Connectivity and upscaling for water and hydrocarbon reservoir simulation (#1600)

Authors

Gonzalez Litardo, Justo

Picighelli, Cristian

Colecchio, Iván

Noetinger, Benoit

Otero, Alejandro

Boschan, Alejandro

Abstract

Reservoir simulation is an essential tool for studying flow and transport phenomena in the earth's subsurface and is widely used in energy and environmental applications. In general, it is computationally expensive to perform simulations on the geological fine grid. Therefore, it is necessary to perform an upscaling stage, which converts the fine grid permeability (k(r)) to the equivalent permeability (Keq) defined on the coarse simulation grid. Because it is the connectivity of the higher permeability classes or components that determines the flow structure in the reservoir, it is crucial to understand how Keq varies under different connectivity scenarios. In this work we use a stochastic approach on 3D synthetic reservoirs in which the values of k(r) have a lognormal distribution, and analyze how the probability density function (PDF) of Keq, and its Gaussian moments, varies with the aggregation scale used for upscaling, as a function of the following connectivity parameters, which allow us to describe a very wide range of connectivity scenarios: correlation length lc, fine mesh variance 𝜎2f , and a type of connectivity structure: low, intermediate and high. The results show that it is not trivial to extend previous results for multigaussian media to more complex connectivity scenarios.