Introduction To Well Calibration Using Zero-Offset Synthetic Modeling
Introduction To Well Calibration Using Zero-Offset Synthetic Modeling
Tutorial
Paradigm®
© 1986–2016 Paradigm B.V. and/or its affiliates and subsidiaries. All rights reserved.
The information in this document is subject to change without notice and should not be construed as a commitment by Paradigm B.V.
and/or its affiliates and subsidiaries (collectively, "Paradigm"). Paradigm assumes no responsibility for any errors that may appear in
this document.
The Copyright Act of the United States, Title 17 of the United States Code, Section 501 prohibits the reproduction or transmission of
Paradigm’s copyrighted material in any form or by any means, electronic or mechanical, including photocopying and recording, or by
any information storage and retrieval system without permission in writing from Paradigm. Violators of this statute will be subject to civil
and possible criminal liability. The infringing activity will be enjoined and the infringing articles will be impounded. Violators will be
personally liable for Paradigm’s actual damages and any additional profits of the infringer, or statutory damages in the amount of up to
$150,000 per infringement. Paradigm will also seek all costs and attorney fees. In addition, any person who infringes this copyright
willfully and for the purpose of commercial advantage or private financial gain, or by the reproduction or distribution of one or more
copies of a copyrighted work with a total retail value of over $1,000 shall be punished under the criminal laws of the United States of
America, including fines and possible imprisonment.
The following are trademarks or registered trademarks of Paradigm B.V. and/or its affiliates and subsidiaries (collectively, "Paradigm")
in the United States or in other countries: Paradigm, Paradigm logo, and/or other Paradigm products referenced herein. For a complete
list of Paradigm trademarks, visit our Web site at www.pdgm.com. All other company or product names are the trademarks or
registered trademarks of their respective holders.
Alea and Jacta software under license from TOTAL. All rights reserved.
Some components or processes may be licensed under one or more of U.S. Patent Numbers 6,765,570 and 6,690,820.
Some components or processes are patented by Paradigm and/or one or more of its affiliates. For a complete list of Paradigm patents,
visit our Web site at www.pdgm.com. All rights not expressly granted are reserved.
Introduction ...................................................................................................... 5
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Workflow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Step 1: Calculate the Initial Synthetic Seismogram ....................................... 9
Select the modeling method and input data . . . . . . . . . . ..... ... .. . .9
Examine the input logs . . . . . . . . . . . . . . . . . . . . . . . . . . ..... ... .. . 10
Define the output dataset . . . . . . . . . . . . . . . . . . . . . . . . ..... ... .. . 11
Calculate the initial synthetic and modify the display . . . ..... ... .. . 12
Modify the wavelet frequencies to match the seismic . . . ..... ... .. . 12
Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... ... .. . 15
Step 2: Calibrate the Synthetic using a Model Wavelet
(Bandpass/Gabor/Ricker) ............................................................................ 16
Define the correlation parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Apply a phase rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Shift the synthetic. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Fine tune the match using Stretch & Squeeze . . . . . . . . . . . . . . . . . . . 20
Decision Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Save the output dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Step 3 (Optional): Refine the Calibration using an Extracted Wavelet........ 26
Extract (estimate) a new wavelet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Save the output logs and wavelet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3
Paradigm®
Introduction
Overview
1 A synthetic seismogram is a seismic trace calculated from well data
(velocity, density and time-depth) and a wavelet. This document explains how
to use the acoustic convolutional modeling method to calculate and calibrate
zero-offset synthetic seismograms in the time domain.
Once the seismogram is calculated, it is inserted into the seismic section. You
then calibrate the synthetic seismogram to improve the match with the seismic.
Options for improving the match between the synthetic and the seismic include:
Change the input well logs used to generate the seismogram.
Modify the wavelet type, frequency range or phase used to create the
seismogram.
Shift, stretch or squeeze the seismogram. (Be careful withthis step.
Stretching and squeezing modifies the checkshot and may adversely
affect log velocity.)
Select different seismic traces in the vicinity of the well that might
provide better correlation.
Workflow
2 This document provides a step-by-step workflow for calibrating your well data to
the seismic data using zero-offset synthetic modeling. This workflow is divided
into two steps:
Step 1: Calculate the Synthetic Seismogram
— Select the Zero-Offset (Acoustic) modeling method and input data
— Examine the input logs
— Define the output dataset
— Calculate the initial synthetic and modify the display
— Modify the wavelet frequencies to match the seismic
— Decision Point: After performing the tasks described in this step,
examine the seismogram to see if it matches the seismic. If the answer
is yes, then your well is calibrated and you are done! If not, proceed to
step 2.
