Linking CMOST with Petrel

Using CMOST v2017.10

Tutorial
Table of Contents

TABLE OF CONTENTS .............................................................................................................1

EXERCISE 1.1 – LINKING PETREL WORKFLOW TO CMOST ................................................2

Generating Petrel Workflow using CMOST Parameters 2

Master Dataset 6
Creating a Project and Study 7
General Properties 8
Fundamental Data 9
Parameterization 9
Pre-Simulation Commands 11
Objective Functions 14
Control Centre 14

1
Linking CMOST with Petrel
Exercise 1.1 – Linking Petrel Workflow to CMOST
A Petrel workflow will be created to include CMOST parameter names. This workflow will be
called when running CMOST and will output newly generated property arrays for porosity as
well as horizontal and vertical permeability. Any study type (Sensitivity Analysis, History
Matching, Optimization, and Uncertainty Assessment) can be used in this workflow.

Generating Petrel Workflow using CMOST Parameters

1. Open the Petrel file PUNQ_GM_Base.pet under the REQUIRED DATA folder.

A workflow has already been created that contains steps for regenerating permeability
and porosity arrays and exporting them to text files (PetrelWorkflow). However, we will
recreate the workflow.

2. In the Workflows tab, right click and select New Workflow

3. Call the workflow “MyWorkflow”, then click Auto generate workflow
Repeat all
Actions on the active 3D grid

2
Linking CMOST with Petrel
 This should apply all the petrophysical modelling steps performed in the model but not
all are necessary for this example. Remove any empty lines and remove options
“Make horizons” and “Scale up well logs”.
We will also add steps to export the property arrays to CMG include files.

The Workflow shown in this tutorial is a simple example. However,

workflows of any complexity may be used. This may include options
such as structural modelling, 3D gridding, well-log upscaling, facies
modeling and petrophysical modeling

In the utilities section select Import Export
Export property to ECLIPSE. Add three
rows of this using the right arrow button

4. On the Models tab, select the property KH [U] then in the Workflow editor window, click
the right arrow button in the first “Export property to ECLIPSE” line

Repeat this for the properties KV and PHIE in the next 2 lines.

3
Linking CMOST with Petrel
 Input the following keywords and path:

Keyword Property Path

PERMI KH [U] $outputPermi

PERMK KV [U] $outputPermk

POR PHIE [U] $outputPor

The keyword represents the CMG keyword for the property. The path is where the file is
saved which will be defined later in CMOST by referring to the variable name given
under the path.
Click Apply

The workflow should be tested at this point by putting in a real file path
in the path section then running the workflow. Fix any warnings or
errors if they occur.
When CMOST is running, Petrel warning or error messages will not
appear since the calculations are run in the background.

4
Linking CMOST with Petrel
 Parameters that will be varied by CMOST will next be defined. Parameters are defined
using a $ character in front of the parameter name in Petrel. The parameter will later be
added into the CMOST interface without the $ character.

5. Double click on the line for Petrophysical Modelling for KH. Click on the Zones button
to make the settings apply to all layers in the model.

Under the anisotropy range, specify $RangeH for the Major and Minor direction. Specify
$RangeV for the vertical direction. Under Azimuth, specify $Azimuth.

Click Apply and OK.

6. Repeat step 6 for KV and PHIE. Use the same parameter names.
7. Click Apply and OK to the workflow editor when complete.
8. Save the project file then exit Petrel.

5
Linking CMOST with Petrel
Master Dataset
A master dataset will be created to be a basis for CMOST to create new datasets when it runs.

9. Open PUNQ_Base_RP.dat from the student solutions folder in CEDIT.

10. Go to File
Save As then save the file as PUNQ_Base_RP.cmm
11. Locate the keyword POR
Highlight the array of values below the keyword (but not the keyword itself). Make sure
to drag the mouse past the “MORE” word to select the entire array.

