Petrel

E&P software platform

Release Notes
Machine Learning for Petrel 2023.3
Version 2023

Release Notes
Version 2023.3.1.0
Copyright Notice
Copyright © 2023 SLB. All rights reserved.
This work contains the confidential and proprietary trade secrets of SLB and may not
be copied or stored in an information retrieval system, transferred, used, distributed,
translated or retransmitted in any form or by any means, electronic or mechanical, in
whole or in part, without the express written permission of the copyright owner.

Trademarks & Service Marks

SLB, Schlumberger, the SLB logotype, and other words or symbols used to identify the
products and services described herein are either trademarks, trade names or service
marks of SLB and its licensors, or are the property of their respective owners. These
marks may not be copied, imitated or used, in whole or in part, without the express
prior written permission of SLB. In addition, covers, page headers, custom graphics,
icons, and other design elements may be service marks, trademarks, and/or trade
dress of SLB, and may not be copied, imitated, or used, in whole or in part, without the
express prior written permission of SLB. Other company, product, and service names
are the properties of their respective owners.
ECLIPSE® is a mark of SLB.
An asterisk (*) is used throughout this document to designate other marks of SLB.

Security Notice
The software described herein is configured to operate with at least the
minimum specifications set out by SLB. You are advised that such minimum
specifications are merely recommendations and not intended to be limiting to
configurations that may be used to operate the software. Similarly, you are
advised that the software should be operated in a secure environment
whether such software is operated across a network, on a single system
and/or on a plurality of systems. It is up to you to configure and maintain your
networks and/or system(s) in a secure manner. If you have further questions
as to recommendations regarding recommended specifications or security,
please feel free to contact your local SLB representative.
Contents

1. Introduction .......................................................................................................................................................................................................................................... 1
Release date ....................................................................................................................................................................................................................................... 1
Version ..................................................................................................................................................................................................................................................... 1

2. Product and feature changes .............................................................................................................................................................................................. 2

3. Platform availability ........................................................................................................................................................................................................................ 2

4. 3rd party components.................................................................................................................................................................................................................. 3

Tensorflow open source software attribution ......................................................................................................................................................... 3
Keras open source software attribution....................................................................................................................................................................... 3

5. UI entry points .................................................................................................................................................................................................................................... 4

6. Custom domain objects............................................................................................................................................................................................................ 5

7. Manipulated domain objects ................................................................................................................................................................................................ 5

8. Known issues...................................................................................................................................................................................................................................... 6

9. Troubleshooting............................................................................................................................................................................................................................... 6
ML Fault Prediction task reports "Success" but it failed.................................................................................................................................. 6
ML Fault Prediction task failed and provided no log ......................................................................................................................................... 7
ML Fault Prediction task failed due to limited memory .................................................................................................................................... 7

10. Help and support information ...................................................................................................................................................................................... 7

 1. Introduction
The plugin works with Petrel 2023.3 or higher.

It provides the Machine Learning Property Modeling, which contains "ML

Property Modeling"; the Machine Learning Assisted Seismic Interpretation
ML Fault Prediction , Fault extraction , and ML Horizon
Prediction (NN) . These features are licensed and require dedicated license
feature on a dongle or a license server.

Release date
October, 2023.

Version
Plug-in version: 2023.3.1.0

1
 2. Product and feature changes
VERSION CHANGES

2023.3.1.0 Released with:

• ML Property Modeling
• ML Fault Prediction
• Fault Extraction
• ML Horizon Prediction (NN)

3. Platform availability
Machine Learning for Petrel is a plug-in for Petrel. You can access and use it if
you install Petrel 2023.3.

The system requirements are the same as for Petrel. For more information, refer
to the "System Requirements" section in the Petrel 2023.3 Installation and
Configuration Guide.

2
 4. 3rd party components
1 Python (Version: 3.10)
2 CUDA (Version: 11.2)
3 CuDNN (Version: 8.1.1)
4 Imageio (Version: 2.31.1)
5 NumPy (Version: 1.25.1)
6 OpenZGY (Version: 0.2.5568903)
7 Overrides (Version: 7.3.1)
8 SciPy (Version: 1.11.1)
9 TensorFlow (Version: 2.10.0)
10 Toolkit (Version: 0.2.5010836)
11 Pyarmor (Version: 8.2.9)
12 Keras (Version: 2.1)
13 nvidia-ml-py3 (Version: 7.352.0)

Tensorflow open source software attribution

Tensorflow was originally created by the Machine Intelligence team at Google
Brain, and is made available with an Apache 2.0 license, via
https://github.com/tensorflow/tensorflow. More details are available on
https://www.tensorflow.org/.
The version of Tensorflow installed alongside Machine Learning for Petrel has
been modified by obfuscating with using PyArmor. No other modifications have
been made.

