eLog Guide

Using the eLog Program ................................................................................................................. 1

December 2006
 Guide to eLog
Using the eLog Program

The eLog program is used to perform manipulations on the logs such as editing, smoothing, and
log correlation. In this section of the guide, we will use eLog to modify some of the logs that
have already been entered into the database.

To start eLog, click the eLog button on the GeoView main window. A window appears,
allowing you to start a new project. Click OK. On the file selection window, call the new
project elog_guide as shown below:
2 eLog

Click OK, and the eLog main window appears:

At the same time, a window appears asking which well we wish to edit:

December 2006
 eLog 3

On this window, select strata_well and click Open.

The eLog window will now contain the log from strata_well:

The eLog window looks very much like the Log Display window from GeoView, with the
addition of a series of editing buttons down the left side.

December 2006
4 eLog

We will do a series of operations on the P-wave log from strata_well. The first thing we will do
is to apply a median filter to the log to remove the high-frequency noise. To do this, click the
Math button. When this window appears, select the Median Filter option, as shown:

The median filter is a math operation using one or more input logs, which will come from
strata_well:

December 2006
 eLog 5

On the following page, click Next >> to accept the defaults because we will write the median
filtered log back to the same well:

Now, select the log to median filter, which will be P-wave_1 from strata_well:

Call the output sonic log P-wave_math as shown, and click Next >>.

December 2006
6 eLog

The window now shows the parameters for this option. Change the Operator Length to 9, as
shown, and click Next, then OK.

When the process has completed, the new filtered log appears:

By default, only the latest P-wave log appears in the eLog window. In order to see both logs, we
need to temporarily modify the display parameters. To do that, click the “eyeball” icon:

December 2006
 eLog 7

On the Layout page of the Parameter window, the important parameter is the option Display
Only Active Logs:

If there is more than one log of a particular type (for example, the P-wave logs in this well), one
of them is selected as having a higher priority. This is called the “Active Log”. By definition,
whenever a process is applied which creates a new log, such as check shot correction or median
filter, the latest log is automatically set to be the Active Log. To see the list of active logs and
possibly modify that list, click the Active Logs button.

The resulting window allows you to define the priority yourself:

We are happy with this priority, so cancel this window.

December 2006
8 eLog

Since we want to see both of the P-wave logs (before and after median filter), we must allow
display of multiple curves by unchecking the Display Only Active Logs toggle:

Check the box for the P-wave_1 log as shown to display the original log adjacent to the median
filtered version:

Click Apply at the bottom of the Parameter window and both logs are displayed in the eLog
window.

December 2006
 eLog 9

From this display, we can see that the median filter has removed much of the high-frequency
jitter from the sonic log. For the rest of the processes, we will be happy to see only one log, so
click Cancel on the Parameter window to reset the eLog window.

The next process we will apply is to create a density log using Gardner’s equation. To do that,
click Transforms and modify the first page as shown below (Select Gardner’s Equation under
the item Select desired operation):

Then click Next >> and modify the second page to select strata_well:

December 2006
10 eLog

Click Next >> to display the third page. The default selection causes the output log to be written
into strata_well:

After clicking on Next >>, the fourth page shows that P-wave_math will be used as the input log.
We will change the default name for the output Density log to Density_gardner:

Then click Next >> to see the transform which will be applied:

December 2006
 eLog 11

Finally, click OK and the density log will be created:

Reading a Seismic File into eLog

Now we will perform the operations of wavelet extraction and log correlation. To do this, we
must first read an external seismic volume.

To start this process, select Data Manager>Import Data>Open Seismic>From SEG-Y File:

December 2006
12 eLog

On the File Selection page, select the file ersk3d.sgy, which will be found in the elog_data
directory.

Click Next >> to get the next page.

The file is a 3D seismic volume with a rectangular grid of inlines and cross lines. This page of
the window defaults to a 3D volume:

After you click Next >>, you see the third page, which asks about the information that is
available in the trace headers:

Note that eLog is capable of reading the Inline & Xline numbers and/or the X & Y coordinates
from the trace headers, but since neither of these is present, we will load the seismic by keying
on the CDP number. In order for this to work, the seismic data must form a uniform, rectangular
grid. For this file, this is the case.

After modifying the window as shown above, click Next >>.

