| |
Data
Analysis
The
project team effort to aggregate and organize the San
Juan Basin regional data for tight gas reservoirs provides
a unique opportunity to make vertical and spatial correlations
to fill in missing data and to generate useful new data
types using various correlation tools that can be mapped
regionally as GIS layers.
Pay Zone Definition
Once the project team has assembled the data into a
useable format, individual well data will be analyzed
to develop vertical correlations of petrophysical properties.
Evaluation of the core data consists of statistical
analysis on a per well basis. This includes simple calculations
of average porosity and permeability, to assembling
vertical-to-horizontal permeability or maximum-to-90
degree horizontal permeability ratios. Figure 1 is an
example of kmax/k90 ratio from seven Dakota wells. The
average ratio for all samples is 2.7:1, with a maximum
ratio of over 30:1.
|
|
|
Fig.
1 Composite Kmax/k90 ratio for Dakota Formation
from seven wells
|
Spatial Analysis
The project team will then utilize Individual well analysis
that will be combined to develop a spatial map of the
variability and/or similarity in the given properties.
For example, the variability in permeability anisotropy
will be investigated and compared to the generic 10:1
ratio currently used throughout the basin.
It is likely that the database will have certain data
that while represented sparsely, or in geographic clusters,
would be of more use if regional estimates could be made.
Geographically limited or sparse data can be correlated
across large areas, both interpolating and extrapolating
data. A potential example is the generation of pseudo-
core porosity using wire line logs over a large area,
employing limited core data to correlate a non-linear
regression. Other examples of possible values to regress
include regional primary fracture orientation, production
indicators, permeability, and water chemistry values.
The short duration of the project shall allow for two
analyses from the list. Previous studies by the investigators
have shown promise in using a variety of advanced computing
techniques to relate wire-line log data to reservoir properties
from core data using artificial neural networks. Other
studies have extrapolated regional properties and production
indicators, and made "best" estimates of reservoir
properties between and away from existing well control.
|
