Published 1994 as part of Computer Applications 3
Copyright © 1994 The American Association of Petroleum Geologists. All Rights Reserved.
 

A geostatistical study of the uppermost Maastrichtian chalk reservoir unit in the Dan field was performed. The ready availability of directional and horizontal wells, uncommon in reservoir characterization studies, allows better inference of the horizontal correlation structure in this special type of reservoir.

The objective was to obtain stochastic images of porosity and permeability properties of the reservoir. Because these parameters are spatially interrelated, a cosimulation algorithm in which these properties are jointly simulated was applied. This algorithm, called the Gaussian collocated cosimulation algorithm, was built on a Markov-type hypothesis, whereby collocated secondary information is assumed to screen out farther away secondary data. This method allows us to perform the direct cosimulation of several interdependent variables integrating several sources of soft information.

The abundantly available wireline log--calculated porosity data from the wells were used as soft data, being completed into the regular grid by using a separate simulation step. The resulting stochastic models reflect both the short- and long-range structures, and the high correlation between porosity and permeability. The fine-scale stochastic images are useful for studies on upscaling techniques.