焊接结构疲劳评估的网格不敏感结构应力方法

1、The Mesh-Insensitive Structural Stress Methodfor Fatigue Evaluation of Welded Structures焊接结构疲劳评估的网格不敏感结构应力方法目录焊接疲劳基本相关术语和定义材料性能的影响残余应力的影响几何不连续性的影响传统方法描述结构应力方法（I）结构应力定义与数值实现焊缝的表示方法与其它方法（如ASME，IIW等）的比较和示例结构应力的度量方法与确认结构应力方法（II）一般的结构应力计算过程边界细节的处理计算实例多轴结构应力状态的特征点焊激光焊演示/工作流程结构应力后处理器试用版主S-N曲线方法基于K的结构应力求解技巧

2、与验证双态裂纹生长模型与验证等效结构应力幅参数与验证基于结构应力的寿命预测过程失效、焊线定义、焊缝表示方法等疲劳测试的含义寿命预测实例其它应用焊喉开裂多轴疲劳低周疲劳与热疲劳电子封装中的焊料疲劳8121631475763717587103115118130135140144158165174185216226234249254268282289Major Awards and Recognitions RecentlyReceived 最近获得的主要奖励与荣誉IIW 2008 Paton PrizeSNAME 2007 Elmer L Hann AwardR&D Magazines R&D 1

3、00 Award (2006)TIME Magazine 2005 Math InnovatorsAviation Week & Space Technology: Aerospace 2004 Laurels AwardSAE 2004 Weld Challenge: Best Life Prediction WinnerAWS 2004 R.D. Thomas AwardSAE 2003 Henry Ford II Distinguished Award for Excellence in Automotive EngineeringASME 2002 G.E.O. Widera Lite

4、rature Award, An Overview of Advanced Weld ModelingCapabilities: Fusion Welding先进焊接建模能力概览：熔化焊An Overview of Advanced Weld ModelingCapabilities: Solid Joining先进焊接建模能力概览：固态连接Fatigue of Weldments: Some Fundamentals焊件疲劳：基本知识Some relevant terminologies and definitions 一些相关术语和定义Whats special about weldmen

5、ts?焊件有哪些特性、 Stress concentration 应力集中 Material properties 材料属性 Residual stresses 残余应力 What decades of research say? 近几十年的研究对此的描述 Limitations of conventional fatigue design methods传统疲劳设计方法的局限 What constitutes a good fatigue parameter?哪些物理量构成好的疲劳参数Terminologies Describing A Typical Welded Joint描述一个典型焊

6、接接头的术语Material zones in a joint接头材料区域Base metal母材（BM）Heat-affected zone热影响区（HAZ）Weld metal焊材（WM）Geometric locations几何位置Weld toe焊趾Weld throat/depth焊喉/焊深Weld root焊根Relevant Fatigue Terminologies andDefinitions I 相关疲劳术语和定义-ICyclic loading循环载荷Nominal stress at a failurelocation, e.g., weld crackinto bas

7、e plate失效位置的名义应力，例如，侵入母材的焊接裂纹处F/（Wt）Mc/lNominal stress range名义应力变化范围F /（Wt）Mc/lRelevant Fatigue Terminologies andDefinitions II相关疲劳术语和定义-IIMaximum stress最大应力Minimum stress最小应力Stress range应力变化范围（）Mean stress平均应力What s Special about Welded Joints?焊接接头的特殊之处Property Heterogeneity材料性质的多样性(母材BM,焊材 WM, 热影响

8、区HAZ)Residual Stresses残余应力Geometric Discontinuities几何不连续性Property Variation in Welded Joints焊接接头材料性质的变化Process-induced induced property heterogeneity焊接过程导致的材料性质的多样性WM strength mismatch by design焊材强度设计的不匹配Hardenable Steel可硬化的钢材BM/WM/HAZ Property Has Little Effect on Fatigue of Welded Joints BM/WM/HAZ

9、的性质对焊接接头疲劳影响很小Comparison of Fatigue Strength between Plain Metal, Notch Bar, and Welded Specimens普通金属、带缺口金属条与焊接试件之间疲劳强度的比较Figure 2: Influence of tensile strength on the fatigue strength图2： 抗拉强度对疲劳强度的影响BM Strengths vs. Fatigue Life (Lewis, 2001)母材强度与疲劳寿命 (Lewis, 2001)TRIP v CMn350 Unwelded Steel TRIP

