English: Advanced Wellbore Transport Modelling Norsk: Avansert modellering av transport i oljebrønnen

One of the biggest challenges in the drilling process is the removal of cuttings particles from the well, also known as hole cleaning. Several factors affect the efficiency of hole cleaning. The most important are the well geometry, the formation being drilled, pipe lateral movement, pipe eccentricity, rotational speed, drilling fluid properties, and the flow rate of the drilling fluid. Good mathematical process models, showing the effect of these parameters on hole cleaning, are therefore important for planning and managing the drilling process. A particular challenge in modeling the cuttings transport process are the non-Newtonian properties of the drilling fluid, meaning that the local viscosity is a function of the variation in the velocity field of the flow. This helps keep the cuttings from falling out of solution, while allowing for pumping at high rates for hole cleaning without causing extreme pressures downhole. Fluid flow, whether Newtonian or non-Newtonian, has typically the characteristic flow regimes of laminar and turbulent flow, and the transitional domain from laminar to turbulent behaviour. When particles are added to the fluid, additional flow characteristics, or flow patterns, may occur, such as stationary cuttings bed, moving bed and heterogeneous or homogeneous flow. Flow regimes and patterns have been investigated in the project through analysis, experiment and modelling. An analysis of the mathematical description of the pipe-flow process using nondimensional parameters was performed. The developed method is applicable for simple studies of particle flow patterns. To determine relevant parameter domains for laboratory investigations and model development, a study was performed of drilling data available in the Norwegian Petroleum Directorate database. The viscous properties of the applied translucent polyanionic cellulose test fluids were further investigated. Results indicated that the four-parameter Cross/Carreau viscosity model was a good fit for such test fluids, and this model was also suitable for CFD simulations applied. Studies of salt impurities in the test fluid showed that the viscosity decreases with increasing salt concentration. Studies of particle flow in laminar and turbulent flow were performed, investigating particle settling, moving cuttings bed and particle clouds. CFD simulations performed with ANSYS Fluent showed good correspondence for experiments with particles in water. However, modelling deficiencies were observed for particle advection and turbulent flow in non-Newtonian fluids. A new discrete phase model was developed and applied, verified with experimental data, accounting for the secondary flows caused by particle-fluid interaction. Studies of turbulent behavior in non-Newtonian flows were performed using high resolution Direct Numerical Simulations, with focus on turbulent friction factors and flow boundary characteristics, based on the viscous properties of the fluid. Such a methodology is applicable for providing input to simplified models. Laboratory studies were performed of particle flow in a horizontal pipe with simulated rotating drill string, using laser and high-speed video equipment for measurement. Analysis of experimental results shows that a cuttings bed influences the low frequency spectrum of pressure oscillations in the pipe. Such effects may also be present in the well and could possibly be used for cuttings bed monitoring during operations. Further, the drillstring rotation and lateral movement caused typical pipe flow patterns of stationary bed and separated moving bed to disappear. As a result of the described experimental and modelling effort, a non-Newtonian CFD model was developed in ANSYS Fluent for simulating cuttings flow in the wellbore with pipe eccentricity and whirl for arbitrary inclination. This has been achieved by combining the Two Fluid Model with the Kinetic Theory of Granular Flows and closures, i.e. required correlations, from soil mechanics to describe the viscous properties of the granular matter, and further using dynamic meshing capability to model lateral movement of the drillstring. This model is applicable for qualitative studies of cuttings transport. A simplified physics-based model for three-layer transport in horizontal pipes was developed, incorporating the effect of flow acceleration and wake effects over cuttings beds, validated by experiments. With pre-processing, this model has the benefit of speed of calculation. Further development is required for applicability to non-Newtonian annular flow with drillstring dynamics. In a parallel effort, an analytical mathematical drillstring dynamics model has been developed with the aim of diagnosing wellbore friction with respect to hole cleaning. Frictional forces on a lab-scale oscillating vertical drillstring have been studied for model validation. A database has been established for sharing of project results.

Det er etablert et godt internasjonalt kontaktnett innen fagfeltet som inkluderer akademiske og industrielle miljøer i Tyskland, Irland og USA. Dette gir godt grunnlag for videre forskning og utvikling. Eksperimentelle data og utviklede metoder er delt gjennom en sharepoint database. Dette muliggjør direkte bruk av resultater blant brukeraktørene i studier og til bruk av modeller for styring av boreprosessen. Utvikling og kompetansebyggingen gjort i prosjektet gir grunnlag for videre utvikling av anvendelser til bruk i planlegging og gjennomføring av boreprosessen. Spesielt fokus for videreutvikling vil være prediksjon av hullrensing for inklinasjoner mellom 60 og 30 grader, hvor eksisterende industrielle modeller har lav nøyaktighet. Videre er resultatene meget relevante til bruk for boring av geotermiske brønner, og vil kunne hjelpe på effektivisering av geotermisk utvikling.

The project is addressing knowledge needs identified by OG21 (TTA3) to be able to increase the number of drilled production wells on the Norwegian continental shelf, and specifically needs related to automation of drilling processes. Today, limitations in process understanding and model accuracy constitute a challenge with respect to achieving accurate and efficient control of the drilling process. Application process automation and diagnostic software tools are an integral part of today's drilling operat ions, and will in the future be even more so as increasingly more complex wells are drilled to improve oil recovery. The project focus is on deeper understanding of the fluid and particle transport processes during drilling using advanced mathematical mo delling supported by experimental data. The choice of modelling methodology shall be guided by the identified needs for improved models in drilling automation systems, with a hierarchy of models ranging from very detailed CFD-models to mechanistic models capable of faster than real-time execution. The project will be built around the three PhD candidates, with the research institutes SINTEF and IRIS contributing to the work packages containing the PhD work. In addition, SINTEF and IRIS scientists will w ork on a work package connecting the three PhD-topics and ensuring that that the end results of the project are targeted at the overall objective of supporting future model development for process automation. Titles for PhD-thesis are suggested as: 1) Cut tings bed interaction with the drill string, 2) Turbulence structure and particle transport in particle loaded Non-Newtonian Fluids, 3) Hole cleaning in drilling of deviated wells with multiphase drilling fluids. Experiments will be performed to generate data sets for analysis and model verification. Already existing process data will be applied where available. Also flow laboratories at the University of Stavanger and at NTNU will be used.