fig.2 Vibration accelerations of the motor with no load or rated load

  • 2 实验设计与采样

  • 2.1 方法提出

  • 图3所示为基于修正代理模型的鲁棒多目标优化基本流程.该方法首先采用抽样方法获得初始样本,然后利用ANSYs Maxwell计算出优化目标值,再根据初始样本集构造初始代理模型,随后算法进入主循环.在循环的每一代中,用当前代理模型代替仿真求解来计算每个个体的适应度值,并利用算法不断搜索优化问题的最优解;将算法求得最优解加入参数干扰进行鲁棒性检验,并修正代理模型.重复运行主循环,直到满足限制条件,获得鲁棒最优解.

  • 图3 鲁棒多目标优化流程

  • Fig.3 Robust multiobjective optimization process

  • 2.2 确定优化参数

  • 在建立电机代理模型之前,首先需要设计实验来选择采样点.目前常用的设计实验方法有Box-Behnken设计(Box-Behnken Design,BBD)[22]、中心复合设计(Central Composite Design,CCD)[23]、拉丁超立方抽样(Latin Hypercube Sampling,LHS)[24]、最优空间填充抽样(Optimal Space Filling Design Sampling,OSF)[25]等.设计参数如图4所示,各参数初始值及采样范围如表3所示.对定子槽参数进行采样,由于BBD和CCD采样水平较少,样本空间填充效果不如LSH和OSF,因此本文采取后两种采样方式来获取实验样本.

  • 图4 定子槽参数

  • Fig.4 Stator slot parameters

  • 表3 定子槽初始值及设计参数

  • Table3 Initial values and design parameters of the stator slot

  • 由图5可以看出,LHS和OSF采样点都可以有效地填充样本空间.为了评价这两种采样方法对本文的适应性,本文引入了AE(Audze-Eglājs)准则[26].

  • AE准则假设在样本空间中具有若干单位质量的点,这些点通过排斥力相互作用,在整个空间累计势能.势能的大小假设为与每对点之间的距离平方成反比,样本空间的总势能为所有对点累计势能之和,评价函数如式(9)所示,E值越小,采样结果的均匀性越好.

  • E=i=1Ns j=i+1Ns 1Lij2.
    (9)
  • 图5 样本点空间分布

  • Fig.5 Spatial distribution of sample points

  • 式中,E为样本点间等效势能,Ns为样本点个数,Lij为两样本点间距离,ij为样本点编号.

  • 通过式(9)得EOSF=4 472.18<ELSH=5 868.535.因此采用OSF采样对代理模型的适应性更强,样本点在样本空间分布更均匀、更充分.

  • 选择合适的采样方式之后进行试验采样,优化参数之间对FaTaTr影响各不相同.此外,参数较多时需要的试验次数也会随之增多,造成计算资源的浪费,并且参数维度越高代理模型精度越低[27].因此,为了缩小设计空间与提高模型精度,先对参数降维.从图5看出,OSF各参数样本点之间为非线性关系,因此引入KPCA算法对7个参数进行非线性降维,以提高代理模型精度.KPCA算法原理为对于样本空间中的点S=(S1S2,···,Sn,···,SkTSn=(s1ns2n,···,snn,···,spn),k=7,p=79,通过高斯核函数将其映射到高位空间,使其线性可分,核函数如式(10)所示,经过试算在MATLAB中设置σ=0.001,得到主元贡献率如图6所示.

  • k(x,y)=e-x-y22σ2.
    (10)
  • 式中,xy为点集,σ为超参数.

  • 由图6a可以看出,当主元个数为4时,主元贡献率为86.375 9%.图6b显示每个参数的贡献率,发现Bs1贡献率低于0.15且Hs1Hs0为零,因此选择Hs2Bs0Bs2Rs 4个优化参数构建代理模型.

  • 上文分析已确定抽样方法与优化参数,将OSF得到的样本点带入ANSYS分析软件,可得到设计变量的响应值,如表4所示.至此,设计实验和抽样部分基本完成,可以根据样本点和响应值建立代理模型.

