Multiple linear regression analysis of comprehensive unit consumption in complex station of Shengli Oilfield

Comprehensive unit consumption in a complex station is affected by physical factors of and process conditions in the station. To achieve comparability of different stations' energy consumption and eliminate the effects caused by processes, the comprehensive unit consumption is analyzed as a greybox model. A multiple linear regression analysis is made on the mathematical relationship between comprehensive unit consumption and key influencing factors of a complex station by considering its actual operations, and then a unit consumption regression model is built. Taking the 36 complex stations in Shengli Oilfield as an example, the model is used in analyzing actual energy consumptions. The results show that the comprehensive unit consumption correlates positively with amount of liquid treated in the complex station and negatively with total water cut and crude oil density. In contrast to volume of liquid treated, total water cut and crude density affect comprehensive unit consumption by 88.4% and 50.9% respectively, that is, they affect comprehensive unit consumption in a descending order of volume of liquid treated, total water cut and crude density. Meanwhile, with measured data, the reasons for higher energy consumption due to such three factors are analyzed. Accordingly, this paper presents that, if possible technically, centralized treatment of feeding liquid and water diversion under ambient temperature (instead of water diversion by heating) are combined to save energy.