Keras open source software attribution

Keras was originally created by the François Chollet and is made available with
an Apache 2.0 license, via https://github.com/keras-team/keras. More details
are available on https://keras.io/.
The version of Keras installed alongside Machine Learning for Petrel has been
modified by obfuscating with using PyArmor. No other modifications have been
made.

3
 5. UI entry points
You can access "User-trained fault Prediction" under Seismic Interpretation ->
Assisted Interpretation -> User-trained fault Prediction.

Figure 1: Machine Learning for Petrel User-trained fault Prediction

You can access Fault Extraction under Seismic Interpretation -> Assisted
Interpretation -> Fault Extraction.

Figure 2: Machine Learning for Petrel Fault Extraction

You can access "NN horizon prediction" under Seismic Interpretation ->
Assisted interpretation -> NN horizon prediction.

Figure 3: Machine Learning for Petrel Horizon Prediction (NN)

You can access "EMBER" machine learning property modeling under Property
Modeling -> Machine Learning -> EMBER.

Figure 4: Machine Learning for Petrel Property Modeling

4
 6. Custom domain objects
Machine Learning for Petrel does not create any custom domain objects in the
Petrel users project.

7. Manipulated domain objects

Module MDO Purpose

User-trained fault The process generates new 3D Seismic cubes to

3D Seismic
prediction store the predicted fault probability cubes.

The fault extraction process generates new 3D

3D Seismic
Seismic cubes to store the evaluation cubes.

The extracted faults are stored in the format of

Fault extraction Points
Points in the Input pane in a new folder.

The extracted faults can be converted to Fault

Fault interpretation interpretation after "Fault point sets post
processing".
The process uses the provided Seismic horizon to
Seismic horizon store attributes, the horizon prediction model, and
ML Horizon updates the Seismic horizon after running.
Prediction (NN)
The process generates new 3D seismic cubes to
3D Seismic
store the horizon confidence factor.

ML Property The process generates new property in the grid to

Property
Modeling store the results.

5
 8. Known issues
1 The license must be selected before Petrel is open. If you go to
File/License module and select the license after, it will not activate the
feature.
2 In this release, you can only run one user-trained Fault Prediction at a time,
once the task is finished, you can run another task.
3 The NN horizon prediction in this release cannot run with fault cube
provided, it will show an error.

9. Troubleshooting
ML Fault Prediction task reports "Success" but it failed
If you have ran a user-trained fault prediction task, but the task fails, it will
appear in the task manager with a status of "Success" and the progress will be
100%, contrary to the result of the fault prediction, shown in figure 5:

Figure 5: User-trained fault-prediction task in Task Manager

In Petrel, the uccess status and progress bar in the Task Manager window
reports the status of the task itself, not the result of the user-trained fault
prediction algorithm.
To see the result of the run from the algorithm, click on and check the
message log for the summary of the task, shown in figure 6:

Figure 6: User-trained fault-prediction with failure summary

The summary will show the success or failure status in the line starting "User-
trained fault prediction is finished". The status is shown in parenthesis, as either:
"(Failed)" or "(Success)"
6
 Alternatively, you can find additional information on the run by checking the log
for the run, which is displayed in the Message Log window when you click
"Log", shown in figure 7:

Figure 7: Log button for user-trained fault-prediction task

ML Fault Prediction task failed and provided no log

If you run user-trained fault-prediction and it fails, but when you click the Log
button for the task it shows nothing in the Message Log window, then the fault
prediction may have failed to even start.
This could be caused by missing installation prerequisites, such as:
• Python 3.10 has been uninstalled or could not be found.
• The Machine Learning Python Environment has been deleted or
damaged.
To fix this, it is recommended to uninstall and install the plug-in again, following
the installation guide.

ML Fault Prediction task failed due to limited memory

If you run a user-trained fault-prediction and it failed with a message included in
the Log which displays a message such as:
"numpy.core._exceptions._ArrayMemoryError: Unable to allocate 10.1 GiB for
an array with shape (3008, 14080, 64) and data type float32", it suggests that
the data set you are trying to train against is too large for your computer.
This means either:
• Your computer system's memory is too small
• Your data set is too large
Try to reduce the size of the data set or improve your computer system's
memory capacity.

10. Help and support information

Machine Learning for Petrel is provided by SLB. For support information,
contact Petrel support at your location or send request to:
https://www.software.slb.com/support

7
*Mark of SLB.
Copyright © 2023 SLB. All rights reserved