Implementation of a thermal model for heat affected zone geometry estimation in wire arc additive manufacturing (WAAM) by FEM analysis and artificial neural network (ANN) integration

Abstract

This article presents the development and validation of a thermal model for the Wire Arc Additive Manufacturing (WAAM) process. The model is based on the Finite Element Method (FEM) and is designed to predict the thermal distribution of the substrate where the welded bead is going to be deposited. To validate the thermal model, an experimentation process was carried out using three different substrate thicknesses (8, 10 and 30 mm), while varying key process parameters that influence heat input (such as deposition rate and travel speed). The simulation results were compared with experimental data by analyzing critical thermal parameters such as the area, height, and width of the heat affected zone (HAZ), as well as the melt pool dimensions. The HAZ geometry obtained from the thermal simulation showed strong agreement with experimental observations, confirming the model’s accuracy. Then, the model was extended to simulate a broader range of travel speed and deposition rate per sample thickness. This allowed for a more comprehensive study of how these parameters affect the thermal response of the substrate. The expanded dataset generated from the simulations was used to train a sequential artificial neural network and then compared with experimentation obtained results. The ANN model demonstrated reliable predictive performance across all cases, with the lowest error observed for the 30 mm substrate thickness due to its reduce extreme temperature gradients and transient thermal fluctuations. The novelty of this work lies in the integration of a validated FEM thermal model with a sequential ANN, enabling rapid and accurate prediction of thermal fields in WAAM process.