Advanced Analytics for Data-Driven Decision Making in Electrical Submersible Pump Operations Management

Abstract

Abstract Increasing demand for cost-cutting measures has seen the petroleum industry taking advantage of new technology to accelerate production and increase ultimate recovery while maximizing reliability of the production system. Electrical submersible pump (ESP) technologies for in-well and subsea boosting systems have evolved to become a critical component in many production operations. However, ESPs are complex dynamical systems whose performance can be degraded by certain faults or events such as gas locking, changes in fluid characteristics, plugged pump, tubing leak and closed valve. With the explosive growth of sensor data, it is no wonder that knowledge discovery has grown in importance in facilitating descriptive analyses (clustering) and predictive analyses (regression and classification) applications to ESP operations management. Interestingly, the development of appropriate algorithms to process ESP signals, detect events and prescribe optimal decisions given ESP real-time data and auxiliary, predictive observations have until now largely been overlooked. This paper presents a decision support system for continuous ESP event detection and providing prescriptive analytics tool to analyze and manage ESP operation performance. The system includes a database management system, a real-time data management architecture, a real-time signals processing engine, forecasting system to forecast instance based on real-time data, outlier detection model, a sequence analyzer for event recognition, a prescriptive analytics to provide recommendations concerning the optimal alternative actions to be executed during events and a graphical user interface to visually analyze the flexibility and prescription model instance. The proposed procedure has been tested in a field to verify the functionalities of the system. The results show that the proposed methodology can be efficiently used in a wide range of electrical submersible pump system operation performance management and surveillance. The results demonstrate the importance of the proposed methodology and adapting it to ESP well production operations.