|
Software
ISIS and NeurISIS
Fracture orientation data is commonly collected through fracture mapping, core
logging, FMI/FMS, and BHTV imaging. ISIS (DOS) and NeurISIS (Windows) provide
(a) set definition based on orientation, geologic, and hydraulic properties,
and (b) distribution fitting. ISIS (Interactive Set
Identification System) uses
a probabilistic neural network (PNN) to divide fracture data into statistically
homogeneous populations. Set definition is based on any combination of fracture
parameters, such as size, orientation, mineralization, aperture, and termination
mode. User-assigned weights are used to define the relative importance to be
assumed. The iterative PNN algorithm has been shown to be reliable even for
strongly overlapping datasets.
| Feature |
Description |
| Sets Identified |
1 to 10 |
| Clustering Algorithm |
Probabilistic Neural Network |
| Clustering Basis |
Maximize Probability of Membership
in Set |
Discriminant Parameters
(Continuous, Class, Boolean) |
Orientation
Size
Mineralogy
Termination Mode
Striation
Infilling
Rock Type
Fracture Type
Geologic Indicators |
| Orientation Distributions |
Fisher
Bingham
Bivariate Fisher
Normal
Constant |
| Goodness of Fit Tests |
Kolmogorov-Smirnov
Chi-Square |
| |
|
|
