混凝土破坏准则 william-warnke模型
- 格式:pdf
- 大小:2.84 MB
- 文档页数:31
天津城建大学学报 第22卷 第1期 2016年2月Journ al of Tian jin Chen gjian Un iversity V ol.22 No.1 Feb. 2016收稿日期:2015-03-09;修订日期:2015-07-06基金项目:国家自然科学基金(51508373);天津市应用基础与前沿技术研究计划重点项目(15JCZDJC39900);天津市高等学校科技发展基金计划项目(20140909)作者简介:王 宁(1989—),女,山东德州人,天津城建大学硕士生. 通讯作者:吕 杨(1984—),男,讲师,博士,从事结构抗震与设计的研究.E-mail :lvyangtju@土木工程钢筋混凝土剪力墙非线性分析模型及损伤准则王 宁,吕 杨(天津城建大学 土木工程学院,天津 300384)摘要:通过LS -DYNA 程序开发了一种剪力墙的宏观模型,基于此模型,提出一种适用于剪力墙的考虑首次超越破坏和能量累积破坏的双参数损伤准则.超越破坏是通过拟合各等级损伤指数与相应应变值的对数函数,累积破坏的极限耗能定义为所开发模型包络线所围成的面积,当应变阈值和累积耗能超过临界值时,可以考虑两者的耦合效应.应用所开发的剪力墙模型,数值模拟了两个剪力墙的低周反复试验,结果表明:所开发的模型能以很低的计算成本较精确地模拟剪力墙滞回关系,并且所提出的损伤准则能很好地跟踪构件损伤发展过程.关 键 词:钢筋混凝土剪力墙;损伤准则;非线性分析模型;超越破坏;累积破坏中图分类号:TU398.2 文献标志码:A 文章编号:2095-719X (2016)01-0017-05钢筋混凝土剪力墙的数值分析方法一直制约着结构精细化分析的发展进程,其中采用实体单元和梁单元分别模拟混凝土和钢筋的微观模型方法具有很高的模拟精度,但这种模拟方法由于计算成本过高,很难用于大型结构分析.宏观模型采用广义力和广义位移建立构件的力学相关关系,将一片剪力墙离散成一个或几个单元,这种模拟方法计算成本低,同时能保证工程上所需的计算精度,因此国内外学者通过试验研究和理论研究,提出了一系列用于模拟钢筋混凝土剪力墙的非线性动力滞回模型:如Paknahad 等[1]提出的三角形单元模型能以很粗略的网格划分获得较高的精度;Mo 等[2]提出一种专门用于分析墙类构件的SMM (softened memb rane model )模型,并基于OpenSees 软件开发了SCS (simulation of concrete structures )平台,数值分析与试验对比表明,该模型具有很高的求解精度;谢凡等[3]提出了一种能同时考虑轴向拉压变形与剪切变形相互作用,以及弯曲变形与剪切变形相互作用的新的多垂杆单元模型;基于非线性梁单元模型,朱杰江等[4]建立了一种非线性剪力墙模型.为了分析剪力墙结构强震作用下的失效破坏过程,还需要建立剪力墙的损伤准则.目前,关于钢筋混凝土柱已经建立了很多构件层次的损伤准则[5-8],最典型的如Park -Ang 双参数模型[6].部分准则也可以近似用于剪力墙结构的损伤分析,但由于强震作用下剪力墙损伤过程主要是水平抗力的退化过程(柱子主要是抗竖向承载力的退化过程),并且剪力墙水平位移由剪切变形和弯曲变形等多种变形形式组成,致使剪力墙破坏点的确定方法与柱类构件有很大的差异,因此还需建立适用于剪力墙构件的损伤准则.笔者在LS -DYNA 程序[9]中开发了一种宏观的剪力墙模型,模型包络线为考虑下降段的5折线,加卸载特性通过一个3次多项式近似模拟;基于此模型提出一种适用于剪力墙构件的双参数损伤准则,准则能考虑首次超越破坏与累积破坏之间的耦合效应,适用于强震作用下剪力墙结构体系损伤过程分析.1 损伤准则1.1 剪力墙非线性分析模型采用LS -DYNA 程序[9]的壳单元模拟剪力墙构件应力应变关系的模型.该模型采用5条折线,分别模拟剪力墙混凝土开裂、钢筋屈服、钢筋强化、钢筋塑性流变以及强度降低等过程,并假定剪力墙·18· 天津城建大学学报 2016年 第22卷 第1期不发生压溃破坏,模型包络线如图1所示.图1 RC 剪力墙剪切应力-应变关系图1中,G 为混凝土剪切模量,σmax 为混凝土最大应力,定义为max c,c s y 2.7(1.9)0.80.5a f f σλρ=−+ (1)式中:λ为剪跨比;s ρ为剪力墙配筋率;c,c f 为受约束混凝土抗压强度;y f 为钢筋屈服强度;a 为考虑边缘框架对剪力墙强度提高的参数,定义为c wmax((0.4),1.0)Aa A =+ (2)式中:A c 、A w 分别为边缘加强构件截面积和墙体截面积.剪力墙构件屈服后混凝土强度会发生退化,后继强度定义为submax max σχσ= (3)式中:χ为考虑剪力墙屈服后强度折减系数.剪力墙极限位移由弹性弯曲、弹性剪切、塑性弯曲和塑性剪切组成,采用文献[10]的方法计算 es eb ps pb Δ=Δ+Δ+Δ+Δ (4)式中:es Δ、eb Δ、ps Δ和pb Δ分别为弹性剪切变形、弹性弯曲变形、塑性剪切变形和塑性弯曲变形,可由下式求得:2eb y e 13l ϕΔ= (5)2w es eb e0.75()h l Δ=Δ (6)pb u p u p e 12l l l φφΔ=+ (7)ps p sPl K Δ= (8)式中:l e 为剪力墙弹性区高度;w h 为剪力墙截面高度;y ϕ为剪力墙屈服曲率;l p 为塑性铰长度;u φ为塑性铰区域的极限曲率;P 为剪力墙峰值承载力;K s 为剪力墙塑性铰区抗剪刚度.由式(5)-(8)求得剪力墙极限位移后,即可求得剪力墙极限应变.模型沿定义的3次多项式[9]曲线进行加卸载,如图2所示,可通过调节多项式参数控制滞回环饱满程度、加卸载刚度等.此外,所开发模型采用显示求解算法,避免了隐式求解算法中负刚度等问题,因此可以实现对剪力墙混凝土开裂、钢筋屈服和强度退化等重要性能的模拟.图2 模型加卸载路径1.2 剪力墙损伤准则地震作用下建筑结构的震害和损伤可以划分为基本完好、轻微破坏、中等破坏、严重破坏和倒塌5个等级[11].表1和图1分别为各等级损伤对应应变值和损伤指数[12].表1 剪力墙损伤准则[12]损伤状态 基本完好 轻微破坏 中等破坏 严重破坏 倒塌 应变 <εsyεsy ~γmaxγmax ~1.5,γmax 1.5,γmax ~2.0,γmax>2.0,γmax 损伤值<0.2 0.2~0.40.4~0.60.6~0.9>0.9已有研究表明,钢筋混凝土结构损伤破坏过程由首次超越破坏和循环累积破坏两部分组成,同时两种破坏的发展过程会相互耦合[13],即随着循环累积破坏的发展,结构发生超越破坏的阈值会降低;反之,随着超越破坏的阈值增加,结构发生累积破坏的界限不断降低.为此,将结构损伤指数d i 分解成考虑首次超越破坏的d ε,i 和考虑循环累积破坏的d E,i ,损伤指数定义为d i =d ε,i + d E ,i (9)式中:d i 为结构第i 时间步损伤指数,根据表1所示损伤等级划分,损伤指数小于0.2时表示结构基本完好,大于0.9时表示结构发生倒塌破坏.天津城建大学学报 王 宁等:钢筋混凝土剪力墙非线性分析模型及损伤准则 ·19·根据表1中剪力墙损伤状态,采用数值拟合的方法,建立剪力墙结构不同损伤等级时应变与损伤值的关系,如图3所示;同时引入参数d ε和d E 来考虑超越破坏和累积破坏之间的相互影响[12],则剪力墙结构超越破坏和累积破坏可以分别定义为 ,u(1)ln(1.4190.983)ii E d εεβαε=−+ (10),uiE iE d E εβα= (11)式中:β为考虑超越破坏与积累耗能对损伤贡献的权重系数;E c 、c ε分别为考虑滞回耗能与损伤阈值相互影响的临界值;E i 、i ε分别为第i 时刻前总滞回耗能和最大应变;E u 、u ε分别为极限滞回耗能和失效应变;εα、E α分别为考虑超越破坏与累积破坏相互影响[14]的系数,定义为c ccu 11i E i i E E E E E E E αβ⎧⎪=−⎨−⎪⎩<≥(12)c ccu 11(1)i i i εεεαεεβεεε⎧⎪=−⎨−−⎪⎩<≥(13)图3 损伤发展过程拟合曲线上述损伤准则中,有β、u E 、u ε、c E 和c ε等5个参数需要确定.由于剪力墙和钢筋混凝土柱首次超越破坏和累积破坏过程相似,而剪力墙模型试验数据离散性很大,因此,参数β、c E 和c ε采用钢筋混凝土柱的试验结果进行确定.在钢筋混凝土柱的Park -Ang 损伤模型[6]中,考虑滞回耗能效应的权重系数通过下式确定(0.4470.0730.240.314)0.7vln hρβρ=−+++×(14)式中:l /h 为剪跨比;n 为柱子轴压比;ρ、v ρ分别为柱子配筋率和体积配箍率.文献[6]中参数β均值大约为0.05,建议值为0.1.此外,在一些改进的Park -Ang 模型中[7-8],参数β取值在0~0.85之间,均值约为0.48.实际应用中,可以通过试验或者数值模拟的方法得到.对于普通钢筋混凝土构件,β可取0.1~0.2.E u 为剪力墙破坏时总滞回耗能,定义为剪力墙模型包络线与应变轴所围多边形的面积,如图1所示.剪力墙破坏点极限应变u ε由剪力墙极限位移计算得到,参见剪力墙模型式(4).当i E (或εi )大于c E (或c ε)时,考虑剪应变阈值与总滞回耗能耦合效应.钢筋混凝土柱低周疲劳试验结果表明[13],钢筋混凝土柱不发生疲劳破坏的界限值约为最大值的25%(μΔ=1与μΔ=4),考虑到强震冲击作用下混凝土构件循环次数和地震动加速度峰值有限,取μΔ=2.5作为疲劳破坏的上限,即E c (c ε)取u E (u ε)的40%.2 算例分析采用上述剪力墙模型和损伤准则,数值模拟文献[15]中编号为HPCW -01和HPCW -02的剪力墙的低周反复试验,数值模拟滞回曲线和试验滞回曲线如图4-7所示.模型参数如表2所示,损伤准则参数如表3所示.图4 剪力墙HPCW -01模拟结果图5 剪力墙HPCW -01试验结果·20·天津城建大学学报 2016年第22卷第1期图6剪力墙HPCW-02模拟结果图7剪力墙HPCW-02试验结果从图4-7可以看出,所开发的模型能很好地模拟剪力墙反复荷载作用下的滞回过程.由图4和图6可以看出,所采用的剪力墙模型能较好跟踪剪力墙真实的力-位移关系:在加载初期,剪力墙基本处于弹性受力阶段,力-位移关系为直线;随着荷载的增大,剪力墙配置的钢筋发生屈服进入塑性流变阶段,此时剪力墙内力不随位移的增加而增大;继续加载,剪力墙内钢筋发生拉断,受压区混凝土压溃,力-位移曲线出现下降;继续加载,剪力墙逐渐破坏失效.由于模型不能模拟混凝土疲劳效应,所以数值模拟中每个荷载步只加载一个循环,并且从图5和图7可以看出,剪力墙力-位移曲线中相同位移加载的三个循环几乎重合,即剪力墙累积破坏不是很明显.表2模型参数试件编号配筋率/%,钢筋屈服强度/MPa剪跨比混凝土抗压强度/MPa极限应变最大应力/MPaHPCW-01 1.587 433.3 2 60 0.021,6 3.26 HPCW-02 1.617 433.3 2 60 0.026,0 3.31表3损伤准则参数试件编号极限应变εu极限滞回耗能/kJ参数β滞回耗能临界值/kJ应变临界值εcHPCW-01 0.021,6 59.40.123.8 0.020,8 HPCW-02 0.026,0 73.0 0.129.2 0.017,3应用所提出的剪力墙损伤准则计算得到两片剪力墙损伤发展过程,如图8-9所示.