An Approach to Gas-Coning Correlations for a Large Grid Cell Reservoir Simulator

Abstract

Summary Generalized gas-coning correlations were developed specifically for a three-dimensional five-layer large grid cell model of the Prudhoe Bay field. These correlations can be used to predict the critical coning rate or to predict the gas/liquid ratio (GLR) of a well after coning has been achieved. Although these correlations have been developed with specific data for the Prudhoe Bay field, the approach and treatment of the parameters should be of use in the evaluation of gas coning in other fields. Introduction Gas-coning behavior can be described accurately in three-dimensional reservoir simulations with sufficient vertical layering. However, for large-scale field simulations, grid blocks are so large that the effects of gas coning are essentially absent. For these cases, it is necessary to use gas-coning correlations. These correlations may be of many forms. Critical coning rate is the maximum oil production rate at which the well produces oil without concurrent production of gas by coning. Critical coning rates may be calculated by standard methods; however, they cannot predict gas production rates if the well intentionally is produced above this rate. Predictions of coning behavior by pseudorelative permeabilities in three-dimensional reservoir simulators have been reported, but grouping of wells into various well types may be necessary. Several other methods for describing well coning behavior have been reported. In all of these methods, the basic purpose of the gas-coning correlations is to predict the gas/oil ratio (GOR) for each well at a given production rate.This study describes a set of gas-coning correlations that have been developed specifically for a three-dimensional five-layer model of the Prudhoe Bay field. The correlations assign a GLR in each well on the basis of the physical parameters of each individual well and the production rate. Although these correlations have been developed specifically for data from the Prudhoe Bay field, the procedure, treatment, and techniques for their development should be valid for more generalized situations.In most previous gas-coning correlations, wells having common characteristics were grouped into a number of well types since it is impractical to simulate individually all the wells in any large field. Gas-coning correlations were developed for each well type, and then either coning behaviors were identical or an interpolation procedure was used for each well of this type. The approach used for these gas-coning correlations was to develop generalized gas-coning correlations that would use individual well parameters rather than to group the wells into various types. By this method it was hoped that each individual well could be modeled more closely. Approach to the Problem Individual coning wells were modeled on a two-dimensional fully implicit radial simulator. Wells were modeled with properties encompassing the range expected from the field and with boundaries that correspond to similar boundaries in the three-dimensional field model. The gas-coning behavior was correlated to one critical parameter - the average oil column height above the perforated interval of the well.We determined the average oil column height above the perforations by first calculating an average oil column height within the drainage area of the well. This oil column height was determined by averaging the oil saturation around the well and calculating an average oil column height as if the gas/oil contact were level. JPT P. 2267^