Full-field simulation of water-based EOR

Oil production on the Norwegian Continental Shelf is declining. At the same time, the sub-sea reservoirs are far from empty. When the production is stopped because it becomes non profitable, a reservoir typically contains half of the original oil in place. Recovery techniques in the North sea consist mainly of water injection: Water is injected to increase the pressure and push the oil out. Water is used because it is inexpensive and available but it is in practice not very efficient. Everyone has experienced that removing fat products that are trapped in the pores of a textile using only water is a nearly impossible task. We need to add a surfactant which diminishes the surface tension between oil and water, for example soap or washing powder. The same type of process can be used in oil recovery: Specific chemicals can be added to the injected water and improve its ability to both detach the oil from the rock and push it forward out of the reservoir. Such recovery methods have been little used on the Norwegian continental shelf because it has been difficult to establish - within an acceptable interval of confidence - that the methods would be successful. The lack of accurate stimulation capabilities is in part responsible for this situation. In this project we have studied and developed numerical tools that can be used to improve the resolution of EOR simulations and, in particular, capture accurately the impacts of the injected chemicals on the recovery process. We have mostly looked at methods that make it possible to increase the number of grid cells - and therefore the accuracy - by reducing significantly the computation time. It includes the use of splitting methods which solve sequentially the pressure system and the transport of fluids and chemical components, multiscale methods for fast computation of pressure updates, improved nonlinear solvers for the transport equations, and use of different forms of additive gridding. We have also studied the use of higher-order discretization methods to reduce numerical diffusion error, in both sequential and fully implicit simulations. Our research is not only documented in journal papers and conference talks but also available as part of an open source reservoir simulation code (www.mrst.no). The results of the project have played a decisive role in the investment and long-term involvement of TOTAL E&P in a new collaboration project with SINTEF, whose ambition is to develop the numerical methods needed for the next-generation compositional full-scale reservoir simulator.

The project has research several new methods for improved field-scale simulation and demonstrated these in terms of open-source software that currently is used in international cooperation. Several of the new methods will be brought to a higher technology-readiness level through an ongoing industry project that aims to develop technology for next-generation compositional simulators. Insight and methods developed in the project will also be used in other parts of SINTEF's ongoing contract research.

Oil production on the Norwegian Continental Shelf started to decrease in 2000. If recovery techniques are not improved, half the original oil in place will likely remain unexploited. Enhanced oil recovery (EOR) techniques that increase the sweep efficiency and enable displacement of immobile oil may be crucial to sustain production from existing fields and increase recovery factors of future developments. Water-based EOR (polymer,surfactant, etc) has been successfully implemented in onshore operations throughout the world, but remains practically unused in Norway. One reason is that business cases are difficult to establish through numerical simulation: EOR effects are often unresolved and masked by numerical diffusion in field-scale simulations. In the project, we will develop new simulation technology that overcomes this problem and implement an integrated, open-source simulator for field-scale EOR simulations. To this end, we will use a combination of two approaches. First, we will develop more efficient nonlinear solvers to reduce computational costs and hence increase the resolution of grid models that can be run within a given time frame. Second, we will develop higher-order discretizations that increase the resolution inside each grid cell without adding substantially to the overall computational cost. The new simulator will be implemented as part of the Open Porous Media (OPM) initiative, which is an open model for the development of efficient, well-maintained, industry-standard software for reservoir engineering. The main drivers behind OPM are Statoil and several Norwegian and German research groups.