Comparative Analysis of the Application of Artificial Neural Network-Genetic Algorithm and Response Surface Methods-Desirability Function for Predicting the Optimal Conditions for Biodiesel Synthesis from Chrysophyllum Albidum Seed Oil

Efficient and sustainable commercial production of biodiesel requires process optimization and modeling using artificial intelligence techniques. Artificial neural network-genetic algorithm (ANN-GA) and response surface methodology (RSM) were used to predict the optimal process conditions for biodiesel synthesis from novel chrysophyllium albidum seed oil. Application of Levenberg-Marquardt algorithm and full factorial central composite design-desirability function analysis were carried out for ANN and RSM respectively. The biodiesel obtained at the optimum conditions was characterized using gas chromatography-mass spectrophotometer (GC-MS), Fourier transform infrared (FTIR), AOAC and ASTM D standard methods. The ANN and RSM correlation coefficients were 0.98 and 0.95 respectively. The best RSM fitted model was second order polynomial while the ANN best architecture topology consists of three layers: input layer with four input variables, hidden layer with ten hidden neurons and an output layer with single output variable. Optimization of the process shows that: RSM optimum conditions for biodiesel yield of 85.91wt% validated at 86.85wt% were 65.62 degrees C, 2.14wt% catalyst concentration, 62.04 min and 5.88 methanol: oil molar ratio with overall desirability of 1.00. ANN-GA gave 9.40% faster reaction time and required 15.88% less energy, 3.76% more alcohol and 16.82% less catalyst with 1.86wt% more biodiesel yield than RSM optimized conditions. Ascertained fuel properties, GC-MS and FTIR characterizations confirmed unsaturation and good cold-flow qualities of the biodiesel. Overall results presented ANN-GA as better than RSM and a reliable modeling and optimization technique for viable and sustainable production of biodiesel from chrysophyllum albidum seed. (c) 2021 Taiwan Institute of Chemical Engineers. Published by Elsevier B.V. All rights reserved.