Prediction of Weld Bead Width Using Artificial Neural Network and Regression Models in Laser Beam Welding

In this paper, a neural network modeling approach is presented for the prediction of weld bead width in Nd:YAG Laser welding process. The data used for the training and checking of the Artificial Neural Networks’ performance are derived from experiments conducted according to the principles of design of experiments (DoE) method. The influential factors considered in the experiments are the laser beam power, the laser beam angle, and the welding speed. Using feed forward back propagation artificial neural networks (ANN) with supervised training a 3_5_1 neural network is developed and is able to predict the laser bead width and tends to be consistent throughout the entire range of values. The experimentally determined weld bead width values are compared with predicted values obtained from regression and ANN model.