Reservoir simulation

Research Leader: Ivar Aavatsmark

The reservoir simulation activities at CIPR are designed to improve the reliability of reservoir models and reservoir model predictions, as well as making the modelling process and the simulations more effective. This includes seismic reservoir characterisation, reservoir simulation, and conditioning models to all available data, including dynamic data (production and pressure data and time-lapse, or 4D, seismic data).

We also aim to include electromagnetic data in an advanced workflow. Conditioning to dynamic data is mathematically an ill-conditioned problem with a non-unique solution, and a good solution requires efficient regularisation and/or uncertainty analysis and multiple runs, using time consuming reservoir simulation models.


Researchers Bakr, Shaaban Elenius, Maria Gasda, Sarah Eileen Hægland, Håkon Kaufmann, Roland Lien, Martha Økland Mannseth, Trond Oliver, Dean Pettersen, Øystein Stephansen, Annette Tai, Xue-Cheng Aavatsmark, Ivar Adjunct researchers Dahle, Helge K. Nordbotten, Jan Martin Nordtvedt, Jan Erik Nævdal, Geir	PhD students Ashraf, Seyed Meisam Fossum, Kristian Gurholt, Tiril Kovalenko, Andrey Muneer, Faiza Sandvin, Andreas Sævik, Pål Næverlid Thulin, Kristian Tveit, Svenn Vasilyev, Leonid Zhang, Yanhui Post Doc researchers Dmitriy Kolyukhin	Master students Ibrahim, Fethia Odeh, Paul Reidar
Key issues
Multipoint Flux Approximation There is special focus on convergence for heterogeneous media and discretisations in three-dimensional near-well regions. Grids and discretisations are compared in order to find a good and robust scheme in the near-well zone. Streamline methods with focus on fractures in 2D Methods have been implemented and tested for fracture models with and without fracture expansion. The two methods behave similarly if the grid resolution is comparable to the fracture density. Computational simulation different approaches have been compared with respect to speed and robustness. In terms of quality and reliability of the results, non-conventional methods appear somewhat less accurate with respect to the phase-state computation. Multiscale methods the focus has been on the relationship between multiscale methods, domain decomposition and upscaling. This has lead to an adaptive multiscale method for multiphase flow. Geomechanical modelling The simulator has been enhanced with options that aid the reservoir engineer in history matching. Changes in compaction tables can be performed by areal allocation, making the history matching process realistic. Inverse problems for reservoir monitoring and dynamic reservoir characterisation As a first result of its kind, a probabilistic error measure for the ensemble Kalman filter analysis step has been developed under the condition of negligible observation errors. Issue in cross-disciplinary research A recurring theme in cross-disciplinary research is the intrinsically disparate scales between disciplines, hindering the direct coupling of different models. Multiscale couplings are not a discipline-specific problem, but a fundamental issue that needs broad, concerted research effort. Our main themes are existence and limitations of single-scale models, multiscale modeling of coupled.

Completed Master Theses
Completed PhD Theses

E-mail cipr@uni.no, Phone + 47 55 58 36 70, Fax + 47 55 58 82 65