Well testing is a common method to understand well performance and production. During a well test, a well is routed to a test separator. Test separators are costly, thus, multiple wells generally share one separator. This leads to sequential testing, leading to intermittent flow insights. This creates challenges on the accuracy of the multiphase flow rates in individual wells and overall production. E.g. for unconventionals, production may change so rapidly that intermittent well testing is insufficient and dedicated test separators are installed on each well.
This paper demonstrates how a self-adjusting virtual flow meter continuously delivers real-time flow rates and supplements intermittent well testing to reduce the need for dedicated test separators and well testing frequency.
Physics-based transient multiphase flow simulations and machine learning are combined into a hybrid modeling approach to deliver accurate virtual flow insights, continuously. The solution automatically self-adjusts to the everchanging operational conditions using available sensor and well test data. Furthermore, it automatically ingests new data, such as new well tests, as they become available. This hybrid modeling approach combines the predictive traits of physics-based modeling with the self-adjusting capabilities of machine learning, always ensuring best performance.
We have demonstrated that we can deliver reliable flow rates for individual wells based on real-time sensor data together with data gained from the intermittent well tests. We observe that well testing can be reduced by up to 70% without compromising accurate self-calibration and fidelity of the VFM. In a scenario where every well has its own dedicated test separator, this could reduce investments in test separator equipment and operational costs associated with well test by 40-60%. Lastly, with physics-based simulations and the self-adjusted VFM model, the operator can experiment virtually with the operations of every well and, for example, analyze how a reduction in flow line pressure or lift gas rate impacts the well flow rates and optimize the production accordingly.
Multiphase virtual flow metering has a long history in offshore operations but has been hugely underexplored within the unconventional onshore space. The ability to self-calibrate and accurately provide multiphase well rates both enable cost savings with respect to well testing and offer a compelling case for a scalable digital solution catering to the short production life of unconventional wells.
