Authors: Peter Dillström Magnus Andersson Iradj Sattari-Far Weilin Zang Iradj Sattari-Far Research 2009:27 Analysis Strategy for Fracture Assessment of Defects in Ductile Materials Report number: 2009:27 ISSN: 2000-0456 Available at www.stralsakerhetsmyndigheten.se Title: Analysis Strategy for Fracture Assessment of Defects in Ductile Materials. Report number: 2009:27. Authors: : Peter Dillström, Magnus Andersson, Iradj Sattari-Far, Weilin Zang and Iradj Sattari-Far. Inspecta Technology AB, Stockholm, Sweden . Date: June 2009. This report concerns a study which has been conducted for the Swedish Radiation Safety Authority, SSM. The conclusions and viewpoints pre- sented in the report are those of the author/authors and do not neces- sarily coincide with those of the SSM. Background SSM has supported research for investigating the role of secondary stresses when fracture assessments are performed for cracked structures made of ductile materials. There are evidences that indicate that some secondary stresses, such as weld residual stresses, are not as important as primary stresses for estimating the safety margin against rupture (measured by the J-integral) for the type of ductile materials which can be found in nuclear power plants. The project was initiated by SKI. Objectives of the project The objective of the project has been to perform numerical analyses of girth welds in pipes of different sizes for estimating the weld residual stresses and investigate how the weld residual stresses behave when cracks are introduced. Based on the results, an analysis strategy has been proposed on how the safety factor against these kinds of secondary stresses can be defined for evaluating pipe rupture in ductile materials. Results - For the studied cases with stainless steel welds or Alloy 182 welds in stainless steel piping, the relative contribution from the weld residual stresses to CTOD or J decreases rapidly for high values of the limit load parameter Lr. For very high values of Lr the analyses indicate that the contribution from the weld residual stresses to fracture becomes negligible. - The precise limit of Lr at which the relative contribution from the weld residual stresses is small is likely to depend on the particular material properties, crack geometry and the weld residual stress distribution. - Based on the analysis result, a new deterministic safety evaluation system is proposed. In the procedure, new safety factors against the fracture toughness are defined where it is distinguished between primary and secondary stresses. Based on the predicted value of Lr at fracture, the safety factor against the local through- thickness secondary stress is lowered. This is assumed also to be valid for thermally induced through-thickness stress gradients. However, no change in safety factor is proposed for global secon- dary stresses such as thermal expansion stresses. SSM 2009:27 Effects on SSM The results of this project will be of use to SSM in safety assessments when cracks are detected in nuclear power plant components. However, SSM wants to further validate the analysis results in this project and the proposed revision of the deterministic safety evaluation system by performing experiments. Such experiments are planned in an ongoing research project. Project information Project Leader at SSM: Kostas Xanthopoulos Project number: 14.42-011210/22094 Project Organisation: Inspecta Technology AB has been managed the pro- ject with Peter Dillström as project leader. Magnus Andersson, Iradj Sattari- Far and Weilin Zang have assisted in the development of the project. SSM 2009:27 SUMMARY The main purpose of this work is to investigate the significance of the residual stresses for defects (cracks) in ductile materials with nuclear applications, when the applied primary (mechanical) loads are high. The treatment of weld-induced stresses as expressed in the SACC/ProSACC handbook and other fracture assessment procedures such as the ASME XI code and the R6-method is believed to be conservative for ductile materials. This is because of the general approach not to account for the improved fracture resistance caused by ductile tearing. Furthermore, there is experimental evidence that the contribution of residual stresses to fracture diminishes as the degree of yielding increases to a high level. However, neglecting weld-induced stresses in general, though, is doubtful for loads that are mostly secondary (e.g. thermal shocks) and for materials which are not ductile enough to be limit load controlled. Both thin-walled and thick-walled pipes containing surface cracks are studied here. This is done by calculating the relative contribution from the weld residual stresses to CTOD and the J-integral. Both circumferential and axial cracks are analysed. Three different crack geometries are studied here by using the finite element method (FEM). (i) 2D axisymmetric modelling of a V-joint weld in a thin-walled pipe. (ii) 2D axisymmetric modelling of a V-joint weld in a thick-walled pipe. (iii) 3D modelling of a X-joint weld in a thick-walled pipe. t. Each crack configuration is analysed for two load cases; (1) Only primary (mechanical) loading is applied to the model, (2) Both secondary stresses and primary loading are applied to the model. Also presented in this report are some published experimental investigations conducted on cracked components of ductile materials subjected to both primary and secondary stresses. Based on the outcome of this study, an analysis strategy for fracture assessment of defects in ductile materials of nuclear components is proposed. A new deterministic safety evaluation system is defined, that more realistically handles the contribution to J or CTOD from secondary stresses. In the new procedure we define new safety factors against fracture described by K and differentiate between SFPrimary (relating to primary stresses) and I K SFSecondary (relating to secondary stresses). The procedure is consistent with the presented analyses and K experimental data. SSM 2009:27 1 TABLE OF CONTENT Page 1 INTRODUCTION....................................................................................................................5 2 ANALYSIS OF INTERNAL CIRCUMFERENTIAL SURFACE CRACKS IN THIN- WALLED PIPES......................................................................................................................7 2.1 Simulation of the welding process and crack growth.................................................................7 2.1.1 Finite element modelling...........................................................................................................7 