Some issues in life assessment of longitudinal seam welds based on creep tests with cross-weld specimens

Abstract

In order to reduce production costs, it is of great interest to use longitudinal seam welds when manufacturing large diameter pipes. The cost reduction can be as high as 30%. However, severe inservice accidents for this type of pipes working in the creep regime have occurred mainly due to mismatch in weldment creep properties. In many cases, creep tests of cross-weld specimens, taken from the seam weld, are used to predict the behaviour of the seam weld, assuming that the creep behaviour of specimen and weldment is equivalent. Experiences of this procedure indicate that further knowledge is required before translation between specimen and component can be made. In the present paper, both full scale seam welded pipes and cross-weld specimens are studied with the damage mechanics concept using finite element, FE, technique. The same mechanical model of multiple material zones is used for the two components. Both the influence of differences in creep properties between the weldment constituents and the size effect of the cross-weld specimen, are studied. It is found that the cross-weld test results can not directly be translated to the full scale component. Factors such as the creep properties and the relative geometry of the weldment constituents and the size of the cross-weld specimen have to be considered when performing creep life assessment.