Use Machine Learning to Optimize Weld Integrity

As a welded part cools down, residual stresses form within the part. The structural performance of a weld depends heavily on the parameters of the welding process, such as heat input, welding paths, and fixture designs. Simulation can model the weld process in order to calculate the residual stress distribution and deformation within welded submarine parts. This article explains how artificial intelligence (AI) and machine learning can be used to accelerate analysis and optimization of welds.