Water-absorption behavior of hybrid biocomposites based on Syagrus romanzoffiana Palm waste and short Sisal \fibers : application of ANN, RSM optimization, and Fick’s Diffusion Model

Abstract

This research advances sustainable production methods by focusing on the reuse of agricultural waste. The absorption behavior of hybrid biocomposites reinforced with biochar derived from Syagrus romanzoffiana palm waste (SrPW) and Agave sisalana fiber (AsF) in an epoxy matrix was examined. The primary objective was to assess how varying AsF content (10, 15, 20, and 30%) affects the dynamic behavior of hybrid biocomposites incorporating a fixed 2% of biochar. Response surface methodology (RSM) and artificial neural networks (ANN), optimized using a genetic algorithm, were employed to evaluate and simulate absorption characteristics. The ANN approach demonstrated superior accuracy and robustness compared to RSM, as indicated by high correlation coefficients (0.9999 in training, 0.9702 in testing, and 0.9970 in validation) between model predictions and experimental results. This predictive method minimizes the need for extensive experimental trials, thereby saving time and resources. The resulting green composite, based on epoxy, palm waste, and sisal, presents an environmentally friendly alternative for potential sustainable applications.