由图8-9可以看出:两片剪力墙损伤发展过程相似,即损伤随加载幅值的增加而逐渐增大,损伤在位移幅值变大处突变,可见本文提出的损伤准则能较好地反映拟静力试验损伤发展过程.图中实线、虚线分别表示损伤准则中考虑和不考虑首次超越破坏与累积效应的相互影响.由此可见,在应变阈值和累积耗能达到临界值之前,两种损伤准则计算得到的损伤发展曲线完全重合;当应变阈值和累积耗能超过临界值之后,考虑相互影响的损伤准则损伤发展增快.图8剪力墙HPCW-01损伤发展过程图9剪力墙HPCW-02损伤发展过程3结 论(1)所采用的剪力墙数值计算模型可以较精确地模拟剪力墙结构受拉开裂、纵筋屈服、纵筋强化、受拉钢筋塑性流变、纵筋拉断直至结构破坏混凝土压溃发展的全过程.(2)提出的钢筋混凝土剪力墙结构的损伤准则能较好地评估结构损伤发展过程,考虑应变阈值与累积耗能之间耦合效应的损伤准则,其模拟结果更合理.(3)所开发的钢筋混凝土剪力墙模型建模方便,计算成本低,能够较好地对剪力墙结构进行强震作用下的损伤分析.天津城建大学学报王 宁等:钢筋混凝土剪力墙非线性分析模型及损伤准则 ·21·参考文献:[1]PAKNAHAD M,NOORZAEI J,JAAFAR M S,et al.Analysis of shear wall structure using optimal membranetriangle element[J]. Finite Elements in Analysis and De-sign,2007,43(11/12):861-869.[2]MO Y L,ZHONG J X,HSU T T C. Seismic simulation of RC wall-type structures [J]. Engineering Structures,2008,30(11):3 167-3 175.[3]谢 凡,沈蒲生. 一种新型剪力墙多垂直杆单元模型:原理和应用[J]. 工程力学,2010,27(9):154-160.[4]朱杰江,郑 琼,田 堃. 非线性剪力墙单元模型的改进及其应用[J]. 上海大学学报(自然科学版),2009,15(3):316-319.[5]WILLIAMS M S,SEXSMITH R G. Seismic damage indices for concrete structures:a state-of-the-art review[J]. Earthquake Spectra,1995,11(2):319-349. [6]PARK Y J,ANG A H-S. Mechanistic seismic damage method for reinforced concrete[J]. Journal of StructuralEngineering,1985,111(4):722-739.[7]王东升,冯启民,王国新. 考虑低周疲劳寿命的改进Park-Ang地震损伤模型[J]. 土木工程学报,2004,37(11):41-49.[8]王宏业. P ark-Ang双参数地震损伤模型的试验统计分析及改进[D]. 大连:大连海事大学,2008. 3-5. [9]LS-DYNA. Keyword user’s manual[M]. Livermore,California:Livermore Software Technology Corpora-tion,2006.[10]张 松,吕西林,章红梅. 钢筋混凝土剪力墙构件极限位移的计算方法及试验研究[J]. 土木工程学报,2009,42(4):10-16.[11]欧进萍,何 政,吴 斌,等. 钢筋混凝土结构的地震损伤控制设计[J]. 建筑结构学报,2000,21(1):63-68.[12]李忠献,吕 杨,徐龙河,等.强震作用下钢-混凝土结构弹塑性损伤分析[J]. 天津大学学报(自然科学与工程技术版),2014,47(2):102-106.[13]刘伯权,白绍良,徐云中,等. 钢筋混凝土柱低周疲劳性能的试验研究[J]. 地震工程与工程振动,1998,18(4):82-89.[14]吕 杨,徐龙河,李忠献,等. 钢筋混凝土柱基于能量阈值的损伤准则[J]. 工程力学,2011,28(5):84-89.[15]田士锋. 高强混凝土剪力墙抗震试验及非线性分析[D]. 西安:西安建筑科技大学,2006:18-19.Nonlinear Analysis Model and Damage Criteriaof Reinforced Concrete Shear WallWANG Ning,LÜ Yang(School of Civil Engineering,TCU,Tianjin 300384,China)Abstract:Based on the LS-DYNA program, a macro shear wall model is developed, in which a two-parameter shear wall damage criteria combined with the first passage damage and cumulative damage is proposed. The first passage damage is a logarithmic function fitted by the damage index and corresponding strain, and the ultimate dissipation energy of the cumula-tive damage is defined as the surrounded area of the envelop curve. When the threshold strain and cumulative energy exceed their critical values, the coupling effect is considered. The shear wall model is used to simulate the cyclic loading test of two shear walls, and results indicate that the model can simulate the hysteresis process precisely with low cost, and the damage criteria can track the damage process of shear wall well.Key words:reinforced concrete shear wall;damage criteria;nonlinear analysis model;first passage damage;cumulative damage。
混凝土强度与破坏准则综述摘要:强度准则是混凝土材料力学行为研究的重要内容. 受骨料及水泥灰的物理和力学性质的影响,混凝土的变形行为非常复杂. 国内外学者对混凝土强度准则的研究已有较长的历史,并提出了不少破坏准则。
本文通过总结前人的文献,从经典强度理论、试验数据经典回归及包络面唯象学描述三个方面来对混凝土的破坏准则进行了简要的述评,并总结了关于混凝土破坏准则研究的最新进展情况。
关键词:混凝土破坏准则破坏面0引言混凝土在复杂应力状态下的强度或破坏准则一直是工程学科中研究讨论的一个重要课题,而混凝土的破坏过程取决于其性质和内部构造、变形的特点和发展程度、微裂纹的特征和扩展过程,以及内部损伤的积累等等。
混凝土强度理论是判断混凝土在复杂应力状态下是否破坏的理论, 是混凝土结构强度计算和设计必需的基础理论, 一些复杂的重大混凝土结构,如水坝、核反应堆压力容器、海洋工程等结构中混凝土处于明显的多轴应力状态。
这些混凝土结构所承受的三向主应力不等,而且可能是压或拉应力的不同组合。
可见混凝土的强度与破坏准则在理论研究、工程应用和有效利用材料等方面具有非常重要的意义.多年来,国内外许多专家学者提出了各种不同的混凝土强度与破坏准则[1-9]。
本文综合以往学者关于混凝土强度准则的文献资料,从三个方面来总结混凝土的强度与破坏准则。
1经典强度理论1.1单参数模型1876年Rankine提出了最大拉应力强度准则即Rankine模型,按照这个强度准则,混凝土材料中任一点的强度达到混凝土单轴抗拉强度时,混凝土即达到脆性破坏,这一点是否有其他法向或剪切应力对该准则没有影响。
Rankine强度准则其破坏面的形状在空间为一正三角锥面,在子午面上为一直线。
1864年Tresca提出当混凝土材料中一点的应力达到最大剪应力的临界值时,混凝土材料即达到极限强度,即Tresca强度准则。
Tresca强度准则的破坏面与静水压力大小无关,其子午线是与等应力轴平行的直线,在偏平面上截面形状是一正六边形。
混凝土损伤本构模型混凝土作为一种重要的建筑材料,在建筑结构中具有重要的作用。
然而,由于外界环境和使用条件的不断变化,混凝土在使用过程中可能会受到损伤,这些损伤可能会导致结构的不安全性。
因此,混凝土损伤本构模型的研究对于建筑结构的安全性具有重要的意义。
混凝土损伤本构模型是指用于描述混凝土材料在受到外部荷载作用后产生的损伤行为的数学模型。
通过研究混凝土在受损状态下的力学性能,可以为工程结构的设计和评估提供重要的依据。
本文将对混凝土损伤本构模型的发展历史、基本原理、研究现状及其应用进行综述,并探讨该领域的未来发展方向。
一、混凝土损伤本构模型的发展历史混凝土损伤本构模型的研究始于上世纪60年代。
最早提出的混凝土损伤本构模型是由Scheel和Lubbock于1961年提出的弹塑性损伤理论。
随后,梁奇等学者在1978年提出了一种考虑混凝土受损状态的本构模型,这为混凝土损伤本构模型的研究奠定了基础。
随着研究的不断深入,人们对混凝土损伤本构模型的要求也越来越高,例如考虑温度、湿度等耐久性因素对混凝土材料的影响。
在本构模型的建立方面,人们不仅关注其数学表达形式,更加重视其实际工程应用的可靠性和有效性。
混凝土损伤本构模型的研究发展历程为混凝土损伤本构模型的研究奠定了基础,同时也为今后的研究提供了重要的借鉴。
二、混凝土损伤本构模型的基本原理混凝土损伤本构模型的基本原理是通过描述混凝土在受到外部荷载作用后产生的损伤和变形过程,从而建立相应的数学模型。
其核心是将损伤参数引入材料的本构关系中,以描述材料在损伤过程中的力学性能。
混凝土损伤本构模型一般包括两方面的内容,即损伤模型和本构模型。
损伤模型用于描述混凝土在受到外部荷载作用后产生的损伤行为,通常采用损伤变量或者损伤指标来描述损伤程度。
本构模型则用于描述混凝土在不同损伤状态下的应力-应变关系,通常采用应力-应变关系的修正形式来描述材料的非线性和损伤效应。
混凝土损伤本构模型的基本原理是将损伤参数引入材料的本构关系中,以描述材料在损伤过程中的力学性能。
混凝土的破坏准则与本构模型混凝土的破坏准则和本构模型是用来描述混凝土材料在外界荷载作用下的破坏行为和力学性能的模型。
破坏准则描述了混凝土在不同应力状态下发生破坏的临界条件,而本构模型描述了混凝土在荷载作用下的应力应变关系。
混凝土的破坏准则和本构模型对于结构设计、材料选择和力学分析等方面起着重要的作用。
混凝土的破坏准则主要包括强度准则和变形准则。
强度准则描述了混凝土的抗拉、抗压、抗剪等强度性能的破坏条件。
常见的强度准则包括最大拉应变准则、最大压应力准则和最大剪应变准则。
