Predictive modelling of flexural behaviour of polymer composites: a machine learning approach through material extrusion

This work presents a novel comprehensive comparative study of different machine learning models on the flexural behaviour of multi-walled carbon nano-tubes reinforced poly lactic acid fabricated through material extrusion. The investigation focused on key printing parameters, including layer thickness, raster orientation, and feed rate. The fabricated specimens were subjected to rigorous flexural testing, followed by fractography analysis to assess the microstructural integrity post-testing. The flexural strength of the specimens exhibited a maximum of 130.935 MPa to a minimum of 60.618 MPa. The flexural testing results’ dataset formed the basis for evaluating the effectiveness of applied eight regression algorithms. With a root mean square error of 1.776 and a mean absolute error of 1.366, the extreme gradient boost algorithm demonstrated the best performance while maintaining the coefficient of determination of 0.99. This analysis emphasizes the potential of integrating machine learning algorithms in expanding predictive methodologies in material science. Such advancements are particularly significant in the realm of additive manufacturing, offering promising avenues for enhancing material performance through informed process parameter selection.