  • 3 代理模型构建

  • 构建代理模型之前,首先定义相应的输入和输出参数.X为训练集的设计参数,Y为训练集的响应值,Y*为预测集的响应值.X=(X1X2,···,Xn,···,XNTXn=(x1nx2n,···,xnn,···,xpn).在本文中n=1,2,···,NN是设计变量的个数,N=4;p=1,2,···,PP是样本点个数,P=25;Y=(Y1Y2,···,Ym,···,YMTYm=(ym1ym2,···,ymm,···,ypm),m=1,2,···,MM是目标个数,M=3;Y*=(Y*1Y*2,···,Y*m,···,Y*M)T,Y*m=(y*1my*2m,···,y*mm,···,y*Mm).

  • 图6 KPCA参数降维

  • Fig.6 KPCA parameter downscaling

  • 表4 样本点的设计变量和响应值

  • Table4 Design variables and response values of sample points

  • 在MATLAB中采用非线性多元回归对表4中的OSF实验数据进行回归,得到各因素与平均转矩Ta、转矩脉动Tr、平均径向电磁力密度Fa之间的连续函数关系.构建的训练集为D1={((XiY1i)│i=1,2,···,25)}、D2={((XiY2i)│i=1,2,···,25)}、D3={((XiY3i)│i=1,2,···,25)},Y1iY2iY3i分别为FaTaTr的响应值.由图5可知LSH采样样本空间填充也较好,因此构建LSH采样的预测变量集X*,代理模型的预测结果如图7所示.

  • 为了评价模型精度,引入决定系数R2来验证代理模型精度:

  • R2=1-i=150 y^i-yi2i=150 yi-y-i2.
    (11)
  • 式中,yi为样本值,y-i为样本平均值,y^i是预测值.

  • 计算得到FaTaTrR2分别为0.997 5、0.966 3和0.935 6,说明代理模型对电机平均径向电磁力、平均输出转矩、转矩脉动的预测准确率分别达到99.75%、96.63%、93.56%,建模精度较好.

  • 完成代理模型的构建后,需要对代理模型进行优化.为此,建立相应的多目标优化模型(Multiobjective Optimization Model,MOOM).为保证电机输出特性,限定电机平均输出转矩降低不超过3%.此外,无刷空心轴电机定子冲片在实际生产中受工艺设计、材料特性、加工误差、装配误差、弹塑性变形等干扰因素的影响,为使优化解具有鲁棒性,因此本文考虑对决策变量添加干扰因子进行求解.由此MOOM优化模型如下所示:

  • minF(x)=fFa'x',f-Ta'x',fTr'x'T,x=Hs2,Bs0,Bs2,Rs,x'=x1+δ1,x2+δ2,x3+δ3,x4+δ4T s.t. TaTa0(1-3%),FaFr0,TrTr0.
    (12)
  • 式中,δ=(δ1δ2,···,δNT为添加在每个决策变量上的干扰向量,Fa0Ta0Tr0分别为平均径向电磁力密度、平均输出转矩、转矩脉动初始值,Fx)为目标空间,fFa'x'f-Ta'x'fTr'x'为相应的目标函数.

  • 本文采取ROMEA对定子槽参数进行减振优化[28].RMOEA是基于分解的鲁棒多目标遗传算法.本文采用边界相交法(Penalty-Based boundary Intersection,PBI)对多目标决策空间进行分解,如图8所示.

  • minPBI=gxλ,Z*=d1+θd2,
    (13)
  • d1=F(x)-Z*Tλλd2=F(x)-Z*-d1λλ
    (14)
  • 式中,λ为基于超平面均匀分布的权重向量,Z*为目标函数最小值所在的参考点,θ为惩罚函数均匀分布的权重向量,d1d2为补足位移.

