The optimization with artificial neural network and response surface methodology models for extracting oil from C. albidum seed for biodiesel production

Mabel Keke, Ozioko Fabian Chidiebere, Rockson-Itiveh David Emoefe and T. F. Adepoju

Abstract

This study aimed to optimize the process of oil extraction from Chrysophyllum albidum (C. albidum) seeds. The research involved quality assessment, seed preparation, and Soxhlet extraction apparatus with n-hexane as a solvent. The utilization of the Box-Behnken design involved implementing a series of three-level factors, leading to a total of 17 experimental runs aimed at attaining optimal oil extraction. The optimal conditions for oil extraction were identified as a 50g sample, 250 mL solvent, and 40 minutes, resulting in a 3.0896% (w/w) yield, the lowest oil yield was 1.5931%(w/w) deviating from projections by response surface methodology (RSM) and artificial neural network (ANN). The oil exhibited a reddish-brown color and various physiochemical properties. The current study did not consider alternative optimization methods such as particle swarm optimization and genetic algorithms when assessing optimal sites. Future research could explore these specific areas. After the optimization methods were validated, the investigation reached the determination that the oil is unsuitable for consumption and possesses significant value within the manufacturing sector.