Machine Learning Approach for Mechanical Property Prediction of a Bio-Epoxy and Glass Fiber Composite Reinforced with Titanium Dioxide Nanoparticles

Glass fiber reinforced polymers (GFRPs) have drawn significant attention given their lightweight, mechanical resistance and tunable properties through constituent selection. Due to environmental concerns, research efforts have focused on incorporating sustainable materials, such as bio-epoxy resins, to reduce the ecological impact of GFRPs. This study characterizes a GFRP containing a bio-epoxy resin matrix, various loadings of titanium dioxide (TiO2) nanoparticles, and a stabilized arrangement of glass fiber. The unreinforced composite exhibited a tensile strength and modulus of 214 MPa, and 13 GPa, respectively, and a flexural strength and modulus of 375 MPa and 14.5 GPa, respectively. The addition of TiO2 produced an improvement in mechanical response for all the composites. The formulation with 1 wt.% TiO2 showed the best tensile response with an improvement of 13% and 14% for its tensile strength, and modulus, respectively; meanwhile, the composites with 2 wt.% TiO2 attained an improvement of 19% and 40% for the flexural strength and modulus, respectively. Scanning electron microscopy (SEM) revealed significant changes in the fracture mechanism of the composites, while energy-dispersive spectroscopy (EDS) confirmed an even nanoparticle distribution. Additionally, machine learning (ML) models were developed to predict the mechanical response as a function of the TiO2 content.