As it is known two-phase gas condensate mixtures include a liquid-form and a vaporized state of condensate. Regarding liquid state, Peters et al.  has conducted detailed research to define the anomalous behavior of the liquid, which forms either by an isothermal decrease in pressure or by an isobaric increase in temperature, and termed this behavior as “retrograde condensation”. Following this it can be understood that the formation of vapor in the same two-phase mixture is the inverse process of retrograde condensation that is to say, this occurs due to either an isothermal increase in pressure or isobaric decrease in temperature. This formation of vapor in two-phase mixture is termed as “retrograde vaporization”. Gas condensate reservoirs initially contain gaseous phase when pressures remain above the dew point. As reservoir pressure further depletes below the dew point, heavier components are condensed out from the reservoir fluid due to which liquid (condensate) forms. Until the pressure reaches a specific value of maximum liquid formation, the reservoir fluid (condensate) starts re-vaporizing with further pressure depletion. This respective liquid buildup and dropout effect turns the gas condensate reservoirs into lower ultimate productivity.
During production depletion of gas condensate reservoir, the overall composition of the reservoir mixture remains constant. The constant composition analysis that was conducted does not accurately depict the real-time behavior of liquid dropout during condensate production. Heavier components, having an increased molecular weight, constitute most of the immobile condensate saturation. The saturation of this condensate (whose molecular weight continues to increase during depletion) is subsequently less than the saturation needed to mobilize the liquid phase. Therefore, the condensate that accumulates in the formation reduces the relative permeability of gas forming a condensate accumulation-condensate blockage. As a result, the productivities of gas and liquid in gas condensate reservoirs are greatly impacted.
Certain gas injection methods are then applied to reverse condensate accumulation-condensate blockage thereby enhancing condensate recovery [2-6] Gas cycling is the most preferable applied economical method used to enhance condensate recovery . The aim is to maintain a certain pressure of gas condensate reservoir to prevent condensate formation or at least minimize heavy condensate formation. Abel et al.  has described two schemes of gas cycling: full pressure maintenance and partial pressure maintenance. Full pressure maintenance involves a vaporizing mechanism in which the injection gas has minimum miscibility pressure (MMP) greater or equal to the dew point of the reservoir; vaporizing drive creates first-contact miscibility with the reservoir fluid (dropout condensate) to vaporize and flow towards production wells. However, partial pressure maintenance has the MMP lower than the reservoir dew point, which works with combined condensing/vaporizing mechanism by creating multiple contacts with the fluid to be displaced .
Boukadi et al.  did simulated work to examine the cyclic gas injection to re-vaporize liquid dropout in a gas field. A compositional simulation model was studied that confirmed the theory of condensate re-vaporization. The results indicated that cyclic gas injection is a viable production method that could improve gas deliverability and increase condensate recovery. Hamidov and Fataliyev  conducted experimental study on the effectiveness of partial gas cycling process in the development of gas condensate reservoir. The study highlights that if a gas cycling (dry gas) injection is applied at pressures near to reservoir dew point, it could possibly yield optimistic condensate recovery.
In this work, using gas cycling on an inverted five-spot well pattern, an analysis is made regarding optimal condensate recovery with production time. Inverted five-spot well pattern is the flooding pattern having one injection well surrounded by four production wells. Generally, flooding pattern has been implemented in EOR projects, where chances are optimal for maximizing oil production .
2 Reservoir model description
Due to complex change in a phase behavior during production depletion, gas condensate reservoirs require numbers of grid blocks to be modeled for maximizing the output of reservoir simulation results. The reservoir has an area of 2640 ft2 and sensitivity was modeled as 27 x 27 x 1 and 54 x 54 x 1 in compositional simulator GEM of computer modeling group (CMG). 54 x 54 x 1 model has larger number of grid blocks and required extensive computational time when comparing with the 27 x 27 x 1 model. However, compared results of condensate saturation for larger number of grid blocks were more accurate than smaller number of grid blocks. Maximum condensate saturation estimated in 54 x 54 x 1 model is approximately 16% which is equivalent to the maximum condensate saturation of CVD test that has been reported in Kenyon’s fluid model  (as PVT data taken in this study is from that model). The low permeability-gas condensate reservoir model is constructed with smaller grid size of 48.85 ft in both I and J directions, and pay-zone is 50 ft thick (K direction). The simulation model is shown in Figure 1 (upside).
