The Application of Neural Networks in Balancing Production of Crude Sunflower Oil and Meal

The aim of the research is to predict specific output characteristics of half finished goods (crude sunflower oil and meal) on the basis of specific input variables (quality and composition of sunflower seeds), with the help of artificial neural networks. This is an attempt to predict the amount much more precisely than is the case with technological calculations commonly used in the oil industry. All input variables are representing the data received by the laboratory, and the output variables except category % of oil which is obtained by measuring the physical quantity of produced crude sunflower oil and sunflower consumed quantity of the processing quality. The correct prediction of the output variables contributes to better sales planning, production of sunflower oil, and better use of storage. Also, the correct prediction of technological results of the quality of crude oil and meal provides timely response and also preventing getting rancid and poor-quality oil, timely categorizing meal, which leads to proper planning and sales to the rational utilization of storage space, allows timely response technologists and prevents the growth of microorganisms in the meal.