2.1.2 Thermal analysis.......................................................................................................................8 2.1.3 Structural analysis...................................................................................................................10 2.1.4 Simulation of crack growth.....................................................................................................11 2.2 A welded pipe subjected to a primary load..............................................................................12 2.2.1 Contribution from residual stresses to J and CTOD using an axial loading...............................12 2.2.2 Relaxation of residual stresses due to unloading......................................................................17 2.2.3 Effect of tangent modulus.......................................................................................................18 2.2.4 Effect of using linear isotropic hardening................................................................................23 2.3 A welded pipe subjected to a thermal load...............................................................................24 2.4 Discussion on the results for thin-walled pipes........................................................................29 3 ANALYSIS OF INTERNAL CIRCUMFERENTIAL SURFACE CRACKS IN THICK- WALLED PIPES....................................................................................................................31 3.1 Geometry................................................................................................................................31 3.2 Material data...........................................................................................................................32 3.3 Element mesh.........................................................................................................................33 3.4 Loading and boundary conditions............................................................................................35 3.4.1 Simulation of the weld process................................................................................................35 3.4.2 Simulation of crack growth.....................................................................................................35 3.4.3 Primary load...........................................................................................................................36 3.5 Results....................................................................................................................................36 4 ANALYSIS OF INTERNAL AXIAL SURFACE CRACKS IN THICK-WALLED PIPES.....42 4.1 Geometry................................................................................................................................42 4.2 Material data...........................................................................................................................42 4.3 Element mesh.........................................................................................................................42 4.4 Loading and boundary conditions............................................................................................44 4.4.1 Simulation of the weld process................................................................................................44 4.4.2 Simulation of crack growth.....................................................................................................44 4.4.3 Primary load...........................................................................................................................44 4.5 Results....................................................................................................................................44 5 EXPERIMENTAL RESULTS................................................................................................51 5.1 Wilkowski and Rudland, Battelle, USA...................................................................................51 5.2 Dong et. al., Battelle, USA......................................................................................................52 5.3 Mohr et. al., Edison Welding Institute, USA............................................................................52 5.4 Sharples et. al., AEA Technology, UK....................................................................................54 5.5 Sharples and Gardner, AEA Technology, UK..........................................................................56 5.6 The IPIRG project, Battelle, USA...........................................................................................57 5.7 Lei et. al., Imperial College, UK..............................................................................................59 5.8 Discussion..............................................................................................................................60 6 A STRATEGY FOR FRACTURE ASSESSMENT OF DEFECTS IN DUCTILE MATERIALS.........................................................................................................................61 6.1 Motivation for a new strategy for fracture assessment of defects in ductile materials ...............61 6.2 Case study 1, a thin-walled pipe containing a circumferential surface crack.............................62 6.3 Case study 2, a thick-walled pipe containing a circumferential surface crack...........................65 6.4 Case study 3, a thick-walled pipe containing an axial surface crack.........................................67 6.5 Case study 4, components with through-wall cracks................................................................69 6.6 Recommendations for a new procedure for fracture assessment of defects in ductile materials.71 6.6.1 A new deterministic safety evaluation system..........................................................................71 SSM 2009:27 3 6.6.2 Application of the new deterministic safety evaluation system.................................................76 7 IMPLEMENTATION OF THE NEW PROCEDURE IN THE FRACTURE ASSESSMENT SOFTWARE PROSACC...................................................................80 7.1 Estimation of safety factors in the new deterministic safety evaluation system.........................80 7.1.1 Safety factors for a normal/upset load event, SF = 3.162........................................................81 K 7.1.2 Safety factors for an emergency/faulted load event, SF = 1.414.............................................81 K 7.2 Choice of options within the ProSACC