METHOD FOR MORE RAPIDLY PRODUCING THE REPRESENTATIVE STOCHASTIC MODEL OF A HETEROGENEOUS UNDERGROUND RESERVOIR DEFINED BY UNCERTAIN STATIC AND DYNAMIC DATA

摘要

Method for more rapidly forming a Gaussian stochastic model or the similar which is representative for a porous heterogeneous medium such as an underground reservoir defined by characteristic fluid displacement data and by punctual observations which are uncertain and are characterised by a probability law. It is based on a method for transforming static information which is given in the form of the probability law into Gaussian punctual pseudo-data. One of the advantages of said transformation technique is that it is fully compatible with a gradual deformation method, thereby making it possible to minimise an objective function J which measures the deviation between dynamic data (for example production data) and corresponding responses which are simulated for a considered reservoir model. Said minimisation is carried out in combination with the realisation of a stochastic model for said reservoir model and for the Gaussian pseudo-data representative for static information. During the minimisation process, combination factors are adjusted and realisations allowing an objective function to be reduced are identified. Said invention can be used, in particular for operating, for example oil pools.