最大拉应变准则认为混凝土的破坏发生在混凝土最大拉应变达到临界值时,而最大压应力准则认为混凝土的破坏发生在混凝土最大压应力达到临界值时,最大剪应变准则认为混凝土的破坏发生在混凝土最大剪应变达到临界值时。
变形准则描述了混凝土在不同应力状态下的应变能力,常见的变形准则包括极限延性准则和极限应变准则。
极限延性准则认为混凝土的破坏发生在混凝土的最大延性达到临界值时,而极限应变准则认为混凝土的破坏发生在混凝土的最大应变达到临界值时。
混凝土的本构模型可以分为线性本构模型和非线性本构模型。
线性本构模型是指混凝土在整个受力过程中满足胡克定律,即应力与应变之间呈线性关系。
线性本构模型常用于结构设计和力学分析中,其优点是计算简单、易于理解和应用。
非线性本构模型是指混凝土在受力过程中出现非线性行为,即应力与应变之间呈非线性关系。
非线性本构模型可以更准确地描述混凝土的力学性能,常用于材料选择和细致的力学分析中。
常见的非线性本构模型包括卓尔金模型、拉勃森模型、屈曲温演模型等。
这些模型根据不同的假设和参数来描述混凝土在不同应力状态下的力学行为。
其中,卓尔金模型是最常用的非线性本构模型之一,它将混凝土的延性和强度性能分别考虑,可以比较准确地描述混凝土的变形和破坏行为。
总的来说,混凝土的破坏准则和本构模型对于混凝土的力学性能描述和结构设计起着重要的作用。
通过研究混凝土的破坏准则和本构模型,可以更好地理解混凝土的破坏机理和力学行为,为混凝土的设计和使用提供科学依据。
三种混凝土应力应变理论模型
凡是在适当的外力作用下物质能流动和变形的性能称为该物质的流变性。
流变学的研对象几乎包括了所有的物质,综合研究了物质的弹性变形,塑性变形和粘性流动。
研究材料的流变特性时,要研究材料在某一瞬间的应力和应变的定量关系,这种关系常用流变方程来表示。
而一般材料的流变方程的建立,都是基于以下三种理想材料的基本模型(或称流变基元)的基本流变方程:(1)胡克(Hooke)固体模型(H-模型)表示具有完全弹性的理想材料;(2)圣维南(St.Venant)固体模型(Stv-模型)表示超过屈服点后只具有塑性变形的理想材料。
(3)牛顿(Newton)液体模型(N-模型)表示只具有粘性的理想材料。
弹性,塑性,粘性和强度是四个基本流变性质,根据这些基本性质,可以导出其他性质。
胡克固体具有弹性和强度,但没有粘性。
圣维南固体具有弹性和塑性,但没有粘性。
牛顿液体具有粘性,但没有弹性和强度。
严格的说,以上三种理想物体并不存在。
大量的物体是介于弹,塑,粘性体之间。
所以,实际材料的流变性质具有所有上述四种基本流变性质,只是在程度上有差异。
因此,各种材料的流变性质可用具有不同的剪切弹性模量G,粘性系数η和表示塑性的屈服应力τy 的流变基元以不同的形式组合成的流变模型研究。
混凝土塑性—损伤本构模型研究一、本文概述Overview of this article混凝土作为一种广泛应用的建筑材料,其力学性能和损伤行为的研究一直是土木工程领域的重要课题。
本文旨在深入研究和探讨混凝土塑性-损伤本构模型,该模型能够更准确地描述混凝土在复杂应力状态下的力学响应和损伤演化过程。
通过对混凝土塑性-损伤本构模型的研究,不仅有助于我们更好地理解混凝土的力学特性,还能为混凝土结构的设计、分析和优化提供理论基础和技术支持。
As a widely used building material, the study of mechanical properties and damage behavior of concrete has always been an important topic in the field of civil engineering. This article aims to conduct in-depth research and exploration on the plastic damage constitutive model of concrete, which can more accurately describe the mechanical response and damage evolution process of concrete under complex stress states. The study of the plastic damage constitutive model of concrete not only helps us better understand the mechanical properties ofconcrete, but also provides theoretical basis and technical support for the design, analysis, and optimization of concrete structures.本文首先介绍了混凝土塑性-损伤本构模型的基本概念和理论框架,包括塑性理论、损伤力学以及混凝土材料的特殊性质。
Constitutive model for the triaxial behaviour of concreteAuthor(en):William, K.J. / Warnke, E.P.Objekttyp:ArticleZeitschrift:IABSE reports of the working commissions = Rapports des commissions de travail AIPC = IVBH Berichte der ArbeitskommissionenBand(Jahr):19(1974)Persistenter Link:/10.5169/seals-17526Erstellt am:22.08.2011NutzungsbedingungenMit dem Zugriff auf den vorliegenden Inhalt gelten die Nutzungsbedingungen als akzeptiert. Dieangebotenen Dokumente stehen für nicht-kommerzielle Zwecke in Lehre, Forschung und für dieprivate Nutzung frei zur Verfügung. Einzelne Dateien oder Ausdrucke aus diesem Angebot könnenzusammen mit diesen Nutzungsbedingungen und unter deren Einhaltung weitergegeben werden.Die Speicherung von Teilen des elektronischen Angebots auf anderen Servern ist nur mit vorherigerschriftlicher Genehmigung des Konsortiums der Schweizer Hochschulbibliotheken möglich. DieRechte für diese und andere Nutzungsarten der Inhalte liegen beim Herausgeber bzw. beim Verlag.SEALSEin Dienst des Konsortiums der Schweizer Hochschulbibliothekenc/o ETH-Bibliothek, Rämistrasse 101, 8092 Zürich, Schweizretro@seals.chhttp://retro.seals.chIABSE AIPC IVBHSEMINAR on:«CONCRETE STRUCTURES SUBJECTED TO TRIAXIAL STRESSES»17th-19th MAY,1974-ISMES-BERGAMO(ITALY)III-lConstitutive Model for the Triaxial Behaviour of Concrete Stoffmodell für das mehrachsiale Verhalten von BetonModlle de Constitution pour le Comportement Triaxial du BitonK.J.WILLAMPh. D.,Project LcaderInstitut für Statik und Dynamik der Luft-und Raumfahrtkonstruktionen University of Stuttgart E.P.WARNKEDipl.-Ing.,Research Associate Institut für Statik und Dynamik der Luft-und Raumfahrtkonstruktionen University of StuttgartSUMMARYThis paper describes different modeis for the failure surface and the constitutive behaviourof concrete under triaxial conditions.The study serves two objectives,the working stressdesign and the ultimate load analysis of three-dimensional concrete components.In the first part a three parameter failure surface is developed for concrete subjected to triaxial loading in the tension and low compression regime.This model is subsequently refined by adding two additional parameters for describing curved meridians,thus extend ing the ränge of appli¬cation to the high compression zone.In the second part two constitutive modeis are formulated for elastic perfectly plastic be¬haviour in compression and elastic perfectly brittle behaviour in tension.Based on the normality principle,explicit ex pressions are developed for the inelastic deformation rate and the correspon¬ding incremental stress-strain relation Thus these modeis can be readily applied to ultimate load analysis using the initial load technique or the tangential stiffness method.Dedicated to the60th birthday