Proportions Modeling

Objectives

The objective of the Proportions Modeling functionality is to estimate local facies proportions from a categorical variable corresponding to the lithotypes. The result is a regular grid containing a macro-variable (one realization per lithotype/facies value). Estimating the local proportions is an optional step to be used prior to facies simulations using the Pluri-Gaussian Simulations (PGS) for example.

An option is offered to add an external constraint (trend) defined at a coarsest scale (typically in oil and gas context, a 2D information derived from the seismic giving the mean proportion of a group of lithotypes on the whole thickness of a layer).

In addition, a new algorithm derived from an innovative approach developed by the School of Mines, based on SPDE to directly estimate proportions from the well data, has been integrated. This method is compatible with local geostatistics (LGS - i.e. local anisotropies). Details about this methodology can be found in:

Nicolas Desassis, Didier Renard, Hélène Beucher, Sylvain Petiteau, Xavier Freulon. A pairwise likelihood approach for the empirical estimation of the underlying variograms in the plurigaussian models. 2015. hal-01213962v2.

This publication can be accessed at the address: https://hal.archives-ouvertes.fr/hal-01213962

The Proportions Modeling procedure is divided into three steps (the second one being optional):

  1. Main Parameters: selection of the input dataset and lithotype variable, definition of the output file, definition of the chosen algorithm (Global VPC, Estimation with or without 2D constraint) and, if the dataset is 3D, a display of Global VPC with some associated parameters.
  2. 2D Constraint and Local Parameters: Hence, this page is skipped if Estimation with 2D Constraint is not selected or if Local Anisotropies is not checked.
  3. Algorithm and Output Parameters: Global VPC smoothing parameters or Estimation parameters, output file options, display (for the first proportion display in Map or 3D Scene) and, if the dataset is 3D, a display of output VPC (after smoothing if using Global VPC or those computed by the Estimation algorithm).