Confronting the Simulation of Fluid Flow in Naturally Fractured Carbonate Karst Reservoirs
Conference Paper
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
AbstractFlow simulation in carbonate reservoirs presents many challenges due to the frequent occurrence of vugs and natural fractures therein. In conventional reservoir simulation practices, fluid flow in vugs and fractures is usually assumed as Darcy flow, and permeability values are estimated for vuggy regions and fractures. Although such estimations are often accomplished in reasonable ways, no physical or mathematical basis exists for them, and even the assumption of Darcy flow itself is questionable.In this paper, we propose a novel workflow for the simulation of fluid flow in naturally fractured carbonate karst reservoirs. The workflow is based on simulation results of a single-phase transient Brinkman model, which provides the correct and complete description of the coupled flow in vuggy and fractured reservoirs by unifying Stokes flow in vugs and fractures with Darcy flow in the rock matrix. The new workflow proceeds through an iterative procedure by increasing the permeability values of vugs and fractures without the disadvantages of accurately estimating them, and attains the final simulation results when a convergence pattern is observed.The novel workflow is implemented and compared with the conventional approaches in commercial reservoir simulators. The workflow is first applied to single-phase flow simulations in a fine-scale 3D geological model which is generated using the multiple-point geostatistical modeling technique, and then extended to immiscible two-phase flow and other multi-phase cases. Simulation results show that in most cases our new workflow yields higher production rate predictions than conventional approaches. This is due to the fact that the permeability values of vugs and fractures estimated by conventional methods are usually lower than the permeability values required for the iterative convergence of the new workflow. The results also have further implications on the history matching process that more focus should be put on fracture geometry, i.e. fracture width and half-length, since the use of fracture permeability alone has lost its physical meaning in the novel workflow according to the Brinkman equation.