  • RMOEA算法优化代理模型分为3部分:第1部分不考虑干扰因子的遗传算法寻优,获得最优解集S、外部存档(Archive),并在目标空间对最优解集S进行分解.如图8所示,小球为解,v为权重向量,例如红色的小球与权重向量v1v2v3相关联,并且每个向量包含4个解;第2部分考虑干扰因子p的鲁棒寻优,该步首先对最优解集S进行鲁棒性检验,然后舍弃PBI值最大的点获得新的存档NewArchive={(p∈Archive│PBIpWmaxPBI)},其中,p为干扰因子,WmaxPBI为PBI值最大的点集,并将其填充到样本空间对代理模型进行修正;第3部分基于新的代理模型求解鲁棒最优前沿.

  • 图7 代理模训练、预测结果

  • Fig.7 Surrogate model training and prediction results

  • 图8 目标空间分解

  • Fig.8 Decomposition in the objective space

  • 在MATLAB中编写RMOEA算法,经过多次试算,设置参数:干扰因子p=0.035、迭代次数m=3 000、干扰测试样本点Thon=1 000、θ=5.求解在短时间内完成,得到多目标遗传算法(MOEA)与RMOEA求解Pareto(帕累托)前沿如图9所示.图9a为原始代理模型的Pareto最优前沿,图9b为干扰因子p=0.035时修正代理模型的鲁棒Pareto最优前沿.从图9 b可以可看出,加入了干扰因子之后,Pareto前沿与原始Pareto前沿几乎保持在相同的位置,此外,鲁棒最优解之间也比原始Pareto前沿更为分散,方便挑选最优解.Pareto前沿的每个解之间为非支配关系,无法评价其优劣,因此本文引入乌托邦点法来挑选最优解.首先,对鲁棒最优前沿进行标准化处理,对平均径向电磁力密度、平均输出转矩、转矩脉动赋予相应权重值.然后,确定乌托邦点P+和反乌托邦点P-,并计算每个点Pi与乌托邦点之间的欧式距离L+,及与反乌托邦点之间的欧式距离L-,如式(15)和(16)所示.最后,通过评价函数(17)确定适应度C,并选择适应度最好的点作为本文最优解.

  • L+=k=1m P+-Pi2,
    (15)
  • L-=k=1m P--Pi2
    (16)
  • C=L+L++L-.
    (17)
  • 式中,k为参数维度,在本文中k=4,Pi为第i个点,L+为第i点与乌托邦点之间的欧式距离,L-为反欧式距离,C为适应度,M为Pareto前沿点的个数.

  • 通过求解适应度,得到最优点P的适应度为C=0.061与乌托邦点相似度最高,此时优化目标P=[293 350.761 8,-654.476 4,34.36%],设计参数X=[3.96,2,9.75,0.9]为算法寻得最优解.

  • 4 结果讨论

  • 通过提出的优化方法,最终确定了电机减振优化定子槽参数.将代理模型参数带回ANSYS Maxwell,得到的优化真实值如表5所示.

  • 表5 基于代理模型优化结果

  • Table5 Optimization results based on surrogate model

  • 图9 Pareto最优前沿

  • Fig.9 Pareto optimal fronts

  • 初始方案和优化方案的比较如图10所示.由图10a、10b与表5可知:电机平均径向电磁力密度从318 706.304 N/m2降至293 350.761 8 N/m2,降低了8%;转矩脉动从66.8%降至34.36%,与优化前相比降低了48.2%;平均输出转矩从657.475 mN·m降至645.476 4 mN·m,降低了1.3%,降幅较小不影响电机输出特性.结果表明,与初始定子冲片设计方案相比,提出的优化方法能够显著降低电机的振动水平.

  • 此外,由图10c、10d可知,电机空载、额定负载在8f1处的振动加速度大大降低,空载时电机的最大振动加速度从3 084 mm/s2降至2 493 mm/s2,降幅为19%,额定负载时最大振动加速度从4 049.4 mm/s2降至2 651.4 mm/s2,降幅为34.5%,有效地抑制了电机的电磁振动.

