Optimization of lap-joint laser welding parameters using high-fidelity simulations and machine learning mode

• DOI: 10.1016/j.jmrt.2023.04.256
• Publication Date: 2023-05-04T06:48:14Z
• Authors: Yung-An Tsai; Yu-Lung Lo; M. Mohsin Raza; Ali N. Saleh; Tzu-Ching Chuang; Cheng-Yen Chen; Chi-Pin Chiu
• Keywords: Weld pool

In lap joint laser welding, a common practice is to conduct trial-and-error experiments using various parameter settings determine processing conditions that enhance the quality of weld. However, these are both time-consuming and expensive. Therefore, in this study, we propose more systematic approach for determining optimal power scanning speed SS316 by highly accurate simulations artificial neural network models. The maps were obtained three criteria: melt pool depth, width, cooling rate, respectively, which screened appropriate criteria could simultaneously minimize porosity, size heat affected zone, residual stress. validity simulation model was confirmed comparing results geometry with experimental data. mean deviations simulated depth width found be only 5.34% 10%, respectively. As result, welds produced parameters exhibited minimal reduced from 1.22% non-penetration zone 0.21% an optimized zone. Additionally, achieved ultimate shear strength up 545.77 MPa, approximately 32% higher than original base metal. effectiveness proposed framework welding has been confirmed.