10、 vs CMn350 and Mild Steel Welded Steel 未焊接钢材 焊接钢材Weld Residual Stress Effects on FatiguePerformance: Weld Sequencing焊接残余应力对疲劳性能的影响焊接顺序Pass Sequence Effects on TransverseResidual Stresses焊道顺序对横向残余应力的影响Travel Speed (Linear Heat Input) Effects Transverse Residual Stresses焊接速度(线性热输入)对横向残余应力的影响Restraint

11、Effects on Transverse ResidaulStresses约束对横向残余应力的影响High Restraint Low Restraint强约束 弱约束Residual Stress Distributions in Some TypicalJoint Types一些典型焊接接头的残余应力分布Then, How to Treat Residual Stress Effects inFatigue Design in Practice?在实际疲劳设计中应如何处理残余应力的影响？Conduct controlled fatigue testing采用控制疲劳测试Effects o

12、n applied mean stresses对施加的平均应力的影响As-welded versus stress-relieved焊态与应力释放Test specimens must contain representative residualstress state in structures试件必须包含结构中主要的残余应力状态Specimen sizing试件尺寸Residual stress induced tri-axiality versus “shake-down”Effects残余应力引起的三轴与失稳（？）的影响Model or Specimen Size Must Be S

13、ufficientlyLarge to Quantify Residual Stress Effects模型或者试件的尺寸应该足够大以量化残余应力的影响A Butt Joint对接接头Plate joints: w/t=6平板接头 w/t=6Residual Stress and Joint Constraint Effectson Local Stress/Strain Behavior残余应力和接头约束对局部应力/应变行为的影响Test Specimen Size Requirements forContaining Weld Residual Stress Effects试件尺寸要求包含

14、焊接残余应力效应Axial Residual Stress轴向残余应力Tube girth weld length: 2.5 sqrt(rt)圆管环焊缝长度: 2.5 sqrt(rt)Tubular Girth Weld管状环焊缝Residual Stress Effects on Fatigue Behavior ofWelded Joints - Experimental Evidence残余应力对焊接接头疲劳行为的影响试验验证Stress ratio is NOT important应力比并不重要Stress range is important应力幅是关键因素Compressive c

15、yclic loading can be equally damaging压缩循环载荷同样可以引起材料破坏Recent S-N Data from HHI Mean StressEffects on S-N Behavior HHI的最新S-N数据平均应力对S-N行为的影响Comparison of Fatigue Test Data with DifferentStress Ratio or Mean Stress Effects: As-Welded在不同应力比或平均应力效应下疲劳测试数据的比较焊态Comparison of Fatigue Test Data with Different

16、 Stress Ratio or Mean Stress Effects: As As-Welded vs Stress-Relieved在不同应力比或平均应力效应下疲劳测试数据的比较焊态与应力释放Stress Relief Post-Weld Heat Treatment(PWHT)焊后热处理(PWHT)的应力释放Uniform PWHT: Certain magnitude of residual stresses still retained in ideal PWHT整体热处理：理想的热处理下仍然存在一定的残余应力Other treatments such as local PWHT其

17、它处理方式比如局部热处理higher residual stresses retained than uniform PWHT比整体热处理方式存在更多的残余应力higher residual stresses than as-welded conditions mayresult可能比焊态条件下的残余应力水平更高Observations Residual Stress Effects onFatigue小结：残余应力对疲劳的影响Adequate specimen size is important in fatigue testing to retain representative resi

18、dual stress states in actual structures为包含实际结构中主要的的残余应力状态，要求进行疲劳测试时要保证足够的试件尺寸Perhaps, the most significant effects of residual stresses in welded joints:焊接接头中残余应力最重要的影响：（？）S-N data not sensitive to applied mean load or mean stress S-N数据对施加的平均载荷或平均应力不敏感Stress range serves as a good parameter for impl

19、icitly incorporating residual stress effects 应力变化范围是一个包含残余应力效应的很好的参数Without adequate control of welding/assembly procedures, any attempts to incorporate residual stresses in fatigue assessment are futile对焊接/组装过程没有足够的控制，任何在疲劳评估中考虑残余应力都没有效果Stress relieved conditions: the improvement in fatigue is not

