Modeling the chemical properties of sesame oil under the influence of pulsed electric field using the artificial neural networks

Abstract

In this study, PEF pretreatment was used to improve the efficiency of screw press method on the properties of extracted sesame seeds oil. Sesame seeds were treated at different PEF intensities (250, 3250 and 6250 Vcm-1) and pulse numbers (10, 30 and 50). Then, the oil was extracted using a screw press at 33 rpm. Some physicochemical properties of the obtained oil including oil extraction efficiency, acidity index, determination of total phenolic compounds and activity of the inhibition of the DPPH free radical were evaluated. The results showed that the oil extraction efficiency initially increased at first but it showed reduction during PEF pretreatment at higher intensities. Increase in the applied PEF intensity and pulse number lead to an increase in the acidity and total phenolic compounds. While the oxidative stability of the oil reduced at the more intensive PEF conditions. However, the antioxidant activity was firstly increased and then decreased during PEF pretreatment. In addition, artificial neural network model was used to predict the effect of different PEF pretreatment conditions on the physicochemical properties of the extracted oil. The best model was the feed forward neural network with sigmoid hyperbolic tangent conduction function, Levenberg – Marquardt training function with 5-6-2 topology.