Lab data will be used in combination with real field data to train ML tools for monitoring of pipe flow, as well as for predicting changes to the flow regime or flow structure based on adjustments to key parameters such as superficial velocities or pipe inclination. Furthermore, a general method/framework for model tuning will be developed. The framework can be used for tuning real-time monitoring systems and/or virtual flow meters, as well as commercial multiphase pipe flow simulators. The target is to move forward from the often subjective, time-consuming and in part arbitrary workflows present in the industry today. In addition, a Stochastic Virtual Flow Meter (SVFM) will be developed tailored to available lab or field use cases employing the model tuning framework and the ML algorithms developed in the project.
The following project elements are considered novel, thus representing important advancements:
- A machine learning (ML) framework for flow monitoring, flow prediction, and automated flow regime identification.
- A general objective method for quantifying pipe flow model performance/quality, addressing important shortcomings of traditional methods.
- A stochastic virtual flow meter, capable of determining flow rates and the associated uncertainties in production systems.
How can such activity have any positive environmental impact? Firstly, it reduces the need for work-overs and drilling of new wells in ongoing production. Secondly, it reduces subjectivity, computational time and uncertainty in simulations leading to reduced investments and operating costs, shorter well delivery time, and accelerated production, all which in turn reduces Green House Gas emissions. The project puts great effort into education and knowledge transfer. NTNU students at the Bachelor, Master and PhD-level will be offered education via ordinary courses at NTNU, and the project participants will be invited to give guest lectures hosted by NTNU open to the public. A postdoctoral research fellow will work across the project work packages. Webinars will be arranged regularly, and the relatively new form of dissemination Massive Open Online Course (MOOC) will be used. The research partners SINTEF, IFE and NTNU continue their longstanding collaborative effort to meet the ambitious national goals for the Norwegian Oil & Gas sector.
Multiflow SUITE is a knowledge building project financed by the Norwegian Research Council (project # 326711), Neptune Energy, Schlumberger, LedaFlow, TechnipFMC and ESSS