20、Significant 应力松弛条件：对疲劳性能的改善不明显Global Geometric Discontinuities整体几何不连续性GMA weldsGMA焊缝RSW or plug weldRSW或塞焊Laser welds激光焊Global versus Local Discontinuities in Welded Joints焊接接头中的整体和局部不连续性Effects of Geometric Discontinuities on FatigueBehavior in Welded Joints: Unique S-N Curve Slope几何不连续性对焊接接头疲劳行为的影

21、响唯一的S-N曲线斜率Unique features for weldments: 焊接件的特性Different slope from smooth barspecimens光滑试件具有不同的斜率Unique slope of about 3唯一的大约为3斜率Different JointGeometries不同的接头几何Effects of Global Geometric Discontinuities:Well-Defined Failure Modes in Welded Joints整体几何不连续性的影响焊接接头合适的失效模式Dominant Modes:主要模式Mode (A)模

22、式（A）Mode (B )模式（B）Mode (A)模式（A）Desirable需要的Easier to analyze更易于分析Mode (B)模式（B）Least desirable很少需要Large data scatter大量数据呈离散性Joint geometr and loading modedetermines which mode dominates接头几何和载荷模式决定哪种失效模式为主Joint Geometry Effects on Nominal Stress Rangeversus Cycle to Failure (Weld Toe Failure)接头几何对名义应力

23、变化范围与失效(焊趾失效)循环次数的影响Stress definition: Nominal stress range (F/A)应力定义：名义应力副(F/A)Thickness Effects厚度的影响Loading Mode Effects Tension versusBending载荷模式的影响拉伸与弯曲Implication: Stress Concentration is the Most ImportantParameter Governing Fatigue Life in Welded Joints推论：应力集中是决定焊接接头疲劳寿命的最重要的参数Dominated by ge

24、ometric and loading conditions取决于几何和载荷条件Joint type/geometry接头类型/几何Loading mode载荷模式Thickness厚度(Weld size, etc)(焊缝尺寸等)Stress concentration determination: deceptively simple应力集中的确定：看起来简单Notch Stress Determination缺口应力的确定“Kt” only well-defined for a given notch radius“Kf”只在给定缺口半径时才有很好的定义Too arbitrary to

25、work in practice太随意以至于无法应用Too small an element size to be practical太小的单元尺寸不切实际Local notch geometry in welded joints: random in nature焊接接头中的局部缺口几何：本质上是随机的Notch Radius=?缺口半径=?Stress Concentration Behavior at WeldedJoints焊接接头处的应力集中行为Difficulties:一些困难Singularity at sharp notch在尖锐缺口处奇异Artificial radius:

26、well, too artificial人为定义的半径：太随意Typical SCF determination methods典型的确定应力集中因子的方法Strain gauge based measurements基于应变片的测量Gauge location/size?应变片位置/尺寸Reference position?参考位置Finite element analysis有限元分析Mesh-size sensitive对单元尺寸敏感Element type sensitive对单元类型敏感How About Extrapolation to the Weld Toe Using Sur

27、face Stresses Hot Spot Stress (HSS) Methods?将表面应力外推到焊趾位置热点应力(HSS)方法e.g., an IIW recommended procedure例如，一种IIW 推荐的方法(a concept used for offshore tubular Structures)一种用于船体管状结构的概念Severe element size/type sensitivity!严重的单元尺寸/类型敏感性Surface Extrapolation Based HSS Method When Using Converged Mesh Plate Joi

28、nts当使用收敛网格时基于HSS方法的表面应力外推板接头 S-N data from the six joint types can be grouped into at least four distinct S-N curves从六种接头类型得到的S-N数据至少可以归纳为四条不同的S-N曲线Detailed stress calculations:详细的应力计算：two distinct types of distributions两种不同的分布类型SCF=?Linear Solid Element Models with Mesh Size 0.1t尺寸约为0.1t的线性实体单元模型Di

29、fference in Surface Stress Gradientsbetween Tubular versus Plate Joints管接头和板接头之间表面应力梯度的区别Stress gradients:应力梯度：Tubular joints: more global管接头：更具整体性Plate joints: more localized板接头：更具局部性Distance from Weld ToeSummary: Important Observations on FatigueBehavior in Welded Joints小结：焊接接头疲劳行为的重要结论Fatigue fai

