Bead geometry prediction and optimization for corner structures in directed energy deposition using machine learning

Abstract

Wire Arc Additive Manufacturing as a Directed Energy Deposition technology has disruptive potential for modern manufacturing. The technology comes with the flexibility and material efficiency of additive manufacturing processes while mitigating the disadvantages through high material deposition rates and high energy efficiency. However, the prevalence of the technology is inhibited by its geometrical inaccuracy and the large induced residual stresses. This work tackles the former problem by capturing the process parameter-geometry relationship using Machine Learning. To do so, multiple mild steel welding beads with varying shape features like corner angles are printed using a Metal Inert Gas welding machine attached to an industrial robot. The cross-sectional profile of the printed beads is measured using a point laser sensor and correlated through a Multilayer Perceptron to input features such as travel speed, wire feed speed, and shape features. By incorporating varying corner angles, a holistic model, not limited to geometry prediction of straight beads only, is trained. Using the model, excess material at the inner angle of corners determined by the overlapping regions of the two adjacent beads can be predicted. By generating a database of possible bead shapes an inverse algorithm that suggests welding parameter combinations that yield a smooth bead shape at the corner is created, thus enabling improvement of the overall accuracy of the Wire Arc Additive Manufacturing process.