Imbibition only relative permeability is commonly used to model water influx in water-drive gas reservoirs, however an aquifer is rarely strong enough to maintain constant pressure support. Continued pressure depletion in the part of the reservoir swept with watercauses expansion and remobilisation of trapped gas behind the waterfront. This paper presents a reservoir simulation study on modelling the expansion and remobilisation of trapped gas due to pressure depletion as secondary drainage flow using relative permeability hysteresis.
Previous studies in literature on relative permeability show the secondary drainage curve during blowdown is below the primary imbibition curve. This is based on field cases and core experimental studies, which establish the existence of a gas remobilisation threshold above residual saturation to reconnect the gas phase. Commonly used hysteresis models by
The conclusion of this study is that the standard formalisms used to model relative permeability hysteresis (Killough, Carlson) should not be used to model trapped gas remobilisation due to blowdown as they do not incorporate a gas remobilisation threshold and a secondary drainage curve underlying the primary imbibition curve. By assuming no mobility threshold above residual gas saturation, the total recovery of residual gas will be overestimated. Instead, by adopting the ODD3P hysteresis model, gas production will be lower and water production higher due to the correct use of secondary-drainage relative permeability curves in a gas reservoir invaded by water. This will lead to a significant improvement in results from reservoir simulation and the subsequentevaluation of trapped gas recovery.
The expansion and remobilisation of residual or trapped gas saturations has a major impact on the prediction and/or matching of production and pressure response from a reservoir. This study intends to understand these impacts and serve as a preliminary guideline in modelling trapped gas expansion and remobilisation as secondary drainage flow, which is applicable to many water-drive gas reservoirs.
Preliminary results obtained from a program of experimental and theoretical studies examining the uncertainties of waterflooding gas-condensate reservoirs are reported. In spite of high trapped-gas saturations (35 to 39%), further aggravated by an unusual type of hysteresis, recoveries of gas and liquids can be increased over those obtained under natural depletion.
Water injection has been suggested as a method of maintaining pressure in gas-condensate reservoirs. This method offers advantages over gas injection: gas can be sold from the start of reservoir production; the injection costs are much lower; the favorable mobility ratio ensures a high sweep efficiency; and the reservoir pressure is maintained without changing the composition, and hence the dewpoint pressure, of the gas. Water injection has not been generally accepted for gas-condensate reservoirs, however, because of the following concerns.
1. The advancing water could trap a significant amount of gas.
2. It may not be possible to remobilize the previously trapped gas during a subsequent depressurization.
3. Three-phase relative permeabilities for conditions where retrograde condensation occurs are virtually unknown and may be unfavorable.
4. Well lift could be a severe problem if there are high water cuts before and during blowdown.
The first three of these factors are concerned with the flow behavior within the reservoir and are addressed in this paper. Well lift is not considered for reasons explained below.
In a pioneer work, Geffen et al. showed that trapped-gas saturations following waterflood are in the same range as the residual oil saturations expected in waterflooded oil reservoirs: i.e., 15 to 50% of pore space, depending on the rock characteristics. They argued that these high values of trapped-gas saturation could substantially reduce the recovery of gas from such reservoirs as a result of their magnitude and permanence.
A large number of gas reservoirs with strong underlying aquifers have been successfully developed, however, and have given moderately high gas recoveries, suggesting that at least some of the trapped gas could be remobilized during a final period of accelerated depressurization. Boyd et al. were able to depressurize the Double Bayou field after it had watered out and thus remobilize some of the residual gas. Four years after the start of the trial, they estimated that an increase in recovery of 10% of the gas initially in place (GIIP) could ultimately be recovered. Of this 10% increase, about 8% was a result of percolation of trapped gas from the watered-out zone. Brinkman found that accelerated depressurization in the Lovells Lake Frio 1 field increased the recovery from 58 to 70 % GIIP. Of the 12 % increase, nearly 10 % was caused by the trapped gas percolating from the watered-out zones. Lutes et al. obtained 8% GIIP from percolation during accelerated depressurization in the Katy field but had expected 20%. They concluded that recovery was restricted by the amount of gas that could percolate out of the waterflooded zones by high pressures arising from percolate out of the waterflooded zones by high pressures arising from unfavorable relative permeabilities.