Additive manufacturing processes by wire deposition such as the Wire Arc Additive Manufacturing process (WAAM) allow manufacturing large mechanical components by adding successive layers of molten metallic wire using an electrical arc. This process raises anincreasing interest and may provide a viable alternative to the conventional manufacturing processes for its many advantages, among which the ability to produce large parts with very high deposition rates in low-cost installations, while avoiding the safety and environmental issues raised by the use of metallic powder. If this process derived from welding is well known, its use at the industrial level requires to better understand the influence of the weldingparameters and the deposition strategies on the residual stress and distortion distributions generated during the manufacturing process. However, the residual strains and stresses, among other quantities of interest, are very difficult and expansive to access experimentally, but can be provided by numerical simulations. Therefore, the proposed research work consists in setting up, calibrating, and validating a finite element model to simulate the WAAM process, to then determine the consequences in terms of stresses and distortions on a 316 stainless steel industrial parts. A macroscopic thermo- mechanical model is implemented using the finite element code Code_Aster. Simulations are carried out for various geometries and with different deposition strategies. The implementation of the model requires the determination of input parameters, such as the parameters of the heat source, that cannot be directly measured or characterized. To calibrate the set-up model, instrumented experimental tests are conducted, using thermocouples, thermal imaging and 3D scan. To fit the thermal model to the experimental data, a bayesian calibration of the parameters based on a surrogate model approach is performed. The finite element results are then compared to multiple test cases experimental data, and they show good agreement. After the validation of the models, different approaches are considered to reduce the calculation time for the simulation of large components.