Clayton V. Deutsch A Review of Techniques for Porosity/Permeability Modeling
(Can We Achieve Objectivity in Reservoir Modeling?)
Clayton V. Deutsch*
University of Alberta, Edmonton, AB, Canada
cdeutsch@ualberta.ca
Geostatistical techniques have become accepted and useful for assigning porosity and
permeability at a small scale in reservoir models. Facies modeling techniques have also evolved,
but the rich variety of facies modeling techniques are not the subject of this presentation.
Assigning porosity and permeability will always be required at some scale; within facies or for an
entire reservoir zone.
There are many issues related to the assignment of porosity and permeability including data
quality, issues of scale, inference of statistical parameters, accounting for geological controls and
calibration with geophysical and production data. The intent of this presentation is to focus on
some alternate methodologies and to discuss the selection of a technique. Most reservoir
properties (facies, porosity, residual saturations,…) are volumetric or mass fractions. Permeability
is quite different; it is a rate constant. This leads to different modeling methods.
Minimum, good and best practice of permeability modeling will be reviewed. There are a variety
of considerations including the geological setting, the fluids, the amount of historical data and the
level of detail being included in the flow simulation. Our primary goal is to create models with a
reasonable structure of high and low values. If a high resolution facies model is available with a
reasonable number of core data within each facies, then a simple phi-k transform may be
appropriate; otherwise, different techniques can be considered.
The last part of this talk will relate to objectivity in numerical modeling. There is no way to attain
complete objectivity, but there are some related considerations that are important: (1) reproduction
of features known to be important for the process and reservoir type, (1) transparency – few
hidden assumptions, (2) accessibility of required input parameters, (3) robustness of the results
with respect to input parameters, (4) little room for mistakes/blunders, (4) internal consistency of
the available data and the model, (5) fidelity of the model with data not used in modeling, (5)
checkability of the results, (6) fairness of the predicted uncertainty when compared to new data.
http://www.cspg.org/conventions/abstracts/2007abstracts/262S0401.pdf
Отправлено: 08.09.08 04:31. Заголовок: Clayton V. Deutsch A Review of Techniques for Porosity/Permeability Modeling