December 2006
 eLog 13

The SEG-Y Format page appears:

This page is where you specify various byte locations for parameters from the trace headers. The
default values are the standard SEGY locations. Because we have specified that the geometry is
Rectangular, the only important locations for this data set are the CDP and Offset locations, as
shown above. Click Next >> on this window.

If you have never read the file ersk3d.sgy before, you will see the following message:

This is because eLog must scan the file to determine general information from the headers. If
you have read the file at least once before, you will not see this message. Click Yes to continue.

December 2006
14 eLog

The final window appears:

Because we have specified the geometry as “Rectangular”, we must now specify the Geometry
layout. For this purpose, we simply need to tell the program that there are 155 Cross-lines. The
number of Inlines will then be calculated automatically. The window will now look like this:

Since we will be using this data set only for log correlation and wavelet extraction, we will not
bother to fill in detailed X & Y coordinate information. Click OK on this window.

December 2006
 eLog 15

When the file has been read, two new windows appear. One window contains the seismic data:

The second window contains this list of the wells in the database, and their location within the
seismic grid:

December 2006
16 eLog

At this point, the Inline and Xline numbers at which the wells have been positioned are not
correct. In order to do that, we would have to specify the correct X & Y locations of the wells in
the GeoView database as well as the correct X & Y geometry of the seismic grid. Since we will
be correlating strata_well, we only need to specify the position of that well correctly.
AVO_WELL does not tie this data set. The strata_well is located at Inline 24 and Xline 75.
Type those numbers in, click in another field, and the window will be updated to look like this:

Click OK.

The Seismic data window now shows a portion of Inline 1. To look at the data near the well,
type the number 24 in the Inline box at the top and click Enter:

December 2006
 eLog 17

After moving the scroll bars, you can see the sonic log from strata_well positioned properly
within the volume:

The seismic data window has many useful viewing capabilities. For now, we will return to the
eLog main window, to continue with the log editing session.

Now that we have loaded the seismic data, we will proceed with Log Correlation. Click the
Correlate button:

December 2006
18 eLog

This window appears:

Log correlation is the process of aligning the synthetic calculated from the well logs with one or
more seismic traces near the well location. In eLog this correlation is always done with a
"Composite Trace”. The Composite Trace is an average of adjacent traces around the borehole
location. If the well is deviated, the averaging follows the deviation path. The window above
determines how the averaging will be done. By selecting the default parameters, we are
averaging traces within +/- 1 inline and crossline of the borehole – in other words, 9 traces.
Click OK to accept this default.

December 2006
 eLog 19

The eLog main window now looks like this:

Log correlation can be thought of as a type of check shot correction, where the depth-time pairs
are provided manually by selecting points on the synthetic and tying them with corresponding
points on the composite trace. The trace shown in blue is the synthetic trace calculated with the
most recently defined wavelet. To see that wavelet, select Wavelet>Display Current Wavelet.

December 2006
20 eLog

We see that it looks like this:

A better wavelet would be one that was extracted from the seismic data. To do this, select
Wavelet>Extract Wavelet. We see that there are two options:

These are “Use Well” and “Statistical”. In order to get the correct phase of the wavelet, we
would prefer to use the well. Because we have not done log correlation yet, the mis-tie between
log and seismic would make that extracted wavelet unacceptable. For this reason, we start by
extracting a statistical wavelet, which will be a zero-phase wavelet with the same amplitude
spectrum as the seismic data.

December 2006
 eLog 21

Select Wavelet>Extract Wavelet>Statistical:

In the window shown above, we have selected a time window of 700 to 1200 ms and limited the
analysis to a range of traces around the well.

December 2006
22 eLog

Click Next >> twice to get the last page:

In this case, we will accept the default values. Click OK to get the extracted wavelet:

December 2006
 eLog 23

The eLog main window has now been redrawn to contain a synthetic created from this extracted
wavelet. We can see that the character tie has improved significantly:

Obviously, there is a time shift between the blue synthetic trace and the red composite trace.
The first step is to correct this shift. Click near the peak on the synthetic at about 750 ms and the
corresponding peak on the composite trace:

Then click Stretch at the bottom of the window.

December 2006
24 eLog

The synthetic has been aligned using the picked values:

The calculated correlation coefficient at the bottom of the window measures the fit
mathematically:

Obviously, the fit is still not good enough, and this is because there is a time-variant stretch
required. The next step is to pick a series of points:

Click Stretch.

