Carruthers, Dan1
(1) The Permedia Research Group Inc, Ottawa, ON
ABSTRACT: Risking and Ranking Petroleum Migration Models
Anyone who has stood before an outcrop knows that all geological models are crude when
compared the real world, incorporating considerable inherent elements of uncertainty.
Petroleum migration modelling, like all modelling, is subject to these uncertainties. As a
result, it is critical that models be evaluated through the execution of multiple
realizations that sample the uncertainty. Without running multiple realizations, the risk
that a model represents a less probable, outlier scenario is higher.
New techniques will be demonstrated to: (1)capture the effects of the dominant
uncertainties in petroleum migration, and (2) apply objective functions to assess model
output against reality
These techniques will include novel stochastic conditioning methods and the application of
distributed processing to generate the required realizations, and the application of
objective functions to discount the realizations that, while still within the specified
scope of uncertainty, have failed to match a given objective (e.g., matching locations of
known fields).
AAPG Search and Discovery Article #90026©2004 AAPG Annual Meeting, Dallas, Texas, April 18-21, 2004.