The Measurement and Modelling of Residual Stresses in a Full-Scale BWR Shroud-Support Mock-Up

Abstract

This paper presents results from a programme of residual stress measurements and modelling carried out in Japan on a full-scale mock-up of the hemi-spherical base of a Boiling Water Reactor (BWR) pressure vessel. The shroud support mock-up consisted of four main parts: pressure vessel, support plate, support cylinder and support legs. The mock-up was manufactured using a combination of ferritic steel and nickel base alloy (i.e. alloys: 82, 182 and 600) in a similar manner to that of the actual component. Overall the mock-up had an outer diameter of 6.6m and a height of 3.4m. The residual stresses generated by the nickel alloy welds during manufacture were measured using the Deep-Hole Drilling (DHD) [1–3] and Sectioning [4, 5] techniques, and modelled using ABAQUS. Presented here are measurement and modelling results from three weld locations within the mock-up: at one location through the “double-bevel” butt-weld joining the top of the support leg to the support cylinder (named H10) and at two locations through the “asymmetric double-V” weld joining the bottom of the support leg to the cladded pressure vessel (named H11a and H11b). The semi-destructive DHD technique was carried out first at all three locations on-site in Japan before the fully destructive Sectioning technique was used. Both techniques measured the biaxial (i.e. mock-up-hoop and -axial) residual stresses. The DHD and Sectioning techniques were not carried out at the exact same locations, rather similar locations due to the axisymmetry of the mock-up. Modelling of the residual stresses generated was undertaken for each weld location using a 2D axisymmetric finite element analysis containing between 40–50 discrete weld beads. The modelled residual stresses were generated using thermal load modelling followed by elastic-plastic mechanical analysis under kinematic hardening rules. Overall there is excellent agreement between the measured and modelled residual stresses at all locations. At all locations the measured peak tensile residual stresses (i.e. H10 = 410MPa, H11a = 260MPa and H11b = 230MPa) were found to be in the hoop direction just below the inner weld cap surface. The modelled peak tensile residual stresses were again found in the hoop direction near the inner weld capped surface, however, they were found to be approximately 155MPa greater than the measured residual stresses, and for locations H10 and H11a similar peaks were found near the outer weld cap surface as well.