In this step you first examine the seismic frequencies you want to model, create
a wavelet and then modify the wavelet to match the seismic frequencies. If the
synthetic seismogram looks good, you can consider your well calibrated. If the
results are not good or you want to see if you can improve the match, you
should continue to step 2.
For help using Templates in the Well Log window, watch the
short video called “Working with Templates in the Well Log
window” in the SeisEarth > Displaying and Editing Wells
category in Paradigm Online University.
dataset
Selecting Output Data
1 Define the output dataset (Parameters tab > Output Data Types). [See table
from Online Help at the end of this step.]
Output data that is not selected will not be available for display or to save.
The default value for Synthetic Seismogram Format is “Seismogram
(convolved)” which results in the SYNT / Synt_Stack:1 - Convolved log
(reflectivity convolved with a wavelet). If you also want to display and/or
save the Reflectivity log, choose “Both Convolved and Full Band
Synthetic Seismograms.” This latter option creates both the convolved
log (synthetic seismogram) and the SYNT_FB / Synt_FB_Stack:1 - Full
Band (reflectivity) log (which has not been convolved with a wavelet).
The Residual log is a useful output because it shows you what seismic
responses are not captured in the synthetic. You can use this log as a
QC tool to help you improve the synthetic. The fewer the deflections in
this log, the more seismic data is captured by the synthetic.
Note Items selected in the Output Data Types panel are not automatically
saved to the database and are only available during the modeling session.
You must explicitly save the data to reuse it in another session.
Notice there is also an Extracted curve you can display in the Normalized
Amplitude Spectrum panel. This curve represents the data adjacent to the
well based on the cross-correlation parameters you define (you will do this in
Step 2). Use this curve also to help you estimate wavelet frequencies.
Without reflectivity information from the well data, the phase estimation can
be unreliable.
The option to include the phase estimation should only be used when you
know the geology and can selected a strong isolated event, such as the water
bottom.
Decision Point
6 1 Examine the seismogram in the Section window. Does it match?
— Yes! The well is calibrated to the seismic and this indicates that the T-D
relationship (checkshot) for the well is good. You do not need to save
the checkshot since it was not modified.
— No! You may need to apply a phase rotation to the wavelet, shift and
modify the checkshot and possibly extract a better wavelet to improve
the well to seismic match. These tasks are covered in the next steps.
Note You should not try to extract a wavelet using methods that
incorporate both well and seismic data (L1, L2, Amplitude-Phase Split)
until you complete Step 2 and have a fairly good match between the well
and seismic.
At this point you have not defined the parameters used to calculate the
cross-correlation coefficient and it is calculated using default parameters
which may not be appropriate for your data. This means you should not rely
on the Cor. Coef or Max Cor. Coef values displayed in the well table in the
Wavelet and Calibration tab. Your determination at this point is strictly visual.
To use correlation parameters to obtain a more quantitative understanding of
the calibration, you need to perform Step 2.
In the previous step you determined the input data and model wavelet for the
synthetic calculation. You also modified the frequencies of the model wavelet to
match the seismic frequencies. If the resulting seismogram does not match the
seismic you can perform several additional adjustments to improve the
calibration using the model wavelet:
Apply a phase rotation to the wavelet
Shift the checkshot
Stretch & squeeze the synthetic to match the seismic
Before performing these adjustments, you should define correlation parameters
used to define the region of well and seismic data to be used to calculate the
correlation coefficient. The correlation coefficient (and error values) can be used
to help you quantify the calibration results to ensure a good calibration.
Define the
1 1 In the Section window, toggle on the Cross-correlation window from the
correlation Display Elements tab.
parameters The Cross-correlation window shows in yellow the interval used to calculate
the correlation coefficient for the Synthetic Seismogram.
Note The Cross-correlation window does not determine the region used to
calculate the synthetic seismogram. You can define this region from the
Parameters tab > Synthetic Calculation Parameters.
2 Open the Well Log window from the Synthetics Utility. You can see the
correlation interval in yellow to the left of the track. (You may need to use the
horizontal scroll bar at the bottom of the Synthetic track, SYNT, to view it.)
3 In the Section window, use the Select tool to drag the top and bottom of the
cross-correlation window to match the range of the input log data or your
zone of interest.
When determining the time range for the Cross-correlation window, keep
in mind that the well and seismic data from this time range will be used to
calculate the cross-correlation value and will be used for the wavelet
extraction calculations (covered in Step 3).
Be sure not to include regions where well data is poor or missing, and be
careful when including regions where seismic data is noisy or has a lot of
structure at the well. Since the time range is used for wavelet calculation, it
should be at least three times the wavelet length (defined in well table).
The number of traces defines the trace at the well plus traces on
either side of the well in all directions. For example, if you select
3, then there will be a total number of nine traces.
