No abstract

This paper aims at introducing a new nodal-integration-based finite element method for the numerical calculation of residual stresses induced by welding processes. The main advantage of the proposed method is to be based on first-order tetrahedral meshes, thus greatly facilitating the meshing of complex geometries using currently available meshing tools. In addition, the formulation of the problem avoids any locking phenomena arising from the plastic incompressibility associated with von Mises plasticity and currently encountered with standard 4-node tetrahedral elements. The numerical results generated by the nodal approach are compared to those obtained with more classical simulations using finite elements based on mixed displacement–pressure formulations: 8-node Q1P0 hexahedra (linear displacement, constant pressure) and 4-node P1P1 tetrahedra (linear displacement, linear pressure). The comparisons evidence the efficiency of the nodal approach for the simulation of complex thermal–elastic–plastic problems.

No abstract

The tungsten inert gas (TIG) welding process is widely used in various industrial fields to achieve high-quality metallic welds. In particular, this process is used for the assembly of primary circuit components in pressurized water reactors. The main limitation of this process is the weak penetration of the welding pool, which does not allow for the welding of large components in one pass, altering its efficiency. [1] Besides, welding operations contribute to modifying material properties and residual stress, both of them constituting input data for the risk analysis of nuclear safety demonstration concerning fatigue degradations, stress corrosion cracking, or mechanical rupture. Numerical simulations are used to characterize, among others, residual stress. Increasing the penetration of the welding pool is thus a major industrial issue requiring the improvement of numerical tools modeling the fundamental phenomena that are involved in the TIG welding process.Among the different physical mechanisms influencing welding pool dynamics, Marangoni flows generated by surface tension gradients at the free surface may play a significant role in the wetting or penetrating behavior of the weld under consideration. [2] As introduced in Section 2, state of the art, the simulation of Marangoni flows requires the knowledge of liquid steel thermophysical properties, such as density ρ and surface tension σ, as a function of temperature. In particular, existing models in the literature have demonstrated their capacities to predict a melt pool if reliable and accurate surface tension data are provided for the steel grade under consideration. [3] These data at liquid state are scarce, and even not available above a temperature threshold around 1800 C, due to technological issues, particularly for austenitic vessel steel AISI 304 L. Moreover, as emphasized hereafter, the strong influence of minor components, such as sulfur and oxygen, on surface tension values, makes such measurements even more challenging.Within this context, the VITI facility is used to provide new data at high temperature T--from melting point up to 2100 C-for the density and surface tension of two grades of AISI 304 L steel, namely 304LP1-low sulfur content-and