Résumé:
The precision of parameter selection in submerged arc welding (SAW)
significantly influences weld quality, strength, and efficiency in industrial
manufacturing. Artificial Intelligence (AI) offers advanced tools for tackling the
complex, non-linear optimization challenges in welding processes, where
traditional trial-and-error methods fall short. This paper introduces the
Adolescent Identity Search Algorithm (AISA), an innovative AI-based, human
inspired optimization technique, to optimize SAW parameters. Implemented in
MATLAB, AISA is applied to minimize bead width—a critical indicator of weld
quality—by iteratively refining welding parameters such as current, voltage, and
welding speed. To validate AISA’s effectiveness, we conduct a comparative
analysis with the Rao algorithm, a recently studied optimization method shown
to enhance welding outcomes. We assess both methods under varying
population sizes and iteration counts, testing two distinct scenarios to examine
their performance and convergence. Results demonstrate AISA’s robustness and
adaptability in optimizing welding parameters, offering a viable alternative to
existing methods and underscoring the growing importance of AI in advancing
manufacturing precision.