of Professor Dr.Drs.h.c.J.H.Argyris.1.INTRODUCTIONOver the last two decades a profound change has taken place with the appearance of digital Computers and recent advances in structural analysis[l],[2],[3].The close symbiosis between Computers and structural theories was instrumental for the development of large scale finite element Software packages[4]which found a wide ränge of application in many fields of eng i nee ring sciences.The high degree of sophistication in structural analysis has clearly left behind many other disciplines,one of them being the field of material science.The proper description of the relevant constitutive phenomena has posed a major limitation on the analysis when applied to complex ope¬rating conditions.In the following a constitutive model is presented for the over load and ultimate load ana¬lysis of three-dimensional concrete structures, e.g.Prestressed Concrete Reactor Vessels and Con¬crete Dams.Considering the size of finite elements in a typical idealization one is clearly deal ing with material behaviour on the continuum level,in which the micro structure of piain and rein¬forced concrete components can be neglected.This scale effect of the analysis allows a macro-scopic point of view according to which material phenomena such as cracking can be simulated bythe behaviour of an equivalent continuum.The objective of this study is twofold:First a mathematical model is developed for the description of initial concrete failure under triaxial conditions.Subsequently,this formulation is applied to construct a constitutive model for the over load and ultimate load analysis of three-dimensional concrete Structures.Alternative ly,the failure surface can be applied to working stress design using relevant safety philosophies.In the first part a three parameter model is developed which defines a conical failure sur¬face with non-circular base section in the principal stress space,thus the strength depends on the hydrostatic as well as deviatoric stress state.The proposed failure surface is convex,continuousand has continuous gradient directions furnishing a close fit of test data in the low compression ränge. In the tension regime the model may be augmented by a tension cut-off criterion.This basic formu¬lation is refined in AppendixII by a five parameter model with curved meridians which provides a close fit of test data also in the high compression regime.Subsequently,a material model is constructed based on an elastic perfectly plastic formu¬lation which is augmented by a brittle failure condition in the tensile regime.In this context equivalent constitutive constraint conditions are developed,based on the"normality"principle, which can be readily applied to the finite element analysis via the concept of initial loads.In the past considerable experimental evidence has been gathered which could be used for the construction of a triaxial failure envelope of concrete.However,most of the data were ob-tained from tests with proportional loading and uniform stress or strain conditions which were distorted by unknown boundary layer effects.For the ultimate load analysis via finite elements these two assumptions are clearly invalid.The non-linearity is responsible for local unloading even if the structure is subjected to monotonically increasing stresses.Moreover,the action of a curved thick-walled structure is control led by non-uniform stress distributions,even if global bending effects and local stress concentrations are neglected for the time being.However,for obvious reasons it is customary to assume that test results from uniform stress-or strain experiments can be used to predict the failure behaviour of structural components subjected to non-uniform stress or strain conditions. One should be aware that this fundamental hypothesis has little justification,except that it is at present the only realistic approach for construct ing a phenomenological constitutive law.The actual mechanism of crack initiation and crack propagation could in fact differ fundamental ly between uniform and non-uniform stress distributions.Considerable test data has accumulated on the multiaxial failure behaviour of mortar and concrete specimens subjected to short term loads.The experimental results can be classified into tests in which either two or three stress components are varied independently.To the first category belong the classical triaxial compression tests on cylindrical specimens(triaxial cell experiments) [5],[6],[7],[8],[9]and the biaxial tension-compression tests on hollow cylinders[lO],[l1J In addition,there is the class of biaxial compression and tension-compression tests on slabs [12]/D3L t14L L15L L16l E17"L t18L[191-The second category contains experi¬ments in which cubic specimens are subjected to arbitrary load combinations[20J,[21J Some of these types of tests are present ly still being