The reservoir rock and fluid properties used in this model are referred from the published data of third SPE comparative gas cycling project . The one layer of low permeability (homogeneous) is taken from Kenyon’s reservoir model for the simulation studies; properties of the reservoir are shown in Table 1. Impact of reservoir heterogeneity on production performance is not considered in this study, however, Muladi and Pinczewski  had carried the simulation study, which reported that reservoir heterogeneity and layering pattern can affect the productivity and performance of vertical gas condensate wells. PVT and compositional data published in Kenyon’s work is used as an input to generate fluid model in CMG Win Prop. The components of reservoir fluid are lumped into 9 pseudo-components. Table 2 lists the pseudo-components descriptions used in this model, and the input for Peng-Robinson equation of state calculations.
Taking two data points of CVD experiment from Kenyon’s fluid model, Figure 2 expresses the simulation result of liquid volume in pressure depletion. It can be seen that condensate starts to form, when the pressure is lower than the dew point (Pd = 3428 psi). Then, condensate volume continues to increase until the pressure reduces to 2800 psi, when the maximum amount of condensate (16%) is achieved. Condensate volume suddenly reduces as the pressure continues to decrease (lower than 2800 psi); this is the representation of liquid re-vaporization.
Production model is designed with four production wells set around an injection well centered in the reservoir. Injection well is set at a distance of 1798 ft from every producer. Production wells are placed at every corner of the reservoir and are 2542ft laterally apart from each other. The production model is shown in Figure 1 (downside). This production model is designed to simulate the application of inverted five-spot gas cycling in the reservoir and perform sensitivity studies on different drainage strategies. Assigning LGR near production wells could yield slightly better simulation results, however, with increasing number of grid blocks the computation time increases resulting in more storage occupancy and delays simulation results . Thus for this model, this small grid size with these number of grid blocks (mentioned above) have been selected to perform the simulation.
As a producer in Kenyon’s third SPE comparative project had been subjected to minimum bottom-hole pressure (BHP) of 500 psi, therefore in this study, the producers (field) are subjected to the minimum BHP constraint of 500 psi. The injection well is subjected to the maximum injection pressure constraint of 3000 psi. The total production time of 10 years is set in this simulation study for every sensitivity analysis. The injection time for every case, with detailed simulation studies, are discussed in the next section.
3 Results and discussion
3.1 Start of gas cycling
The low permeability-gas condensate reservoir model with four production wells at defined field production constraint has been observed with the productivity life of 1215 days (3.33 years at constraint 500 psi min. BHP). 18% gas-rate of initial production rate (312 MMSCF/D) is observed in short span of 14 days followed by long decline curve up to abandonment limit (1215 days). This gas production rate of primary depletion study is shown in Figure 3(a). This refers that at first, the volumetric reservoir is sharply depleting with factor of 19% in the initial 14 days due to its small size (evident in Figure 3(b) and Figure 3(d) in which GOR is sharply increasing (to 18.4 MMSCF/MBBL) as pressure depletes sharply). A gradual decrease in production occurs, that is, production is at slower rate than the initial 14 days, this decline in gas production rate is due to retrograde condensation that has occurred in the reservoir. The condensate buildup effects of up to 16% approximate condensate saturation with pressure depletion from dew point to 2800 psi has affected the gas production rate (GOR increases continuously (from 12.5 MMSCF/MBBL to 29.7MMSCF/MBBL) within 14 to 218 days as shown in Figure 3(d)). Then a further more rapid decline in gas production rate is observed that is due to condensate dropout effects in the reservoir after 2800 psi pressure depletion (shown in Figure 3(c)); GOR decreases continuously after 218 days production time (Figure 3(d)). Therefore, the high decline rate (gas production) of around 87% in first 35 days and around 93% in first 120 days, is not only due to a small size of the volumetric low permeability gas condensate system, but also due to the accumulation of condensate in the formation. Thus at 304 days of natural depletion, the field gas production rate is 98% declined, leaving both heavy and light hydrocarbon components in the reservoir.
Continuous gas cycling with inverted five-spot pattern (injection well centered with four production wells around) is applied in the simulation model at different start times: 1 month, 4 months and 9 months after natural depletion. The total production time in this sensitivity work is set to 10 years and their results are compared with 10 years of primary depletion. The results are shown in Figure 4. Compared with the 10 years of primary depletion case in which condensate recovery reaches 28.2%, the injection application of gas cycling with inverted five-spot pattern can increase the condensate recovery. From 10 years simulation, result shows that condensate recovered up to 72.4%, 70.2% and 69.3% respectively for the cases if injection starts after 1, 4 and 9 months of natural depletion, respectively.