December 2006
 eLog 25

This time the check shot windows appear:

Whenever there is more than one point in the correlation picks, this is interpreted as a type of
check shot, and these windows appear.

To accept this correlation, click OK on the Check Shot Analysis window.

December 2006
26 eLog

The synthetic trace now shows a much-improved correlation:

with a correspondingly larger numerical correlation value:

Now that the log has been correlated better, we can extract a new wavelet using the log. Select
Wavelet>Extract Wavelet>Use Well:

December 2006
 eLog 27

On the first window, select strata_well as the one to use for wavelet extraction:

Click Next >>.

On the next page, set the time window from 700 to 1200 ms:

December 2006
28 eLog

Click Next >> twice. On each subsequent page, accept all the defaults.

Click Next>> to see the history which will be saved with this wavelet:

December 2006
 eLog 29

December 2006
30 eLog

Finally, click OK to get the extracted wavelet:

Click the Frequency tab on this window to see its frequency spectrum:

December 2006
 eLog 31

The eLog window shows an improved correlation with a larger numerical correlation of about
0.86:

Now that the correlation has been completed, click OK at the bottom of the eLog window. This
window appears, showing you the name of the name correlated sonic log:

Click OK on this window to accept the suggested name.

December 2006
32 eLog

Saving the Edited Logs

We have now completed the editing of the well log. At this point, the changes are contained
only within the eLog project. This means that the GeoView main window still contains only the
original logs. We can still choose to discard these edits if we wish. To save both the eLog
project and the edited logs, select Project>Save. The following question appears:

We want these changes to be stored permanently in the GeoView database, so the answer to this
is “Yes”. This list appears, showing all the new logs that have been created and the names under
which they will be saved:

Click OK on this window to save all the new logs. Go back to the Well Explorer window, and
you can see that the new logs P-wave_math, P-wave_corr and Density_gardner have been
inserted into strata_well.

December 2006
 eLog 33

The Crossplot Option in eLog

We will now look at the Crossplot option in eLog. To do this, we will first display AVO_WELL
from the database. To do this, click Open Well to get the list as shown:

Select AVO_WELL and click OK.

December 2006
34 eLog

This will load AVO_WELL into the eLog window. The resulting window looks like this:

Next, select General from the Crossplot sidebar menu to bring up the Well Log Crossplot
window.

December 2006
 eLog 35

Select <Measured Depth> from the list of Available types and click Add >>. Repeat these
steps to select P-wave, Density, Gamma Ray, SP, and Resistivity so that the window appears as
shown:

Note that <Vertical Depth> will be selected automatically since the Processing domain type is
the “Vertical Depth Domain”.

December 2006
36 eLog

Click Next >> to get the second page of the Well Log Crossplot window, which allows you to
choose the wells. Fill in this window as shown below, to choose AVO_WELL:

Click Next >> to get the third page of the Well Log Crossplot window, which defines the
curves that will appear on the X and Y axes.

Modify the window to plot the P-wave log on the X-axis and the Density log on the Y-axis. The
completed window looks like this:

December 2006
 eLog 37

Click Next >> to get the next page, which allows you to choose among logs of the same log type
in addition to selecting the amplitude units.

Accept the defaults by clicking on Next >>.

The next page of the window allows you to select the depth range for cross plotting. Fill in a
Start Depth of 550 m and End Depth of 700 m, as shown below:

December 2006
38 eLog

Click Next, then click OK, and the following cross plot will appear, where Density, in g/cc, has
been cross plotted against P-wave travel time, in μs/m:

The default color key shows Vertical Depth. If we wish, we can use one of the other logs to
create the color bar, creating a type of 3-dimensional plot. To do this, select New Plot>P-wave
vs Density (primary), to get the Data Specification window.

Fill in the window as shown below, by changing Item # for Color to 6 (notice that this is the
Column number of the Gamma Ray curve from the list in the top half of the window):

December 2006
 eLog 39

Click OK to get the same cross plot with a new color scheme.

Now, select New Plot>P-wave vs Density (primary) and change the color key back to Vertical
Depth by changing Item # for Color back to 5.

December 2006
40 eLog

Now, select Transform>P-wave vs Density (primary). Fill in the Transform Data window as
shown below, by changing the X and Y axis values to Log(x) and Log(y):

December 2006
 eLog 41

Click the OK button to bring up the transformed cross plot:

We will now perform a regression analysis on the transformed axes.