software.....................................................................82 8 CONCLUSIONS AND RECOMMENDATIONS...................................................................83 9 ACKNOWLEDGEMENT.......................................................................................................85 10 REFERENCES.......................................................................................................................86 APPENDIX A. THE J-INTEGRAL AND CTOD AS FRACTURE PARAMETERS.............................89 APPENDIX B. CALCULATION OF CTOD........................................................................................93 APPENDIX C. LIMIT LOAD OF 2D MODEL....................................................................................95 APPENDIX D. LIMIT LOAD OF 3D MODEL....................................................................................96 SSM 2009:27 4 1 INTRODUCTION Structures may fail because of crack growth both in welds and in the heat affected zone (HAZ). The welding process itself induces residual stresses in the weld and HAZ, which contribute to crack growth. The mechanism of growth can be sub critical i.e. IGSCC or fatigue, or critical growth initiation. Two examples of crack growth in welds are studied by Kanninen et al. [1981] and by Hou et al [1996]. So far, it is not self evident which characterising fracture parameter to use when the crack is subjected to weld-induced stresses. The results presented in the scientific literature indicate difficulties in using the J-integral as fracture parameter when residual stresses are present. Some researchers have proposed to use CTOD (Crack Tip Opening Displacement) as a suitable fracture parameters when secondary stresses are also present in the component, see for instance Kanninen et al. [1981] and Hou et al. [1996]. In this report both the J-integral and CTOD (obtained by the 90°-interception construction) are evaluated using finite element analysis, and their validity and usefulness as suitable fracture parameters are discussed. The main purpose of this work is to investigate the significance of the secondary stresses for defects (cracks) in ductile materials within nuclear applications. The treatment of weld-induced stresses as expressed in the SACC handbook by Andersson et al [1996] is believed to be too conservative for ductile materials. This is because of the general approach does not to account for the improved fracture resistance caused by ductile tearing and furthermore there is experimental evidence that the contribution of residual stresses to fracture diminishes as the degree of yielding increases to a high level. Green et al [1993, 1994] and Sharples et al [1993, 1995, 1996] showed in a series of experiments that at low load levels, i.e. small L, the influence of the residual stresses was large, but r near plastic instability (L = 1) weld-induced stresses were of little importance. Available procedures r for flaw assessments, such as the ASME XI code and the R6 procedure [Milne et al, 1988] treat this issue differently. For instance, the ASME XI code does not consider weld-induced residual stresses in some materials e.g. stainless steel welds. Neglecting weld-induced stresses in general, though, is doubtful for loads that are mostly secondary (e.g. thermal shocks) and for materials which are not ductile enough to be limit load controlled. Several references regarding simulation of the welding process can be found in the literature, see for instance Brickstad and Josefson [1996] and Hou et al. [1996]. The purpose of this study is to investigate the significance of the secondary stresses (mainly weld residual stresses has been studied) for cracks in pipes of ductile materials. Both thin-walled and thick- walled pipes are studied. This is done by calculating the relative contribution from the weld residual stresses to CTOD and the J-integral. Both circumferential and axial cracks are analysed. However, the conclusions from this study are also valid for others components made of ductile materials. Three different crack geometries are studied here, by using the finite element method (FEM). - The first analysis is a 2D axisymmetric modelling of a V-joint weld in a thin-walled pipe. The crack introduced in this analysis is a complete circumferential internal surface crack in the centre of the weld. The results of this analysis are given in Chapter 2 of this report. - The second analysis is similar to the first analysis, but for a thick-walled pipe. The results of this analysis are given in Chapter 3 of this report. - The third analysis is a 3D modelling of an X-joint weld in a thick-walled pipe. The crack geometry in this analysis is a semicircular axial internal surface crack in the centre of the weld. The results of this analysis are given in Chapter 4 of this report. SSM 2009:27 5 Each crack configuration is analysed for two load cases: - Only primary (mechanical) loading is applied to the model. - Both secondary stresses and primary loading are applied to the model. The steps in the analyses of the cases subjected to both primary and secondary stresses are given below: - The welding process is simulated to introduce weld residual stresses (secondary stresses) in the model of the pipe. - The crack is introduced within the model. - A primary load is applied. For this “combined” loading case, the fracture mechanics parameters J and CTOD are calculated. The relative contribution from the secondary stresses (mainly weld residual stresses in this study) is calculated in this report according to the following definitions: CTOD CTOD CTOD  combined primary contributionfromresidualstresses CTOD primary (1-1) J J J  combined primary contributionfromresidualstresses J primary where - CTOD ,J CTOD and J calculated from load cases with both weld residual combined combined stresses and primary loading. - CTOD ,J CTOD and J calculated from load cases with primary loading only. primary primary Details of these analyses are previously reported by Delfin et al. [1997] and Andersson and Dillström [2004]. This report covers the main parts of these two reports. Also presented in this report are some published experimental investigations conducted on cracked components of ductile materials subjected to both primary and secondary stresses. Based on the outcome of these results, an analysis strategy for fracture assessment of defects in ductile materials of nuclear components is proposed in Chapter 6 of this report. Implementation of this analysis strategy into the fracture assessment software ProSACC is given in Chapter 7 of this report. SSM 2009:27 6