The integrity of structures in nuclear power plants has to be assessed to meet given safety criteria. For a better understanding of the in-service loads in welded areas of PWRs components, the residual stresses resulting from the welding process have to be evaluated. For that purpose, numerical simulation of welding has proved efficient. However, in an industrial context, simple models have to be used to stay reliable, brief and easy to use.
Among all the steps required to run a suitable computation of the welding, this paper focuses on the calibration of the heat source. Two complementary approaches are applied to define an equivalent heat input, which coarsely include complex phenomena in the weld pool. The engineering practice is driven by a know-how: simplified method are used to get an efficient result in short modelling time. On the other hand, EDF R&D develops tools which quantify the difference between models and lead to an objective choice. The mix of both approaches enables to consolidate the engineering practice.