Denise N. Pfeffer¹, Dennis R. Prezbindowski², Blake B. Sherman², Michael W. Waite³
(1) Texaco Upstream Technology, Houston, TX
(2) Texaco, Houston, TX
(3) Saudi Arabian Texaco, Kuwait

Abstract: Deterministic and Stochastic Methods of Distributing Reservoir Properties in a Complex Carbonate Reservoir – Cretaceous, Ratawi Oolite in the Partitioned Neutral Zone of Saudi Arabia and Kuwait

Complex porosity and permeability relationships are common in carbonate reservoirs, even within individual lithofacies. These relationships make it difficult to accurately predict and distribute reservoir properties in a static geological model because of varying degrees of spacial uncertainty. Measured reservoir properties can be distributed to nonsampled areas using deterministic or stochastic techniques. Deterministic techniques range from simple distance weighted interpolation to linear and non-linear transformations of distributed properties. Geostatistical approaches utilize stochastic techniques such as Sequential Gaussian simulation and Cloud Transforms. Comparisons of deterministic and stochastic approaches for predicting permeability and distributing reservoir properties were conducted in a Lower Cretaceous, carbonate reservoir in the Partitioned Neutral Zone of Saudi Arabia and Kuwait. Cumulative density function graphs (CDFs) were used to compare permeability predicted from deterministic and stochastic techniques with measured values from core. This 1-dimensional comparison using CDF plots demonstrated that the stochastic technique (cloud transform) more closely matched the permeability distribution derived from core measurements for individual wells as well as globally. In addition to more accurately reproducing the permeability distribution, the cloud transform provided the means to quantify the inherent uncertainty associated with the reservoir’s complex porosity and permeability relationships. Quantitative and graphical comparisons of 3-dimensional, deterministic and stochastic porosity and permeability models demonstrated that significant differences are observed in fluid flow for areas associated with particular wells or for regional-area fluid flow. The stochastic technique proved to be a more accurate method for distributing observed complex reservoir properties.

AAPG Search and Discovery Article #90914©2000 AAPG Annual Convention, New Orleans, Louisiana