Year
2014
Degree Name
Doctor of Philosophy
Department
School of Mechanical, Materials and Mechatronic Engineering
Recommended Citation
Darmadi, Djarot B., Residual stress analysis of pipeline girth weld joints, Doctor of Philosophy thesis, School of Mechanical, Materials and Mechatronic Engineering, University of Wollongong, 2014. https://ro.uow.edu.au/theses/4160
Abstract
Residual stress is one of the important factors that should be considered when assessing the integrity of welded structures because it is well known that residual stress may lead to failure in the weld joint. Residual stress may lead to stress corrosion cracking in a corrosive environment and has resulted in failures in energy pipelines.
Residual stress and stress corrosion cracking in the region of the girth welded joint in pipelines was investigated in the current work. The emphasis was placed on residual stress modeling in pipeline girth welds since it was judged to be an essential prerequisite for stress corrosion cracking.
In this thesis, the residual stress around girth welds was studied with FEM using ANSYS (APDL) to give a deep insight into the formation of residual stress. A coupled Thermo-Metallurgical-Mechanical analysis was carried out because the metallurgical analysis of the high strength steel used extensively in pipelines cannot be ignored. The final residual stress prediction using ANSYS was verified with experimentally to test the validity of the FEM model.
The FEM analysis is presented step by step which commenced with a simple thermal analysis, followed by a thermo-mechanical analysis and finally a coupled thermometallurgical- mechanical analysis. The weld configuration also started from a simple bead-on-plate leading on to multi-pass weld on plate and multi-pass girth weld respectively. Various moving heat source models were used, commencing with a simple point heat source, and then a uniform surface heat source and finally Goldak‟s volumetric heat source.
For the complex model the FEM results were validated with the experimental method to ensure that the model was correct. A new heat source model that combined Goldak‟s volumetric heat source and uniform temperature load was used and it provided good results. The Goldak‟s heat source model was used to represent the heat transferred to the base metal and previous bead (for multi-pass welding), whilst the uniform temperature load represents melted filler metal (droplets) that form the weld bead. A new programming technique using a database which was composed using the ANSYS standard files was also developed. It was found that using the database provided at least three advantages: greater flexibility, faster computing time and the ability to interchange data.
role of solid state phase transformation (SSPT) was also studied in terms of volumetric change due to the atomic packaging factor (APF), alteration of the mechanical properties or transformation plasticity. To assess the contribution of each aspect: volume, mechanical properties and transformation plasticity, FEM with volumetric change (mechanical properties and transformation plasticity excluded), FEM with mechanical properties change (volumetric change and transformation plasticity excluded) and FEM with transformation plasticity (volumetric and mechanical properties excluded) were carried out. These contributions were studied by comparing the results of the FEM with and without SSPT consideration, SSPT with volumetric change only, SSPT with alterations to the mechanical properties only and SSPT with transformation plasticity only. The martensitic SSPT was divided into two classes, namely prime martensite and aged martensite and these prime and aged martensite modes were included in the FEM model which is a new feature. The concept of prime and aged martensite is not new one, but the literature did not indicate that this degree of detail had previously been included in FEM programs. Some papers do use SSPT but only prime martensite is modeled. It was found that the results which considered SSPT were closer to the experimental results for ferritic steel welding.
In the earlier chapters, all above aspects were simulated in plate welding and validated using secondary data. Using the same logic an FEM model was programmed on a multi pass girth weld joint. Phase transformations that were included in the model were austenite and martensite (both: prime and aged martensite) SSPT. A reduction factor (Kd) was used to model the yield stress of filler metal at elevated temperature. Using Kd v in the FEM model of welding may be considered as a new feature. All of the modeled aspects have been validated or verified in the plate welding moreover the temperature history and residual stress distribution obtained from FEM prediction in the multi pass girth welding were validated by the experimental results.
DC-LSND (Direct Cooling – Low Stress No Distortion) control is one of suggested stress mitigation techniques for welding. The DC-LSND typically applied on thin plate welding and the main goal is to reduce distortions. There are no published papers found that quantify the effect of DC-LSND on the residual stresses of multi pass thick pipe girth welds. In this work the effect of DC-LSND on the residual stress in a multi pass girth weld joint of pipe with 8mm thickness was studied through FEM analysis. From FEM predictions the application of DC-LSND reduced maximum residual stress by around 35%. From this result, it is suggested that DC-LSND could be applied to the multi pass butt girth welding of thick pipe although the practicality of this procedure would need to be assessed against its potential benefits.
Overall it can be said that the FEM model in this thesis provides an accurate prediction of residual stress and also a deeper understanding of its development particularly in pipeline girth welds in ferritic steels. It is in line with one of the recent research trends in the welding area: improving the basic understanding of the welding process and the resulting residual stresses.
FoR codes (2008)
0910 MANUFACTURING ENGINEERING, 0912 MATERIALS ENGINEERING
Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.