processed[22],[23],[24JSo far few attempts have been made to utilize this experimental evidence for construct inga mathematical model of the triaxial failure behaviour of concrete.A comprehensive study of this problem was undertaken in[25],for which similar conclusions were reached in[26J,[27].All three modeis fall into the class of pyramidal failure envelopes which have been examined extensive ly within the context of brittle material modeis as general izations of the Mohr-Coulomb criterion[28].In the same publication different modifications of the Griffith criterion are discussed,which have also been applied in[20]to model the failue surface of cubic mortar specimens in the tension-compression regime.None of these previous studies on failure envelopes was directed towards the non-linear analysis of concrete structures.To this end a number of rather simple material formulations were reviewed in[29],[30],[31]and applied to the ultimate load analysis of different concrete structures.TRIAXIAL FAILURE SURFACEIn the following a mathematical model is developed for the triaxial failure surface of con¬crete type materials.Assuming isotropic behaviour the initial failure envelope is fully describedin the principal stress space.Figure1shows the triaxial envelope of concrete type materials.The failure surface is basically a cone with curved meridians and a non-circular base section.The limited tension capacity is responsible for the tetrahedral shape in the tensile regime,while in compression a cylindrical form is ultimate ly reached.For the mathematical model only a sextant of the principal stress space has to be considered, if the stress components are ordered according to S,>Ct>Gs The surface is conveniently represented by hydrostatic and deviatoric sections where the first one forms a meridianal plane which contains the equisectrix S.«6,.»Gh as an axis of revolution The deviatoric section lies in a plane normal to the equisectrix,the deviatoric trace being described by the polar coordinatesr,ösee Fig. 2.Basically,there are four aspects to the mathematical model of the failure surface:1Close fit of experimental data in the operating ränge.2.Simple identification of model parameters from Standard test data.3.Smoothness-continuous surface with continuously varying tangent planes.4.Convexity-monotonically curved surface without inflection points.Close approximation of concrete data is reached if the failure surface depends on the hydrostatic as well as the deviatoric state,whereby the latter should distinguish different strength values according to the direction of deviatoric stress.Therefore,the failure envelope must be basically a conical surface with curved meridians and a non-circular base section.In addition,in the tensile regime the failure suface could be augmented by a tension cut-off criterion in the form of a pyramid with triangulär section in the deviatoric plane.Simple identification means that the mathematical model of the failure surface is definedby a very small number of parameters which can be determined from Standard test data, e.g. uniaxial tension,uniaxial compression,biaxial compression tests,etc.The description of the failure surface should also encompass simple failure envelopes for specific model parameters.In other words,the cylindrical von Mises and the conical Drucker-Prager model should be special cases of the sophisticated failure formula tion.Continuity is an important property for two reasons:From a computational point of view,it is very convenient if a single description of the failure surface is valid within the stress space under consideration.From the theoretical point of view the proposed failure surface should havea unique gradient for defining the direction of the inelastic deformations according to the1normolity principle1.The actual nature of concrete failure mechanisms also supports the conceptof a gradual change of strength for small variations in loading.Geometrical ly,the smoothness condition implies that the failure surface is continuous and has continuous derivatives.Therefore,the deviatoric trace of the failure surface must pass through r,and ru with the tangents4:,and tr at9-O*and0-CO°,see Fig. 2.Recall thatfor Isotropie conditions only a sextant of the stress space has to be considered,O^O*60*.Convexity is an important property since it assures stable matenal behaviour according to the postulate of Drucker[32],if the"normolity"principle determi es the direction of inelastic deformations.Stability infers positive dissipation of inelastic work during a loading cycle according to the coneepts of thermodynamics Figure3indicates that convexity of the overall deviatoric trace can be assured only if there are no inflection points end if the position vector satisfies the basic convexity conditionJL>J_where r;r(0.:*,i2o",z4o#)Ti.