--> Abstract: Improved Facies Modeling Using Multipoint Geostatistics in the Brent Reservoir Interval, Total Alwyn North Field, UK North Sea, by Stewart Clark, David McCormick, Tuanfeng Zhang, David Brethaut, Mirna Slim, and Neil Hurley; #90078 (2008)
[First Hit]

Datapages, Inc.Print this page

Improved Facies Modeling Using Multipoint Geostatistics in the Brent Reservoir Interval, Total Alwyn Previous HitNorthNext Hit Previous HitFieldNext Hit, UK Previous HitNorthNext Hit Previous HitSeaNext Hit

Stewart Clark2, David McCormick1, Tuanfeng Zhang1, David Brethaut1, Mirna Slim1, and Neil Hurley1
1Schlumberger-Doll Research, Cambridge, MA
2Geoscience Research Centre, TOTAL E&P UK PLC, Aberdeen, United Kingdom

We have modeled facies in Total’s Alwyn Previous HitNorthNext Hit Previous HitfieldNext Hit, UKCS, to validate multipoint geostatistics (MPS) for modeling complex geological facies heterogeneity. MPS property modeling methods can integrate multiple data types, create stochastic conditional simulations of facies, and potentially overcome the main limitations of commonly used property modeling methods. MPS methods use 3D training images to explicitly quantify the idealized subsurface texture interpreted to exist in the reservoir.

The upper Brent reservoir interval of the Total Alwyn Previous HitNorthNext Hit Previous HitFieldNext Hit in the UK Previous HitNorthNext Hit Previous HitSeaNext Hit contains a complex mix of siliciclastic coastal plain and shallow marine facies. The dataset has 15 wells with 5-8 facies per zone. We modeled zones with approximately one million cells each. The facies mosaics comprised meandering fluvial and associated levee, crevasse splay, and floodplain in the coastal plain; deltaic and shallow marine facies were modeled using channels on a coastal plain background connected to lobate deltas on a background of shallow marine. The stochastic realizations of facies were controlled by a combination of these environment-specific training images, probability-of-facies grids, and other soft constraints. The final realizations match net-to-gross, azimuth, and spatial trends in facies.

In this real Previous HitfieldTop case study the MPS technique was successful in modeling complex 3D facies mosaics efficiently and quantitatively, whilst honoring multiple data types including well, seismic, and trend information. MPS methods are computationally competitive with established property modeling methods. The explicit use of the training image is a powerful tool to explicitly and quantitatively embed geological interpretation into geological models.

 

AAPG Search and Discover Article #90078©2008 AAPG Annual Convention, San Antonio, Texas