Performance Evaluation of Fiber Waste-Reinforced Polymer Using Backpropagation Neural Network Model and Finite Element Analysis

Sustainable and renewable polymer composites are a remarkable feat, hinged on being eco-conscious by nature. This study aims to enhance global support in pursuing the potential of utilizing bio-composites in strengthening concrete structures. Beam-column joint, the most critical element in a building, is considered for studying concrete performance by incorporating fiber wastes for strengthening applications using fiber-reinforced polymer. This paper focuses on utilizing natural fiber wastes from abaca and coconut to manufacture polymer composites with epoxy resin, using a hand lay-up technique, designed to strengthen unreinforced beam-column joints. The fabricated polymers have undergone experimental investigation to obtain their strength properties for concrete retrofitting. Researchers conducted finite element analysis for 25 exterior joint specimens, applying different polymer configurations theoretically. Artificial neural network was utilized to create predictive strength models using the dataset obtained from numerical simulation, which included minimum, maximum, and average values. According to the analysis, the wrapping arrangement and thickness greatly influenced the concrete behavior when subjected to cyclic loading. Experimental data and numerical investigation with neural network analysis validate the hypothetical assumption to make fiber wastes an alternative to conventional material in developing fiber-reinforced polymers for structural applications.