Mechanical behaviour and microscopic analysis of epoxy and E-glass reinforced banyan fibre composites with the application of artificial neural network and deep neural network for the automatic prediction of orientation

This paper deals with the testing of tensile and flexural behaviour of epoxy-reinforced natural fibre composites, for which Banyan fibres have been selected as the natural fibre. Variations are made in the orientation of the fibres to determine which orientation made the composite the strongest. The fibre strands are arranged in different orientations, such as the uniaxial, biaxial and criss-cross arrangements, to differentiate between the orientations and observe which arrangement exhibited the best mechanical behaviour. The fibres are initially washed with 0.5% weight/volume (w/v) NaOH solution, following which specimens of the composites are made using wooden moulds designed according to ASTM standards. Biaxial layers of E-glass are incorporated into the matrix in an attempt to enhance the mechanical properties of the specimen. The variances observed in the Young's modulus are analysed to understand the factors that majorly impacted it. For a better understanding of the results, the chemical functional groups and the microstructure of the samples are analysed with the aid of Fourier-Transform Infrared Spectroscopy (FTIR), Field Emission Scanning Electron Microscopy (FESEM) and X-Ray powder Diffraction (XRD). Additionally, predictive models are simulated using Artificial and Deep Neural Networks to recognise patterns in the data, by varying specific parameters. The results obtained indicated that Banyan fibre composites can replace conventionally-used materials and serve real-world purposes better.