The complex thermo-mechanical processes that occur during the manufacturing (by welding) of large components involve important volume changes that cause residual stresses and distortions in the final geometry of the assembly, being their prediction prior to the manufacturing one of the most interesting challenges. Today, it is possible to accurately estimate the final deformation of an assembly produced by welding process and predict the geometrical distortions by means of simulation techniques using Finite Element Analyses. The implementation of the thermal shrinkage model in ANSYS® software has allowed the development of the RED-WeldS (Rapid Evaluation of Distortions in Welded Structures). This tool allows developing a fast and accurate enough approach for predicting welding distortions. On the other hand, one of the main challenges generally faced by the welding engineer is the determination and use of the appropriate welding sequence to achieve the minimum deformations and obtain the final welded assembly within tolerances. The implementation of Artificial Intelligence algorithms within RED-WeldS for the management and analysis of the possible welding sequences has enabled the tool to predict the optimal sequence under two assumptions: a) optimal solution, minimum distortion of the welded assembly and b) acceptable solution, distortion below the tolerance allowed by the user. Finally, the implemented algorithm is applied to a mock-up assembly in the laboratory with 18 welds, including different number of passes. Then, the distortion measurements and simulation results for the optimal welding sequence were compared to validate the RED-WeldS tool and its optimization capability. This development has been carried out in the framework of the project "Strategic Positioning in Virtual Models and Digital Twins for Industry 4.0 (MIRAGED)" (Expte: CER-20190001), funded by the Centre for the Development of Industrial Technology (CDTI) for which IDONIAL has been distinguished as "Cervera Centre of Excellence" in relation to Deep learning, Artificial Intelligence.