Optimizing Wear Behavior of Epoxy Composites Using Response Surface Methodology and Artificial Neural Networks

Growing industrial demand in present competitive scenario has raised the need for new materials with high strength to weight ratio to replace the conventional materials. This work discusses the multi-objective optimization of wear performance of plain epoxy, epoxy/E-glass fiber and epoxy/E-glass fiber/carbon particles composites prepared by simple hand lay-up process with clamp loading. Wear test carried in pin on disc setup, with design of experiments based on response surface methodology with speed, load and sliding distance as numerical factors at three levels and sample as categorical. Optimization carried, aiming towards reduction of performance measures like Weight loss (W-L) and Coefficient of friction (mu) followed by development of an artificial neural model for predicting the wear performances. The experimental and the artificial neural network (ANN) predicted values are found very close and thus the ANN model developed can predict the W-L and coefficient of friction both within and across the design space. POLYM. COMPOS., 40:2812-2818, 2019. (c) 2018 Society of Plastics Engineers