Abstract

Glass fiber reinforced polyester (GFRP) composites are widely used in applications that require high durability and excellent corrosion resistance. In Vietnam, lightweight GFRP materials are primarily used in the production of boat hulls and bows, automobile and motorcycle chassis, as well as various industrial and household devices due to their high corrosion resistance. Accurately predicting the wear rate of GFRP composites is of significant practical importance as it provides important guidance for the experimental design and production of these materials. In this study, artificial intelligence (AI) techniques, specifically artificial neural networks (ANN) and fuzzy inference systems (FIS), were used to predict the wear rate of GFRP composites. Both methods are well-suited to modeling nonlinear systems. Experimental results from trained ANN and adaptive neuro-fuzzy inference system (ANFIS) models were used to determine the manufacturing parameters that yielded the lowest wear rate for GFRP composites. Specifically, a wear rate of 18.241 × 10^-6 mm³/Nm was obtained using a composition of 30 wt.% CaCO₃, 8 wt.% P, 11 wt.% GF, 5 wt.% GB, and 1 wt.% Al, with a compressive load of 10 N and a rotation speed of 200 rpm.