Datarock Knowledge Base

Product descriptions

Fractures

This document details Datarock’s products Fracture Frequency and Spacing.

Dependent Models

The outputs of the following models are used to determine Fracture Frequency and Spacing:

Model Name	Model Type
Fracture Detection and Classification	Object Detection
Drillers Break	Object Detection

Data Processing

The outputs of the Fracture Detection & Classification and Drillers Break models are combined to create an understanding of natural fractures, allowing fracture frequency to be calculated.

Detection of Fractures and Mechanical Breaks

Several fracture types are detected by the Fracture Detection and Classification model, and are summarised in the table below.

In addition, drillers break marks are detected by the Drillers Break model to differentiate between natural and mechanical breaks. If a drillers break mark is located within 5cm of a fracture detection, the fracture will be flagged as mechanical.

The following images show example detection and classifications of fractures. In the second row, first fracture from the left, it can be seen that no fracture shown due to the presence of the drillers break marks.

$frac-1$

$frac-2$

$frac-3$

$frac-4$

Fracture Counts

To calculate the fracture frequency, Datarock calculates an Equivalent Fracture Count for a given interval using the equivalent counts shown below. In addition to counting each fracture detection, the equivalent fracture counts are applied to incoherent zones.

Fracture Type	Description	Equivalent Fracture Count
Measurable	A fracture, with single interface and clear shadow line that allows structural measurements automatically approximated.	1
Displaced	A fracture, with single interface, however either one or both pieces of core have dislodged/ rotated from their original position creating a physical gap between fracture edges.	1
Jigsaw	Multiple fractures that intersect each other with pieces well fitted together or slightly displaced	2 fractures / 100mm
Broken	Multiple fractures that intersect each other that do not fit together well with larger spaces between them between them	4 fractures / 100mm
Rubble	Multiple pieces of small rocks, pebbles, fine rubble and/or powder between two pieces of core	5 fractures / 100mm

The resultant equivalent fracture count is then used to calculate the fracture frequency and spacing for given intervals.

Product Configuration Options

The equivalent fracture counts can be configured via the Fracture detection tab access via the settings cog next to the name of your project page.

Once in this tab, you can then edit the equivalent fracture counts for Jigsaw, Broken and/or Rubble to better match your own equivalent fracture counts (see below). These are preset to the default counts Datarock uses as described above.

Note: you will need to click Submit to save the edits and then re-run the Fractures product to process the results using the new equivalent fracture counts.

prod-config

Output Intervals

Default interval length: 1.0m

Customisable interval available: Yes, via uploading sample table to platform (see User Data below)

User Data

User data may be uploaded to the platform via csv in the following format:

· HoleID_sampling_intervals_fractures.csv

CSV file to contain the following headers:

File Header	Description
depth_from	Start of interval
depth_to	End of interval
fractures	Number of natural fractures counted on site

Data Output

Results from this class of models can be obtained using the Download artefacts option from the Actions button in the Model Review tab of Datarock. The available CSV files include the following:

· ProjectID_HoleID_fracture_frequency_by_metre.csv

· ProjectID_HoleID_fracture_frequency_by_user_intervals.csv

· ProjectID_HoleID_fracture_location.csv

The first two CSV files contain the following headers:

File Header	Description
hole_id	Customer’s Hole ID
depth_from_m	Start of interval (metres)
depth_to_m	End of interval (metres)
depth_from_ft*	Start of interval (feet)
depth_to_ft*	End of interval (feet)
interval_metres	Length of interval in metres
fracture_count	Number of simple fractures detected
equivalent_fracture_count	Fracture_count plus equivalent fracture count for jigsaw, broken and rubble zones
mechanical_fracture_count	Number of detected fractures within threshold distance from detected drillers break marks
driller_break_mark_count	Number of drillers break marks detected
fractures_per_metre	Calculation of fracture_count divided by interval_metres
equivalent_fractures_per_metre	Calculation of equivalent_fracture_count divided by interval_metres
timestamp	Time of model prediction
fracture_spacing	Inverse of equivalent_fractures_per_metre
fracture_spacing_category	Categorised description of the fracture spacing
version	A model version identifier

*Only included if project depths are in feet.

ProjectID_HoleID_fracture_location.csv provides data for each individual fracture and contains the following headers

File Header	Description
hole_id	Customer’s Hole ID
fracture_uuid	Universally unique identifier for the fracture
depth_from_m	Start depth of the fracture bounding box (metres)
depth_to_m	End depth of the fracture bounding box (metres)
depth_from_ft*	Start depth of the fracture bounding box (feet)
depth_to_ft*	End depth of the fracture bounding box (feet)
depth_centre	Centre depth of the fracture bounding box. This is used as the actual depth of the fracture
fracture_class	The class of fracture as predicted by the object detection model
probability	The probability, or model’s confidence, of the fracture detection
type	Natural or mechanical fracture
row_uuid	Universally unique identifier for the row that contains the fracture
row_url	Datarock URL that contains the fracture
bounding_box_scaled	Coordinates in the image of the depth registered fracture bounding box
inference_timestamp	Time of model prediction
version	A model version identifier

*Only included if project depths are in feet.

Product Limitations

Limitations

Comments

Fracture count based on assumed equivalent fracture count in broken zones

This method is based on counting individual simple, single interface fractures, and assumed equivalent fracture counts for broken zones. It is recommended that these assumptions be verified based on customer’s logging schema.

Reliance on drillers break mark detection

Mechanical breaks are very difficult to classify from a photo, therefore Datarock’s Fracture analysis relies on these breaks being visually identified, usually with a cross adjacent to the break.

If mechanical breaks are not identified, the resulting fracture counts will generally be higher than the site logging.

Document Version

Version	Date	Author	Rationale
1	6 October 2022	S Johnson	Initial release
2	25 January 2023	S Johnson	Update mechanical break description
3	29 January 2024	S Johnson	Update to include depths in feet
4	02 May 2024	N Pittaway	Update to include editing equivalent fracture counts