Document Type : Original Article
Authors
1 University of Benin
2 University of Benin, Nigeria
Abstract
High production costs, environmental concerns, and optimization complexities limit conventional biodiesel viability. To address this, a novel bio-based heterogeneous catalyst was synthesized by impregnating Coconut Shell BioChar (CS-BC) with copper-doped Calcined Periwinkle Shell (CPS). The morphology and stability of the catalyst was assessed using SEM, EDX, FTIR, XRD, and BET. The catalyst was applied to transesterify an optimized blend of Waste Cooking Oil (WCO) and Animal Fat (AF). The process was modeled using Support Vector Regression (SVR), Artificial Neural Networks (ANN), Random Forest (RF), and Extreme Gradient Boosting (XGBoost), then globally optimized via the Manta Ray Foraging Optimization (MRFO) algorithm. XGBoost outperformed the other models, achieving a coefficient of determination (R2) of 0.899 with the lowest error metrics. SHAP analysis identified reaction temperature as the most influential feature. The MRFO-optimized conditions were a reaction temperature of 66.3°C, catalyst loading of 6.37 wt.%, reaction time of 104.6 minutes, and a methanol-to-oil ratio of 8.84:1. Under these optimized conditions, a maximum predicted biodiesel yield of 81.9% was achieved, which closely matched the experimentally validated yield of 82.3 ± 0.8%. The synthesized catalyst demonstrated excellent stability, maintaining a 97.1% biodiesel yield after five consecutive reaction cycles, highlighting its potential for sustainable, large-scale application. Physicochemical analysis confirmed that the produced biodiesel complies with ASTM D6751 and EN 14214 standards. This integrated framework of waste-derived feedstocks, green catalysis, and metaheuristic optimization provides a sustainable, data-driven approach to biofuel production.
Keywords
Main Subjects
)