Enhancing mechanical and thermal properties of isophthalic polyester resin composites reinforced with graphene oxide and nanosilica using RSM and ANN

Abstract

This study examines the mechanical and thermal characteristics of isophthalic polyester (IP) resin composites reinforced with graphene oxide (GO), nanosilica (NS), and their hybrid combinations. Composites with different filler concentrations of 0.05, 0.1, 0.3, and 0.5 wt percentages were assessed by tensile, flexural, impact strength, and flammability tests. Structural properties were examined via X-ray diffraction (XRD). The findings indicate that incorporating GO and NS improves the mechanical properties of IP resin composites, with the hybrid composite at 0.3 wt% attaining peak performance. The hybrid composite at 0.3 wt% demonstrated a 59.47 % enhancement in tensile strength and an 82.16 % augmentation in flexural strength relative to pure IP resin. Moreover, the 0.3 wt% hybrid composites exhibited enhanced fire resistance, signifying a significant decrease in flammability. XRD analysis validated the effective integration of GO and NS into the IP resin matrix. Mechanical properties were predicted using two computational approaches: artificial neural networks (ANN) and response surface methodology (RSM). The RSM model precisely predicted tensile strength (R² > 0.9736) and flexural strength (R² ≥ 0.9736). The ANN model demonstrated remarkable accuracy, with correlation coefficients above (R > 0.890) for tensile strength and (R > 0.999) for flexural strength in training, testing, and validation, highlighting its effectiveness in capturing data variability. The comparison of the models found that the ANN model exceeded the RSM in predictive accuracy, as demonstrated by a robust correlation between experimental and anticipated values. The exceptional mechanical properties and fire resistance of hybrid IP resin composites make them suitable for high-performance structural applications in the automotive, construction, and aerospace industries.