30、lures follow a few distinct failure modes, e.g., toe crack, throat crack, etc.疲劳失效遵从几种截然不同的模式，例如，焊趾裂纹，焊喉裂纹等BM/WM/HAZ properties are not important within a broad class of materials对这一类材料来说？，BM/WM/HAZ 材料属性是不重要的Mean stress effects are not significant平均应力的影响不显著Presence of high residual stresses高残余应力的存在S

31、tress range should be used应当使用应力变化范围S-N curves exhibit a distinct slope ( 1/3)S-N 曲线表现为一个唯一的斜率( 1/3)Dominated by crack propagation决定于裂纹扩展Inherent crack-like discontinuities activated by global stress concentration characteristics at welds由整体应力集中导致的内在的类裂纹不连续性焊缝处的特性？Stress concentration is most import

32、ant 应力集中是最重要的Local notch stress effects (e.g., weld bead, toe profile, under cuts, etc): random in nature and captured in S-N data 局部缺口应力效应 (例如，焊道, 焊趾轮廓, 切割痕等): 本质上是随机的并且可以反映在S-N数据中A reliable method is needed for characterizing global stress concentration (overalljoint geometry effect)为反映整体应力集中的特性，需

33、要一种值得信赖的方法 (整体的接头几何效应)Considerations for SCF Characterization对SCF特性的一些考虑Local stresses cannot be readily determined for most practicalapplications对于大多数的实际应用，局部应力并不容易得到 Singular behavior at a sharp notch尖锐缺口处的奇异性Random nature of weld toe/root local details for assuming a definite radius对于假定的一个半径值来说，焊

34、趾/焊根处的局部细节本质上是随机的 Conceptually, there should exist a global stress parameter：理论上来说，应当存在一个整体应力参数： Dominated by overall joint geometry excluding the local effects排除局部效应，而决定于整体接头几何形状 Load mode, etc.载荷模式等The simplest global stress parameter: nominal stress definition perstrength of material, if applicab

35、le最简单的整体应力参数：如果适合的话，名义应力以材料强度来定义A Brief Overview of Conventional Stress Concentration Determination Methods传统的应力集中计算方法简介Global stress parameter based:整体应力参数基于Nominal stress名义应力Extrapolation based spot stress基于热点应力的外推Local notch stress parameter based基于局部缺口应力参数Classical Weld Classification Approach B

36、asedon Joint Geometry 基于接头几何的传统焊缝分类方法 Also referred to as “weld category approach”, “cartoonapproach”, “fatigue design rules”(Gurney, 1967)也称为“焊缝分类方法”，“图表方法”，“疲劳设计准则”(Gurney, 1967) Nominal stress S-N curves (mean and -2 ) in log-logscale from fatigue tests:由疲劳测试得到的对数坐标中的名义应力S-N曲线(平均和-2) Different jo

37、int geometries不同的接头几何 Mostly applicable for weld toe failure mode大部分适用于焊趾失效模式 Limited considerations for weld throat failure modes很少考虑焊喉失效模式Weld Classification Approach: Mean S-N CurveBS 7608 of (Ref. Plate Thickness: 16mm)焊缝分类方法：BS 7608的平均S-N曲线（参考板厚：16mm）A total of 8 S-N curves provided based lab s

38、pecimen tests实验室试件测试一共提供了8条S-N曲线 Fatigue evaluation procedure:疲劳计算过程 Determine which curve (“B”-“W”) is applicable to a joint of concern确定哪条曲线适合于这种接头 Calculate nominal stress (range)per strength of material由材料力学计算名义应力（变化范围） Calculate mean life directly fromthe S-N chart, if constant amplitude loadin

39、g如果是等幅载荷，直接由S-N图计算平均寿命 Miners rule summation if variable amplitude loading如果是变幅载荷，根据Miner准则进行累积In Practice, A Family of Infinite S-N Curves is Needed,e.g., IIW s Fatigue Design Recommendations (2004)实际应用中，需要一族无限条S-N曲线，例如，IIW的疲劳设计建议（2004）FEA-Based Fatigue Design and Life Prediction基于有限元的疲劳设计与寿命预测Nomi