Right click and select Create Include File

Name the file Porosity (“.inc” will automatically be added to end of the file name) and
also check the box for Create CMOST Parameter

Click OK. Call the parameter outputPor.

6
Linking CMOST with Petrel
 This should same name as was specified in Petrel but without the $
character.
Parameter names are case sensitive.

12. Repeat step 3 for PERMI and PERMK. Title the files PermeabilityI and PermeabilityK
respectively. The parameter names will be outputPermi and outputPermk respectively.

The section should look as follows when complete:

13. Save the file. It should already be renamed to PUNQ_Base_RP.cmm.

Creating a Project and Study

A project in CMOST groups together all studies that might be done on similar models.

14. Open CMOST and select File
New
Project

15. Name the project PUNQ_Base and select PUNQ_Base_RP.dat from the Student
Solutions folder.
16. Click OK.

7
Linking CMOST with Petrel
A study contains all of the input information as well as the results for a particular task. Before
creating a study, you must have run the base dataset and have the simulation results of the
dataset.

17. Click on the New Study button to create a new study. Name the study PUNQ UA.
Define PUNQ_Base_RP.dat as the base dataset. Uncheck the checkbox for
‘Automatically create master datasetH’ since this file has already been created using
CEDIT. Select Uncertainty Assessment as the task type then click OK.

Any other study type could be selected (e.g. History Match, Optimization, etc.)

General Properties

18. Click the Browse button next to Master Dataset. Select the file PUNQ_Base_RP.cmm
that was created with CEDIT.

8
Linking CMOST with Petrel
Some files that are used with CMOST should be located in the study directory. CMOST may ask
to copy the files to that directory. Click Yes.

Fundamental Data
Fundamental data defines the plots that one would like to view from the
simulation results. Often most of the simulation results files are removed after
they have been analyzed by CMOST. Fundamental data saves some of the
simulation results information so that it can be reviewed even if the original
files have been removed.

This section will be skipped in this example.

Parameterization
In the parameterization section, we will select the variables in our simulation
model that we want to adjust. For this study, we have already created our
master dataset so we can import the parameter names directly from the file.

The import option does not read the parameters defined in Petrel
workflow(s) so you must insert these variables yourself.

19. On the Parameters page, click the Import button to load the parameters from the master
dataset. These will be the names of the include files.
20. The parameters that were used in Petrel workflow must be added separately. Click
Insert 3x and name the parameters RangeH, RangeV, & Azimuth. These names must
match the names specified in Petrel workflow without the $ character. The names are
also case sensitive.

9
Linking CMOST with Petrel
 21. Apply the following ranges to each of the new parameters and give each parameter a
Normal distribution:
Parameter Lower Limit Upper Limit Default Value
RangeH 1000 3000 2000
RangeV 15 35 25
Azimuth 10 30 20

For the output parameters, change their source to Formula and define the following formulas:

outputPermi:

StudyDirectory+"\\"+ExperimentName+"_Permi.inc"

outputPermk:

StudyDirectory+"\\"+ExperimentName+"_Permk.inc"

outputPor:

StudyDirectory+"\\"+ExperimentName+"_Por.inc"

10
Linking CMOST with Petrel
Where “StudyDirectory” is a predefined variable that refers to the location where the file is saved
and “ExperimentName” is the file name of the dataset that is created but does not contain a file
extension.

22. Save the study

Pre-Simulation Commands
The presimulation command will define the commands that will be sent to
Petrel. This will instruct petrel to run the workflow(s) that we defined earlier
that will regenerate the property arrays and save them as CMG include files.

23. Go to the Pre-simulations Commands page. Click Insert
Run Petrel Silently

Click the first Browse button and locate the path for the Petrel Executable (e.g. C:\Program
Files\Schlumberger\Petrel 2015\Petrel.exe)

Click on the second Browse button and locate the petrel file (.pet) that was accessed earlier in
the required data folder.

