Computational Intelligence for Welding and Joining

Computational intelligence is advancing welding and joining by enhancing both process design and in-process control. This presentation highlights two complementary developments that together enable more robust, intelligent, and cost-effective welding.

On the modeling side, a novel high-performance welding simulation solver is developed to make large-scale welding simulations practical. The solver enables rapid prediction of key outcomes such as temperature, stress, and distortion, allowing engineers to design and optimize welding processes faster and at lower cost, reducing the need for trial-and-error experiments.

In parallel, an AI-driven approach is presented for real-time welding monitoring and control. A machine vision system captures weld pool dynamics and extracts key features using machine learning based models. These features feed a closed-loop control system that adaptively adjusts process parameters, improving stability, automation, and weld quality.

Together, these advances enable intelligent welding from design to execution.