Integrating machine learning and response surface methodology for analyzing anisotropic mechanical properties of biocomposites

Abstract

ABSTRACTThis study enhances the anisotropic mechanical properties of banana fiber-epoxy composites by optimizing fiber loading, orientation, and treatment using Response Surface Methodology (RSM) and Artificial Neural Network (ANN). RSM suggests optimal values: fiber loading at 33 wt%, NaOH treatment at 6.8 wt%, and fiber orientation at 15 degrees. This material has exceptional mechanical characteristics, including a maximum tensile strength (TLS) of 31.72 MPa, a maximum flexural strength (FLS) of 42.86 MPa, and a maximum impact strength (IPS) of 38.56 kJm-2. ANN effectively predicts strengths with high R2 scores of 0.969, 0.984, and 0.954 for tensile, flexural, and impact strengths. Incorporating batch normalization and dropout layers enhances robustness. The study concludes that NaOH treatment and fiber orientation significantly impact the composite's anisotropy.KEYWORDS: ANNbiocompositesfiber orientationalkali treatmentanisotropic behaviormechanical propertiesresponse surface methodology (RSM) AcknowledgementsThe authors would like to acknowledge the scheme of Innovation, Technology Development, and Deployment (1819) of the Department of Science and Technology (DST) - Delhi.Disclosure statementNo potential conflict of interest was reported by the author(s).Author contributionAuthor 1: Corresponding AuthorAuthor 2: Research GuideAuthor 3: Machine Learning Prediction model developed.