This study predicts melt-pool depth and width in laser welding of 316L stainless steel using a multilayer feed-forward neural network. Trained on experimental data designed with the Taguchi method, the 3–10-10–10-2 model achieved a correlation coefficient of 0.99995 between predicted and experimental results. The approach can support parameter selection and help reduce welding defects.

The full study is available through the source link.

#NeuralNetwork #LaserWelding #316L