Details SeisWare - Velocity as a Background |
This way takes a few more steps. Velocity segy files are normally large in size and very few of us will actually store them. What we typically store is a card deck containing the velocity information (time rms pairs). I'm going to show how to take a velocity card image file and create your own velocity segy file! This takes a few steps with the help of BP's FreeUSP software program velin. It takes a time rms velocity card deck and interpolates these data to fill in all time samples with a velocity values!
Here are the real steps that need to be taken:
- Reformat the velocity tables into a suitable format for the USP program velin
- Convert the SeisX/SeisWare segy file into USP format
- Change the USP binary and trace headers from one record of 679 traces to 679 records of one trace (line is 2d).
- Create USP velocity velocity file exactly the same size in traces and samples as the stack file above
- Convert the USP velocity file back to SeisWare format with perl script usp2sx.pl
- Add shotpoints and XY's back into the velocity file with SeisWare
VELOCITY INFORMATION (FROM PROCESSING DATUM)
CDP # # PAIRS
TIME | RMS VELOCITY
40
9525 16
0.0489860 2920.103
0.2948324 3136.598
0.4535310 3362.114
0.6117027 3706.842
0.8076851 3996.667
0.9624081 4026.809
1.1377608 4103.493
1.2724162 4228.093
1.4014084 4352.941
...
I then looked at the documentation for
USP program velin and
I noticed that the -flat format was the easiest to reformat the data into.
All I have to do is covert velocity in seconds to milli seconds, add some commas... I wrote an awk script
to do this called vel_cnc2flat.ak. It's quite simple and is only
a fel lines of code. Here's how to run it:
awk -f ~ekeyser/ak/vel_cnc2flat.ak < cnc.txt > cnc_vel.txtThe input and output files look like:
- Input file cnc.txt
- Output file cnc_vel.txt
2 Convert the SeisX/SeisWare segy file into USP format
Create a text file containing the names of the SeisX files you want to convert using program sx2usp.pl. For only one file I will use the simple unix command:ls orig.MIG.0.sgy > junk ~ekeyser/Perl/sx2usp.plFor many files I execute the following unix commands and scripts:
ls *sgy > junk ~ekeyser/Perl/sx2usp.plHere's the output from this program:
~ekeyser/Perl/sx2usp.pl
-Working on orig.MIG.0.sgy
1+0 records in
1+1 records out
Big Endian si=2000 ns=1501 format=1 start_x=0 start_y=0 end_x=679 end_y=0 sxflag=92
Line Trace SPID CDP UTMX UTMY
Data not SeisX -- cannot re scale
0 0 670.25 9405 513261 6191324
0 0 670.50 9410 569132 6113238
0 0 670.75 9415 561320 6132952
0 0 671.00 9420 569750 6113266
0 0 671.25 9425 513280 6191321
0 0 671.50 9430 569132 6113295
0 0 671.75 9435 561321 6113210
0 0 672.00 9440 569132 6113224
0 0 672.25 9445 561320 6113239
-fini
Note the sxflag of the number 92, this means the data is SeisX 4.0 format. The Big Endian label means
the data came from the unix side of things. Little Endian means it came from the Intel (PC) side. Either
way the conversion program works. Next step is to fire up a USP Xterm (Window) environment shell
and check the data. Here's how we do this at EnCana.
/usr/global/startup/usp.kshThis opens a window. Let's check that the man pages are working. Lets check out the documentation for probram scan we use to check out data.
bash-2.03$ uman scan
User Commands scan(1)
NAME
scan - module to scan traces
SYNOPSIS
scan [ -Nntap ] [ -V ] [ -v ] [ -c ] [ -t ] [ -V ] [ -H
] [ -S ] [ -nsns ] [ -nene ] [ -rsirs ] [ -reire ] [ -npnp
] [ -iidec ] [ -bnbin ] [ -wihdrwd ] [ -Wword ] [ -R ] [
-Bbyte ] [ -Llength ] [ -M ] [ -mt ] [ -Ffmt ] [ -? ]
DESCRIPTION
scan allows a scan of the contents of the SIS data file,
providing all sorts of information concerning its various
contents including the historical line header. This infor-
mation under certain conditions takes the form of mnemonic
references to header values. A list of the line and trace
header mnemonics is proovided at the end of this manual
page.
scan gets processing controls from the command line. Reason-
able defaults are set up. In addition, if no file name is
given, the binary trace data are expected to be on the stan-
dard input.
Command line arguments
-N ntap
Enter the full path of the file containing the data
set. If not specified, input is expected on the stan-
dard input. If standard input is not specified, and
there is no input, e.g., program run in background,
expect a very ungracious crash. (default standard
input)
This is a very important program, read the documentation as it contains
a list of all header words.
scan -Norig.MIG.0.usp | more
line header values after default check
# of bytes in line header= 1004
# of samples/trace = 1501
sample interval = 2
s.i. multiplier = 1.00000E-03
traces per record = 679
records per line = 1
format of data = 3
luin = 2
input data set name = orig.MIG.0.usp
ns (starting trace) = 1
ne (ending trace ) = 679
rs (starting record) = 1
re (ending record) = 1
Group interval =
OAC Line = orig
Number ampl. histogrm bins= 10
NOTE: a bin = (trcmax-trcmin)/nbin
verbose output = F
rec trc gi di si_x10 src_pt sp/di li dist static
RecNum TrcNum RecInd DphInd SrcLoc SoPtNm SrcPnt LinInd DstSgn StaCor
1 1 0 9405 4702 470 470 0 0 0
1 2 0 9410 4705 470 470 0 0 0
1 3 0 9415 4707 470 470 0 0 0
1 4 0 9420 4710 471 471 0 0 0
1 5 0 9425 4712 471 471 0 0 0
1 6 0 9430 4715 471 471 0 0 0
Our data is converted properly but we have to
create a different record for each trace. We will now correct this.