  • 与传统的电机减振研究相比,本文基于修正代理模型对电机电磁振动进行了系统的研究.针对不同电机的实际要求(如输出转矩、电磁力、转矩脉动),利用本文提出的优化方法对不同参数进行采样,采用修正代理模型进行数据建模,并结合算法寻找最优解,可以有效地降低计算成本、降低设计难度,获得具有鲁棒性的最优解.在实际的工业应用过程中将大大减少无用样机,并且可以帮助设计人员准确地找到最合适的减振参数,具有重要的现实意义和经济价值.

  • 5 总结

  • 本文以一台400 W应用于激光雷达扫描仪的永磁无刷直流空心轴电机为研究对象,提出一种可推广应用的电机多目标优化方法.通过优化方法,完成了空心轴电机定子槽减振优化分析,得到定子槽的优化方案.本文的主要结论如下:

  • 1)提出一种基于代理模型的多目标优化方法,与一般多目标优化不同,加入了参数干扰,进行鲁棒性检验并修正代理模型,从而使得代理模型Pareto最优前沿更为分散,更符合实际情况而不仅是数学模型,最优解具有鲁棒性.

  • 图10 优化前后结果比较

  • Fig.10 Comparison of results before and after optimization

  • 2)通过AE准则,从4种采样方法中选择最适合本文的OSF方法进行采样.通过计算3个优化目标训练值和预测值的决定系数,验证代理模型的高精度;建立鲁棒多目标优化数学模型,选择RMOEA算法对代理模型进行修正,获得鲁棒Pareto前沿.最后通过乌托邦点法,在Pareto前沿中挑选最优点P.

  • 3)通过参数化建模、初始仿真设置、设计实验、代理模型建立和模型优化等步骤,优化了电机定子槽参数.在不影响电机输出转矩的前提下,有效地削弱了电机的转矩脉动与平均径向电磁力密度.与优化前相比,转矩脉动和平均径向电磁力密度分别降低48.2%、8%,电机空载、额定负载最大振动加速度分别降低19%和34.5%,有效地降低了电磁振动.

  • 参考文献

    • [1] El-Refaie A M.Fractional-slot concentrated-windings synchronous permanent magnet machines:opportunities and challenges[J].IEEE Transactions on Industrial Electronics,2010,57(1):107-121

    • [2] Magnussen F,Lendenmann H.Parasitic effects in PM machines with concentrated windings[J].IEEE Transactions on Industry Applications,2007,43(5):1223-1232

    • [3] Cassat A,Espanet C,Coleman R,et al.A practical solution to mitigate vibrations in industrial PM motors having concentric windings[J].IEEE Transactions on Industry Applications,2012,48(5):1526-1538

    • [4] Hong J F,Wang S M,Sun Y G,et al.Piecewise stagger poles with continuous skew edge for vibration reduction in surface-mounted PM synchronous machines[J].IEEE Transactions on Industrial Electronics,2021,68(9):8498-8506

    • [5] Wang S,Li H F.Effects of rotor skewing on the vibration of permanent magnet synchronous motors with elastic-plastic stator[J].IEEE Transactions on Energy Conversion,2022,37(1):87-96

    • [6] Jung J W,Kim D J,Hong J P,et al.Experimental verification and effects of step skewed rotor type IPMSM on vibration and noise[J].IEEE Transactions on Magnetics,2011,47(10):3661-3664

    • [7] Peng C,Wang D H,Feng Z K,et al.A new segmented rotor to mitigate torqueripple and electromagnetic vibration of interior permanent magnet machine[J].IEEE Transactions on Industrial Electronics,2022,69(2):1367-1377

    • [8] Wang S M,Hong J F,Sun Y G,et al.Effect comparison of zigzag skew PM pole and straight skew slot for vibration mitigation of PM brush DC motors[J].IEEE Transactions on Industrial Electronics,2020,67(6):4752-4761

    • [9] Jean L B.Vibroacoustic analysis of radial and tangential air-gap magnetic forces in permanent magnet synchronous machines[J].IEEE Transactions on Magnetics,2015,51(6):1-9

