Influence of fibre direction on the mechanical properties and artificial intelligence-based performance prediction of hybrid abaca-jute amino composites

Abstract

In this study, the Hot-press method was used to create hybrid laminates using an equal amount of jute and abaca fibers with various fiber orientations [0/0/0/0/0/0, 0/45/0/0/45/0, 0/90/0/0/90/0, and +45/−45/+45/+45/−45/+45] embedded in amino resin. To assess the impact of fiber orientation on natural fiber hybrid eco-fiber composites, as per ASTM standard the tensile, flexural, impact, and ILSS tests were carried out. Additionally, conducting an in-depth statistical examination through the utilization of Machine Learning techniques like Artificial Neural Networks (ANN) to make predictions. The experimental results revealed that the voids and defects have a significant impact on composite density. (0/45/0/0/45/0)The fiber-oriented composite (Composite B) had a higher experimental density and void content than the other composites. The Composite A (0/0/0/0/0/0) outperformed other composites in terms of tensile strength, flexural load-bearing capacity, and impact strength by 5% to 29%. When compared to other composites, the Composite A fiber-oriented composite had the highest ILSS value (3.35 MPa). SEM analysis revealed a good interface between abaca-jute ecofibers and amino resin in the composite. ANN effectively predicts strengths with high R2 scores of 1, 0.98, and 0.99 for tensile, flexural, and impact strengths. The model’s interaction coefficients are close to 1, indicating strong correlative and predictive value.

Keywords

• abaca fiber
• phenoplast composites
• jute fiber
• artificial intelligence