With combined observations from gas production decline rate and the effect of starting time of gas cycling, it can be seen that gas cycling is more effective when starting at earlier period of after 1 month of natural depletion. This is when production rate decreases to around 86% in this case (production rate observed: 41.87 MMSCF/D) as compared with injection after 4 and 9 months natural depletion (in which production rate decreased around 93% and 98%, respectively). Applying gas cycling after 4 month of natural depletion (production rate observed: 21.2 MMSCF/D) yields lower incremental condensate recovery and injection process might not produce cost-effective outcomes (in terms of condensate sale). For the case of starting gas cycling after 9 months of natural depletion, the reservoir pressure is depleted around 88% with 97% gas production decline (production rate observed: 6.7 MMSCF/D); thus the injection gas is unable to create effective miscibility-contact with lost hydrocarbon liquid components. This proves the displacement efficiency of injection gas is ineffective in vaporizing and displacing more lost condensate in the case of gas cycling that starts at 9 months after natural depletion; the same may be applicable to 4 months case as both are yielding approximately same value of 70% condensate recovery.
Excluding capital and operating costs (CAPEX and OPEX), the profit for every case is also investigated to evaluate their cost-effectiveness in window of 1215 days production time. In this investigation, low oil and gas prices of 0.276 CNY/MBBL (equivalent 0.04 USD/MBBL) and 0.0138 CNY/MMSCF (equivalent 0.002 USD/MMSCF), respectively are set to conclude which is the optimistic gas cycling case. Figure 5 and Table 3 shows the results and critical summary, respectively. It can be seen from Table 3 that gas cycling starts at 1 month after natural depletion has slightly greater condensate production with 68.18% recovery than 4 and 9 months case (condensate recovery of 66.18 and 65.16, respectively). And 1 month case yields 1.036 and 1.109 times more revenue than 4 and 9 months cases, respectively because of larger net gas production volume of 29.486 BSCF as opposed to 28.362 BSCF and 25.654 BSCF, respectively. This is because gas cycling (for optimizing condensate recovery) started earlier thereby maximizing gas productivity by improving gas deliverability (gas mobility and relative permeability). In the case of gas cycling starting 9 months after natural depletion does not yield profitable revenue as compared with other two cases.
3.2 Gas cycling injection period
This section reports the simulation study of various injection times of 100, 400 and 900 days. Their results are then compared with primary depletion (10 years) to determine the efficiencies of respective cases. Applying inverted five-spot gas cycling, the impact of injection time on condensate recovery is also studied. Gas cycling (injection) is applied at 4 months after the natural depletion. While during injection, four wells in the reservoir are kept on continuous production at defined field constraint of 500 psi min. BHP for 10 years of production time. A schematic of production-injection model is shown in Figure 1 (downside).
With defined production (drainage) strategies, Figure 6(a) shows that: comparative to shorter injection time of 100 and 400 days, greater injection time of up to 900 days has more variable injection rate throughout injection time that sustains for longer the reservoir pressure (as shown in Figure 6(b)). In Figure 6(a), injection rate after reaching 9.33 MMSCF/D is then witnessed to continuously decline in case of 100 days injection (also considered as initial injection plateau for 400 and 900 days injection times). In the case of 400 days injection period, injection rate is observed to be approximately constant (this can also be considered as middle injection plateau for 900 days injection time). After that 400 days of injection plateau is observed, injection rate can be seen slightly increasing at first and then follows a cyclic increase and decrease pattern up to injection stop time of 900 days. To explain these three injection plateaus, Figure 6(b) shows that reservoir pressure in the first injection plateau (100 days) seems to increase when injection starts. This is because of the vaporizing drive of gas cycling injection to recover lost condensate. Reservoir pressure in the second injection plateau (400 days) is initially observed to only slightly increase in production time; however this does prove that vaporizing drive also exists. Later on, a slight decrease in reservoir pressure with production time proves that a cyclic-switch in between vaporizing drive and combined condensing-vaporizing drive in the reservoir takes place. The third injection plateau, initially observed to slightly decrease reservoir pressure, is because of change in reservoir dew point pressure that could have occurred at the time. Due to this change in dew point pressure, the vaporizing drive can transform into combined condensing-vaporizing drive. Thus, a slight increase in injection rate is evident in Figure 6(a), which also depicts that post the injection-span of 580 to 687 days, the reservoir pressure appears to be constant (see Figure 6(b)).
Raising the conclusion from Figure 7 when comparing results of injection times of 100, 400 and 900 days with the primary depletion case, condensate recovery reaches up to 38.38%, 59.52% and 65.34%, respectively in contrast to 28.2%. Hence the application of inverted five-spot gas cycling is found to be more effective in 900 days injection (rather than 100 and 400 days injection) to maximize the condensate recovery.
