Modeling and Optimization of Tensile Properties of Epoxy Biocomposites Reinforced with Washingtonia robusta Waste and Biochar Using Response Surface Methodology, Artificial Neural Networks, and Multi-Criteria Decision-Making

Abstract

The current study examined the tensile properties of epoxy biocomposites reinforced with untreated and NaOH-treated Washingtonia robusta waste (WRW) and biochar considering different fiber weight fractions (10%, 20%, 30%), NaOH concentrations (2%, 2.5%, 3%), and treatment durations (4, 12, 24 h). The potential of WRW/biochar composites for sustainable applications, particularly in the automotive sector, was highlighted. The maximum tensile strength (35.69 MPa) and Young�s modulus (7.67 GPa) were achieved at 30% WRW treated for 4 h with 3% NaOH. These improvements are attributed to better interfacial bonding and fiber-matrix adhesion. To model and optimize the mechanical behavior, Response Surface Methodology (RSM), Artificial Neural Networks (ANN), and a Multi-Criteria Decision-Making (MCDM) method based on TOPSIS were applied. ANN provided higher predictive accuracy (R2?=?0.9993 for tensile strength, 0.9819 for Young�s modulus) compared to RSM. Optimization results indicated ideal conditions of 29.35-29.41% WRW, 11.06-11.24?h treatment time, and 2.99-3% NaOH, based on desirability function RSM and genetic algorithm ANN optimization. The integration of ANN, RSM, and TOPSIS-MCDM provided a comprehensive optimization framework, confirming the potential of WRW/biochar composites for eco-efficient engineering applications, such as in the automotive sector.