DECISION MAKING MODEL FOR PETROLEUM PRODUCT DENSITY DETERMINATION UNDER CONDITIONS OF UNCERTAINTY

Abstract

Background When storing petroleum products in the reservoir tank, the following parameters change: 1. temperature - due to diurnal fluctuations in ambient temperature or oil heating during the cold period; 2. density - due to the light fractions evaporation; 3. level - due to evaporation, changes in temperature and density; 4. pressure - with increasing pressure in the gas space. The oil product density is determined by oil meter (aerometer). To this purpose, a petroleum product sample is taken from the container by a selective bucket, an oil meter is immersed in it and the density is determined on the upper scale and the temperature is determined on the lower scale. The oil product density can also be determined by calculation using the oil product temperature in the tank. The sought-for oil product density at a given temperature is determined through the oil product density at 20 0C and the temperature correction of the change in the oil product density with a temperature change of 1 0 C . This does not take into account the parameters of excessive pressure and the oil product level height. To create modern methods for determining the oil products density, mathematical models describing various processes in tanks must be developed and specified. At the same time, information accumulation process takes place; empirical and expert models should be supplemented and refined with the new experimental data accumulation obtained under model and experimental conditions. In this connection, the development of a model for determining the oil products density based on modern methods of artificial intelligence is an actual scientific and technical task. Aims and Objectives Creation of a base of fuzzy production rules for determining the oil products density in tanks. The choice of the characteristic function for the input linguistic variables fuzzification and the determination of the degree of veracity of all input conditions. Activation and accumulation of output linguistic variables, determination of the membership function for conclusions and the truth degree of the output linguistic variables. Determination, as a defuzzification result, of the oil product density in uncertainty conditions. Results A rules database was created for determining the oil products density. The fuzzification procedure was implemented under all the input linguistic variables conditions. The veracity conditions degree in the fuzzy products rules was determined. The activation procedure was carried out and all the values of the veracity degrees under the conclusions for each rule were found. As a defuzzification result, the quantitative output linguistic variables value was determined.