40、nal stresses: difficult to extract in most cases名义应力：在大多数情况下难以提取 Cartoon-based S-N curve definitions: difficult to apply基于图表的S-N曲线的定义：难以应用 Solution?求解？ Almost all research has been focused on how to refinesurface extrapolation procedures几乎所有的研究都集中在如何完善表面外推过程上 The objectives:目标 Reduced the number of

41、S-N curves needed for fatigue design for engineering structures减少对工程结构进行疲劳设计所需的S-N曲线的数量 Improve the mesh-sensitivity in HSS calculations在热点应力计算时改善网格敏感性Extrapolation-Based Hot Spot StressApproach Basic Definition and Assumptions基于外推法的热点应力方法基本定义与假定Developed in 70s for offshoretubular joints发展于70s船体结构的

42、管接头 Used membrane + bendingdecomposition as an argument forexistence of such a stress definition使用膜力+弯曲分解作为这种应力定义存在的证明？ Recommends extrapolations using surface stress extrapolationwithout justifying the linkage建议使用without justifying the linkage表面应力外推方法进行外推？ Extrapolation positions should be outside

43、of the region dominated by local notch stress外推位置应当在局部缺口应力控制区域之外Surface Extrapolation Based Hot Spot Stress Methods基于表面外推的热点应力方法 Also referred to as:也被称为 Hot spot structural stress热点结构应力 Structural stress结构应力 Geometric stress几何应力 More attractive in principle than nominal stress:从原理上比名义应力更吸引人 Applica

44、tions where nominal stress cannot be defined应用于名义应力无法定义的地方 Potentially reduced the number of S-N curves needed潜在地减少了所需S-N曲线的数目 Adopted by various fatigue design codes: IIW Recommendations, AWSD1.1, API RP2A, BS 7608, Class Societies Fatigue Guidance,Eurocode, etc.被多种疲劳设计标准采用：IIW Recommendations, AWS

45、D1.1, API RP2A, BS 7608, Class Societies Fatigue Guidance,Eurocode,等 Recent EN 13445, PD5500 etc for pressure vessel and pipingapplications最近在 EN 13445, PD5500 等压力容器与管道标准中得到应用Extrapolation Based HSS Approach (IIW,04)基于外推的热点应力方法(IIW,04)Notch Stress Methods缺口应力方法 Notch stress is ill-defined if assumin

46、g sharp notches如果是尖锐的缺口，缺口应力很难定义 Some researchers are investigating the use of notch stress with a fictitious notch radius in FE model一些研究人员正在研究有限元模型中假定缺口半径下的缺口应力的使用 Classical ASME fatigue procedure is by definition is a notch stress approach经典的ASME疲劳设计方法从定义来看是一种缺口应力方法 relying empirical methods to d

47、etermine FSRF (Kf)确定FSRF(Kf)时依赖于经验方法 cant be directly and consistently used for FEA based fatigue design in general不能直接和一致地应用于一般的基于有限元方法的疲劳设计 Kt is not equal to KfKt不等于Kf Kf: joint type dependent and normalized w.r.t nominal stressesKf：依赖于接头类型和名义应力的正则化 Linearized “stress intensity” definitions in AS

48、ME are not the nominal stressesASME中线性化的“应力强度”的定义不是名义应力Kf is also Dependent Upon Cycles to Failure When Scaling from Smooth Bar Data 从光滑试样数据进行缩放时，Kf还要依赖于失效循环数ASME Fatigue Design Curve DefinitionASME疲劳设计曲线定义Used smooth bar mean S-N data使用光滑试样的平均S-N数据Displacement controlledtests位移控制测试 Expressed using

49、either nominal strain orpseudo-elastic nominalstress amplitude使用名义应变或者伪弹性名义应力幅来表达Apply 2/20 rule应用2/20规则Nominal Stress Range Versus N Pressure Vesseland Pipe Weld Data versus ASME Fatigue Curves名义应力变化范围与N压力容器和管道焊接结构与ASME疲劳曲线In Summary: The Two Major Issues Must beResolved for Reliable FE-Based Fatig