    • [10] Zhang J X,Zhang B Y,Feng G H.Influence of pole and slot combination on torque characteristics and radial force of fractional slot permanent magnet machines[J].IEEJ Transactions on Electrical and Electronic Engineering,2021,16(8):1055-1066

    • [11] Lin F,Zuo S G,Wu X D.Electromagnetic vibration and noise analysis of permanent magnet synchronous motor with different slot-pole combinations[J].IET Electric Power Applications,2016,10(9):900-908

    • [12] Xing Z Z,Wang X H,Zhao W L.Optimization of stator slot parameters for electromagnetic vibration reduction of permanent magnet synchronous motors[J].IEEE Transactions on Transportation Electrification,2022,8(4):4337-4347

    • [13] Zhao W X,Zhu S D,Ji J H,et al.Analysis and reduction of electromagnetic vibration in fractional-slot concentrated-windings PM machines[J].IEEE Transactions on Industrial Electronics,2022,69(4):3357-3367

    • [14] Yang I J,Lee S H,Lee K B,et al.A process to reduce the electromagnetic vibration by reducing the spatial harmonics of air gap magnetic flux density[J].IEEE Transactions on Magnetics,2021,57(2):1-6

    • [15] 曹永娟,冯亮亮.基于响应面法的轴向磁场永磁记忆电机多目标优化设计[J].南京信息工程大学学报(自然科学版),2021,13(5):620-627.CAO Yongjuan,FENG Liangliang.Multi-objective optimization design of axial-flux permanent magnet memory motor based on response surface method[J].Journal of Nanjing University of Information Science & Technology(Natural Science Edition),2021,13(5):620-627

    • [16] 华逸舟,刘奕辰,潘伟,等.基于改进粒子群算法的无轴承永磁同步电机多目标优化设计[J].中国电机工程学报,2023,43(11):4443-4452.HUA Yizhou,LIU Yichen,PAN Wei,et al.Multi-objective optimization design of bearingless permanent magnet synchronous motor using improved particle swarm optimization algorithm[J].Proceedings of the CSEE,2023,43(11):4443-4452

    • [17] 王群京,郑耀达,刘先增.基于结构参数优化的电机振动噪声的抑制研究[J].电气工程学报,2023,18(2):16-25.WANG Qunjing,ZHENG Yaoda,LIU Xianzeng.Research on suppression of motor vibration and noise based on structural parameter optimization[J].Journal of Electrical Engineering,2023,18(2):16-25

    • [18] Wang S,Li H F.Reduction of electromagnetic vibration and noise in permanent magnet motor for EVs by optimizing design of rotor based on GPR-PSO model[J].Journal of Electrical Engineering & Technology,2020,15(3):1231-1243

    • [19] Ma C G,Li Q,Liu Q H,et al.Sound quality evaluation of noise of hub permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2016,63(9):5663-5673

    • [20] Ma C G,Li Q,Deng L W,et al.A novel sound quality evaluation method of the diagnosis of abnormal noise in interior permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2017,64(5):3883-3891

    • [21] Yan B,Yang Y B,Wang X H.Design of a large capacity line-start permanent magnet synchronous motor equipped with hybrid salient rotor[J].IEEE Transactions on Industrial Electronics,2021,68(8):6662-6671

    • [22] Aoki S,Hibi T,Ohsugi H.Markov-chain Monte Carlo methods for the Box-Behnken designs and centrally symmetric configurations[J].Journal of Statistical Theory and Practice,2016,10(1):59-72

    • [23] Lei G,Chen X M,Zhu J G,et al.Multiobjective sequential optimization method for the design of industrial electromagnetic devices[J].IEEE Transactions on Magnetics,2012,48(11):4538-4541

    • [24] Dalbey K R,Karystinos G N.Generating a maximally spaced set of bins to fill for high-dimensional space-filling Latin hypercube sampling[J].International Journal for Uncertainty Quantification,2011,1(3):241-255

    • [25] Kailkhura B,Thiagarajan J J,Rastogi C,et al.A spectral approach for the design of experiments:design,analysis and algorithms[J].The Journal of Machine Learning Research,2018,19(1):1214-1259

