Modeling of the Behavior of a Welded Joint Subjected to Reverse Bending Moment at High Temperature

Abstract

The paper is concerned with the modeling of the behaviour of welds when subjected to severe thermal and mechanical loads where the maximum temperature during dwell periods lies in the creep range. The methodology of the life assessment method R5 is applied where the detailed calculations are carried out using the linear matching method (LMM), with the objective of generating an analytic model. The linear matching method has been developed to allow accurate predictions using the methodology of R5, the UK life assessment method. The method is here applied to a set of weld endurance tests, where reverse bending is interrupted by creep dwell periods. The weld and parent material are both Type 316L(N) material, and data were available for fatigue tests and tests with 1 and 5h dwell periods to failure. The elastic, plastic, and creep behavior of the weld geometry is predicted with the LMM using the best available understanding of the properties of the weld and parent material. The numerical results are translated into a semi-analytic model. Using the R5 standard creep/fatigue model, the predicted life of the experimental welds specimens are compared with experimental data. The analysis shows that the most severe conditions occur at the weld/parent material interface, with fatigue damage concentrated predominantly in the parent material, whereas the creep damage occurs predominantly in the weld material. Hence, creep and fatigue damage proceed relatively independently. The predictions of the model are good, except that the reduction in fatigue life due to the presence of the weld is underestimated. This is attributed to the lack of separate fatigue date for the weld and parent material and the lack of information concerning the heat affected zone. With an adjustment of a single factor in the model, the predictions are very good. The analysis in this paper demonstrates that the primary properties of weld structures may be understood through a number of structural parameters, defined by cyclic analysis using the linear matching method and through the choice of appropriate material data. The physical assumptions adopted conform to those of the R5 life assessment procedure. The resulting semi-analytic model provides a more secure method for extrapolation of experimental data than previously available.