Smart Alarming for Intelligent Surveillance of Electrical Submersible Pump Systems

Abstract

Abstract Electrical submersible pump (ESP) technology has evolved to become a critical component in many production operations and well productivity enhancement. However, one of the important challenges in real-time ESP-enabled well management is the implementation of intelligent systems that can assist human operators in making control decisions. Modern technological advances have resulted in increasingly complicated processes that present considerable challenges in performance analysis and well management for successful operation of electrical submersible pumps. Given the size, scope, and complexity of modern engineered electrical submersible pump systems, it is becoming significantly more difficult for engineers to anticipate, diagnose and control serious abnormal events in a timely manner. Failure of the operator to exercise the appropriate mitigation actions often has an adverse effect on the process safety quality, run life, surface hardware and downhole equipment. Hence, there exist considerable incentives to develop intelligent alarm systems for automating electrical submersible pump system parameters estimation and optimization. The difficulties associated with implementing intelligent alarms and the opportunities for improvements are even greater in the advanced wells equipped with electrical submersible pumps due to complex flow and transport process challenges. In this paper, a state estimator is implemented in the form of data assimilation algorithm using a variety of data-driven models and continued ESP operations performance properties measurements. The data assimilation3 estimates continuously the state variables of the data-driven ESP models to provide a feedback to an online intelligent alarm monitoring system. The intelligent alarm surveillance system workflow combines streaming surface controller data, well head data and sensor data to intelligently define thresholds and determine when a given measurement is out of range and human intervention is needed. Further multi-signal data analysis is employed to characterize given events and perform dynamical optimization (based on define objective functions and operations constraints) for recommending real-time controller set point updating and corrective actions during real time ESP operations. Such optimization framework has the potential to improve production while simultaneously providing cost savings by reducing remote human intervention and the deployment of personnel to field locations. The web-based alarm surveillance system has been successfully tested in multiple fields to verify the functionalities of the alarming system. Numerous abnormal events were identified in the field and faults signatures and trends were stored in the knowledge database along with the corresponding alarm mitigation strategy. The smart alarming produces superior results in several case studies performed on multiple Permian basin wells and fields. This new smart alarming approach will greatly help in the increasing real time artificial lift ESP management over existing conventional basic ESP alarm monitoring methods.