Well Test Analysis of Multiple Matrix-to-Fracture Fluid Transfer Fluid Transfer in Fractured-Vuggy Reservoir

Abstract

This paper describes techniques to identify vugs in a real pressure buildup data. The reservoir under study is an undersaturated carbonate reservoir dominated by vugs and karst. We analyse the 80 hours of pressure buildup following a 100 hour of pressure drawdown. We observe the pressure derivative behavior cannot be described using current NFR models such as dual porosity and triple porosity models. This is because some pressure derivatives are obscured by reservoir boundaries and the matrix-to-fracture fluid transfer does not behave as it were in a multiple porosity reservoir. In this heterogeneous reservoir, we believe this corresponds to the number of matrix inter-porosity flow parameter, and thus degree of reservoir heterogeneity. If we force the data using dual porosity or triple porosity models, the error will be high (>15%), because the pressure derivative cannot follow all the transition periods.We use combined analytical and numerical methods to tackle this problem. First the dual porosity analytical solution is invoked to estimate the reservoir parameter (ω, λ and k). Next, a numerical triple porosity model is set up to represent the reservoir. The reservoir parameters initially follow analytical results as initial guesses. History matching procedures are performed to compare DST data with the calculated bottom-hole pressure. When these agree, reservoir flow capacity and storage capacity can be determined for fracture, matrix and vug respectively. We found volumes of vug cannot be ignored in a triple porosity system. This technique can be used to quantify vugs or karst from well testing when analytical dual porosity or triple porosity model are unable to handle all matrix-to-fracture inter-porosity flow due to reservoir heterogeneity. Using this technique, we believe not all well test data in carbonate reservoirs (especially for high k.h) can be analyzed using current analytical models. Ignoring this, the estimated permeability is significantly different from the actual.