Need: The stresses that develop across a weld are important since they may lead to weld cracking, may limit additional service stresses that can be applied and may influence the risk of brittle fracture, stress corrosion cracking or other aspects of service behavior. There has been significant progress in determining stresses around welds due to external loads such as lifting. There has, however, been considerable work published (mainly from Japan) on stresses in flat plates due to welding, and the associated theories are reasonably well developed. In particular, the concept of ‘intensity of restraint* has been introduced to provide some kind of measure of the ‘stiffness’ of a joint before it is welded. This enables the stresses that develop to be related to the type of joint (thickness, size and configuration of plates, etc.). In developing these ideas, it has been necessary to measure experimentally the magnitude of the contractions that occur across welds in unrestrained and restrained welds together with the stresses developed. Much data for flat plate are available but little for pipe welds.
Result: The Welding Institute of Canada has carried out a program for the Pipeline Research Committee of PRCI aimed at developing the restraint concepts for girth welds of pipes. The approach has been to measure the weld contractions and reaction stresses in a bench-top experiment using a 12-inch diameter pipe rotated beneath the welding head. A self-restraining system was designed in which shrinkage of the pipe was restrained by a smaller internal pipe attached with end pieces. Since such a system would be in equilibrium, restraining forces could be measured by compression of the central pipe as well as tension in the external pipe.
Benefit: The program was intended to provide basic data on shrinkage and stresses in pipe welding to determine whether any of the models developed for flat plate were applicable. The experiments were designed to show the effects of restraint level and welding parameters.
