唐東峰+游世輝
摘 要:實(shí)際工程結(jié)構(gòu)設(shè)計(jì)往往在確定性范疇內(nèi)進(jìn)行,所得結(jié)構(gòu)存在較大失效可能性.基于此,提出一種基于可靠性的連續(xù)體動(dòng)態(tài)拓?fù)鋬?yōu)化方法,將結(jié)構(gòu)可靠性分析方法嵌套到連續(xù)體拓?fù)鋬?yōu)化中.考慮了結(jié)構(gòu)幾何尺寸和材料體積的不確定性,并用高斯分布來(lái)度量.將結(jié)構(gòu)可靠度作為約束嵌套到連續(xù)體拓?fù)鋬?yōu)化中,屬于二次嵌套優(yōu)化問(wèn)題,但計(jì)算效率低下,不適于工程應(yīng)用.提出一種解耦策略將結(jié)構(gòu)可靠性分析從連續(xù)體拓?fù)鋬?yōu)化中解耦出來(lái),使結(jié)構(gòu)可靠性分析與動(dòng)態(tài)拓?fù)鋬?yōu)化為兩個(gè)獨(dú)立的優(yōu)化循環(huán),大大提高了計(jì)算效率.建立以結(jié)構(gòu)基頻最大為優(yōu)化目標(biāo),滿足一定體積約束和可靠度要求的優(yōu)化問(wèn)題,利用各向同性材料懲罰模型(SIMP)和移動(dòng)漸進(jìn)方法(MMA)求解該優(yōu)化問(wèn)題.所提方法可以得到滿足不同可靠度要求的一系列最優(yōu)結(jié)構(gòu),并用標(biāo)準(zhǔn)算例驗(yàn)證其有效性.
關(guān)鍵字:可靠性;不確定性;動(dòng)態(tài)拓?fù)鋬?yōu)化;解耦
中圖分類號(hào):TM30 文獻(xiàn)標(biāo)志碼:A
Reliability-based Structural Dynamic Topology Optimization Method
TANG Dongfeng1,2,YOU Shihui1
(1.College of Civil Engineering and Mechanics,Xiangtan University,Xiangtan 411105,China;
2. School of Information and Electrical Engineering,Hunan University of Science and Technology,Xiangtan 411201,China )
Abstract:Actual engineering structures are often designed using deterministic parameters,which may lead to high failure probability. This paper proposed a reliability-based structural dynamic topology method,in which structural reliability analysis was incorporated into the topology optimization procedure. The geometry dimensions and material volume were considered as uncertain parameters,and it was assumed that they obey a Gaussian distribution. It is a two-nested optimization problem when the structural reliability analysis is considered as constraints into the topology optimization,which results in low efficiency and cannot be used in practice. To this end,a new decouple strategy was proposed to decouple the reliability analysis from the topology optimization procedure. In this case,structure reliability analysis and dynamic topology optimization become two independent optimization cycles,and the computational efficiency is improved enormously. The design problem was then constructed so as to maximize the first eigenfrequency and to meet the volume and reliability requirement. SIMP and MMA were combined to successfully solve the design problem. The proposed method can produce various topologies that satisfy different reliability requirement,and its validity is demonstrated by one benchmark example.
Key words:reliability; uncertainty; dynamic topology optimization; decouple
連續(xù)體拓?fù)鋬?yōu)化方法旨在滿足一定約束條件下尋求材料最優(yōu)分布,相對(duì)于尺寸優(yōu)化[1]和形狀優(yōu)化[2],它有更多的設(shè)計(jì)自由度,同時(shí)也是更具挑戰(zhàn)性的研究領(lǐng)域.自從1988年Bendse和Kikuchi[3]提出基于均勻化法的結(jié)構(gòu)拓?fù)鋬?yōu)化理論以來(lái),連續(xù)體拓?fù)鋬?yōu)化方法[4-10]經(jīng)過(guò)近30年的發(fā)展,已經(jīng)成功應(yīng)用到各種領(lǐng)域,如航天[11]、碰撞[12]、汽車[13]和橋梁[14]等.其中,動(dòng)態(tài)連續(xù)體拓?fù)鋬?yōu)化問(wèn)題[15-17]更具有挑戰(zhàn)性,其難點(diǎn)在于優(yōu)化過(guò)程中需要克服局部模態(tài)和頻率交換現(xiàn)象,研究相對(duì)較少.