The traces are averaged to create a single trace, identified as
the EXTRACTED seismic log which can be displayed in the Well
Log window (if it was selected in the Output Data Types list).
The Extracted seismic trace is used as input for the cross correlation and
wavelet extraction calculations. Selecting a higher number of traces will
create a smoother Extracted seismic trace and likely increase the correlation
value.
The amplitude spectrum of this extracted trace can also be displayed in the
Normalized Amplitude Spectrum panel.
Paradigm® 15.5® Step 2: Calibrate the Synthetic using a Model Wavelet (Bandpass/Gabor/Ricker) 17
Paradigm®
Apply a phase
2 1 Examine the symmetry of cross-correlation curve to determine if you are
rotation working with zero-phase data or non-zero phase data (example, minimum
phase, reverse polarity, etc.).
The shape of the cross-correlation curve shows how well the wavelet (in this
case a zero-phase bandpass) models the phase of the seismic data.
If the curve appears extremely asymmetrical then you should correct the
polarity of the wavelet by applying a phase rotation. This should improve the
cross-correlation values shown in the Well Table.
Zero‐phase wavelet with
Zero‐phase wavelet with 90‐degree phase rotated
zero‐phase seismic seismic
2 To calculate the phase of the seismic data using the zero-phase wavelet,
right-click in the Wavelet and Calibration tab and select Estimate Phase
Rotation of Wavelet.
Note You can test the sensitivity of the phase estimation in this area by
modifying the parameters for the cross-correlation window (see green
information box below).
3 Once you determine the appropriate phase, you can either automatically
apply the calculated phase or enter your own value and apply it to the
wavelet. The Phase column will be updated in the Well Table.
To automatically apply the estimated phase rotation to the wavelet:
a After estimating the phase rotation, click Apply to recalculate the
synthetic.
To apply a user-defined phase rotation to the wavelet:
a If you want to apply a user-defined value, use the slider bar at the
bottom of the Phase Spectrum panel (or enter the value in the text field).
b Click Apply to recalculate the synthetic with the user-defined phase
rotation applied to the wavelet.
4 Check that the cross-correlation value improved and also examine the
synthetic seismogram in the Section window and the Well Log window.
Applying a phase rotation will affect the time shift required to maximize the
correlation (Shift Max Cor.). This is why it is good practice to first estimate the
phase and then apply the shift.
Paradigm® 15.5® Step 2: Calibrate the Synthetic using a Model Wavelet (Bandpass/Gabor/Ricker) 19
Paradigm®
2 At the bottom of the Synthetics Utility, on the Wavelet and Calibration tab,
click the Shift to Max. Cross Correlation button.
Note Be careful to only apply the time shift if it seems reasonable (<~24 ms).
You can also enter a time shift manually into the Well Table and apply it.
Paradigm® 15.5® Step 2: Calibrate the Synthetic using a Model Wavelet (Bandpass/Gabor/Ricker) 21
Paradigm®
Decision Point
5 Determine if the well is now calibrated sufficiently for your purposes. Use the
following methods to help you with this determination:
— Visual examination of the synthetic against the seismic in the Section
window.
— Visual examination of the residual vs. seismic log in the Well Log
window.
Note Check that important reflectors are modeled in the synthetic and
don’t remain in the residual.
— Quantitative evaluation based on parameters in the well table
Cor Coef: >0.6. The higher the better (range is 0-1).
PEP: The higher the better. This measures the proportion of seismic
energy modeled by the synthetic (range is 0.05-1).
NMSE: The lower the better. This measures the error between the
synthetic and seismic (range is 0-1).
If the well looks calibrated, you can now save the output data (go to the next
task).
If you would like to further improve the match using an extracted wavelet,
calculate a more realistic wavelet for a workflow such as inversion, or are
struggling to get a good correlation and want to use a wavelet extracted from
the seismic, continue to Step 3.
Only data for the data types you selected in the Output Data Types panel of
the Parameters tab are available to save to the database. It is not necessary
to save them all. The logs you save will be written to the well database.
Table 1-1 Suggested uses for some of the zero-phase modeling output data.
Set Name (+
Log Name Suggestions for use
suffix)
TIMEDEPTH or TWTIME To display well data (markers and
CHECKSHOT logs, etc.) at the proper location
on the seismic data and to use in
any calculations which use both
well data and seismic data and
require domain conversion.
Paradigm® 15.5® Step 2: Calibrate the Synthetic using a Model Wavelet (Bandpass/Gabor/Ricker) 23
Paradigm®
New log sets are automatically created and listed in red. Newly created or
modified logs are also listed in red in the new sets. By default, these sets have
a suffix taken from the Unique set name suffix, the User-Defined Comment
and the Automatic Suffix at the bottom of the Input/Output tab of the
Synthetics Utility.