    • [26] Smidova M,Sadilek V,Elias J,et al.Audze-Eglājs criterion for orthogonal and regular triangular grids[C]//1st ECCOMAS Thematic Conference on Uncertainty Quantification in Computational Sciences and Engineering,UNCECOMP 2015.Crete Island,Greece,2015∶664-679

    • [27] Cai X W,Gao L,Li X Y.Efficient generalized surrogate-assisted evolutionary algorithm for high-dimensional expensive problems[J].IEEE Transactions on Evolutionary Computation,2020,24(2):365-379

    • [28] Beyer H G,Sendhoff B.Robust optimization:a comprehensive survey[J].Computer Methods in Applied Mechanics and Engineering,2007,196(33/34):3190-3218

  • 参考文献

    • [1] El-Refaie A M.Fractional-slot concentrated-windings synchronous permanent magnet machines:opportunities and challenges[J].IEEE Transactions on Industrial Electronics,2010,57(1):107-121

    • [2] Magnussen F,Lendenmann H.Parasitic effects in PM machines with concentrated windings[J].IEEE Transactions on Industry Applications,2007,43(5):1223-1232

    • [3] Cassat A,Espanet C,Coleman R,et al.A practical solution to mitigate vibrations in industrial PM motors having concentric windings[J].IEEE Transactions on Industry Applications,2012,48(5):1526-1538

    • [4] Hong J F,Wang S M,Sun Y G,et al.Piecewise stagger poles with continuous skew edge for vibration reduction in surface-mounted PM synchronous machines[J].IEEE Transactions on Industrial Electronics,2021,68(9):8498-8506

    • [5] Wang S,Li H F.Effects of rotor skewing on the vibration of permanent magnet synchronous motors with elastic-plastic stator[J].IEEE Transactions on Energy Conversion,2022,37(1):87-96

    • [6] Jung J W,Kim D J,Hong J P,et al.Experimental verification and effects of step skewed rotor type IPMSM on vibration and noise[J].IEEE Transactions on Magnetics,2011,47(10):3661-3664

    • [7] Peng C,Wang D H,Feng Z K,et al.A new segmented rotor to mitigate torqueripple and electromagnetic vibration of interior permanent magnet machine[J].IEEE Transactions on Industrial Electronics,2022,69(2):1367-1377

    • [8] Wang S M,Hong J F,Sun Y G,et al.Effect comparison of zigzag skew PM pole and straight skew slot for vibration mitigation of PM brush DC motors[J].IEEE Transactions on Industrial Electronics,2020,67(6):4752-4761

    • [9] Jean L B.Vibroacoustic analysis of radial and tangential air-gap magnetic forces in permanent magnet synchronous machines[J].IEEE Transactions on Magnetics,2015,51(6):1-9

    • [10] Zhang J X,Zhang B Y,Feng G H.Influence of pole and slot combination on torque characteristics and radial force of fractional slot permanent magnet machines[J].IEEJ Transactions on Electrical and Electronic Engineering,2021,16(8):1055-1066

    • [11] Lin F,Zuo S G,Wu X D.Electromagnetic vibration and noise analysis of permanent magnet synchronous motor with different slot-pole combinations[J].IET Electric Power Applications,2016,10(9):900-908

    • [12] Xing Z Z,Wang X H,Zhao W L.Optimization of stator slot parameters for electromagnetic vibration reduction of permanent magnet synchronous motors[J].IEEE Transactions on Transportation Electrification,2022,8(4):4337-4347

    • [13] Zhao W X,Zhu S D,Ji J H,et al.Analysis and reduction of electromagnetic vibration in fractional-slot concentrated-windings PM machines[J].IEEE Transactions on Industrial Electronics,2022,69(4):3357-3367

    • [14] Yang I J,Lee S H,Lee K B,et al.A process to reduce the electromagnetic vibration by reducing the spatial harmonics of air gap magnetic flux density[J].IEEE Transactions on Magnetics,2021,57(2):1-6

