Exploration & Production Geology

Advances in Multipoint Facies Simulation (MPS)
Page 1 of 1

Author:  Jerry_Cofton [ Wed Nov 26, 2014 11:22 pm ]
Post subject:  Advances in Multipoint Facies Simulation (MPS)

I have been coming across more and more reservoir study reports that seem to incorporate MultiPoint Statistics techniques for Facies modeling. For me it has always been somewhat of a mystery how this algorithm works and most articles I find online seem to explain it in very 'mathmatical' or 'statistical' ways. If you could help shed a simple light on the application, workings and results of this algorithm I would be very appreciative.

I am also interested in why companies seem to revert to MPS over the more conventional Object modeling, Kriging or Stochastic Gaussian Simulation.

thanks in advance!


Author:  WangR [ Mon Apr 13, 2015 8:47 am ]
Post subject:  Re: Advances in Multipoint Facies Simulation (MPS)

The No.1 reason why MPS is better than traditional 2-point-statistics methods is that it can replicate geological geometry better. Good algorithms,in this case, replicate the shapes on the training images, which are provided by geologists out of experience most of the time. Former algorithms, such as snesim, still use normal equation as core mathematics, which is also at the core of traditional geostatistics methods. Latest ones gradually abandoned the strict mathematical principal, utilizing approaches from computer vision. As a result, you would find the appearances of the realizations and of the training images quite similar. To my knowledge, 2-point-statistics concern about probabilistic correctness, MPS appearance likeness.

Page 1 of 1 All times are UTC + 1 hour
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group