盡管傳統(tǒng)連續(xù)拓?fù)鋬?yōu)化方法可以得到性能優(yōu)越的設(shè)計(jì)結(jié)果,但是,其未考慮結(jié)構(gòu)參數(shù)的不確定性.不確定性是工程結(jié)構(gòu)的固有屬性,在結(jié)構(gòu)設(shè)計(jì)時(shí)不容忽略,否則設(shè)計(jì)出的產(chǎn)品會(huì)存在較大的失效風(fēng)險(xiǎn),甚至可能造成災(zāi)難性的后果[18-19].考慮結(jié)構(gòu)不確定性的優(yōu)化設(shè)計(jì)方法一般分為兩類:穩(wěn)健性優(yōu)化設(shè)計(jì)方法[20]和基于可靠性優(yōu)化設(shè)計(jì)方法[21].前者為了降低不確定性變量對(duì)結(jié)構(gòu)響應(yīng)的敏感性,而后者使結(jié)構(gòu)優(yōu)化設(shè)計(jì)增加一個(gè)可靠度約束.基于可靠性的優(yōu)化設(shè)計(jì)方法可以得到滿足設(shè)計(jì)人員一系列不同可靠度要求的結(jié)構(gòu).而基于可靠性的連續(xù)體拓?fù)鋬?yōu)化方法的研究,由于其設(shè)計(jì)變量數(shù)量龐大,且功能函數(shù)往往為隱式,更具有挑戰(zhàn)性.Kharmanda等[22]首次將結(jié)構(gòu)可靠性分析引入到連續(xù)體拓?fù)鋬?yōu)化中,研究發(fā)現(xiàn)所提方法設(shè)計(jì)出的結(jié)構(gòu)比傳統(tǒng)確定性方法所得結(jié)構(gòu)更加可靠.在其研究基礎(chǔ)上,該領(lǐng)域越來(lái)越受到國(guó)內(nèi)外學(xué)者的關(guān)注,出現(xiàn)大量研究成果[23-27].最近,Zhao等[28]提出一種高效的基于可靠性連續(xù)體拓?fù)鋬?yōu)化方法,利用隨機(jī)響應(yīng)面顯式表達(dá)結(jié)構(gòu)失效功能函數(shù).Liu等[29]結(jié)合一次二階矩法和描述函數(shù)拓?fù)鋬?yōu)化方法探討了一種可以獲得光滑邊界且滿足一定可靠度要求的優(yōu)化方法.Jalalpour和Tootkaboni[30]考慮材料的不確定性,提出了一種高效的基于可靠性的連續(xù)體拓?fù)鋬?yōu)化方法,假設(shè)材料服從相關(guān)對(duì)數(shù)正態(tài)分布,并利用隨機(jī)攝動(dòng)法近似得到結(jié)構(gòu)響應(yīng),使結(jié)構(gòu)功能函數(shù)近似顯式化.endprint
然而,上述研究均是靜力學(xué)方面的研究,動(dòng)力學(xué)方面的研究甚少.動(dòng)態(tài)拓?fù)鋬?yōu)化問(wèn)題研究本身就較靜力學(xué)拓?fù)鋬?yōu)化繁瑣,將可靠性分析與動(dòng)態(tài)拓?fù)鋬?yōu)化相結(jié)合更加具有挑戰(zhàn)性.工程中存在大量承受動(dòng)載荷的結(jié)構(gòu),這些結(jié)構(gòu)較之靜力下的結(jié)構(gòu)更容易失效,因此,考慮結(jié)構(gòu)可靠性的動(dòng)態(tài)連續(xù)體拓?fù)鋬?yōu)化方法的研究具有重要理論意義和工程實(shí)際價(jià)值.
