Modeling the Extraction Yield of Anthole from Fennel Essential Oil through Response Surface Methodology and Artificial Neural Network

In this study, experimental findings related to the efficiency of a solid-liquid extraction were modeled through Response Surface Methodology (RSM) and Artificial Neural Network (ANN). Anethole is the main active pharmaceutical ingredient in the essential oil of fennel seeds, and it was extracted in contact with 70% ethanol as a solvent in a new modified Rotating Disc Contactor (RDC) column. The extraction yield of Anethole was considered as a response factor on which the impact of three variables was investigated, including fennel particle size, rotor speed, and solvent-to-solid ratio. The experiments were conducted with finely chopped fennel particles in sizes of 1.7, 1, and 0.3 mm, the rotor speed of 90, 135, and 180 rpm, and solvent-to-solid ratios of 10, 15, and 20. The obtained yields were modeled through RSM and were also simulated with the ANN method. The result of GC-MS and GC analyzers as well as simulation findings showed that by decreasing the size of fennel seeds, increasing the solvent-to-solid ratio, and rotor speed, the yield of the extraction process was enhanced. In addition, the correlation coefficient for the RSM and ANN were 0.9604 and 0.9955, respectively; which proves a high accuracy of ANN modeling in comparison with RSM.