Sunday, June 20, 2010

One more word on monitoring the Deepwater Horizon well.


Following up on the recent entries on BP's Deepwater Horizon well and the need for its monitoring, it occurred to me that most petroleum engineers would recognize the words "compressive sensing" as being the technique that uses the l_1 sparsity inducing norm to reconstruct signals for seismic waves probing. It isn't so. Compressive Sensing is really about mixing the right type of probing waves (seismic, ultrasonic,...) or nuclear sources (neutron, gammas,...). By this I mean several sources located in different locations being fired in parallel (not sequentially). The way you mix them is what compressive sensing is really about. Much of the work that started since 2004 on the subject should allow interested parties to get the information they are looking for much faster. In effect, if mixing is done right, compressive sensing should provide some sorts good resolution in space or time of the changes happening under the seabed floor while the relief well is being dug.

To whom it may concern: this blog is being read by more than a thousand specialists from universities and industry everyday. If there is an interest I am sure that some of them could sign an NDA to provide guidance on how to do this right. Also, since the summer session is coming up, it is a good opportunity to get the academic folks to be working full time on this.

Image of burning methane at the Deepwater site courtesy of David Valentine (via Scientific American)

1 comment:

flex said...

You are absolutely right and this is a topic of great interest to the oil & gas community. See for some recent work on this

http://slim.eos.ubc.ca/index.php?module=articles&func=display&catid=18&aid=317

and other material on our website

http://slim.eos.ubc.ca

Printfriendly