基于此,本文提出一種基于可靠性的動(dòng)態(tài)連續(xù)體拓?fù)鋬?yōu)化方法,嘗試解決自由振動(dòng)結(jié)構(gòu)拓?fù)鋬?yōu)化時(shí)考慮結(jié)構(gòu)可靠度的難題.實(shí)際工程制造誤差可能造成結(jié)構(gòu)幾何尺寸和體積的不確定性,假設(shè)這些不確定性變量服從高斯分布,利用一次二階矩法計(jì)算結(jié)構(gòu)的可靠度(可靠性指標(biāo)),并作為約束參與到動(dòng)態(tài)拓?fù)鋬?yōu)化循環(huán)中.而結(jié)構(gòu)可靠性分析也屬于一個(gè)優(yōu)化過(guò)程,因此,整個(gè)優(yōu)化過(guò)程屬于二次嵌套優(yōu)化.考慮到連續(xù)體拓?fù)鋬?yōu)化設(shè)計(jì)變量數(shù)量龐大,導(dǎo)致計(jì)算效率十分低下,無(wú)法實(shí)際應(yīng)用,為突破該瓶頸,提出一種解耦策略,將二次嵌套優(yōu)化循環(huán)解耦成兩個(gè)獨(dú)立的單獨(dú)優(yōu)化循環(huán),大大提高了計(jì)算效率.建立以結(jié)構(gòu)第一階固有頻率最大為優(yōu)化目標(biāo),以體積和可靠度指標(biāo)為約束的連續(xù)拓?fù)鋬?yōu)化問(wèn)題,用各向同性材料懲罰模型懲罰連續(xù)設(shè)計(jì)變量,并利用移動(dòng)漸進(jìn)方法求解該優(yōu)化問(wèn)題.所提方法可以得到一系列滿足設(shè)計(jì)人員不同可靠度需求的拓?fù)浣Y(jié)構(gòu),具有重要的工程意義.
1 動(dòng)態(tài)拓?fù)鋬?yōu)化方法
工程結(jié)構(gòu)受到動(dòng)態(tài)載荷時(shí),設(shè)計(jì)人員往往希望結(jié)構(gòu)固有頻率偏離驅(qū)動(dòng)頻率,以防止共振發(fā)生.未考慮阻尼時(shí),自由振動(dòng)系統(tǒng)的微分方程為:
M+Ku=F=0(1)
式中:M為系統(tǒng)質(zhì)量矩陣;K為系統(tǒng)剛度矩陣;u和分別為系統(tǒng)位移和加速度;F為系統(tǒng)所受外力,自由振動(dòng)系統(tǒng)中其為一個(gè)零向量.
對(duì)式(1)進(jìn)行拉普拉斯變換:
MU(s)s2+KU(s)=0(2)
令s=jω,可以得到:
[K-ω2iM]ui=0(3)
式中:ωi為系統(tǒng)第i階固有頻率;ui是響應(yīng)的特征向量.
固有頻率可以寫(xiě)成:
ω2i=kimi=uTiKuiuTiMui(4)
式中:ki和mi分別為模態(tài)剛度矩陣和模態(tài)質(zhì)量矩陣.
在動(dòng)力學(xué)拓?fù)鋬?yōu)化研究中,比較常用的研究目標(biāo)是使結(jié)構(gòu)的第一階固有頻率最大,并受一定的體積約束,即:
to find:ρ1,…,ρe,…,ρN
maximizeρ:ω21=uT1Ku1uT1Mu1
subject to:[K-ω21M]u1=0,
:∑Ne=1veρe≤V0,
:0<ρmin≤ρe≤1,e=1,…,N.(5)
式中:ρe為單元密度;ρmin的設(shè)定是為了防止發(fā)生奇異解,一般取ρmin=0.001;N為單元個(gè)數(shù);ve和V0分別為單元體積和設(shè)計(jì)允許的最終體積.
