Development of a Local Grid-Refinement Technique for Accurate Representation of Cavity-Completed Wells in Reservoir Simulators

Summary

In this paper the development of a three-dimensional, single-, and two-phase fluid flow model to study flow performances of cavity-completed wells is presented. A static local grid refinement (LGR) is embedded into the model in order to adequately capture the flow behavior around cavity completions using a reasonable number of grid cells. The model adopts an approach that is significantly different from the current practice in which cavity-completed wells are incorporated into computations by assigning large wellbore radii and/or large negative skin values to a conventional wellbore model.

The model utilizes the residual equation rather than a wellbore equation to determine the flow into cavity-completed wells. Cavity-completed wells are described as a collection of fine rectangular grid blocks. This representation is referred to as the cavity approach. The constructed fine grid lines are terminated at the coarse/fine grid block interfaces by implementing the LGR technique developed. The technique subdivides selected coarse grid blocks into fine rectangular grid blocks in all three dimensions that are known as window blocks. The flow equations generated for window and coarse grid cells are then solved simultaneously by using the preconditioned biconjugate gradient stabilized iterative method.

First, a radial-cylindrical wellbore is approximated using the cavity representation. The model results from single phase, slightly compressible fluid flow simulation with the cavity approach are, then, compared against the analytical solutions. In these comparisons, very good matches are obtained for both flow rate and sandface pressure specifications at the wellbore. The model is further extended to single phase, compressible, and two-phase fluid flow conditions in both conventional and coalbed reservoirs. The model results are, in this case, compared with the existing numerical models that utilize Peaceman's wellbore model. Again, close agreement in flow response is obtained for a series of comparisons. Finally, the cavity approach is tested by modeling flow performances of irregularly shaped, cavity-completed wells in conventional and coalbed reservoirs.

During the creation of cavity completions in coal seams, it is hypothesized that fracture networks in certain directions are generated around cavities. The effects of fractures are usually incorporated into a simulation study by modifying the permeability values around the wellbore. The cavity approach presented in this paper provides not only a better approximation to the shape of the cavity but also a better representation of the altered permeability values around the cavity since fine grid cells are already in place around cavity completions.

Introduction

Cavity completion refers to an enlargement of an open-hole interval and was a common type of well completion technique in the early stages of petroleum industry. Explosives such as solid and liquid nitroglycerine were used to create cavity completions in open-hole-completed wells in tight sand reservoirs. Starting from the early 1960's hydraulic fracturing replaced cavity completions.

Recently, cavity completions have been revisited as alternative stimulation techniques to hydraulic fracturing of cased holes in coalbed methane reservoirs. Explosives, however, are no longer in use. Instead, cavity completions are created by injecting air or an air/water mixture into an open-hole interval at a high pressure, and suddenly releasing the wellbore pressure. The sudden release in wellbore pressure causes friable, low strength coal and rock particles to slough into the wellbore.^1,2 The injection/blowdown sequences are repeated until the shape of the cavity is formed. A 3D representation of the cavity as a function of depth and azimuth can be obtained by the use of a sonar log.³ The volume of removed reservoir rock particles can also be used to confirm the log data. When this information is coupled with the additional descriptive information available from a downhole camera, it will be possible to reconstruct the actual shape of the cavity accurately at the surface to be incorporated into the numerical model.

It is hypothesized that tensile and shear fractures are created during cavity completion process due to the sudden increase and decrease in wellbore pressure.^3-6 Tensile fractures are produced as a result of the injection process where wellbore pressure is greater than minimum in situ stresses around the wellbore. During blowdown process, shear fractures are created as a result of the reduction in wellbore pressure. Limited laboratory data indicate that tensile fractures may extend 100 to 200 ft and shear fractures may extend up to 25 ft from the wellbore in each direction.^3,5 With a cavity completion in coalbed reservoirs, besides increasing wellbore diameter and removing possible drilling damage, fracture networks around the cavity are generated. These artificial fractures are believed to be responsible for major increases in flow rates. Consequently, cavity completions are characterized by an increase in wellbore diameter and a permeability enhancement zone around cavities.^7-9 In current studies, increase in wellbore diameter is accounted for by using a large negative skin factor and fracture networks are modeled by an alteration of the permeability around the wellbore.