Function window cross plots

Petrel 2012 and beyond releases

Page 1 of 13
Introduction

The objective of this document is to provide the user a simple and comprehensive overview of the function
window functionalities.

By following the step by step workflow procedure, the beginner will be able to create simple to more
complex cross plots in Petrel function window.

This procedure is illustrated with examples that use log data. The steps can be extended to the other data
type listed at the beginning of the document.

Table of Contents

Function window .................................................................................................................... 3

Which data type ...................................................................................................................... 4
Cross plot of Logs .................................................................................................................. 4
Example ...................................................................................................................................................... 5
Resampling log workflow steps...................................................................................................................... 6

Linear regression .................................................................................................................... 8

Filter Management .................................................................................................................. 9
RawCrossplot ............................................................................................................................................ 12

Page 2 of 13
Function window
In Petrel, the Function window is the plot frame to use to do cross correlations. To do so, go to the menu Window >
insert a new Function window.
The new Function window should open automatically in Petrel interface. If it doesn’t, it can be selected from the
Windows Pane – Cf. Picture below.

In the Windows Pane, the Function window object has different features attached to it – they can be accessed by
expanding the Function window structure. Each feature can be selected or deselected according to the user
preferences and has its own Settings as every other Petrel object.

Note: From the picture below, you can see that the Settings of the Function 1 viewport shows common
features with the Function tree showed in the previous Picture, i.e. Axis, Header & Legend.

Page 3 of 13
Which data type
Different data type can be used to create cross plots
- Logs
- Well Attributes
- Log point sets’ attributes
- Well Tops sets attributes
- Surfaces objects
- Functions objects
- 3D model properties

Cross plot of Logs

Once the function window opened, go to the Input Pane > Expand the main Wells folder > Expand the Global well log
folder > select the logs you want to correlate by using the tick box.
Note: the first log selected will be marked as X and the second one as Y. If a third log should be added as
variable, Petrel will recognize it as Z.
In order to see the correlation in the Function window, the boreholes of interest must be selected one by one from the
Wells folder tree or selected all at once if all of them are part of the study.

The previous picture shows a correlation of a porosity log with a gamma ray log, for two boreholes only, i.e. B4 & B9.

Page 4 of 13
Important: It’s not rare to have logs with different sample rate. In this case, the sample rate of the log
selected as X is used by Petrel as being the base of the cross plot. Thus, in the background, the log Y is
resampled to be able to be correlated with the X log.
This Petrel behavior can lead to very different correlation according to the order of selection of the logs to be
correlated – as showed in the example below.

Example

Two logs – log Total Clays & IGR such as follow. Total Clays log has very few points compared to IGR; its sample
rate is less precise than IGR’s one.

As a result, when the Total Clays log is selected as X, the number of correlation points with IGR log in the Function
window will be restricted to the number of sample points of the Total Clays log. In this example, Total Clays has only
40 sample points.

On the other hand, if IGR log is selected as X, the number of correlation points with Total Clays log in the function
window will be higher due to the fine sample rate of the IGR log. In this case, the Total Clays log is being resampled
in the Background by Petrel to be able to correlate the two logs.

Cf. Following picture.

Page 5 of 13
 Resampling log workflow steps

In order to avoid any misinterpretation between the logs depending on their order of selection, it is always possible to
proceed with a resampling operation.
For instance, we could want to resample the Total Clays log by using the sample rate of the IGR log. To do so:
- Right-click on the Global well log folder > select Calculator > the Global Well log calculator should pop up.
- In the Formula bar, create a Total Clays resampled log such as the picture below.
- Set the appropriate template for the output log in the Attach new to template field.
- In the Sample MD field > choose from log option and select the IGR log from the pull down menu list.
- Press the Enter button to run the process.

Page 6 of 13
When logs have the same sample rate, no matters the order of selection: the number of points in the function window
will be identical, as showed in the Picture below.

In order to visualize the correlation point that are in a specific zone – between two markers – it is possible to create a
zone log from the Well tops set (Right-click on the Well Tops set > Insert/update zone log > zone log is stored in the
Global Well log tree) or any other discrete log as soon as the zone of interest is referenced correctly.
In this example, the zone log is used as the third variable and marked as Z. Cf. Picture below.

The remaining grey points are the correlation points outside of the zone of interest.

Page 7 of 13
Linear regression
When the Function window is open, the function bar is updated with appropriate tools icons. Two main icons are

provided and allow the interpreter to create either linear regressions or nonlinear regression curve from
the cross plot window.
Note: unless a filter is activated, the regression functions are created using all the cross plot points visible.

In the case of a linear regression, when clicking on the icon, the following window should pop up automatically
to set up some parameters of the function to be generated.

Information preserved in the

function settings > Info tab > Comments sub tab

Function X range preserved in the

Statistics tab of the function object

The function object created is stored at the bottom of the Input Pane tree. By the time the function is generated, it
should be automatically plotted in the function window.

The function can be edited:

- Graphically in its Settings > function tab
- Via the spreadsheet associated to the function object (Right-click on the function > Spreadsheet)

Page 8 of 13
The nonlinear function creation follows the exact same steps but needs to use the icon in the function bar.

Filter Management
In Petrel it is possible to create filters from the function window. In order to do so, Petrel function bar provides
different filter type tools, as followed:

- Select using 2D rectangle filter

- Select using freehand draw

- Select using 1D range on X axis

- Select using 1D range on Y axis

Every filter created is stored in the Input Pane tree > Filter folder > User sub folder. They can be renamed and
customized from individual Settings.

Page 9 of 13
The filtered points can be converted in a Boolean log. To do so, right-click on the filter in the filter folder > select the
convert to Boolean log option > the Boolean log created is stored under the Global well log tree.

Note: The Boolean log is a discrete log with two ‘facies’ codes, i.e. True (1) where the points are included in
the filter, False (0) where the points are outside the filter. The Boolean filter can be used for further log
operations and can be plotted in a well section window.

Page 10 of 13
Multiple filters can be combined together as shown in the following picture. They need to be selected from the Input
Pane.

Once the filtered points have been selected, it is possible to create a linear regression or nonlinear regression curve
over only these points – Cf. Linear regression previous part.

Page 11 of 13
 Raw Cross plot

Filter combination can be saved as one independent point cloud object in Petrel Input Pane tree: it is a Raw Cross
plot object.

Here are the steps to be followed to create a Raw Cross plot:

-Filter folder > Settings of each of the filter selected > Style tab > select the Remove points outside filter option >
Apply. The non-filtered points should be removed from the display as shown in the following picture.

The function bar provides a tool called ‘Create raw crossplot’ and presented by the icon . By clicking on this tool,
a window should pop up and ask the user to rename the raw crossplot being created. Cf. following picture.

Page 12 of 13
The new raw cross plot is stored at the bottom of the Input Pane tree. It is completely independent from the Wells
data selected as well as from the filter selected. It has its own settings and any further operations can be performed
using it.

Linear regression analysis and/or non-linear regression analysis can be performed as described previously in this
document.

Page 13 of 13