利用各向同性材料懲罰模型對(duì)連續(xù)密度進(jìn)行懲罰,文中分別對(duì)單元?jiǎng)偠染仃嚭蛦卧|(zhì)量矩陣進(jìn)行懲罰,懲罰策略如下:
Ke=(ρe)p1K0(6)
Me=(ρe)p2M0(7)
式中:Ke和Me分別為單元?jiǎng)偠染仃嚭蛦卧|(zhì)量矩陣;K0和M0分別為實(shí)體單元?jiǎng)偠染仃嚭蛦卧|(zhì)量矩陣;ρ1和ρ2為值均大于1的懲罰系數(shù),為了在優(yōu)化過(guò)程中抑制中間密度的產(chǎn)生,傳統(tǒng)方法視ρ1和ρ2為相同的值,但是研究發(fā)現(xiàn),這樣對(duì)中間密度的抑制并不理想.因此,本文采用一種新的懲罰策略,取ρ1和ρ2為不同的值,即:
當(dāng)ρ≤ρe≤1,p1=3;
當(dāng)ρmin≤ρe≤0.1,p1=0.01;
當(dāng)0<ρmin≤ρe≤1,p2=1.
問(wèn)題(5)是一個(gè)非凸優(yōu)化問(wèn)題,可用移動(dòng)漸進(jìn)方法(MMA)[31]進(jìn)行高效求解.MMA將初始非凸優(yōu)化問(wèn)題近似為許多獨(dú)立的、小的凸優(yōu)化子問(wèn)題,并對(duì)之進(jìn)行高效求解.利用MMA需要首先得到目標(biāo)函數(shù)對(duì)設(shè)計(jì)變量的一階偏導(dǎo)數(shù):
ω21ρe=uT1Kρeu1uT1Mu1-uT1Mρeu1uT1Ku1(uT1Mu1)2
=uT1Kρe-ω21Mρeu1uT1Mu1(8)
將材料懲罰策略代入式(8)中,得到:
ω21ρe=(ue1)[p(ρe)p1-1K0-ω21p2(ρe)p2-1M0]ue1uT1Mu1(9)
以式(9)為梯度信息,MMA可以快速且高效求解式(5)的動(dòng)態(tài)拓?fù)鋬?yōu)化問(wèn)題.但是,不確定性是結(jié)構(gòu)設(shè)計(jì)中無(wú)法避免的,例如,由于制造誤差造成的結(jié)構(gòu)尺寸不確定性、材料自身的不確定性和結(jié)構(gòu)使用過(guò)程中加載力的不確定性等.而上述動(dòng)態(tài)拓?fù)鋬?yōu)化問(wèn)題并未考慮結(jié)構(gòu)的不確定性,所設(shè)計(jì)出的結(jié)構(gòu)往往存在較高的失效概率.因此,文章嘗試將結(jié)構(gòu)不確定性引入到動(dòng)態(tài)拓?fù)鋬?yōu)化中,設(shè)計(jì)出滿足一定結(jié)構(gòu)可靠度要求的結(jié)構(gòu).
2 基于可靠性的動(dòng)態(tài)拓?fù)鋬?yōu)化方法
基于可靠性的連續(xù)體拓?fù)鋬?yōu)化方法有較多學(xué)者進(jìn)行了研究,但大多數(shù)局限于靜力學(xué)結(jié)構(gòu)的優(yōu)化設(shè)計(jì).將可靠性分析與動(dòng)態(tài)拓?fù)鋬?yōu)化問(wèn)題相結(jié)合更具有挑戰(zhàn)性.
