Experimental, Numerical, and ANN-Based Study of Free Vibration of Glass Fibre and MWCNT-Reinforced Laminated Structures

This study develops a predictive model for the natural frequencies of multi-walled carbon nanotube (MWCNT)-reinforced woven glass fibre-metal laminates (FMLs) using Artificial Neural Networks (ANNs). Laminates were fabricated via the open mould technique, and material properties were obtained through uniaxial tensile testing per the ISO 527–5 standard. Experimental modal analysis (EMA) and finite-element simulations in ABAQUS using S58R shell elements with engineering constants showed strong agreement. The ANN was trained on experimental and numerical data to predict the natural frequencies under varying laminate sequences, aspect ratios, and thickness ratios. The obtained results indicate that increasing the aspect ratio, side-to-thickness ratio, and fibre orientation angle reduces the natural frequencies of FML. The first mode of natural frequency increased by 83.80%, 84.85%, and 89.52% for CFCF, CCCF, and CCCC boundary conditions, respectively, compared to CFFF. Conversely, increasing fibre angles led to reductions in the natural frequency of 1.662–7.385% for CFFF, 0.027–1.146% for CFCF, 0.175–0.705% for CCCF, and 0.036–0.148% for CCCC. The ANN model demonstrated high accuracy and efficiency, supporting its use in the design and optimization of advanced composite structures.