Optimization of Gas-Well Completion and Production Practices

Abstract

In water drive gas reservoirs, wells can be completed with a long perforated interval and produced at high rates to minimize abandonment pressure and maximize recovery. Alternatively, the perforations can be limited to the top of the productive interval and the well produced at a low rate in an effort to prevent coning which results in high abandonment pressures if the strength of the aquifer is adequate to support reservoir pressure. This study uses a reservoir simulation coning model to evaluate these two conflicting completion and production practices. The impact of completion interval, gas production rate, and reservoir permeability were evaluated. Ultimate gas recovery was found to be largely insensitive to variations in perforated interval and production rate in high permeability systems. Ultimate water production, however, was found to increase at high gas rates and lengthened perforated intervals. In lower permeability systems, ultimate gas recovery was found to increase significantly as production rates were increased, while ultimate water production was actually observed to fall. Sensitivity analysis of vertical to horizontal permeability ratio, fluid density contrast, relative permeability, and formation dip did not alter these conclusions. The conclusion that elevated production rates can be expected to have no detrimental impact on ultimate gas recovery suggests that gas rates should be maximized in low water disposal cost situations. This finding favors the completion of an interval sufficiently long to maximize production rate and thereby insure that gas recovery and present value of gas reserves are maximized. In high water disposal cost situations, however, it should be recognized that this strategy might result in elevated water production in high permeability systems. Introduction Many natural gas reservoirs have some level of pressure support from aquifer encroachment. In gas reservoirs with active water encroachment, it is well known that lowering reservoir pressure with fluid withdrawal rates in excess of aquifer encroachment rates can significantly increase ultimate recovery. In a reservoir produced with a single well, the well would most certainly not be at the very crest of the structure. If the reservoir pressure could be lowered by high withdrawal rates, it would be possible to actually drain up-dip gas that would otherwise be unrecoverable. In a multi-well reservoir situation, drainage and competitive situations could make the decisions on how a well of this type should be completed and produced of paramount importance. In either situation, the possibility of "out-running" the aquifer can significantly add to the ultimate recovery from the reservoir. Unfortunately, the decision to produce a water-drive gas reservoir at high production rates is not as clear-cut as it might seem from the perspective of reservoir engineering. Logic and experience indicate that this type of well will ultimately cease production as the result of excessive water production. It is well known that coning of water will occur if the well is produced at high rates. Clearly, water will be produced sooner as the perforated interval is extended toward the gas-water contact. From this point of view, the prudent course of action would be to delay the occurrence of water in the well stream by perforating the very top of the interval and producing at a restricted rate. Selection of optimum completion and production practices for water-drive gas reservoirs requires an understanding of water coning behavior and the implications of water production on well and reservoir performance. Coning behavior during the production of oil and gas has been extensively investigated. The vast majority of the literature on this subject reports the behavior of water or gas coning into an oil completion. Investigations of coning in gas-water systems are much less numerous in the literature. Kabir 1 reported the most extensive evaluation of water coning in a gas-water system in 1983. Kabir used a radial reservoir simulation model with 10 blocks logarithmically distributed in the radial direction and 14 layers in the vertical direction. This model was extremely course by modern standards. Also, reservoir simulation technology had not yet been developed to allow the inclusion of wellbore friction and loading models.