2.1 問(wèn)題提出
基于可靠性的動(dòng)態(tài)拓?fù)鋬?yōu)化方法旨在提高所設(shè)計(jì)結(jié)構(gòu)的可靠度水平,當(dāng)結(jié)構(gòu)可靠度作為拓?fù)鋬?yōu)化的一個(gè)約束時(shí),利用可靠性指標(biāo)法,優(yōu)化問(wèn)題(5)變?yōu)椋?/p>
to find:ρ1,…,ρe,…,ρN
maximizeρ:ω21=uT1Ku1uT1Mu1
suject to:[K-ω21M]u1=0,
:g1=∑Ne=1veρe-V0≤0,
:g2=β0j-βj(ρe,μγ)≤0,j=1,…,m,
:0<ρmin≤ρe≤1,e=1,…,N,(10)
其中,結(jié)構(gòu)可靠性指標(biāo)的求解如下:
maximizeμγ:βj(η)=‖η‖suject to:Gj(d,η)=0(11)endprint
式中:g1和g2為不等式約束,分別控制體積和結(jié)構(gòu)可靠度;β0j和βj分別為第j個(gè)結(jié)構(gòu)可靠性指標(biāo)目標(biāo)值和第j個(gè)結(jié)構(gòu)可靠性指標(biāo);m指結(jié)構(gòu)系統(tǒng)一共有m個(gè)可靠度約束;μγ為結(jié)構(gòu)不確定性變量均值;Gj(·)為在標(biāo)準(zhǔn)正態(tài)空間中第j個(gè)結(jié)構(gòu)的功能函數(shù),當(dāng)Gj(·)>0時(shí)認(rèn)為結(jié)構(gòu)處于可靠狀態(tài),當(dāng)Gj(·)<0時(shí)認(rèn)為結(jié)構(gòu)處于失效狀態(tài),Gj(·)=0時(shí)為結(jié)構(gòu)極限狀態(tài);‖·‖指向量的范數(shù);η為不確定性變量在標(biāo)準(zhǔn)正態(tài)坐標(biāo)系中的形式.
利用伴隨法可以求得式(10)優(yōu)化問(wèn)題的敏感系數(shù),然后利用MMA進(jìn)行求解.然而,該問(wèn)題求解屬于二次嵌套優(yōu)化問(wèn)題,即結(jié)構(gòu)動(dòng)態(tài)拓?fù)鋬?yōu)化每次迭代循環(huán)都需要判斷該次迭代結(jié)果是否滿足所需結(jié)構(gòu)可靠度要求,而結(jié)構(gòu)可靠性分析也是一個(gè)優(yōu)化問(wèn)題.直接求解計(jì)算效率十分低下,不適合實(shí)際工程應(yīng)用.因此,文章提出一種解耦策略,將二次嵌套優(yōu)化問(wèn)題轉(zhuǎn)化為兩個(gè)單獨(dú)的優(yōu)化循環(huán),大大提高了計(jì)算效率.
2.2 解耦策略
不確定性變量用Y度量,Y=(Y1,…,Yt),結(jié)構(gòu)擁有t個(gè)不確定變量,其均值和方差分別為μγ和σ2Y.這些不確定性變量對(duì)結(jié)構(gòu)可靠性分析具有重要影響,然而,它們對(duì)結(jié)構(gòu)固有頻率并不一定有很大的影響.因此,需要首先計(jì)算目標(biāo)函數(shù)對(duì)不確定性變量的敏感性,選擇對(duì)目標(biāo)函數(shù)影響大的不確定性變量,去除影響小的不確定性變量.這里使用有限差分法[32]:
ω21μY1=δω21δμY1=ω21(μY1+δμY1)-ω21(μY1)δμY1(12)
式中:μYtμYt=0.005.
確定對(duì)目標(biāo)函數(shù)貢獻(xiàn)大的不確定性變量后,需要確定可靠性指標(biāo)對(duì)不確定性變量的梯度信息以每個(gè)迭代步更新可靠性指標(biāo).不確定性變量首先轉(zhuǎn)化為標(biāo)準(zhǔn)正態(tài)分布:
η=Y-μYσY(13)
此時(shí),結(jié)構(gòu)的可靠性指標(biāo)為:
β=∑η2t(14)
可靠性指標(biāo)對(duì)不確定變量的一階導(dǎo)數(shù)可以顯式表達(dá):
ε=βηt=ηtβ(15)
利用式(15)的梯度信息可以在每一迭代步更新結(jié)構(gòu)可靠性指標(biāo)的值.
