A good friend of mine (Sunawar Kunaifi) showed me his script on VSP (borehole seismic) processing with SeismicUnix (CWP, Colorado Schol of Mines), we exchanged email back and forth, and discussion with him was very good and very enlightening, so that I can come up with a set of script to do VSP modeling and VSP processing, for the basic and simple VSP data. The VSP tutorial from EnsiklopediSeismik.blogspot.com is also very helpful, so my scripts is kind like wrap things up that has not yet been covered by that tutorial.
I set up an information page here LatihanVSPSU (http://www.gm2001.net/latihanvspsu/). I also start to learn and begin to like Git, I use github as my code repository, visit http://www.github.com/fahdi104/LatihanVSPSU.
Synthetic modeling is not using SeismicUnix though, since I can’t find a way or understand which SU program that can do elastic modeling using certain geometry. I’m using Ela2D from Joachim Falk (University of Hamburg) for elastic synthetic modeling.
*Update April 2013* Apparently, the Ela2D (or me) does not do the VSP modeling correctly. The waveforms looks slower than it should. So when I tried to processed synthetic data, the reflector that should come out as flat, is not appear what it should be. Anybody can help me with?
The script is still far away from perfect. If I processed the synthetic data, the results looks weird, and also in real dataset. I mean weird in terms of looks like the amplitude data is clipped/distorted. The other thing is I couldn’t find a good deconvolution to be applied with the data. Of course Seismic Unix has this predictive error filtering/predictive deconvolution module, but my exercise shown the result is not that good.
*Update April 2013* I updated a few things on the scripts, fixed some of the bugs, I also add the transit time correction calculation using the standard cosine corrections.
If you have sometimes, maybe you can see have a look at my scripts, and give some suggestion to me.