Figure 8 shows the condensate saturation and pressure response of different times; i.e.; before injection (after 4 months natural depletion), 100, 400 and 900 days of injection. Common response in gas condensate reservoirs is also observed in the case of primary depletion of 4 months that shows that condensate saturation remains high near production wells (approximately 26.8% in Figure 8(a)). As a result lower pressure of 964 psi to 882 psi is witnessed near wellbore (see Figure 8(b)) with approx. 1600 psi average pressure throughout the reservoir. Average condensate saturation of the reservoir after 4 years primary depletion is reported approximately 15%. To study the remedy for lost condensate, the results obtained after 900, 400 and 100 days of injections (gas cycling) are studied and shared. Their analysis, when compared, expresses that 900 days injection can reduce more condensate saturation-near wellbore (to 16% - Figure 8(g)) as compared to 400 days (to 24% - Figure 8(e)) and 100 days (to 25.2% - Figure 8(c)). Improved condensate saturation (average) of the reservoir for respective cases at their end of injections is reported up to 9% (900 days), 10% (400 days) and 15% (100 days), respectively. Hence condensate saturation response supports the fact (as shown in Figure 7) that injections (gas cycling) for longer time of 900 days can sweep more and recover most condensate up to 65.34% recovery factor.
At 400 days injection period, pressure response-near wellbore is found to be 1105-943 psi (see Figure 8(f)) with average reservoir pressure of 1900 psi; this case yields optimistic pressure response value at the end of injection when compared with 100 days (1097-927 psi near wellbore (Figure 8(d)) with 1800 psi average reservoir pressure) and 900 days (1086-930 psi near wellbore (Figure 8(h)) with 1700 psi average reservoir pressure) gas cycling cases. Besides the fact that injection for longer times can recover maximum lost condensate, a thought can be drawn from this analysis that possibly 400 days injection time may be more economical than 900 days injection in terms of ultimate productivity of the gas condensate reservoir. Thus economic analysis of every case with primary depletion is conducted within the window of 1765 days production time. As stated in the previous section, oil and gas prices of 0.276 CNY/MBBL and 0.0138 CNY/MMSCF, respectively are set in this investigation, however, costs for capital and operating expenditures are not included in the analysis. Results of production assets in 1765 production time are shown in Figure 9 and economic analysis is summarized in Table 4.
It can be seen from Table 4 that when comparing with primary depletion case, gas cycling for 400 days, though increases cumulative profit up to good mark of 54% approximately, but 900 days gas cycling turns out a more optimistic result with approx. 67.8% cumulative profit yield. Thus, while applying injection (gas cycling) for longer time of 900 days, the low permeability-gas condensate reservoir when allowed to deplete with BHP constraint (flow-rate constraint not applied) has the potential to recover greater amount of lost condensate in maximum production time. Simulation results shared in this section conclusively express that the application of inverted five-spot gas cycling (flooding) is effective in vaporizing lost condensate and recovering it to the maximum depending on chosen drainage strategy to be applied.
The accumulated condensate saturation in volumetric gas condensate system is usually lower than the critical condensate saturation. This causes trapping of lost amount of condensate in reservoir pore volume.
With rapid pressure depletion, trapped condensate is lost not only due to condensate accumulation-condensate blockage courtesy of high molecular weight heavy condensate residue, but the sharp decline in gas production rate observed is also because of smaller size of low permeability gas condensate system.
To devise the most economical and optimal way for recovering lost condensate has always been a challenging goal. Thus, gas cycling at higher pressure of 3000 psi is applied to alleviate such a drastic loss in resources. The gas cycling-miscible injection is more effective in creating miscibility, condensing and vaporizing mechanisms in between gas and lost condensate, which causes more condensate recovery.
In this paper, starting time of gas cycling and injection period are the parameters used to influence condensate recovery of a five-spot well pattern which has a constraint injection pressure of 3000 psi and production wells are constraint at 500 psi min. BHP. Starting injection times of 1 month, 4 months and 9 months after natural depletion are applied in the first study limited to 1215 days production time. The results of this conclude that condensate recovery in 1 month case is respectively 2% and 3% greater than 4 (66%) and 9 (65%) months cases. However economically, the 1 month also produces greater revenue than 4 and 9 months, since net gas production volume of 1 month is greater than 4 and 9 months. This proves to be more economical as both time and effort are spared in condensate recovery of 1 month.
