Numerical Modelling of Repair Welding Using a Temper Bead Process: Application to Dissimilar Metal Welds

For nuclear reactor applications, AREVA NP has to perform junctions between ferritic low alloy steel heavy section components and austenitic stainless steel piping systems. For Gas Tungsten Arc Welding (GTAW) of dissimilar metal weld (DMW) narrow gap, AREVA NP has developed special manufacturing procedures guaranteeing high quality standards and resistance in service. Since a decade, AREVA NP is developing the numerical simulation of welding to have a better understanding of involved physical phenomena and to predict residual stresses. In spite of the large thickness of Pressurized Water Reactor (PWR) components, the distortion issue may also be important. Narrow gap welding requires indeed a close control of the groove width. This paper presents numerical simulations performed by AREVA NP on 14^″ narrow gap DMW mock-ups as part of a research project carried out internally. The simulations focus on the predictions of microstructure and residual stress distribution. The analysis simulates the main steps of the mock-up manufacturing procedure. Multi pass welding simulation reproduces the deposit of each bead by thermo-metallurgical and mechanical calculations. A special attention has been paid on the buttering of the ferritic side. Generally a post weld heat treatment (PWHT) is carried out after the buttering of the ferritic side in order to relieve residual stresses. For some repair operations, a PWHT is not feasible. Thus a temper bead process can be used. During this process, a large part of the previous heat affected zone is tempered to guarantee a limited hardness and to reduce the risk of cold cracking. The results in terms of microstructure and stress obtained with the two techniques are compared. With the temper bead process, the final level of hoop stresses in the heat affected zone (HAZ) of the buttering remains significant as stresses are not relieved by viscous effects implied during PWHT. Nevertheless the temper bead process has a positive effect on the material hardness as the proportion of tempered phase is higher. One of the objectives of this task is to compare the numerical results with measurements. This comparison is not only a validation of numerical simulation of welding but also a way to investigate the relevance of residual stress measurement by Deep Hole Drilling (DHD). Calculated stresses are globally in good agreement with measurements made by DHD. A comparison with axial shrinkage is also made for validation of the modelling methodology.