Optimization of hempseed oil extraction by n-hexane

• Solvent extraction of hempseed oil under different extraction conditions.
• Evaluating the influence of process variables on the hempseed oil yield.
• Prediction of the hempseed oil yield by second-order polynomial and ANN models.
• Optimization of hempseed oil yield using RSM and ANN-GA models.
• ANN-GA model is more effective than RSM.

The influence of extraction temperature, solvent:seed ratio and extraction time on hempseed oil yield was investigated. The oil extraction was optimized using the response surface methodology (RSM) coupled with a 33 full factorial experiment with two replications and an artificial neural network (ANN) model. For the former method, the second-order polynomial equation and the analysis of variance were used for evaluating the significance of the influence of the extraction variables on hempseed oil yield and for optimizing the extraction process. The ANN model was combined with a genetic algorithm (GA) for the process optimization. The low mean relative percent deviation (MRPD) between the experimental data and predicted oil yield obtained by using RSM and ANN models showed that both models were suitable for modeling the extraction process. The ANN model was more accurate because of lower MRPD value (±1.0%) than the second-order polynomial equation (±2.3%). The established optimal extraction conditions for achieving the maximum hempseed oil yield based on the ANN-GA methodology were extraction temperature 70 °C, solvent:seed ratio of 10:1 and 10 min extraction time, under which the predicted and actual hempseed oil yields were 29.56 g/100 g and 29.06 ± 0.11 g/100 g, respectively. Unsaturated linoleic, linolenic and oleic fatty acids were the main fractions of the fatty acids in the hempseed oil, although the saturated fatty acids were also present.