50、ue Evaluation总之，为得到可信的基于有限元的疲劳评估结果，两个问题必须解决 What stress to use? Which S-N curve to use? 使用什么应力？ 使用哪条S-N曲线？What Constitutes a Good Stress-Based Fatigue Parameter?哪些物理量构成好的基于应力的疲劳参数 Necessary conditions必要条件 Consistency in calculation:计算一致性 Good mesh-insensitivity好的网格不敏感性 Ability to measure global stre

51、ss concentration effects seen inTests 能够度量试验中捕捉到的整体应力集中效应 Sufficient conditions:充分条件 Effectiveness in S-N data correlation:S-N数据相关的有效性 Different joint geometries不同的接头几何 Different loading modes不同的载荷模式 Different plate thicknesses, etc不同的板厚，等 Robustness for practical applications实际应用的稳健性The Structural

52、Stress Method Part I结构应力方法第一部分 Requirements for an effective fatigue parameter对有效的疲劳参数的要求 The new structural stress definition新的结构应力的定义 Mechanics basis力学基础 Equilibrium arguments平衡条件 SS determination procedures结构应力计算过程 Simple FEA-based procedures简单的基于有限元方法的过程 Measurement procedures测量过程 Validation by

53、correlating S-N data 与S-N数据相关的证明 Similarities and differences between the SS and other global stress parameters结构应力与其它整体应力方法的异同Symmetry and anti-symmetry对称与反对称 Demo of manual SS calculation procedures using FE model使用有限元模型，手动计算结构应力的过程示例 Straight and curved lap fillet welds直线和曲线搭接焊缝 In-plane gusset a

54、ttachments面内搭接附件 SS measurements and calculations from FE model基于有限元模型的结构应力度量与计算 Shear locking effects剪切锁死的影响Requirements for a FE Based Stress Parameter Definition for Fatigue Evaluation疲劳评估对基于有限元应力参数定义的要求 Consistency in stress determination:计算应力时的一致性 e.g., good mesh-insensitivity例如，良好的网格不敏感性 Effec

55、tiveness in S-N data correlation:与S-N数据相关的有效性 Robustness for practical applications实际应用的稳健性The Structural Stress Definition for a 2D Problem Traction Based2D问题的结构应力定义基于拉伸情况Structural Stress: Equilibrium Equivalent结构应力：平衡等效Notch Stress: Self-Equilibrating缺口应力：自平衡Traction-Based Structural Stress Defin

56、ition in 3D3D问题基于拉伸的结构应力定义In general, three structural stress components exist corresponding to the linear representations of three traction components一般情况下，相应于三个拉伸分量的线性表示，存在三个结构应力分量 Normal法向 In-plane shear面内剪切 Transverse shear横向剪切Section A-A in a 2D Cross-Section Section A-A-C-C in a 3D Cross-Secti

57、on2D横截面中的截面A-A 3D横截面中的截面A-A-C-C Equilibrium Considerations in Numerical Implementation在数值实现中考虑的平衡条件 Displacement based FE procedures:基于位移的有限元方法 Nodal forces and displacements are the most fundamental quantities节点力和位移是最基本的物理量 Equilibrium conditions are always guaranteed in terms of nodal forces at no

58、des, but not in terms of stresses以节点上的节点力表示的平衡条件总能得到保证，然而应力却不能保证 The equilibrium-equivalent structural stresses can be extracted using:平衡等效结构应力可以提取出来，通过使用： Balanced nodal forces from finite element output有限元输出的平衡节点力 The key step: “work equivalent” based transformation fromnodal force/moments to line

59、 force/moments关键步骤：基于“功等效”将节点力/弯矩转换为线力/弯矩Balanced Nodal Forces from Finite Element Solutions从有限元解获得的平衡的节点力 “GPFORCE” in NASTRANNASTRAN中的“GPFORCE” “NFORC” in ABAQUSABAQUS中的“NFORC” “NLOAD” in ANSYSANSYS中的“NLOAD” Property of these forces: equilibrium conditions with respect to each element and all elem

60、ents together are exactly satisfied这些力的性质: 无论是对于每一个单元还是所有单元，平衡条件总能精确满足From Nodal Forces/Moments to SS Simple Calculation Method for Shell/Plate Models从节点力/弯矩到结构应力板壳模型的简单计算方法Extract nodal force/moment (M/F) w.r.t each element along a weld line沿着焊线对每一个单元提取节点力/弯矩（M/F）Obtain line force/moment (m/f) dist