Maya Yermekova, Zeshan Hyder, Carl Kemp, Saket Srivastava, Agustin Gambaretto, Yury Pazniak, Svetlana Pivtoratskaia, Ambica Agarwal, and Amey Ambad
SLB
Sucker rod pumps (SRPs) stand as the foremost artificial lift (AL) technique globally, with advancements dating back to the development of the wave equation in the 1960s. Leveraging edge-based technologies, a workflow has been devised, building upon the foundation of existing Pump-Off Controller (POC) capabilities. This workflow seamlessly integrates machine learning (ML) based dynamometer card classification for real-time event detection with forward-thinking logic to autonomously optimize SRP operation setpoints. Operating within an Industrial Internet of Things (IIoT) framework, high-frequency dynamometer cards and pump data are analyzed.
The workflow incorporates two distinct sets of controls tailored for SRPs. The first set, Fast Loop Mitigation controls, utilizes outputs from an ML algorithm to promptly process and classify surface and downhole dynamometer cards in real time (Bruns, Sharma and Whitley 2022). These mitigations are specifically designed to identify and address common SRP issues such as flatlining, fluid pound, gas interference, and tagging as they arise.
The second set of controls, known as the Production Optimization (POPT) Algorithm, collects operational data and monitors it through a dynamically shifting time window (Gambaretto, et al. 2024). This window provides a relevant snapshot of operational history, enabling the algorithm to synthesize trends into performance indicators. These indicators are then utilized to forecast the most optimal pump operating setpoint for enhanced efficiency and productivity.
Results from testing underscore the significant benefits achieved through the synergistic application of both control systems. On tested wells, inferred production saw a notable average increase of 15%, while runtime exhibited an average uptick of 3%. Concurrently, cycling decreased by an average of 29% by maintaining optimal pump fillage (Gambaretto, et al. 2024). Moreover, stable operating conditions were observed throughout.
This approach represents a paradigm shift, integrating the mitigation of short-term potential pump damaging issues, with the long-term POPT algorithm for maintaining optimal well production. By leveraging the strengths of both algorithms, it offers a holistic optimization strategy that addresses immediate challenges while aligning with long-term objectives, thereby providing a comprehensive solution for well optimization.