為了將結(jié)構(gòu)可靠性分析解耦出來(lái),首先給出一個(gè)初始猜測(cè)值η,將該猜測(cè)值代入式(14)中計(jì)算可靠性指標(biāo).若該可靠性指標(biāo)大于目標(biāo)值,利用式(15)梯度信息更新其數(shù)值,直到滿足目標(biāo)值為止.此時(shí),得到一個(gè)新的η*,進(jìn)而可以得到在物理坐標(biāo)系下的不確定性變量:
Y*=η*σY+μγ(16)
上式得到的是一個(gè)確定性向量,使用其作為拓?fù)鋬?yōu)化的輸入量所得到的拓?fù)浣Y(jié)構(gòu),可以滿足所需結(jié)構(gòu)可靠度要求.這樣,結(jié)構(gòu)可靠性分析成功地從拓?fù)鋬?yōu)化循環(huán)中解耦出來(lái),大大提高了計(jì)算效率.需要指出的是,由于文章只關(guān)心線彈性結(jié)構(gòu)的優(yōu)化設(shè)計(jì),故該解耦策略的精度可以接受.使用解耦策略的算法流程如圖1所示.
3 數(shù)值算例
用一個(gè)標(biāo)準(zhǔn)算例來(lái)驗(yàn)證文章所提方法的有效性,同時(shí)說(shuō)明動(dòng)態(tài)拓?fù)鋬?yōu)化設(shè)計(jì)考慮結(jié)構(gòu)可靠性的重要性.這里只是為了驗(yàn)證算法的有效性,所有單位均無(wú)量綱處理.同時(shí),使用了防止優(yōu)化過(guò)程中棋盤(pán)格現(xiàn)象的策略[33],得到邊界光滑的拓?fù)浣Y(jié)構(gòu).
兩端固定梁的動(dòng)態(tài)拓?fù)鋬?yōu)化設(shè)計(jì)問(wèn)題,如圖2所示,長(zhǎng)和寬分別為100和20,假設(shè)均服從正態(tài)分布,且離差均為0.02.初始域被離散為100×20個(gè)矩形單元.設(shè)計(jì)域材料的屬性為:彈性模量為1,泊松比為0.3,密度為1,體積約束設(shè)為0.9,假設(shè)服從正態(tài)分布量,離差為0.02.
確定性動(dòng)態(tài)拓?fù)鋬?yōu)化所得結(jié)果如圖3所示,拓?fù)浣Y(jié)構(gòu)邊界光滑,便于實(shí)際結(jié)構(gòu)制造.圖4表明利用MMA求解動(dòng)態(tài)拓?fù)鋬?yōu)化問(wèn)題收斂性良好,迭代106步后收斂,結(jié)構(gòu)的第一階固有頻率為0.574 7,圖中也給出了結(jié)構(gòu)的第二階和第三階固有頻率.
然而,在實(shí)際制造加工中誤差無(wú)法避免,由誤差導(dǎo)致的結(jié)構(gòu)不確定性,在圖3結(jié)構(gòu)中并未考慮,該結(jié)構(gòu)在實(shí)際工程服役過(guò)程中往往會(huì)存在較大的失效概率.利用文中所提方法優(yōu)化設(shè)計(jì)該梁結(jié)構(gòu),設(shè)定結(jié)構(gòu)可靠性指標(biāo)值β0=4.0,利用一次二階矩可得結(jié)構(gòu)的失效概率為pf=Φ(-β0)=0.000 032,所得拓?fù)浣Y(jié)構(gòu)(圖5所示)有很高的可靠度,可以很好地應(yīng)對(duì)復(fù)雜多變的不確定性服役環(huán)境,如制造誤差等.所得拓?fù)浣Y(jié)構(gòu)明顯與圖3不同,可以更好地抵抗結(jié)構(gòu)不確定性的變化.圖6顯示所提算法收斂性良好,目標(biāo)函數(shù)迭代90步收斂為0.520 3.需要指出的是,圖5中結(jié)構(gòu)最終材料體積為0.916,大于確定性優(yōu)化結(jié)果的09,這是由于結(jié)構(gòu)需要更多的材料來(lái)提高自身的可靠度,這也和文獻(xiàn)[26,27]研究結(jié)果一致.
