Manufacturing Parameters Optimization and Modeling of PxGyEz Hybrid Composite for High Flexural Strength Using Taguchi Robust Experimental Design Technique and General Regression Analysis

Abstract

With the continuous need for sustainable, environmentally friendly, and low-cost systems, processes, and materials, natural fibers have been a major topic of discussion in the materials science community as it has increasingly found acceptance in material development as an alternative to synthetic fibers due to environmental concerns. Although many studies have been carried out in this regard, the optimal flexural performance of pineapple leaf fiber/synthetic fiber hybrid reinforced composite has not been studied even with its promising application in aviation, health, and fitness, marine, etc. In this study, a Pineapple Leaf fiber (PALF)/Glass fiber Epoxy hybrid composite PxGyEz (with x, y, and z representing the volume fraction of pineapple leaf fiber, the volume fraction of glass fiber, and fiber length respectively) was developed and its flexural properties optimized and modeled with regards to the variable manufacturing parameters of x, y, and z respectively. For the quality characteristics (flexural strength) investigated, the Minitab®19 software was used to analyze the Taguchi robust experiment design technique on a higher the better basis. The optimum combination of the control factors was found at x = 20%, y = 20%, and z = 25mm. The optimized composite P20G20E25 possessed a flexural strength of 144.4994MPa which was only a 7.17% deviation from the predicted optimum flexural strength. Analysis of variance showed that the PALF had the highest contribution of 23.97% to the flexural strength of the PxGyEz, glass fiber 7.13%, and fiber length 12.79%. SEM Images of the PALF, glass fiber, and the fractured surface of the optimized material P20G20E25 revealed the surface structure which explained their different contribution to the flexural strength of the materials. An equation for the prediction of the flexural properties of PxGyEz was derived from the regression model and it had an approximately 77.57% agreement with experimentation.