Maximizing waste cooking oil biodiesel production employing novel Brotia costula derived catalyst through statistical and machine learning optimization techniques

The limitations of homogeneous catalysts associated with biodiesel production such as non-reusability, emulsification, and saponification have led to the increased cost of biodiesel synthesis. To overcome these limitations, the current work aims to investigate the applicability and effectiveness of green catalyst made from Brotia costula shells (BCS) for waste cooking oil biodiesel production. The catalyst was synthesized by calcinating BCS at 800 degrees C-1000 degrees C for 4 h. Catalyst characterization showed that BC900 (BCS calcined at 900 degrees C) contained 98.15 CaO wt% and exhibited the smallest grain size of 39.23 nm with the highest specific surface area of 14.28 m(2)/g and highest catalytic activity. Optimization of biodiesel synthesis was carried out using response surface methodology - Box-Behnken design (RSM-BBD), artificial neural network (ANN) coupled with genetic algorithm (GA), and adaptive neuro-fuzzy inference system (ANFIS) coupled with GA. Maximum experimental biodiesel yield of 97.78 +/- 0.59 % at optimum conditions of methanol: WCO molar ratio 9.73: 1, reaction time 2.12 h, catalyst dose wt% 7.59 and reaction temperature 69.3 degrees C was obtained with ANFIS-GA optimization model. The prediction accuracy of the ANFIS model (R2 = 0.988) was better than those of the RSM-BBD (R-2 = 0.982) and ANN (R-2 = 0.955) models. The synthesized catalyst exhibited high catalytic activity up to four cycles of reusability (yield > 88 %). It is environment-friendly, inexpensive, renewable, and performs on par with catalysts synthesized from other mollusk shells. Using this catalyst for biodiesel production will contribute to the biorefinery's ability to explore sustainable raw materials and may lower overall biodiesel cost.