To create log sets with a user-defined suffix, toggle on Save As. Toggle on
Save As Suffix and enter the suffix you want. Select the data you want to
save and click Save. In the Well Data Manager you will find the logs saved to
a set name with the user-defined suffix.
If you toggle on Save As, you have several options to name the sets.
Automatic version advancement & Create New Logs in current
set. This option results in a new version of the log name in the same
set name that the original log was created in.
Save As Suffix & Create New Sets. This option results in a new set
name comprising the original set name plus this suffix. Use this option
if you don't want to use the Unique set name suffix that is listed at the
bottom right of the Synthetic Utility > Input/Output tab. [The Unique
set name suffix is generated by the software and cannot be changed
in the Synthetics Utility. Set names can be changed in the Well Data
Manager > Sets tab.]
Automatic version advancement & Create New Sets. This option
results in a new set name comprising the original set name and
unique set name suffix with a new version number.
Paradigm® 15.5® Step 2: Calibrate the Synthetic using a Model Wavelet (Bandpass/Gabor/Ricker) 25
Paradigm®
Once you have a good match between the synthetic and the seismic (Cor Coef.
> ~0.6), you can perform additional work to try to improve the wavelet using
information from the well and the seismic data. To do this you use wavelet
extraction.
Extracted wavelets allow a non-constant phase and are thus more realistic (they
model a more complex signal) than the model wavelets used in Steps 1 and 2. If
you plan on saving and using the wavelet for seismic inversion and other related
workflows, it is recommended to perform this optional step.
If your purpose is just to tie the well data to the seismic so that the well data
displays properly on the seismic time sections, then this step is not necessary.
Before using any wavelet extraction methods which use both well and seismic
data (L1, L2 and amplitude-phase split), confirm that the correlation is
sufficient to obtain a reliable result (Cor Coef. >~0.6).
Note The Seismic Only (Zero-Phase) wavelet extraction method does not
include well data and therefore does not require a good correlation between the
synthetic and seismic. If you are having difficulty obtaining a good correlation
using a Model wavelet in Step 2, try using the Seismic Only extraction method to
calculate a wavelet. To do this:
a Reset the checkshot (right-click on the synthetic seismogram and select
Reset) to remove the stretch & squeeze and time shift applied to the
checkshot in Step 2.
b Right-click on one of the panels in the Wavelet and Calibration tab and
select Automatic Wavelet Extraction > Seismic Only (Zero Phase).
c Recalculate the synthetic using the original checkshot and seismic only
wavelet.
d Apply a phase rotation, time shift and stretch & squeeze to improve the
match.
Extract (estimate) a
1 1 Right-click on one of the panels in the Wavelet and Calibration tab and
new wavelet select Automatic Wavelet Extraction to display the estimation methods.
Select one of the following methods:
— Amplitude-Phase Split
— General (L1)
— General (L2 “Roy White”)
Note The Amplitude-Phase Split (Global) method should be used only
when you are trying to find a single global wavelet from multiple calibrated
wells. This is usually part of a seismic inversion workflow and is not covered
in this document.
Notice that the Wavelet changes and the phase spectrum of the wavelet is
displayed in the Phase Spectrum panel. For a perfectly zero-phase wavelet,
you would expect a straight horizontal line at zero indicating zero phase for all
frequencies. Phase adjustment for different frequencies is usually required to
get the synthetic to match the seismic.
2 Click Apply to recalculate the synthetic seismogram with this new wavelet.
3 Examine the synthetic seismogram against the seismic in the Section
window. Examine the seismogram and the residual in the well log window
and notice that the Max Cor. Coef. and the Cor. Coef values in the well table
change.
Note If the Max Cor. Coef. and the Cor. Coef values are different, it indicates
that if you use this wavelet you will need to apply another shift to obtain the
best correlation. The shift to apply is indicated in the Shift Max Cor. column of
the table. You can also see it in the Cross Correlation panel.
4 Repeat steps 1 and 2 in this task to see the effect of different wavelets on
the synthetic. Determine which extraction method provides the highest Max.
Cor. Coef. value. Use the correlation value and an examination of the
wavelet itself in the Wavelet panel to determine which wavelet is the best for
your needs.
Note You may find that you get a very good correlation but the wavelet has
strange sidelobes and phase changes across the different frequencies. In
this case, you would not select this wavelet, even though it has a high
correlation value.
5 Once you select the best wavelet, click the Shift to Max. Cross Correlation
button to adjust the checkshot to maximize the correlation.
Paradigm® 15.5® Step 3 (Optional): Refine the Calibration using an Extracted Wavelet 27
Paradigm®