Interview with Torgeir Vanvik,
Principal Project Engineer at Turbulent Flux.
Torgeir, what is the value of VFM for unconventional field operators?
VFM provides continuous flow rates from wells, improving well performance monitoring and reducing costs associated with flow rate measurements. Early identification of production anomalies, increased overall production outputs and reduce testing costs are key benefits.
Who are your clients and how do you work with them?
We work with leading operators in the Permian field in the USA to reduce costs and improve production efficiency. Testing well flow rates is required for measurement and operational issues. To test, a well is routed to a test separator or multiphase flow meter but installing dedicated test separators on each well is costly. We work with clients’ technical production and IT teams to implement VFMs on their wells, starting with a few wells for operator confidence before scaling up.
Is sensor data availability a challenge in unconventional wells?
Data from production conditions, such as pressure and temperature, can be challenging for flow rate calculation methods. Our hybrid software technology does well even with limited data.
Why is the hybrid solution important for VFM?
Our hybrid VFM combines physics-based simulations with machine learning for improved accuracy and detailed understanding of the system. It can make accurate predictions in a wide range of conditions.
How do you cope with changing production conditions in unconventional wells?
Our VFM has a self-adjusting capability, which eliminates the need for manual re-calibration based on reference measurements. It addresses changes in phase compositions as they occur.
What are the pros and cons of data-driven and physics-driven approaches?
Machine learning (ML) has the potential to revolutionize fluid flow predictions, but it requires significant amounts of data and re-training for new operating conditions. Physics-based approach can be computationally expensive but requires little data and extrapolates outside known data. Our hybrid approach combines these methodologies and works on unconventional wells with fluctuating production outputs.
Here is a link to a relevant paper we published in 2022 : https://onepetro.org/URTECONF/proceedings-abstract/22URTC/1-22URTC/D011S015R003/489232?redirectedFrom=PDF
How do you scale the solution over large well counts?
We designed our technology to be scalable, with automated workflows reducing manual work and a self-adjusting capability reducing maintenance work. A REST API is available for easy integration with clients’ IT infrastructure.
Thank you for your insight, Torgeir.
