This paper discusses the development and application of a stochastic three-dimensional reservoir model for gas-water flow. The model treats the occurrence of natural fractures and shales in the reservoir as random, and as such the discontinuities may be arbitrarily distributed among the reservoir grid blocks. A simple shale management scheme that does not require the formulation of a sand-shale permeability relationship is used. However, the developed model requires the estimation of effective flow properties for the fractured reservoir nodes. By this, we attempt to simulate the effective permeability measured by the slope of a fractured reservoir build-up curve. The model may be used to compute the reservoir performance of a fractured, lenticular sand gas reservoir as well as that of a conventional blanket sand. Results of demonstration problems based on published data on Paludal zones 3 and 4 of the Mesaverde formation in western Colorado suggest that the effects of sand lenticularity may not be as marked as have previously been calculated by modifying sand thickness by sand lense factor.
The U.S. Geological Survey estimates Piceance Basin gas in place at 440 tcf, including 68 tcf potentially recoverable. In order to develop these microdarcy fields to their economic production capacities, unconventional methods of production are required.
The effort to develop such efficient production techniques is the principal objective of the U.S. Department of Energy's Multiwell Experiment (MWX) located in the Rulison Field in the Piceance Basin of northwestern Colorado. The MWX field project is aimed at characterizing the lenticular, tight gas sands that are present in this region. The reservoir and stimulation information acquired from this field research was used in the hydraulic stimulation of the Paludal zones 3 and 4 of the Mesaverde formation. As the preliminary results suggest, an optimum performance was not realized from the stimulation program. This an indication of the difficulties encountered in developing noncontinuous, tight gas sands.
In situations such as the MWX problem, or in cases where conventional well methods are unreliable or impractical, computer-based simulators may be used to access the influence of each pertinent reservoir parameter on gas production. The effects on reservoir performance of such discontinuities as shales, natural fractures and faults are especially suitable for study by the use of numerical reservoir simulators. Such simulators would devise approximate, yet realistic, methods for generating idealized reservoir representations for subsequent evaluation of proposed field development schemes.
There are simulation tools, for noncontinuous gas reservoirs in existence. The few papers cued represent only a partial list of the publications in this area. One of the major problems of modeling noncontinuous tight gas sands stems from the fact that the sands. exhibit many internal permeability variations. Property correlations developed for a particular sand lense may not be appropriate for other distributed lenses.