4 結(jié) 論
文章將結(jié)構(gòu)可靠性分析引入到連續(xù)體動(dòng)態(tài)拓?fù)鋬?yōu)化中,提出一種基于可靠性的連續(xù)體拓?fù)鋬?yōu)化方法.同時(shí),為了提高計(jì)算效率,提出一種新的解耦策略,將結(jié)構(gòu)可靠度計(jì)算約束從拓?fù)鋬?yōu)化循環(huán)中解耦出來(lái),大大提高了計(jì)算效率.得出以下結(jié)論;
1)考慮了由于制造誤差等造成的幾何尺寸和體積不確定性,設(shè)計(jì)出的結(jié)構(gòu)可以滿足設(shè)計(jì)人員的可靠度要求,可以更好地應(yīng)對(duì)復(fù)雜的不確定性的服役環(huán)境.
2)設(shè)計(jì)出的結(jié)構(gòu)相對(duì)于確定性優(yōu)化結(jié)構(gòu)擁有更多的材料,擁有更高的可靠度.
3)所提算法有很好的收斂性,迭代過(guò)程平穩(wěn),并得到邊界光滑的拓?fù)浣Y(jié)構(gòu).
參考文獻(xiàn)
[1] LI C,KIM I Y,JESWIET J. Conceptual and detailed design of an automotive engine cradle by using topology,shape,and size optimization[J]. Structural and Multidisciplinary Optim-ization,2015,51(2); 547-564.
[2] TOLEDO R,AZNREZ J J,GREINER D,et al. Shape design optimization of road acoustic barriers featuring top-edge devices by using genetic algorithms and boundary elements[J]. Engineering Analysis with Boundary Elements,2016,63:49-60.endprint
[3] BENDSE M P,KIKUCHI N. Generating optimal topologies in structural design using a homogenization method[J]. Computer Methods in Applied Mechanics and Engineering,1988,71(2):197-224.
[4] SIGMUND O. A 99 line topology optimization code written in Matlab[J]. Structural and Multidisciplinary Optimization,2001,21(2): 120-127.
[5] BENDSE M P,SIGMUND O. Topology optimization:theo-ry,methods,and applications[M]. Springer Science & Business Media,2013:100-102.
[6] WANG M Y,WANG X,GUO D. A level set method for structural topology optimization[J]. Computer Methods in Applied Mechanics and Engineering,2003,192(1): 227-246.
[7] XIE Y M,STEVEN G P. Basic evolutionary structural optimization[M]. London:Springer,1997:74-75.
[8] 隋允康,楊德慶,王備. 多工況應(yīng)力和位移約束下連續(xù)體結(jié)構(gòu)拓?fù)鋬?yōu)化[J]. 力學(xué)學(xué)報(bào),2000,32(2): 171-179.
SUI Yongkang,YANG Deqing,WANG Bei. Topological optimization of continuum structure with stress and displacement constraints under multiple loading cases[J]. Acta Mechanica Sinica,2000,32(2): 171-179. (In Chinese)
[9] ZHOU M,LAZAROV B S,WANG F,et al. Minimum length scale in topology optimization by geometric constraints[J]. Computer Methods in Applied Mechanics and Engineering,2015,293: 266-282.
[10]GUEST J K. Optimizing the layout of discrete objects in structures and materials:a projection-based topology optimization approach[J]. Computer Methods in Applied Mechanics and Engineering,2015,283: 330-351.
[11]LPEZ C,BALDOMIR A,HERNNDEZ S. Deterministic versus reliability-based topology optimization of aeronautical structures[J]. Structural and Multidisciplinary Optimization,2016,53(4): 907-921.
[12]DUDDECK F,HUNKELER S,LOZANO P,et al. Topology optimization for crashworthiness of thin-walled structures under axial impact using hybrid cellular automata[J]. Structural and Multidisciplinary Optimization,2016,54(3):415-428
[13]張偉,侯文彬,胡平. 基于拓?fù)鋬?yōu)化的電動(dòng)汽車白車身優(yōu)化設(shè)計(jì)[J]. 湖南大學(xué)學(xué)報(bào):自然科學(xué)版,2014,41(10):42-48.
ZHANG Wei,HOU Wenbin,HU Ping. The body in white optimization of an electric vehicle using topology optimization[J]. Journal of Hunan University:Natural Sciences,2014,41(10):42-48.(In Chinese)
[14]劉志文,辛亞兵,陳政清. 鋁合金橋面板合理斷面形式拓?fù)浞治龊蛢?yōu)化[J]. 湖南大學(xué)學(xué)報(bào):自然科學(xué)版,2010,37(1);11-16.
