Author(s)

J. E. Capilla, J. Rodrigo & J. J. Gómez-Hernández

Abstract

Modeling flow and transport in fracture rocks requires in many cases an explicit representation of fractured zones in the model, even when using a continuum media approach. Hydraulic conductivity (K) is usually the most relevant parameter characterizing fractures. This parameter usually exhibits a high degree of spatial variability being direct measurements of it very scarce, when available. Thus the application of stochastic simulation techniques to obtain K fields in fractures is one approach that allows dealing with both uncertainty and variability of K. In practice, the simulated fields also should be conditioned to pressure measurements requiring the application of sophisticated inversion techniques capable to deal with the identification of parameters in the fractures and rock matrix. Based on a 3D real case study, this paper addresses this problem presenting a way to model K in fractures and a simulation technique specifically designed to obtain K fields conditional to pressure data in this type of fractured formation. It is assumed that K in fractures may or may not be multiGaussian. Using non-parametric statistics, the method allows the reproduction of K fields in which zones of extreme value of K may appear highly interconnected. The method allows the assumption of independent stochastic processes representing different fracture families corresponding to different tectonic and erosion processes. For that purpose a multi-parameterization approach is embedded in the inversion technique known as Conditional Probabilities method.

Keywords