Reservoir Modeling Using Multi-Point Statistics (MPS), Berkine Basin, Algeria
Delgado, Jose F.; Djettou, Farid; Noureddine, Benourkhau; Hachelaf, Houari; Kerboub, Zohra
The Menzel Lejmat North (MLN) field and its satellites are located in the Berkine Basin of Algeria approximately 150 miles (250 km) south-east of Hassi-Messaoud and are part of a multi-field development located in the Sahara Desert of Algeria. This work presents a case study of one of the ConocoPhillips-operated fields at MLN. The MLN field is data rich, with most fault blocks containing a cored well and full log suite, seismic data and abundant testing and production data.
MPS is a facies modeling technique based on multiple-point statistics instead of the conventional variogram models founded in 2-points statistics, where algorithms that only account for 2-point statistics cannot reproduce some features that are captured by higher order statistics (e.g. curvilinear channels). But one of the main challenges is acquiring the correct training image with the right scale and representative of the 3-D complexity and heterogeneity of the reservoir. A novel geomodeling workflow and methodology is presented as an alternative to merge and optimize these sets of multiscale data within a geological conceptual model using three different sets of training images (TI): one coming from a local detailed sedimentological interpretation of core data, the second one coming from a more regional sequence stratigraphy and sedimentological study and the third one coming from modern analogs using satellite images, georeferencing and tracing the plan-view analog images within the geomodeling software. This methodology helped capture the potential spatial variety due to the different interpretation or creation of the training image and also reduce the uncertainty in facies spatial distribution.
After the TI were created and successfully tested, the fluvial-channel facies model for the Late Triassic (228 million years ago) TAG-I formation was populated using multipoint statistics (MPS) simulation, and integrated with some bedscale heterogeneities to capture complex spatial relationships between the facies. The benefits of improving reservoir architecture and connectivity by using this facies modeling technique were that the new set of stochastic geomodels became a better tool to predict well performance, accurately predict flow properties and match the productivity of the wells while doing the history match, compared with previous facies models techniques (i.e: variogram-based geostatistics).
AAPG Search and Discovery Article #90163©2013AAPG 2013 Annual Convention and Exhibition, Pittsburgh, Pennsylvania, May 19-22, 2013