3 Change the USP binary and trace headers from one record of 679 traces to 679 records of one trace
Use program utop to correct the binary (line) header first and then use program rnum to renumber our data. Use scan again to verify the data. Notice we are piping the output of utop directly into rnum and saving the output in the file corr.MIG.0.usp.
utop -Norig.MIG.0.usp -L 1 -R 679 |\
rnum -Ocorr.MIG.0.usp -hwRecNum -sw1 -iw1 -R
scan -Ncorr.MIG.0.usp | more
line header values after default check
# of bytes in line header= 1097
# of bytes in output line header= 1120
# of samples/trace = 1501
sample interval = 2
s.i. multiplier = 1.00000E-03
traces per record = 1
records per line = 679
format of data = 3
luin = 2
input data set name = corr.MIG.0.usp
ns (starting trace) = 1
ne (ending trace ) = 1
rs (starting record) = 1
re (ending record) = 679
Group interval =
OAC Line = orig
Number ampl. histogrm bins= 10
NOTE: a bin = (trcmax-trcmin)/nbin
verbose output = F
rec trc gi di si_x10 src_pt sp/di li dist static
RecNum TrcNum RecInd DphInd SrcLoc SoPtNm SrcPnt LinInd DstSgn StaCor
1 1 0 9405 4702 470 470 0 0 0
2 2 0 9410 4705 470 470 0 0 0
3 3 0 9415 4707 470 470 0 0 0
4 4 0 9420 4710 471 471 0 0 0
5 5 0 9425 4712 471 471 0 0 0
Yes, that is what we want, each trace has a different record number.
3 Create USP velocity velocity file exactly the same size in traces and samples as the stack file above
| |
Program velin will create a velocity file of the same dimensions as the file corr.MIG.0.usp and place the output in the file vel.MIG.0.usp.
Comments are V for Verbose and the time, rms velocity file is read from the text file cnc_vel.txt.
The data format is the -flat format and our velocities are indexd by cdp depth index (di) DphInd.
Note that the cdps are not required to be sequential! (SeisWare does require this)
velin -N corr.MIG.0.usp -Ovel.MIG.0.usp -V -v cnc_vel.txt -flat -hw DphInd scan -Nvel.MIG.0.usp | moreProgram scan will tell you the RecNum now has the values of the DphInd and the Shot Points are all reset to zero use usp program xsd to look at your data. My display is up and to the right. Does it look like a velocity field? (File) (Open data) change end record to 679 (otherwise you will see only a single trace)
Here are the commands if you want velin to compute the interval velocity for you:
velin -N corr.MIG.0.usp -Ovel.MIG.2.usp -V -v cnc_vel.txt -flat -hw DphInd -vout I -force scan -Nvel.MIG.0.usp | moreAll we have needed to add is a -vout to specify Interval Velocity and a -force to ensure that RMS velocity inversion will do kill the program.
What to do if all the above doesn't work:
- use the scan program to inspect the data sets. Set the -V very verbose options to see what the header words really are.
- Read the messages stored in the text files after each job is run. These files look like VELIN.5454 This file also contains the processing flow to date. It's stored in a variable length descriptive header.
- If you have too many messages, use the usp cammand rmprint to erase all these messages!
- Hire opengeosolutions to solve your problems. The are very good at setting up the usp software in the first place. A few support dollars will go a long way.
- Send me an email and I can probably help. I'm free but some time you get what you pay for!
4 Convert the USP velocity file back to SeisWare format
Convert the velocity file back to SeisWare with with perl script usp2sx.pl.
ls vel.MIG.0.usp > junk
~ekeyser/Perl/usp2sx.pl
-Input file = vel.MIG.0.usp
-Output file = vel.MIG.0.sgy
PrcDat OACLin JobNum NumTrc NumRec SmpInt NumSmp Fmt UnitSc
-Green word 1180
"07/21/05" "orig " " Job num" 1 679 2 1501 3 0.001
Trace SrPtXC SrPtYC RecNum TrcNum Line Trace SrcLoc CDP
1 0 0 9405 1 0 9405 0.00 0
679 0 0 12795 1 0 12795 0.00 678
Yup, shot points SrcLoc are now all zero as are the UTMX and UTMY. These need to be fixed in SeisWare.
5 Add shotpoints and XY's back into the velocity file with SeisWare
Your best bet is to re load the velocity line into SeisWare, The usp file above does not contain the information to make a totally comptable SeisWare file. You can fix up the numbers but it is easier to just re load your data. The easy way to add the XY's is to strip out XY's from our initial line and then merge them into the velocity segy file.ls orig.MIG.0.sgy > junk ~ekeyser/perl/sx2segp.pl orig.MIG.0.sgy orig MIG fini --This script creates a flat file called:
bash-2.03$ more navigation.segp B12345 570.2 512361.0 6123924.0 B12345 580.0 571238.0 6123233.0 B12345 590.0 512337.0 6112347.0 B12345 600.0 577989.0 6123923.0 B12345 610.0 578123.0 6191233.0 ...Use (Seismic) (Import) segp files to import the XY information. Remember to make sure your internal line name for the segy data is the same as in the navigation file.
Finished -- you are now read to display velocity and your seismic section in the same display!