Modeling the cooking process during the extraction of oil from Sunflower seeds using artificial neural networks on an industrial scale

The most common methods of extraction of oil from oils seeds are pressing and solvent methods that the most effective method of extracting sunflower oil, like other seeds with high oil content such as rapeseed, is mechanical press and then solvent extraction. In this research, to model the process of oil extraction from sunflower seeds on an industrial scale, three levels cooking temperature (70, 80 and 90 ° C) and three levels of moisture of the output seeds from the cooker (7, 7.5 and 8 %) was used and the amount of oil acidity, the contentt of oil, protein and moisture of meal and the percentage of insoluble fine partical in oil were studied. To predict the change's trend the artificial neural network in MATLAB R2013a software was used. By studying the various networks of back propagation feed forward network with topologies 2-10-5 with a correlation coefficient of more than 0.999 and the mean squared error of less than 0.003 and with using sigmoid hyperbolic of tangent activation function, the Levenberg–Marquardt learning algorithm and learning cycle of 1000 were specified as the best neural model. The results of the optimized and selected models were evaluated and these models with high correlation coefficients (over 0.96), were able to predict the change's trend.