*r(ö*£3,iSo^soo)W Continuity infers compatibility of the position vectors and the slopes et0«O*and0¦GOD. Consequently,there are at least four conditions for curve-fitting the deviatoric trace withinO**0*GOt In addition,the convexity condition implies that the curve should have no inflection points in this interval,thus the approximation can not be based on trigonometric functions[30]or Hermitian Interpolation.If the curve should also degenerate to a circle for r,-r^then an elliptic approximation has to be used for the functional Variation of the deviatoric trace.The ellipsoidal surface assures smoothness and convexity for all position vectors r satisfy ing|rl4r<rL(2)The geometric construction of the ellipse is shown in Fig.4,the details of the derivation are given in the Appendix I.The half axes of the ellipse a,b are defined in terms of the position vectorsErt-4.r,(3)a.r,1'-Sr.ty*Lr%¦4.r,-BrtThe elliptic trace is expressed in terms of the polar coordinates r,B by(r>-r.*)cos9*rt(ir.-r«)&(tf-tf)o»9?Sif-Ar.rj(4a) rf8x1ft,with the angle of similarity9^,^gt"2-g3>(4b) aIn the following the deviatoric trace is used as base section of a conical failure surface with the equisectrix as axis of revolution.A linear Variation with hydrostatic stress generates a cone with straight line meridians.In this case the failure surface is defined in principal stress space by a homogeneous expansion in the"average11stress components$>Ä«:«.and the angle of similarity9IdW^MJ.i^^J-i(5)The average stress components6«,t«.represent the mean distribution of normal and shear stresses on an infinitesimal spherical surface.These values are normolized in the failure condition eq.(5)by the uniaxial compressive strength f^The stress components are defined in terms of principal stresses by(6)^-jL[(».-*£-(*-*£>+t*»-**]*These scalar representations of the state of stress at a point are related to the stress components on the"octahedral"plane60^o by**s*o(7a)**«ff t0The average stress components also correspond to the first principal stress invariant I,and the second deviatoric stress invariant I%according to**-T*.(7b)».-nfi^-RW«^For material failure,-f(>)*0the following constraint condition must hold between the average normal stress and the average shear stresst-rt^D-Ht]i«.1(8)The free parameters of the failure surface model fc H and f«.are identified below from typical concrete test data,such as the uniaxial tension test«f t the uniaxial compression testf4a and the biaxial compression test^Introducing the strength ratios ocft,«c0*z*ft/fco(9) the three tests are characterized by6.TEST Wf«,«/?«.er.*,-ft+--(¥->6Ä»feo3K CO***xei*^"£fc"*««R-O*n(10)Substituting these strength values into the failure condition eq.8,the model parametersare readily obtainedn<*o Af«*»-**«Tu**£«r+ä**oätjt*«««?<*.-«z(11)The apex of the conical surface lies on the equisectrix at*»«*The opening angle<P of the cone varies betweenandtan<y,«-^«+0-GO°(12)(13) The proposed three parameter model is illustrated in Fig.5for the strength ratios0(O'and42*o.i The hydrostatic and deviatoric sections indicate the convexity and smoothness failure envelope.The proposed failure surface degenerates to the Drucker-Prager model of a circular cone ifIn this case the conical failure surface is described by the two parameters2and r0±_**^-^-1i.i of the(14)(15)The single parameter von Mises model is obtained/if in addition7.2.-t>oo(16)[21].In this case the Drucker-Prager cone degenerates into a circular cylinder whose radius is defined byr.fo (17)with the strength ratios**m **"¦(18)Figure 6shows a comparison between the failure surface and experimental data reported inClose agreement can be observed in the low pressure regime for the strength ratios *oc'-Äand *i«o.i?In the high compression regime there is considerable disagreement mainly along the compressive branch.Therefore,the three parameter model is refined in the Appendix II by two additional parameters,extend ing the ränge of application to the high compression regime.This five parameter model establishes a failure surface with curved meridians in which the generators are approximated by second order parabolas along 0s O*and 0*Go with a common apex at the equisectrix,see also Fig.11Figure 7shows the biaxial failure envelope of the three parameter model for three differentstrength ratios otu*l&*^»o-U &u~)o,**.