Prediction Model of Unit Consumption for Oilfield Water Injection Based on the Grain of Association Rule and Chaotic Time Series

Abstract

The water injection system, which is prone to failure and high energy consumptions, consists of many different parts, perform complex system behaviors and has a high degree of correlation with the environment and strong coupling with each other. In the view of these characteristics, in this paper, based on the big data including temperature, pressure, flow rate, working system of the equipment and energy consumption datas accumulated during water-flooding productive process, a granularity model of water flooding dynamic data is set up to predict energy consumption with multiple targets like water injection system efficiency, energy consumption per unit amount of liquid and multi-variable time sequence. A chaotic energy consumption forecasting model with multi-variable time sequence is established according to different time granularity (day, month and year, etc different space granularity (well group, block and field, etc.) and different ways of water injection pattern (general water injection and separate zone water injection, etc Take a certain water injection plant as an example, the grain of association rule algorithm is constructed to study the relationship between the energy consumption factors of water injection system, and to predict energy consumption change affected by water injection production parameters, yet early warnings of water injection system efficiency and energy consumption are gained by predicting the production parameters.