Approach for weld bead geometry prediction to optimize process planning in wire arc additive manufacturing applications

Abstract

To enable the industrial use of wire arc additive manufacturing (WAAM), it is crucial to provide effective work preparation through process planning systems such as computer-aided-manufacturing (CAM) software. However, current CAM software systems are limited regarding deposition, path manipulation and simulation. To overcome these issues, this paper presents an artificial neural network (ANN) driven methodology for the prediction and integration of effective weld bead geometry information. The approach comprises three phases encompassing the steps of automated data processing, data-based modelling of weld bead geometry and the utilization of this additional information through CAM integration.