The main goal of the work is to review and describe the algorithms currently being developed in order to obtain reliable well operation parameters using machine learning (ML) methods, existing constraints and results achieved. The work covers a number of ML applications. The methodology of bottomhole pressure modeling is described, which, in the absence or in case of failure of pressure sensors at an ESP inlet, allows more reliably, in comparison with a conventional approach, to determine the bottomhole pressure in wells and thereby improve the efficiency of selecting well candidates for well interventions. To assess well flow rates, another key indicator, an ESP instant flow ate simulator has been developed that allows obtaining information about the instantaneous production rate at any time, thus increasing the accuracy of cyclic wells measurements, and promptly implementing the necessary corrective measures. Approaches to modeling the impact of injection on the production well rates have also been considered.
The paper is devoted to the estimation of energy efficiency of wells equipped with ESP at different operating conditions in Western Siberia. As a measure of energy efficiency, the overall efficiency of the wells equipped with ESPs was used. The ESP efficiency can be defined as the ratio of the useful hydraulic power developed by ESP to electric power consumed by the ESP on the surface. Corporate production monitoring systems have evolved significantly in recent years in Gaspromneft [1]. Large amount of production data has been gathered and processed, including well operation and energy consumption data. It allows to analyze ESP's energy efficiency for each well. This analysis has been focused on comparison of different ESP operating modes – continuously operating mode (COM) – standard ESP working mode and periodic short-term activation (PSA) mode [2]. ESP in PSA mode starts and stops frequently – an average start stop period less that one hour. The analysis showed that at the current time the information gathered during the monitoring of wells is sufficient to conduct energy audit of the equipment at the level of ESP efficiency evaluation. It makes it possible to identify the dependencies between equipment operation parameters and evaluate the operation efficiency under various operating conditions. Energy efficiency estimation can be used to make decisions for the well stock efficiency improvement. For example, in the course of the analysis, a number of results were revealed: the efficiency in the operation of a well in a periodic short-term activation mode (PSA) in comparison with a continuously operating mode (COM) on a low rate wells is higher by an average of 20%; With the increase in water cut, the EPS's energy efficiency is increasing, both in terms of PSA and COM; The average EPS's energy efficiency decreases with increasing rotating frequency; with an increase in run life, the efficiency of the ESP system is smoothly reduced by an average of 5% in three years, after a gain of 1,300 days, the rate of decrease in efficiency is significantly increased; on COM wells with a motor load of less than 40% efficiency is dramatically reduced, so it is recommended to transfer this fund to the PSA; there is a potential for increasing efficiency and oil production at a low rate wells by using ESP with larger nominal rate in PSA mode and reduced the ratio of operation and accumulation periods. Finally, the efficiency analysis can be used to rank ESP manufacturer by energy efficiency in specific conditions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.