LIU Zhiwen,XIN Yabing,CHEN Zhengqing. Topologic analysis and optimization of reasonable cross-section of aluminum alloy bridge decks[J]. Journal of Hunan University:Natural Sciences,2010,37(1):11-16.(In Chinese)
[15]LIU J,WEN G,QING Q,et al. An efficient method for topology optimization of continuum structures in the presence of uncertainty in loading direction[J]. International Journal of Computational Methods,2017,14(2):1750054.endprint
[16]LIU J,WEN G,XIE Y M. Layout optimization of continuum structures considering the probabilistic and fuzzy directional uncertainty of applied loads based on the cloud model[J]. Structural and Multidisciplinary Optimization,2016,53(1): 81-100.
[17]LIU J,WEN G,HUANG X. To avoid unpractical optimal design without support[J]. Structural and Multidisciplinary Optimization,2017: 1-7.
[18]LIU J,QING Q,DENG Y,et al. Fatigue reliability study on T-welded component considering load shedding[J]. Fatigue & Fracture of Engineering Materials & Structures,2015,38(7):780-788.
[19]LIU J,WEN G. Continuum topology optimization considering uncertainties in load locations based on the cloud model[J]. Engineering Optimization,2017: 1-20.
[20]BEYER H G,SENDHOFF B. Robust optimization-a comprehensive survey[J]. Computer Methods in Applied Mechanics and Engineering,2007,196(33):3190-3218.
[21]VALDEBENITO M A,SCHUЁLLER G I. A survey on approaches for reliability-based optimization[J]. Structural and Multidisciplinary Optimization,2010,42(5):645-663.
[22]KHARMANDA G,OLHOFF N,MOHAMED A,et al. Reliability-based topology optimization[J]. Structural and Multidisciplinary Optimization,2004,26(5): 295-307.
[23]JUNG H S,CHO S. Reliability-based topology optimization of geometrically nonlinear structures with loading and material uncertainties[J]. Finite Elements in Analysis and Design,2004,41(3): 311-331.
[24]MAUTE K,F(xiàn)RANGOPOL D M. Reliability-based design of MEMS mechanisms by topology optimization[J]. Computers & Structures,2003,81(8): 813-824.
[25]KANG Z,LUO Y. Non-probabilistic reliability-based topology optimization of geometrically nonlinear structures using convex models[J]. Computer Methods in Applied Mechanics and Engineering,2009,198(41):3228-3238.
[26]SILVA M,TORTORELLI D A,NORATO J A,et al. Component and system reliability-based topology optimization using a single-loop method[J]. Structural and Multidisciplinary Optimization,2010,41(1): 87-106.
[27]NGUYEN T H,SONG J,PAULINO G H. Single-loop system reliability-based topology optimization considering statistical dependence between limit-states[J]. Structural and Multidisciplinary Optimization,2011,44(5): 593-611.
[28]ZHAO Q,CHEN X,MA Z D,et al. Reliability-based topology optimization using stochastic response surface method with sparse grid design[J]. Mathematical Problems in Engineering,2015,2013(1):1-13.
[29]LIU J,WEN G,ZUO H Z,et al. A simple reliability-based topology optimization approach for continuum structures using a topology description function[J]. Engineering Optimization,2016,48(7): 1182-1201.
[30]JALALPOUR M,TOOTKABONI M. An efficient approach to reliability-based topology optimization for continua under material uncertainty[J]. Structural and Multidisciplinary Optimization,2016,53(4): 759-772.
[31]SVANBERG K. The method of moving asymptotes—a new method for structural optimization[J]. International Journal for Numerical Methods in Engineering,1987,24(2):359-373.
[32]LIPNIKOV K,MANZINI G,SHASHKOV M. Mimetic finite difference method[J]. Journal of Computational Physics,2014,257(1): 1163-1227.
[33]POULSEN T A. A simple scheme to prevent checkerboard patterns and one-node connected hinges in topology optimization[J]. Structural and Multidisciplinary Optimization,2002,24(5): 396-399.endprint