    • [15] 曹永娟,冯亮亮.基于响应面法的轴向磁场永磁记忆电机多目标优化设计[J].南京信息工程大学学报(自然科学版),2021,13(5):620-627.CAO Yongjuan,FENG Liangliang.Multi-objective optimization design of axial-flux permanent magnet memory motor based on response surface method[J].Journal of Nanjing University of Information Science & Technology(Natural Science Edition),2021,13(5):620-627

    • [16] 华逸舟,刘奕辰,潘伟,等.基于改进粒子群算法的无轴承永磁同步电机多目标优化设计[J].中国电机工程学报,2023,43(11):4443-4452.HUA Yizhou,LIU Yichen,PAN Wei,et al.Multi-objective optimization design of bearingless permanent magnet synchronous motor using improved particle swarm optimization algorithm[J].Proceedings of the CSEE,2023,43(11):4443-4452

    • [17] 王群京,郑耀达,刘先增.基于结构参数优化的电机振动噪声的抑制研究[J].电气工程学报,2023,18(2):16-25.WANG Qunjing,ZHENG Yaoda,LIU Xianzeng.Research on suppression of motor vibration and noise based on structural parameter optimization[J].Journal of Electrical Engineering,2023,18(2):16-25

    • [18] Wang S,Li H F.Reduction of electromagnetic vibration and noise in permanent magnet motor for EVs by optimizing design of rotor based on GPR-PSO model[J].Journal of Electrical Engineering & Technology,2020,15(3):1231-1243

    • [19] Ma C G,Li Q,Liu Q H,et al.Sound quality evaluation of noise of hub permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2016,63(9):5663-5673

    • [20] Ma C G,Li Q,Deng L W,et al.A novel sound quality evaluation method of the diagnosis of abnormal noise in interior permanent-magnet synchronous motors for electric vehicles[J].IEEE Transactions on Industrial Electronics,2017,64(5):3883-3891

    • [21] Yan B,Yang Y B,Wang X H.Design of a large capacity line-start permanent magnet synchronous motor equipped with hybrid salient rotor[J].IEEE Transactions on Industrial Electronics,2021,68(8):6662-6671

    • [22] Aoki S,Hibi T,Ohsugi H.Markov-chain Monte Carlo methods for the Box-Behnken designs and centrally symmetric configurations[J].Journal of Statistical Theory and Practice,2016,10(1):59-72

    • [23] Lei G,Chen X M,Zhu J G,et al.Multiobjective sequential optimization method for the design of industrial electromagnetic devices[J].IEEE Transactions on Magnetics,2012,48(11):4538-4541

    • [24] Dalbey K R,Karystinos G N.Generating a maximally spaced set of bins to fill for high-dimensional space-filling Latin hypercube sampling[J].International Journal for Uncertainty Quantification,2011,1(3):241-255

    • [25] Kailkhura B,Thiagarajan J J,Rastogi C,et al.A spectral approach for the design of experiments:design,analysis and algorithms[J].The Journal of Machine Learning Research,2018,19(1):1214-1259

    • [26] Smidova M,Sadilek V,Elias J,et al.Audze-Eglājs criterion for orthogonal and regular triangular grids[C]//1st ECCOMAS Thematic Conference on Uncertainty Quantification in Computational Sciences and Engineering,UNCECOMP 2015.Crete Island,Greece,2015∶664-679

    • [27] Cai X W,Gao L,Li X Y.Efficient generalized surrogate-assisted evolutionary algorithm for high-dimensional expensive problems[J].IEEE Transactions on Evolutionary Computation,2020,24(2):365-379

    • [28] Beyer H G,Sendhoff B.Robust optimization:a comprehensive survey[J].Computer Methods in Applied Mechanics and Engineering,2007,196(33/34):3190-3218

  • 地址:江苏省南京市宁六路219号    邮编:210044

    联系电话:025-58731025    E-mail:nxdxb@nuist.edu.cn

    南京信息工程大学学报 ® 2024 版权所有  技术支持:北京勤云科技发展有限公司