Response Surface Methodology (RSM)- Artificial Neural Networks (ANN) Aided Prediction of The Impact Strength of Natural Fibre/Carbon Nanotubes Hybrid Reinforced Polymer Nanocomposite

Abstract

Numerous engineering issues have been addressed using Artificial Intelligence (AI) approaches, such as Machine Learning and Artificial Neural Networks (ANN). In this study, Plantain fiber (PF) and multiwall carbon nanotube (MWCNT) were used to create a PF/MWCNT reinforced hybrid nanocomposite (PFRHN) using epoxy resin as the base polymer. To advance adhesion and contact between the fiber/matrix, an oxidizing agent solution of potassium permanganate in acetone (KMnO4-Acetone) was applied to modify the fiber surface. To predict and maximize the impact strength of the hybridized nanocomposite, Response Surface Methodology (RSM) via Box-Behnken Design (BBD) and hyper-parameter optimization in a single-layer-perceptron ANN, with a range of 1-8 neurons in the hidden layer was utilized. The model predicted an impact strength of 44.54 KJ/m2. To verify the viability of the statistical experimental analysis, impact strength testing was carried out for pristine and hybridized composites under optimal conditions. Results showed an average impact strength of 24.32 KJ/m2 and 45.21 KJ/m2 for pristine and hybridized nanocomposite. PFRHN impact strength was significantly increased over pristine epoxy composite. The RSM-ANN technique has been shown in this study to be an effective way of reaching ideal mechanical property values in the shortest time, lowering the costs of production and conserving resources.