Applying Simple Geostatistical Techniques to a Routine Production Geology Problem--A Case Study: ABSTRACT

Abstract

A production geology reservoir description was shown to represent poorly the known dynamic reservoir behavior. The permeability field was originally generated from a general porosity-permeability law applied to a contoured porosity field. This resulted in unrealistically high permeability values in some layers and early water breakthrough in flow simulations. Furthermore, known well values of permeability were not honored. To improve the model, a geostatistical approach was used. The first step involved the use of a porosity-permeability law based on total layer porosity values. That is, with no cutoff applied, thereby allowing obtainment of values of k closer to the known reservoir values. This alone produces a permeability field, which although better represents the absolute permeability values is still too smooth-preferential pathways occurring as artifacts of contouring. A second step, therefore, involves the [open quotes]unsmoothing[close quotes] of the permeability field. At each point of the permeability field the permeability values are resampled from a distribution centered around the original value. This is a form of Monte-Carlo replacement, conditioned to well data and honoring the original trend to the data. In this case no correlation was included due to lack of information. This simple Gaussian simulation approach provides multiple realizations based onmore » deterministic information. The advantages are clear, more realistic images (no artificial pathways), improved match of water breakthrough, and the honoring of all deterministic data (wells and trend). A subsequent step was the incorporation of further [open quotes]soft[close quotes] information. Geological analysis suggested that there was a degradation of reservoir properties along the east-west axis. This new information was rapidly assimilated into the model to produce final images of the reservoir.« less