December 2006
42 eLog

Select Regression>Least squares, to bring up the following window:

Accept the defaults by clicking on OK.

December 2006
 eLog 43

The resulting cross plot has the linear regression line drawn on it, and the equation is shown in
the legend.

Next, click Cross-Section to produce a display of the two log curves, P-wave and Density. Use
the View>Zoom option or the magnifying glass icon to zoom in on the curves:

December 2006
44 eLog

We next want to select certain zones on these curves from clusters on the cross plot. To do this,
select Zones>Add to get the Add Zone window shown here:

The first color is a dull gray. To change this to a brighter color, like red, click once on the color
itself to bring up the color editor, shown next (the appearance of this window will depend on
whether you are running the program on a PC or under Unix):

December 2006
 eLog 45

December 2006
46 eLog

On the color editor, click once on the red color (second down from the top left corner), and you
will see the Color/Solid box change to red. Click OK to save this color. Now, go back to the
cross plot itself. Following the instructions at the top of the Add Zone window, draw an ellipse
around the points as shown. Click Apply on the Add Zone window to save zone1, as shown
here:

Note that your ellipse may be slightly different. Now, add a second ellipse on the set of points to
the right of the first ellipse. Click Apply again to get something similar to the next plot:

December 2006
 eLog 47

Now that the two ellipses have been drawn, look at the two well log traces again. Notice where
the elliptical zones are present on this plot, as shown below:

We will choose to keep only the points in the second zone. To do this, select Filter>Filter
points, and fill out the window as shown below, by choosing Remove Points Outside Zone
and adding zone2 to the Selected List:

December 2006
48 eLog

Click OK to filter the points, and the result will look like this:

Now, recalculate the regression fit, using the Regression>Least squares option and clicking
OK to choose the defaults. The new fit looks like this:

December 2006
 eLog 49

What is the significance of the two linear regressions that we performed on the cross plotted
data? Note that the first was done on the overall dataset, and the fit did not appear to have a high
correlation value. The second was done on a more restricted set of points, and had a better
looking fit. Now, recall Gardner’s equation for the fit between density and velocity:

ρ = aV b , where ρ = density
V = P-wave velocity
a, b = constant values

Although the above equation is nonlinear, we can linearize it by taking the logarithm:

log(ρ) = log(a) + b log(V)

Now you can see why we applied a logarithmic transform to each axis. However, we also used
transit-time rather than velocity units, where:

1, 000, 000
V= , Δt = transit-time in μs/m
Δt
or
log(V) = log(1,000,000) - log (Δt) = 6 - log (Δt)

Substituting gives us:

log(ρ) = log(a) + 6b – b(log (Δt) )

Let us write the least-squares fit from the cross plot as:

log(ρ) = c(log (Δt) ) + d

Note that your c and d values will probably not be the same as here, since your elliptical region
will be slightly different than the one shown. Equating the a and b values from Gardner’s
equation with the c and d values from the program, we get:

b = -c

a = 10^[d - 6b] = 10^[d + 6c]

The regression line on the cross plot above was as follows:

y = -0.444688x + 1.51878, where y = log(ρ)

December 2006
50 eLog

So we have:

c = -0.444688, d = 1.51878
and therefore:
b = 0.444688

a = 10^[1.51878 – 2.66813] = 10^[-1.14935] = 0.0709

Thus, we can write Gardner’s equation as:

ρ = 0.071 V0.445

To perform this transform, go back to the eLog window containing AVO_WELL and click
Transforms. Fill out the Transforms window as shown below (i.e. apply Gardner’s Equation):

December 2006
 eLog 51

Click Next >>, and fill out the next window as follows, by selecting AVO_WELL:

Click Next >>, and fill out the window as follows, selecting AVO_WELL as the Output Well:

Click Next >>, and fill out that window as follows, selecting P-wave_chk to be the Input
P-wave log and Density_Gardner as the Output Density log:

December 2006
52 eLog

Click Next >> and fill out the final window with the computed coefficients, 0.071 and 0.445, as
shown below:

Click OK and the resulting density log will look like this:

December 2006
 eLog 53

This is the end of the GeoView & eLog tutorial. To close down the eLog program, select
File>Exit Project on that window.

You will be prompted to save your project. If you want to review the tutorial data later, click
Yes:

December 2006