An Evaluation of Geologically Based Inter-Well Permeability Modeling Techniques Using Outcrop Data and the 'ECLIPSE' Reservoir Simulator

LOWDEN, BEN D., JONATHAN J. M. LEWIS, T. SEZGIN DALTABAN, and JOHN S. ARCHER, Imperial College, London, U.K.

The prediction of inter-well scale geological (permeability) variation remains a difficult problem. In this paper, an extensive permeability data set from a shallow marine, clastic sand-body is used to evaluate the accuracy of various permeability infill techniques.

The experiment used a permeability dataset 500 m long and 5 m high. This two-dimensional permeability section comes from a coarsening upwards, shallow marine sequence in which permeability varies from less than 250 md at the base to over 5 darcies at the top. Over a thousand permeability values were collected regularly over the unit using an electronic field minipermeameter. Directional core-plugs were taken along orthogonal axes relative to the bedding from a number of large sediment blocks in order to input permeability anisotropy ratios into the numerical simulator model. Other reservoir variables, such as porosity, capillary pressure, mobility ratio, and hydrocarbon viscosity, were kept constant.

A water flood simulation was performed based on one injector well and one producer well, each placed at opposite ends of the dataset. The results from the simulation of the original dataset were then compared with the results from the simulation of a number of permeability patterns produced by routine permeability infill techniques. The results indicate that many of the techniques widely used in industry produce quite different predictions of hydrocarbon recovery and water breakthrough, though it is encouraging that some techniques are far better than others at approaching the true permeability pattern and predicting the true reservoir performance.

AAPG Search and Discovery Article #91004 © 1991 AAPG Annual Convention Dallas, Texas, April 7-10, 1991 (2009)