The petrel license profile defines which licenses will be used. This can be found in Petrel
through File
License Modules

11
Linking CMOST with Petrel
The workflow refers to the workflow that was defined earlier in Petrel. We previously named this
MyWorkflow.

If you a situation where you have multiple workflows defined, you can run
them all by selecting the “All” button next to Petrel workflow name.

Additional command line switches can be defined if necessary. We will leave this line blank.

In some instances, the block indices (cell numbering) are reversed between Petrel and CMG. If
this is the case, choose the option to reverse index directions. You will be required to define the
grid dimensions if this is the case. In this example, leave both checkboxes unchecked.

Click the Insert button under “Petrel workflow parameters”. Check the checkboxes for RangeH,
RangeV & Azimuth then click OK.

12
Linking CMOST with Petrel
Click the Insert button under “Result files from Petrel Workflow” and check the checkboxes for
outputPermi, outputPermk & outputPor.

The window should appear as follows when complete:

13
Linking CMOST with Petrel
Objective Functions
Objective Functions determine which results CMOST will analyze. The
selection will vary depending on the objective of the study (e.g. History Match
Quality for history matching studies). An arbitrary basic simulation result
objective function will be added to complete the required inputs for CMOST.

24. Go to the Basic Simulation Results page

25. Click on the top Insert button
26. Rename the objective function to GOR. Set the Origin Type to Groups and FIELD-PRO
for Origin Name and set the Property to Gas Oil Ratio SC

Control Centre
The Control Centre defines how to run the simulations and allows you to start
and stop the CMOST engine. You will be required to select which computers
to run the simulations on, the simulator version, number of processors, etc.

27. Click on the Engine Settings page. The study type should already be defined as
Uncertainty Assessment and the Engine Name should be set to Monte Carlo
Simulation using Proxy. Leave other settings as the default.
Note that the default action in an uncertainty assessment study is to remove simulation
files (including .inc files) after CMOST reads the results.
As an alternative, Monte Carlo Simulation using Reservoir Simulation can be useful
in determining cumulative probability percentiles (e.g. P10, P50, P90) for the realizations
that were generated.
Different information will appear if another study type was specified when CMOST study
was first created. Any study type can utilize this workflow.

28. Click on the Simulation Settings page. If you have set up multiple schedulers through
Launcher, they will be listed on this page. The ‘Local’ scheduler refers to your current
computer. Make sure that the scheduler Local has been set to Active by checking the
checkbox. Set the Max Concurrent Jobs to 4.

14
Linking CMOST with Petrel
 29. Make sure that the most recent version of IMEX has been selected.
Sometimes simulations may run into numerical problems and may take an extremely
long time to finish. CMOST can automatically kill simulations that take much longer than
expected to complete. Since our simulation should complete in a few minutes, we will set
the Max Run Time per Job to 1 hour.

There may still be a warning on the experiments table. This is because we introduced a new
parameters but haven’t yet specified which value has been applied in the base dataset. A
description of the warning can be found by clicking on the Validation tab at the bottom of the
window.

30. Go to the Experiments Table page. Right click on the first row of data and choose
Resolve Reuse Pending
Resolve Selected Experiment. Specify the default value
for each parameter:

15
Linking CMOST with Petrel
 31. This completes the required input for the CMOST study. Save the study.
32. Click on the Control Centre page. Click the button to start the CMOST engine.
CMOST will now begin to create simulation datasets and run them with the simulator.

As CMOST is running, Petrel is called to create new include files by running the workflow. A
dataset file is also created then run through the simulator automatically using Launcher. When a
simulation completes, the results are read by CMOST and saved.

Each dataset will have a different realization that is generated:

The dataset and include files for each case will be saved to the .cmsd folder. The options
selected on the control centre section of CMOST will determine how many datatsets and include
files are kept (e.g. default for history match is to keep best 5 sets of files)

16
Linking CMOST with Petrel
17
Linking CMOST with Petrel