Intelligent Hybrid-Layer Friction Drill Joining Using a Coupled Finite Element and Deep Learning Framework for Cast Al380 Enhancement

Joining cast aluminum alloys remains a significant challenge due to their inherent brittleness and their propensity to crack under thermo-mechanical loading. This limitation is particularly evident in Al380, which exhibits severe petal formation and unstable bush development during friction drilling. To overcome these challenges, this study investigates the friction drill joining (FDJ) of Al380 using three distinct upper-sheet materials – Cu, Al6061, and AISI304 – with the aim of improving heat distribution, reducing crack initiation, and enhancing bush formation. A combined experimental, numerical, and data-driven methodology was adopted. Experimentally, friction drilling test was conducted under certain process conditions to validate the influence of upper-layer material on heat generation, material flow, and crack initiation. Complementary thermo-mechanical finite element simulations were developed using ABAQUS to evaluate temperature evolution, von Mises stress, and bushing formation behavior within the joint region, thereby elucidating the mechanisms governing improved formability in hybrid-layer assemblies. Experimental and numerical analyses revealed that the presence of an upper sheet significantly modified the heat generation and stress distribution within the joint zone, reducing crack initiation and improving material flow. A feed-forward artificial neural network (ANN) composed of multiple layers developed using the MATLAB Deep Learning Toolbox was further employed to predict bushing length, peak temperature, and stress response across different process conditions. The results demonstrate that integrating AL380 with dissimilar metallic upper layers effectively overcomes its brittleness and expands the applicability of friction drill joining for cast alloys, offering a promising route toward reliable lightweight multi-material assemblies.