Abstract
Numerical modeling of naturally fractured vuggy reservoirs presents many challenges due to the coexistence of three very different kinds of media and their complex interaction on multiple scales. With current computing capabilities, conventional fine-scale single-porosity models are not practical for large-scale reservoir simulation. A multi-continuum reservoir model is presented as an effective approach to modeling fractured vuggy reservoirs on the coarse scale. This model consists of four different porosity systems, i.e. the matrix, fractures, isolated vugs and connected vugs. This study investigates mass exchange between different porosity systems with the final objective of developing new transfer functions that can be used as an application to upscaling fractured vuggy reservoir models.
A well-designed procedure is proposed to obtain the novel transfer functions in multi-continuum fractured vuggy reservoir models. Different realizations of the vug-filled matrix blocks are generated to show the effect of vug fraction, distribution and connectivity on multi-phase fluid flow. For interporosity flow between vugs and the other two media, new formulations of the shape factors that incorporate the effect of vug spatial variation are developed respectively. The dominant mechanisms of multiphase fluid exchange between each two porosity systems (matrix-isolated vugs, fractures-connected vugs) are discussed separately. New transfer functions for multiphase flow in the multi-continuum fractured vuggy model to capture the complex flow mechanisms and emulate the results of the fine-grid model are provided. In addition, a transmissibility multiplier table is introduced as another connection term for the transfer functions to improve the accuracy of upscaling solution. Finally, a new upscaling approach by incorporating the proposed transfer functions into multi-continuum models is presented.
This paper provides a new insight to the complex fluid exchange among three different media in fractured vuggy reservoirs. Results show that the new upscaling methodology helps to reduce the size of simulation model and improve the computational speed significantly, while providing an accurate representation of the fine-scale results.