A Computational Tool for the Geometric Modeling of Naturally Fractured Carbonate (Karst) Petroleum Reservoirs

Resumo

One of the main reasons why the modeling of carbonate (karst) reservoirs is a challenging task is the
geometric representation required to capture the complex geological structures in these reservoirs, such as fractures,
vugs, cavities and caves. Mesh generation from a geometrical model is thus affected by the choices made for vugs
and fractures representations. An alternative approach is to assign geological features to the volumes in a fine scale
mesh as physical properties. These discrete models are required when, for instance, Stokes-Brinkman’s equations

are adopted to model multiphase flow simulation in carbonate karstic oil reservoirs, found often in Brazilian Pre-
Salt. In this paper, we propose a computational tool for generation of karst reservoir scenarios by creating a

geometrical model of the geological features and then computing the mesh volumes inside vugs, cavities, caves and
fractures. Vugs are represented by randomly distributed and possibly overlapping ellipsoids in a three dimension
space. They might be connected by fractures, modeled as cylinders, ellipsoids or boxes. Face connectivity between
mesh volumes in a fracture is assured. The code was written in Python using the NumPy library and the IMPRESS
(Intuitive Multilevel Pre-processor for Smart Simulation), an “in-house” pre-processor used for mesh management.
Example scenarios were generated to evaluate the correctness and robustness of the developed computational tool.