Experimental Evaluation and RBFNN Prediction of Mechanical Properties in Madar Fiber–Groundnut Shell Powder Reinforced Epoxy Composites

Abstract

This study investigates the mechanical response of hybrid composites fabricated from Madar fiber, groundnut shell powder, and epoxy resin. The reinforcement and particulate filler were incorporated into the epoxy matrix to develop composites with improved load-bearing characteristics. Mechanical performance was examined through tensile, flexural, impact, and interlaminar shear strength tests, while scanning electron microscopy (SEM) were used to assess fracture morphology and interfacial characteristics. A radial basis function neural network (RBFNN) were additionally developed to estimate tensile and flexural strengths from the material composition. The network employed radial basis neurons to capture nonlinear relationships between the input variables and mechanical responses. Among the investigated compositions, MGE-2 containing 20 wt.% Madar fiber, 15 wt.% groundnut shell powder, and 65 wt.% epoxy resin exhibited the most favorable performance. MGE-2 attained tensile strength of 61.78 MPa and tensile modulus of 3.26 GPa, together with a flexural strength of 94.36 MPa and flexural modulus of 4.78 GPa. Corresponding impact strength and interlaminar shear strength reached 20.68 kJ/m² and 21.37 MPa, respectively. The RBFNN provided reliable predictions, yielding test R² values of 0.9204 and 0.9447 for tensile and flexural strengths, respectively.