Three porosity types, matrix, vugs and fractures, are usually present in naturally fractured, vuggy carbonate reservoirs. The vugs are generally considered connected either to the matrix or to the fractures in numerical simulation. One of the challenges of modeling these reservoirs is the partitioning of the porosities into two components since dual porosity reservoir simulators can only handle two rock components, namely matrix system and fracture system. The goal of this study is to characterize the vugs in these systems, and to determine the pore volume compressibility for the simulation of vuggy, naturally fractured reservoirs.
Sequential laboratory experiments were designed and conducted to determine the amount of secondary porosity in the core samples. A combination of capillary pressure (centrifuge and mercury injection experiments) and NMR experiments was used to determine the vug or secondary porosity of the samples from pore size and T2 (relaxation time) distributions. The results were compared and reconciled with those from porosity logs and image logs. Pore volume compressibility tests and compaction tests were used to determine the compressibility of each sample. The compaction and pore volume compressibility tests were evaluated at different net stresses to determine the influence of vuggy porosity in the samples.
The composite (sample) compressibility versus porosity plots showed that pore volume compressibility increases as secondary porosity increases. This dependence, which is strong at low net stresses, gradually disappears as the reservoir pressure decreases or the net stress increases. This implies that the effects of secondary porosity on pore volume compressibility of the total system are minimal at low reservoir pressures. Based on these results, pore volume compressibility was generated for various values of effective and secondary porosity and served as input data to the numerical flow models. The development of the vug porosity, pore volume compressibility and the construction of the simulation flow model are described in this paper.