Modelling the Toughness of Nanostructured Polyhedral Oligomeric Silsesquioxane Composites Fabricated by Stereolithography 3D Printing: A Response Surface Methodology and Artificial Neural Network Approach

Abstract

In this study, stereolithography (SLA) 3D printing was used to prepare toughened composites by facile blending of chemically compatible nanoscale polyhedral oligomeric silsesquioxanes (POSS) to commercial photoreactive resin. Due to the complex nature of 3D printing, the mechanical performance of the final parts cannot be simply determined or even estimated until they are manufactured and tested. Thus, response surface methodology (RSM) and artificial neural network (ANN) were used to build regression models for determining the toughness of fabricated composites as function of toughener (POSS) amount and printing conditions (layer thickness and annealing temperature). The influence of the mentioned process parameters on toughness were investigated through a 17-run three-factor three-level Box-Behnken RSM design (BBD). The same experimental design was also used to acquire a data set for ANN. Finally, both the modeling methodologies were compared by coefficient of determination (R2) and residual distribution values. Results reveal that ANN possesses a better data fitting and predictive power as compared to RSM.