Gas Lift Anomaly Detection in Unconventional Fields Using Expert System Techniques

Abstract Production engineers managing unconventional gas lifted wells face ever-changing operational conditions with newly drilled wells coming online, on-going well workover, facility upsets and artificial lift configuration changes. Optimizing gaslift operations involves delivering optimal gaslift injection to wells while respecting operational constraints but also detecting misbehaving wells early to mitigate underperformance. Detecting anomalous behavior in gaslift wells is onerous due to the dynamic nature of the operations and the deluge of field measurements that need to be analyzed. This paper will cover a novel approach to detecting anomalous behavior, suboptimal production, and integrity issues in gaslift wells in unconventional fields. The approach relies on using expert system techniques in conjunction with physics-based models to detect these conditions and notify engineers to take remedial action. The paper will also cover how this novel approach was built into a scalable and extensible solution called the Gas Lift Anomaly Detection (GLAD) tool.