*o.o%and *u-l-&j **-o.is A comparison with test data from [18]^2l]indicates that the shear strength is overestimated consi-derably because of the acute intersection with the biaxial stress plane.However,if we consider the dominant influence of the post-failure behaviour on the structural response [30J,there is little reason for further refinements of the initial failure surface model.3.CONSTITUTIVE MODELIn the following the previous model of the failure envelope is utilized for the developmentof an elastic perfectly plastic material formula tion in compression.The constitutive model is sub¬sequently augmented by a tension cut-off criterion to account for cracking in the tension regime.In both cases it is assumed that the normo lity principle determines the direction of the inelastic deformation rates for ductile as well as brittle post failure behaviour.3.1Elastic Plastic FormulationInviscid plasticity is the classical approach for describing inelastic behaviour via incremen-tal stress-strain relations.The constitutive model is based on two fundamental assumptions,an appropriate description of the material failure envelope and the definition of inelastic deformation rates e.g.via the normo lity principle.a.Yield ConditionThe yield surface serves two objectives,it distinguishes linear from non-linear andelastic from inelastic deformations.The failure envelope is defined by a scalar function of stress,$(J5»o indicating plastic flow if the stress path intersects the yield surface.For concrete type of materials the yield condition can be approximated by the three parameter model shown in Fig.5or more accurately by the five parameter model developed in the Appendix II.b.Flow RuleFor perfectly plastic behaviour the yield surface does not change its configurationduring plastic flow,hence the stress path describes a trajectory on the initial yield surface,whilethe inelastic strains increase continuously.In this case the inelastic deformations do not contri-bute to the elastic strain energy,thus the inner product of plastic strain and elastic stressrates must be zeron**-°09)-In other words,the plastic strain rate must be perpendicular to the yield surfaceV^(2°) where the normal n is the unit gradient vector of the yield surface9j/9*(21)mExplicit expressions of2f/d9are developed in Appendix III for different yield surfaces.The normal defines the direction of the plastic strain rate,the length of which determines the loading parameter a The normality condition follows from Drucker's stability postulate which assures non-negative work dissipation during a loading cycle,also infernng convexity of the yield surface.For perfectly plastic behaviour the material stability is"indifferent"in the small, corresponding to the"neutral"loading condition for which initial yield and subsequent flow is governed by*«>-0and fC*>-o(22) The consistency condition implies that?(«>-Tt*"°(23) This Statement is clearly equivalent to the normality principle stated in eq.(19).c.Incremental Stress-Strain RelationsIn the following an elastic perfectly plastic consitutive model is derived using the previous Statements and the kinematic decomposition of the total deformationsy-C+Ylf,and TF"*+V(24) The linear elastic material behaviour is given by the rate formulation of generalized Hooke's law*Ei-£(i-^(25) Substituting the stress rate into the consistency condition,eq.(23),we obtain£«-n4E(JHW(26) This expression yields for the undetermined loading parameter ahtE(T-n\)-o(27)and hereby±i'WTT"E*(28) x The dot indicates the rate of change.9. The plastic strain rate follows from eq.(20)(29) The incremental stress-strain relations are obtained by Substitut ing y*v into the expression ofthe stress rate,eq.(25)Note the linear relationship between the stress and deformation rates in eq.(30)&F Y od The tangential material law Tr is defined byFor perfectly plastic behaviour,T depends only on the elastic properties and the instantaneous stress state via II The second term of eq.(32)represents the degradation of the material Constitution due to plastic flow.3.2Elastic Cracking FormulationSmall tensile strength is the predominant feature of concrete-type materials.In the following a simple constitutive model is developed for perfectly brittle behaviour in the tensile regime.In analogy to the elastic plastic formulation the elastic cracking model is based on two fundamental assumptions,a tension cut-off criterion for the prediction of cracking and an appro-priate description of inelastic deformation rates e.g.via the normality principle.a.Crack ConditionThe tension cut-off criterion distinguishes elastic behaviour from brittle fracture,i.e. Separation of the material constituents due to excess tension.To this end it is assumed that the scale of Observation justifies a continuum approach.For concrete-type materials cracking may be predicted by the single one parameter model based on the major principal stress^t**)<5-.