Sampling Gas-Condensate Wells
- W.D. McCain Jr. (Cawley, Gillespie and Assocs. Inc.) | R.A. Alexander (Cawley, Gillespie and Assocs. Inc.)
- Document ID
- Society of Petroleum Engineers
- SPE Reservoir Engineering
- Publication Date
- August 1992
- Document Type
- Journal Paper
- 358 - 362
- 1992. Society of Petroleum Engineers
- 4.6 Natural Gas, 5.5 Reservoir Simulation, 4.3.4 Scale, 5.2.1 Phase Behavior and PVT Measurements, 2.2.2 Perforating, 4.1.9 Tanks and storage systems, 5.2 Reservoir Fluid Dynamics, 4.1.5 Processing Equipment, 4.1.2 Separation and Treating, 5.2.2 Fluid Modeling, Equations of State
- 4 in the last 30 days
- 1,343 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 5.00|
|SPE Non-Member Price:||USD 35.00|
The industry has a procedure for stabilizing and sampling retrograde-gas(gas-condensate) wells; however, no investigation of the quality of the samplesresulting from this procedure has been published. During sampling, bottomholeflowing procedure has been published. During sampling, bottomhole flowingpressure (BHFP) typically is less than the dewpoint pressure of the pressure(BHFP) typically is less than the dewpoint pressure of the original reservoirgas. This causes condensate liquid to build up in the reservoir around thewellbore. This paper presents the results of a study of the sampling procedureand of the buildup and stability of the condensate ring around the wellbore. Aprocedure designed to give the best chance of obtaining a representative sampleis presented.
A gas sample from a retrograde-gas (gas-condensate) reservoir almost alwaysis obtained by sampling the gas and liquid from the separator and recombiningthe samples at the producing gas/liquid ratio. It is important that thewellhead pressure and the gas and liquid production rates be stable before andduring sampling. Also, the gas production rate during this time must be largeenough to remove all liquid continuously from the production string. Duringsampling, it is likely that the BHFP will be less than the dewpoint pressure ofthe original reservoir gas. If so, there will be a pressure gradient in thereservoir near the well where pressure is less than dewpoint pressure. pressureis less than dewpoint pressure. How can a recombined surface samplerepresentative of the original reservoir gas be obtained under theseconditions? This paper answers that question. Sampling problems, such as humanerror, measurement bias, and incorrect laboratory recombination, are notconsidered here. Rather, this paper addresses the conditions necessary for theproduction stream to be of the correct composition so that obtaining a goodsample is possible. Radial compositional simulation was used to examine thepressure and saturation distributions in the reservoir, the producing GOR, andthe composition of the total production stream at the surface during sampling.("GOR" is used in this paper because of its common usage in thepetroleum industry, the liquid is actually condensate.) The equation of state(EOS) was tuned with laboratory data. Three retrograde gases (gas condensates)with different compositions were used in the study. The richest gas caused thelargest buildup of condensate around the wellbore. The conclusions of thisstudy are independent of gas composition, however, so the results presented arebased on only one of these gases Several different sets of relativepermeability data were used. All were for water-wet systems. The positions ofthe endpoints and shapes of the relative permeability curves affected thequantity of condensate around the wellbore but did not affect the conclusionsof this study. Thus, the results reported here are based on one set of relativepermeability data. Several combinations of initial reservoir pressure anddewpoint pressure of original reservoir gas were examined. Only the worst caseof initial reservoir pressure slightly greater (15 psi) than dewpoint pressureis presented. psi) than dewpoint pressure is presented. SimulationProcedure
The Soave-Redlich-Kwong (SRK) EOS was used in the compositional simulationsCompositional analyses through C30 were available for all gases used.Components between C7 and C 30 were grouped into four pseudocomponents, iso-and n-butanes and pentanes were combined, and the small amounts ofnonhydrocarbon components were combined with the appropriate hydrocarbons. Thisresulted in a 10-component mixture. The EOS was tuned toconstant-composition-expansion and constant-volume-depletion data. The a and bof methane and the four heavy pseudocomponents and the binary interactioncoefficients between methane and each of the four heavy pseudocomponents wereadjusted in the manner suggested by Coats pseudocomponents were adjusted in themanner suggested by Coats and Smart. Agreement between the results of the tunedEOS and the laboratory data was excellent, the reservoir conditions duringsimulation were well within the limits of the data used in tuning, andreservoir conditions were well removed from the critical point of the mixture.Thus, the characterization of the reservoir fluid was adequate for the purposeof this study. Dewpoint pressure of the original reservoir gas was 5,170 psia,and the initial reservoir pressure at the top of the reservoir psia, and theinitial reservoir pressure at the top of the reservoir was 5,185 psia. Theendpoint of the three-phase relative permeability to liquid condensate occurredat a condensate saturation of 8.5% and a gas saturation of 75.0%. Irreduciblewater saturation was 16.5%. The reservoir discussed in this paper was a radial160 acres made up of five layers of various thicknesses. Vertical andhorizontal permeabilities were equal, The permeability-thickness product washigh, about 700 md-ft. Sensitivity runs indicated, product was high, about 700md-ft. Sensitivity runs indicated, however, that when the gas production ratewas normalized to percent of capacity, the value of the permeability-thicknesspercent of capacity, the value of the permeability-thickness product wasimmaterial to the conclusions of the study. product was immaterial to theconclusions of the study. The grid pattern was radial 5 layers by 12 segments.The solution was implicit with D4 Gaussian elimination. The results presentedin this paper are for a well perforated across the entire interval. Wells withpartial completions acted like layered reservoirs and should be considered assuch for the purposes of planning a fluid sampling program. purposes ofplanning a fluid sampling program. Results
Fig. 1 shows the total producing GOR (separator plus stock-tank gas) andquantity of heptanes plus in the recombined surface samples for a wellproducing at a constant rate of about 15 % of capacity. Although the productionrate is low and constant, a sample representative of the original reservoir gascan be obtained only during the first 30 days of production. After 30 days, theloss of condensate from the gas in the reservoir results in a decrease ofsurface liquid. This causes the producing GOR to increase from the initialvalue of 6,633 scf/STB and the quantity of heptanes plus in the recombinedsurface samples to begin a steady decline away from the value of 7.04% in theoriginal gas. The composition of heptanes plus is selected as the criterion ofa good sample for two reasons. First, heptanes plus is the component mostaffected by loss of condensate in the reservoir. Second, the composition andproperties of heptanes plus strongly affect the properties of the recombinedsurface sample. At 330 days, the gas production rate was reduced to 5% ofcapacity in an attempt to get a good sample. This caused an instantaneouschange in producing GOR. However, GOR did not stabilize; rather, it continuedto increase steadily. The quantity of heptanes plus in the recombined samplesdid not recover to the correct value. Obviously, the cutback did not result ina good sample. Once the opportunity of obtaining a good sample is lost,reducing the rate or even shutting in will not improve the chance of getting agood sample (as shown later). This shows that sampling must take place early inthe production of a well. The remainder of this paper focuses on the earlyproduction period.
Condensate Ring. Fig. 2 shows the buildup of the condensate ring around thewellbore as pressure falls below the dewpoint.
|File Size||929 KB||Number of Pages||5|