This program shall support the following Program 2 objectives

1. Embed model-based decision-support tools for RTRM into new work processes

2. Develop methods and prototype technology for RTRM

• Methods for predicting the flow response of alternative well trajectories instantly – as basis for a draw-and-evaluate tool for well trajectory planning.
• Methods for rapid updates of the reservoir model or geological model for the Norne field based on 4D seismic, production data, and other available data. The goal is to facilitate integration of new data, e.g. obtained during drilling, and to reduce risk and uncertainty due to mismatch between model and measurements.
• Fast simulation technology with seamless model-reduction capabilities (model complexity selected semi-automatically at run-time to fit time-constraints and available computer resources). The goal is to enable increased use of flow information for decision support in operations with short time horizons.
• Methods for closed loop reservoir management. The goal is to provide suggestions for future reservoir management which may include rates as well as well locations.

4. Communicate results with industrial partners

5. Publish and present results.

6. Educate PhDs and master students.

8. Establish strong collaboration with Stanford University and TU Delft

9. Develop a benchmark data base

Simulator technology

Multiple simulation runs are usually necessary in RTRM to compute optimized solutions. However, even for models of moderate size, simulation time may be a severe bottle-neck when traditional simulators are used. This has been addressed in the research-community by introducing various model-reduction techniques that decrease the number of simulations needed of the full model during a given optimization workflow. In addition to such model reduction techniques, we propose to use a novel simulation framework consisting of a multiscale pressure solver and various fast transport solvers (streamlines, non-uniform coarsening, flow-based grids, finite volume methods with optimal ordering of cells etc.) to speed-up simulation time. Under certain simplifying assumptions, the method may be orders of magnitude faster than traditional methods, without sacrificing faithfulness to the underlying geomodel. In addition, it inherently lends itself to seamless model reduction strategies. The goal is to provide fit-for-purpose models with high resolution only where needed. Since the multiscale methodology is under development, established simulators (Eclipse and/or GRPS) will also be actively used.

Schematic of how the IO centre simulation relates to ECLIPSE.

History matching, time-lapse seismic and closed loop reservoir management

To minimize risk and uncertainty of future production it is important to keep the geological reservoir description and the reservoir simulation model up to date, i.e., as close as possible to how the geophysicists and reservoir engineers conceive the true state of the reservoir at all times. The approach in this program is twofold. First, methods that use 4D seismic data to directly estimate e.g. water saturation using a seismic-geological model will be developed . Second, history matching methods to assimilate different types of data (production data, 4D seimic data, drilling data) will be developed. The combination of very fast simulator technology and data assimilation methods will be studied since provides new opportunities for history matching techniques.

The reservoir model is used for assisted decisions support. The goal is to develop a robust method to assist positioning of wells and well-paths for optimized reservoir management using a combination of discrete and continuous optimization strategies. The long-term goal is to combine this with optimization of production rates and injection rates, as well as artificial lift mechanisms like electric submersible pumps and gas-lift. It should be noted that decisions vary throughout the life-time of an asset. Artificial lift is important during the decline phase of an asset while well planning is important much earlier on.

Work processes

Through visits in participating oil companies, document studies, interviews and investigation of work- and decision making processes, we seek to identify key workflow-challenges related to field optimization. These studies will provide input and context to the development of technology and methods within the program. In particular, based on observations from the Norne field we will try to target relevant reservoir management decisions and try to develop technology to simplify and accelerate associated work processes. We will attempt to convert successful results into decision support tools that can be embedded into operational use.

Pilot case

A pilot study (using the Norne field) will be used as one test bed for RTRM. This field is suitable since it has been in production for about 10 years. It has high quality 4D seismic data, good quality production data, and a geo-model which lends itself nicely to the use of streamline technology. Further, the organization is positive towards a cooperation.

Cooperation with IO-center industrial partners

The Norne field operated by StatoilHydro is the first field case for Program 2.1. This case will be a key component in Program 2.1 in 2008-2009. Other cases will be identified later.

Results will be disseminated through workshops and publications. A workshop series with internal and industrial speakers started in October 2007 and will be a continuous platform for discussions and dissemination.

Cooperation with IO-center academic partners

Important research on RTRM is and has been performed within the ISAPP center at TU Delft, the Smart Field center at Stanford University and IRIS. This includes methods for history matching and optimization. The research in Program 2.1 will seek close cooperation with these research centres, and resources will be allocated for researchers at the cooperating centres.

A workshop is planned at Stanford University to define a common field case as a benchmark problem. This will be used for assessing methods and comparing results developed at the ISAPP center, the Smart Field center and the IO-center.

Cooperation with the Department of Applied Mathematics, University of Oslo has been established on streamline simulator technology.

RTRM decision support should ideally consist of efficient and geologically sane techniques for model-updating, efficient and robust optimization algorithms and techniques for monitoring uncertainty propagation. In current approaches, minimizing the number of forward (and adjoint) simulations have been emphasized, since simulation typically constitutes the main computational workload. The use of more flexible simulation techniques, which may give meaningful results in seconds or minutes rather than hours, opens up for a wider range of techniques for all components of RTRM. Also, faster simulations/seamless model-reduction approaches may allow subsurface modelers to work more directly on one earth model shared between geologists and reservoir engineers. This minimizes the risk that geological information and knowledge is lost when model-updating is performed on an upscaled simulation model. Ultimately, such a shared model rather than being a particular realization should contain geological uncertainty, which in turn should be reflected in optimized solutions for well locations, well scheduling etc.