-^e wifh*>i*^i>^i(33) where C\corresponds in general to the uniaxial tensile strength ft The failure surface is shown in Fig.8,which indicates the pyramidal shape and the triangulär base section in the deviatoric plane.Alternative ly,the tension cut-off condition could also be expressed in terms of the three parameter model of the previous section or the five parameter model developed inthe Appendix II.b.Fracture RuleFor ductile behaviour in the post failure ränge the inelastic deformation rate due to cracking is derived exactly along the formulation of an elastic plastic solid.The ductilepost failure behaviour forms an upper bound of the actual soften ing behaviour[30J,which may develop in concrete components due to reinforcements,dowel action and aggregate interlock.In the following/the case of perfectly brittle post-failure behaviour is discussed,since it requires slight modifications of the previous constitutive model for an elastic perfectly plastic solid. In analogy to elasto-plasticity the inelastic deformations due to cracking tlc do not contribute to the elastic strain energytYI.5-°/iü\10.This normality principle corresponds to the flow rule of plasticity stating that the inelastic strain rate due to cracking is perpendicular to the plane of fractureT|c-nX(35) For the maximum stress tension cut-off criterion the normal vector f\is defined by the direction of the major principal stress;thus in the principal stress spacem »f/»««(36)n|9$/9«M^iwhere£,is the unit vector«,-[l,o,o,o,o.°i(37) For perfectly brittle behaviour the loading parameter X is determined from the sofrening condition|C«}-0and£C«^*-*t(360 In this case the consistency condition infers that|t**t(39)c.Inelastic Strain IncrementsIn the following an expression is derived for the inelastic deformation rotes due to cracking.Substituting the stress rate expression eq.(25)into the consistency condition eq.(39)&*-**(*-M w we obtain an expression for the undetermined loading parameter A«*6(*-«.X)--«e.(") and hereby*-«Fg-*,(*,*T+<)(«)Note the equivalence to the elastic plastic formulation in eq.(28)except for the release of^due to brittle softening.The resulting inelastic fracture strain rate follows from eq.(35)(43)The first portion of this expression can be used to construct incremental stress-strain rela-tions in analogy to the elastic plastic formulation,see eq.(30).This part would correspondexactly to a ductile cracking model in which the major principal stress is kept constant at thetensile strength S,^e The corresponding tangential material law would become transversely isotropic with zero stiffness along the major principal axis.Additional cracking in other directions can be considered according ly.The second portion of eq.(43)represents the sudden stress release due to brittle fracture,*\, which is projeeted onto the structural level by a single initial load step in the analysis.11.3.3Transition ProblemThe previous rate formulation for elastic plastic and brittle behaviour is valid in a diffe¬rential sense only.In a numericaj environment clearly finite increments prevail during numerical Integration of the rate equations[33J,[34J.This approximation problem is magnified by the sudden transition from elastic to plastic or elastic to brittle behaviour.In the latter case the dis-continuity of the process is further increased due to the immediate stress release if the failure condition has been reached.Clearly,the success of the numerical technique depends primärily on the proper treatment of the transition problem for finite increments.Consider the most general case of a finite load step shown in Fig.9.At the outset we assume that the stress path has reached point A for which^C«*V°indicates an elastic state. Due to the finite load increment a fully elastic stress path would reach point B penetrating the yield surface at C for proportional loading.The condition$C^O>°violates Tne constitutive constraint condition^«^°(45) and suggests two strategies for numerical implementation.a.^P^^^J.^6.^^*J£n MethodAssuming proportional loading the load increment is subdivided into two parts,an elastic portion for the path A-C and an inelastic portion governing the behaviour after the failure surface has been reached at C.The evaluation of the penetration point C reduces to the geometric problem of intersecting a surface with a line,a task which is non-linear for curved failure envelopes.The computation of the stress trajectory on the yield surface involves the numerical integration of5^1F if(46)since the tangential material law varies with the current state of stress.In addition we have to assume that the inelastic strains increase proportional ly from ycto<jf£.In numerical calculations additional corrections are required at each iteration step to place the stress path back onto the yield surface[37]b.Normal Penetration MethodIn this scheme we assume that the elastic path reaches the yield surface at the inter-section with the normal ns The evaluation of the foot point D reduces to the geometric-prob¬lem of minimizing the distance between B and the failure envelope,see Fig.9<A-(«^«^(«^«J"*Minimum(47)The extremum condition is used to determine the components of C^by solving the linear system of equations.subjected to the constraint conditionft«*)*ö(49)Note that the loading parameter X is proportional to the distance d,thus the length of the inelastic deformation increment is determined from。