Machine learning to predict interface adhesion between dissimilar materials in multi-material extrusion additive manufacturing

Material extrusion (ME) additive manufacturing is renowned for its ability to simultaneously extrude two materials, enabling the fabrication of assembly-free structures (i.e., smart structures). However, multi-material extrusion (M-MEX) of diverse materials presents significant challenges, necessitating numerous trials to identify the optimal material combinations (e.g., soft and stiff) and set of process parameters to maximize interface adhesion. At present, there is no known automated tool capable of predicting interfacial adhesion between dissimilar materials, highlighting a significant gap in the current state of the art. This study addresses this gap by developing a machine learning (ML)-based tool to predict adhesion between dissimilar materials as a function of process parameters. The ML algorithm was trained using four representative materials categorized as extremely soft, soft, stiff, and composite: pairs of materials were jointly extruded, varying several process parameters (e.g., extrusion order, speed of material 1, and speed of material 2 for a total of 144 samples. Three ML algorithms, tailored for small data sets, were tested, optimized, and validated by 3D printing eight combinations of untested materials. The support vector regression (SVR) model was found to be the most effective in predicting interface adhesion, achieving an average prediction accuracy of 88.53%. The SVR model was subsequently employed as a design for M-MEX tool, demonstrating its potential by predicting the optimal material combinations (within a given set) and process parameters to maximize interfacial adhesion. The proposed ML-based approach is envisioned to elevate multi-material 3D printing from a trial-and-error-based process to a systematic, knowledge-driven approach by eliminating manual trials for maximizing adhesion, thus enhancing the efficiency and cost-effectiveness of the 3D printing process.