Comparative Analysis and Predictive Modeling of Wear Performance of Glass- and Bamboo Fiber-Reinforced Nanoclay–Epoxy Composites Using RSM and ANN

This research contributes to the field of materials engineering through an analysis of the wear performance of both glass fiber-reinforced epoxy composites (GFEC) and bamboo fiber-reinforced epoxy composites (BFEC). This study aims to assess the wear performance, defined by mass loss, of the composites under various factors: load, speed, time, nanoclay content, and composite type. Specimens are subjected to wear tests by a pin-on-disc tribometer. Composite wear performance is studied through Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) as predictive models. Experimental runs are planned based on the Box–Behnken design of RSM to present a regression model that is then checked with an ANOVA analysis; afterwards, training and testing are performed using an ANN model to improve predictive accuracy. As per the results, GFEC exhibits lower mass loss compared to BFEC. For both of the composites, the mass loss is drastically reduced by the addition of nanoclay. The addition of nanoclay has more pronounced effects on BFECs than on GFECs. ANN predictions are found to be better in agreement with the experimental values compared to those derived from the RSM model. Scanning Electron Microscopy (SEM) analysis provides insight into wear mechanisms. This study demonstrates the effectiveness of a statistical and machine learning approach in optimizing wear performance in composite materials.