Stochastic Evaluation of Fluvial to Marginal Marine Sealing Facies

S.A. Barboza¹, R. Alway², T. Akpulat¹, W.L. Esch², and P. Hicks, Jr.^1
1ExxonMobil Upstream Research Company, P.O. Box 2189, Houston, TX 77252-2189
²ExxonMobil Exploration Company, 13401 North Freeway, Houston, TX 77060

Hydrocarbon bed-seal analysis is particularly difficult in fluvial and marginal marine settings. Although these settings often contain abundant low-permeability, clay-rich lithologies that are associated with high-quality seals, their presence does not guarantee an effective seal because individual beds of the sealing lithologies within the stratal unit may be continuous only over short length scales. Discontinuities due to lateral facies changes, truncation of seal lithologies by channel incision, or crevasse-splay facies may provide cross-stratal migration pathways across the sealing interval, which reduce the effective seal capacity and lead to capillary seal failure before a trap can be completely filled.

To evaluate probable bed seal capacities in such a setting (Paleozoic marginal marine), we have calculated 75 statistically equivalent object-based geologic models of a heterogeneous bed seal based upon facies interpretations derived from wireline logs, cores, and cuttings. The flow properties (porosity, permeability, and capillary entry pressure) were then fit to Gaussian distributions based on core plugs measurements taken within the key facies. The object-based geologic models were then populated with flow properties by randomly selecting from the derived Gaussian normal distributions. Two realizations of the rock property distributions were performed for each of the 75 object-based geologic models resulting in 150 separate seal property distribution models.

We used these seal property distribution models as stochastic inputs to Permedia MPath, a percolation-based hydrocarbon migration software program, and conducted a flow simulation on each. This approach returns a probability distribution of bed seal capacities from which cumulative frequency diagrams and percentile rankings may be derived for a particular geologic characterization. Channel amalgamation by incision, juxtaposing levee overbank or channel facies formed the most common type of cross-stratal leak point. A factor of 10 difference in the P50 (percentile rank = 50) bed seal capacity was observed by varying the net volume fraction of channels from 10% to 30% of the seal. Higher seal capacity realizations were found to have a low density of channels near the structural crest, a variable outweighing all other stochastic parameters. These results highlight the importance of careful evaluation of bed seal stratigraphy in fluvial and marginal marine settings.

 

AAPG Search and Discover Article #90066©2007 AAPG Hedberg Conference, The Hague, The Netherlands

 

AAPG Search and Discover Article #90066©2007 AAPG Hedberg Conference, The Hague, The Netherlands