Copy the page URI to the clipboard
Elcoate, C. D.; Dennis, R. J.; Bouchard, P. J. and Smith, M. C.
(2005).
DOI: https://doi.org/10.1016/j.ijpvp.2004.08.003
Abstract
Full 3-dimensional (3-D) simulation of multi-pass weld repairs is now feasible and practical given the development of improved analysis tools and significantly greater computer power. This paper presents residual stress results from 3-D finite element (FE) analyses simulating a long (arc length of 62°) and a short (arc length of 20°) repair to a girth weld in a 19.6 mm thick, 432 mm outer diameter cylindrical test component. Sensitivity studies are used to illustrate the importance of weld bead inter-pass temperature assumptions and to show where model symmetry can be used to reduce the analysis size.
The predicted residual stress results are compared with measured axial, hoop and radial through-wall profiles in the heat affected zone of the test component repairs. A good overall agreement is achieved between neutron diffraction and deep hole drilling measurements and the prediction at the mid-length position of the short repair. These results demonstrate that a coarse 3-D FE model, using a ‘block-dumped’ weld bead deposition approach (rather than progressively depositing weld metal), can accurately capture the important components of a short repair weld residual stress field. However, comparisons of measured with predicted residual stress at mid-length and stop-end positions in the long repair are less satisfactory implying some shortcomings in the FE modelling approach that warrant further investigation.