Development of a comprehensive methodology for the forecast of effectiveness of geological and technical measures based on machine learning algorithms

Abstract

The main part of hydrocarbon production in Russia is represented by old oil and gas producing regions. Such areas are characterized by a significant decrease in well productivity due to high water cut and faster production of the most productive facilities. An important role for such deposits is played by stabilization of production and increase of mobile reserves by improving the development system. This is facilitated by various geological and technical measures. Today, an urgent problem is to increase the reliability of the forecast of technological and economic efficiency when planning various geological and technical measures. This is due to the difficulty in selecting candidate wells under the conditions of the old stock, the large volume of planned activities, the reduction in the profitability of measures, the lack of a comprehensive methodology for assessing the potential of wells for the short and long term. Currently, there are several methods to evaluate the effectiveness of geological and technical measures: forecast based on geological and field analysis, statistical forecast, machine learning, hydrodynamic modeling. However, each of them has its own shortcomings and assumptions. The authors propose a methodology for predicting the effectiveness of geological and technical measures, which allows one to combine the main methods at different stages of evaluating the effectiveness and to predict the increase in fluid and oil production rates, additional production, changes in the dynamics of reservoir pressure and the rate of watering of well production.