文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > 工程科技 > 机械/仪表 > A 99 line topology optimization code written in Matlab
侵权投诉

A 99 line topology optimization code written in Matlab

拓扑优化的经典,99行MATLAB程序实现整个拓扑优化,是学习拓扑优化概念及编程的绝好教材。

fullysolvedlarge-scalecomplianceminimizationprob-lemsusingadual-approachbutotherapproachesbasedongeneticalgorithmsorothersemi-randomapproachesrequirethousandsoffunctionevaluationsevenforsmallnumberofelementsandmustbeconsideredimpractical.Apartfromabovementionedapproaches,whichallsolvewellde nedproblems(e.g.minimizationofcom-pliance)anumberofheuristicorintuitionbasedap-proacheshavebeenshowntodecreasecomplianceorotherobjectivefunctions.Amongthesemethodsareso-calledevolutionarydesignmethods(seee.g.XieandSteven1997;Baumgartneretal.1992).Apartfrombe-ingveryeasytounderstandandimplement(atleastforthecomplianceminimizationcase),themainmoti-vationfortheevolutionaryapproachesseemstobethatmathematicallybasedorcontinuousvariableapproaches“involvesomecomplexcalculusoperationsandmathe-maticalprogramming”(citationfromLietal.1999)andtheycontain“mathematicalmethodsofsomecomplex-ity”(citationfromZhaoetal.1998)whereastheevo-lutionaryapproach“takesadvantageofpowerfulcom-putingtechnologyandintuitiveconceptsofevolutionprocessesinnature”(citationfromLietal.1999).Twothingscanbearguedagainstthis.First,theevolutionaryapproachesbecomecomplicatedthemselves,oncemorecomplexobjectivesthancomplianceminimizationareconsideredandsecond,asshowninthispaper,the“math-ematicallybased”approachesforcomplianceminimiza-tionaresimpletoimplementaswellandarecompu-tationallyequallye cient.Furthermore,mathematicalprogrammingbasedmethodscaneasilybeextendedtoothernon-complianceobjectivessuchasnon-self-adjointandmultiphysicsproblemsandtoproblemswithmultipleconstraints(e.g.Sigmund1999).Extensionsoftheevolu-tionaryapproachtosuchcasesseemmorequestionable.ThecompleteMatlabcodeisgivenintheAppendix.Theremainderofthepaperconsistsofde nitionanddiscussionoftheoptimizationproblem(Sect.2),com-mentsabouttheMatlabimplementation(Sect.3)fol-lowedbyadiscussionofextensions(Sect.4)andaconclu-sion(Sect.5).

2

Thetopologyoptimizationproblem

Anumberofsimpli cationsareintroducedtosimplifytheMatlabcode.First,thedesigndomainisassumedtoberectangularanddiscretizedbysquare niteelements.Inthisway,thenumberingofelementsandnodesissimple(columnbycolumnstartingintheupperleftcorner)andtheaspectratioofthestructureisgivenbytheratioofelementsinthehorizontal(nelx)andtheverticaldirec-tion(nely).1

1

Namesintype-writerstylerefertoMatlabvariablenamesthatdi erfromtheobvious(seetheMatlabcodeintheAppendix)

121

Atopologyoptimizationproblembasedonthepower-lawapproach,wheretheobjectiveistominimizecompli-ancecanbewrittenas

min N x:c(x)=UTKU=(xe)puTek0ue subjectto:V(x)

e=1 V0

=f :KU=F , :0<x≤x≤1

min(1)whereUandFaretheglobaldisplacementandforcevectors,respectively,Kistheglobalsti nessmatrix,ueandkearetheelementdisplacementvectorandsti nessmatrix,respectively,xisthevectorofdesignvariables,xminisavectorofminimumrelativedensities(non-zerotoavoidsingularity),N(=nelx×nely)isthenumberofelementsusedtodiscretizethedesigndomain,pisthepenalizationpower(typicallyp=3),V(x)andV0isthematerialvolumeanddesigndomainvolume,respectivelyandf(volfrac)istheprescribedvolumefraction.

Theoptimizationproblem(1)couldbesolvedusingseveraldi erentapproachessuchasOptimalityCriteria(OC)methods,SequentialLinearProgramming(SLP)methodsortheMethodofMovingAsymptotes(MMAbySvanberg1987)andothers.Forsimplicity,wewillhereuseastandardOC-method.

FollowingBendsøe(1995)aheuristicupdatingschemeforthedesignvariablescanbeformulatedasxnewe

=

max(xmin,x e m) ifx eBηe≤max(xmin,xe m), xeBηe

ifmax(xmin,xe m)<xeBηe<min(1,xe+m), min(1,xe+m)ifmin(1,xe+m)≤xeBη

e,

(2)wherem(move)isapositivemove-limit,η(=1/2)isanumericaldampingcoe cientandBeisfoundfromtheoptimalityconditionas

cB xe

e= V,

(3)

λ xe

whereλisaLagrangianmultiplierthatcanbefoundbyabi-sectioningalgorithm.

Thesensitivityoftheobjectivefunctionisfoundas c

x= p(xe)p 1uTek0ue.(4)

e

上一页第2页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email