The second study is conducted by varying injection duration. Three durations are selected: 100 days, 400 days and 900 days keeping a production time of 1765 days. Out of all these three, 900 days duration produces a maximum condensate recovery (65.34%) along with greater revenue (¥594 Million). Compared to this 100 days and 400 days injection duration yield 38.38% and 59.52% condensate recovery, respectively, generating revenues of ¥257 Million and ¥415 Million respectively. Evaluating the time it takes to produce greater revenue of all these three, it can be concluded that the time, resources and energy invested in the 900 days far outweighs the possible economic benefits it can provide. At 400 days, the difference in condensate recovery is only 6% as compared to 900 days and also the difference in revenue is not that significant. Therefore, at an industrial level 400 days injection period could be considered economically beneficial.
From this study, it is proven that the application of gas cycling on five-spot well pattern greatly enhances condensate recovery when compared to natural depletion and this enhancement prevents any loss that previously occurred.
We would like to thank the support from the National Natural Science Foundation of China (No.51490654, No.61573018, No.51504276), Shandong Provincial Natural Science Foundation, China (No. ZR2016EL09), “China Important National Science & Technology Specific Projects” under Grant 2016ZX05025001-006, “863 Important Project” under Grant 2013AA09A215, the Fundamental Research Funds for the Central Universities” under Grant 15CX05035A and 17CX05002A, and Applied basic research projects of Qingdao innovation plan (16-5-1-38-jch).
Kurdi M., Xiao J., Liu J., Impact of CO2 Injection on Condensate Banking in Gas Condensate Reservoirs, In SPE Saudi Arabia Section Young Professionals Technical Symposium (19-21 March 2012, Dhahran, Saudi Arabia), https://doi.org/10.2118/160923-MS.
Pires A.P., Correa A.C.F., Mohamed R.S., Sousa R. Jr., Optimization of lean gas injection in gas-condensate reservoirs, In SPE Eastern Regional Meeting (18-20 September 1995, Morgantown, West Verginia, USA), https://doi.org/10.2118/31004-MS.
Taheri A., Hoier L., Torsaeter O., Miscible and Immiscible Gas Injection for Enhancing of Condensate Recovery in Fractured Gas Condensate Reservoirs, In EAGE Annual Conference & Exhibition incorporating SPE Europec (10-13 June 2013, London, UK), https://doi.org/10.2118/164934-MS .
Sanger P.J., Hagoort J., Recovery of gas-condensate by nitrogen injection compared with methane injection, J. Society of Petroleum Engineers, March 1998, 03(1), https://doi.org/10.2118/30795-PA .
Abel W., Jackson R.F., Wattenbarger R.A., Simulation of a Partial Pressure Maintenance Gas Cycling Project with a Compositional Model, Carson Creek Field, Alberta, J. Petroleum Tech., January 1970, 22(1), https://doi.org/10.2118/2580-PA .
Zick A.A., A Combined Condensing/Vaporizing Mechanism in the Displacement of Oil by Enriched Gases, In SPE Annual Technical Conference & Exhibition (5-8 October 1986, New Orleans, Louisiana), https://doi.org/10.2118/15493-MS.
Boukadi F.H., Al-Wadhahi M., Al-Bemani A., Al-Maamari R., Al-Hadrami H., Mobilizing Condensate in Gas Reservoirs-A Numerical Simulation Study, J. Petroleum Science & Tech., 2007 25(4), 517-533, http://dx.doi.org/10.1080/10916460500295405. Crossref
Hamidov N.N., Fataliyev V.M., Experimental study into the effectiveness of the partial gas cycling process in the gas-condensate reservoir development, J. Petroleum Science & Tech., 2016, 34(7), 677-684,https://doi.org/10.1080/10916466.2016.1160112. Crossref
Wardlaw N.C., Chapter 10 - Factors affecting oil recovery from carbonate reservoirs and prediction of recovery. In: G.V. Chilingarian S.J.M., Rieke H.H. (Eds.), Developments in Petroleum Science, 44(2), 867-903, Elsevier, 1996. CrossrefGoogle Scholar
Kenyon D., Third SPE Comparative Solution Project: Gas Cycling of Retrograde Condensate Reservoirs, J. Petroleum Tech., August 1987, 39(8), https://doi.org/10.2118/12278-PA.
Muladi A., Pinczewski W.V., Application of Horizontal Well in Heterogeneity Gas Condensate Reservoir, In SPE Asia Pacific Oil and Gas Conference & Exhibition (20-22 April 1999, Jakarta, Indonesia), https://doi.org/10.2118/54351-MS.
Orangi A., Nagarajan N.R., Unconventional Shale Gas-Condensate Reservoir Performance: Impact of Rock, Fluid, and Rock-Fluid Properties and their Variations, In Unconventional Resources Technology Conference (20-22 July 2015, San Antonio, Texas, USA), https://doi.org/10.15530/URTEC-2015-2170061.
About the article
Published Online: 2017-08-03