Experimental and Numerical Investigation on The Elastic Properties of Natural Fiber Composites

In recent years, the application of natural fibers as reinforcement in composite structures has received increasing attention due to their advantages of low cost, environmental friendliness and favorable biocompatibility over synthetic fiber composite materials. The present work is an investigation on the tensile properties of palm as well as luffa natural fiber composites (NFC) in high density polyethylene (HDPE), polypropylene (PP), epoxy, and ecopoxy (BioPoxy 36) matrices, taking into consideration the effect of fibers volume fraction (Vf) variation. Finite element analysis i.e. representative volume element (RVE) models with unidirectional and chopped random fiber orientations, as well as analytical simulation i.e. Rule of Mixture (ROM), Halpin-Tsai, Chamis, and Nielsen approaches were utilized for predicting the elastic properties. Tensile test following ASTM D3039 standard was conducted. Artificial Neural Network (ANN), Multiple Linear Regression (MLR), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Support Vector Machine were implemented for defining the design space upon the considered parameters and evaluating the reliability of these machine learning approaches in predicting the tensile strength of natural fibers composites. Furthermore, biopoxy 36 with 0.3 luffa fibers exhibited the highest tensile strength. Finite element analysisfindings profusely agreed with the experimental results. ANFIS machine learning (ML) tool showed least prediction error in predicting tensile strength of natural fibers composites.