| Popis: |
In this paper we describe the Fenja Digital-Twin solution, its automated advisory workflow, and how it enhances operational excellence and further strengthens the operational health, safety, and environment (HSE) vision of Neptune Energy. The Fenja development is a subsea tieback to the Njord-A platform located in the Norwegian Continental Shelf. Fenja wax and hydrate management strategies are based on active heating of the world's longest electrically traced heated pipe-in-pipe (ETH PIP) as the primary layer of protection. The Fenja development demands a reliable and robust solution to address flow assurance challenges and support operational activities planning (production management, shutdown, startup, virtual metering, hydrate/wax formation preventions, ETH pipeline operation, and chemical tracking). A digital twin solution deployed for the Fenja development uses the integrated power of machine learning and data analytics with multiphase flow simulation, enabling enhanced efficiency and transforming operations. The Keep-Warm and Warm-Up advisory workflows help reduce carbon dioxide (CO2) emissions to the environment by optimizing the ETH pipeline power consumption and empowering the operational decision-making process for operations. Leak detection advisory (LDA) uses data analytics, signatures analysis, and pattern recognition to rapidly detect potential leaks of fluids from hydrocarbon pipeline. Combining digital twin with data analytics makes detecting leaks more precise and avoids false alarms. A faster and highly precise LDA advisory can potentially minimize a great deal of hydrocarbon loss to the environment in a real leak event. In the test environment for the factory acceptance test, the LDA was able to produce an alarm within 1 hour after the onset of a small leak (1% of pipe diameter) and within 15 minutes for a moderate (>3% of pipe diameter) to large leaks. A vast amount of real-time distributed temperature sensing data (at every 3-m distance, a total of 34,000 data points) along the entire length of the ETH pipeline and riser are continuously collected offshore and then fed into the advisory module. A customized automated workflow is developed to process this enormous amount of near real-time data to be used in calibration and to provide useful insights to the end user. New methods, using machine learning and data clustering techniques, are used to develop specific automated advisories. The digital twin with automated advisory runs automatically and needs minimal to no intervention, providing operational awareness and actionable insights to decision-makers in real time. The integration of a "replay" functionality to rerun historical actions performed in the field and "look-ahead" functionality for future predictions in a collaborative environment makes the Fenja Digital-Twin solution an essential tool for operations. |
