Oil IT interview - Indy Chakrabarti, Paradigm

Paradigm Senior VP tells Oil IT Journal about working in the downturn and the revamped collaboration with Dassault Systèmes on linking geological and mechanical modeling. He also explains how data compression is helping the interpreter, with help from Hue CEO Diderich Buch.

How is Paradigm doing in the downturn?

Well it is no longer business as usual! But we are keeping busy as witnessed by our deals with GE, Dassault and Hue. With Dassault, we are targeting what we call ‘reservoir-driven production risk management,’ adding geomechanical modeling to our portfolio with Dassault’s Abacus. This has been coupled with our Skua modeling flagship. Re-frac workflows are a current industry preoccupation and the Skua/Abacus combo lets you see fractures propagate in the vicinity of the well bore after an initial frac.

We didn’t know that Skua could do fine grained well bore modeling.

It is first and foremost a reservoir-scale modeler but its grid refinement capability can be used to study fault reactivation as producing fluids lubricate fault planes in the vicinity of a well. Such phenomena can damage the well bore or open up zones of lost production.

Are there anchor clients or sponsors of this work?

The project has only just kicked off. We are in discussion with potential clients but this is our own development, not a consortium. Also this is not packaged software, rather a software and services bundle. We also couple the macro-scale reservoir model to Abaqus for full field geomechanics for well placement and drill planning.

Doesn’t Geolog do geomechanics?

Yes, it always has had geomechanical functionality and we have built significantly in the last release with more work on image logs for fracture detection. A full workflow would be from Geolog to Skua for building the reservoir model and for data exchange with Abaqus.

Tell us more about the deal with Hue.

Data compression saves on disk space and raises interpreter productivity as less time is spent waiting for data while roaming through a seismic cube. But geophysicists don’t want to sacrifice resolution. We have checked Hue’s technology and determined that at 3-4x data compression there is no discernable loss in resolution at all, out to six decimal places. For quick-look or QC we can even go to 30-40x. Actually, conventional interpretation systems often use brute force compression that truncates 64 bit data to 32 or even 16 bits. Now that does affect resolution!

Which products use the compression?

All the Epos-based seismic products including Skua and Seis Earth.

It used to be thought that seismics was not compressible*!

There have been attempts over the years but most were too aggressive and lossy. Hue lets you do accurate work at 3-4x.

Things might be different with pre-stack data as imperceptible details may stack into signal…

That’s what we are working on now, using Hue’s technology on pre stack data.


Additional input from Hue CEO Diderich Buch.

Paradigm is the first of the big providers of seismic-oriented geoscience applications to adopt our technology. Today, storage and network performance lags behind acquisition and computation. There may be cases where lossless compression is needed but most seismic usage clearly does not. Minor losses due to compression don’t affect phase, frequency or amplitude for imaging and quantitative interpretation. Several oil companies and seismic/software vendors have arrived at the same conclusion, and it’s safe to assume that processing shops and oil companies will also adopt our technology. This saves on average 80% on disk space, maximizes interpreter productivity, optimizes your networks and makes sure your CPUs are continuously fed with data.

Our next target will be the national data repositories which currently only store stacks, not the gathers. NDRs could provide a significantly more valuable service if the gathers were available and more data was available live. Turn-around time for re-processing and interpretation would also improve significantly.

* OITJ December 1996.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.