Artificial neural network modeling-coupled genetic algorithm optimization of supercritical methanol transesterification of Aegle marmelos oil to biodiesel

This study aims to enhance the transesterification process to convert supercritical methanol (SCM) to biodiesel by applying artificial neural network-coupled genetic algorithm (ANN-GA) optimization. The process parameters selected for ANN-GA optimization were temperature of 230-350 °C, methanol-to-oil molar ratio of 24:1 to 48:1, and time of 5-25 min. The initial set of experiments were designed using the central composite design for ANN modeling. The best ANN topology with an optimal number of hidden neurons was obtained by heuristic analysis of coefficient of determination (R) and mean square error values. The R values (1, 0.99986, 0.9998 and 0.99997) for training, validation, testing, and overall prove the precision of the ANN model. Moreover, hypothesis tests were accomplished to enumerate model fitness. The GA-optimized process conditions for SCM transesterification using ANN as the fitness function were temperature of 325.5 °C, methanol-to-oil molar ratio of 41:1, and time of 23 min. Additionally, relative importance analysis was executed to ascertain the most influential variable on the SCM process. The major fuel properties of Aegle marmelos biodiesel were analyzed and found to comply well with American Society for Testing and Materials (ASTM) D6751 standards.