Tensile performance of additively manufactured short carbon fibre-PLA composites: Neural networking and GA for prediction and optimisation

Abstract

Despite the vast applications of fibre-reinforced polymer composites in multiple domains, conventional fabrication is laden with many difficulties, thus bringing focus on Additive manufacturing technologies. The aim of this study is to evaluate the effect of infill percentage, layer thickness and carbon fibre layer position on the mechanical tensile performance of Fused Deposition Modelling (FDM) printed short Carbon Fibre-reinforced Polylactic Acid composites. Experimentation is based on full factorial design, with elastic modulus and maximum tensile stress as the response variables. Two distinct models for output are developed using an multivariate regression analysis and Generalised Regression Neural Network and are further validated through three set of randomised experiments. Genetic Algorithm is used to optimise the outputs, the results of which agree well with experiments. Changes in the carbon fibre layer position in the fabricated composites are found to have a visible and significant effect on the mechanical performance of the composites.