akp a bio 552014 air 16000

7/21/2019 Akp a Bio 552014 Air 16000

http://slidepdf.com/reader/full/akp-a-bio-552014-air-16000 1/11

_____________________________________________________________________________________________________

*Corresponding author: E-mail: [email protected];

Advances in Research5(5): 1-11, 2015, Article no.AIR.16000

ISSN: 2348-0394

SCIENCEDOMAIN internationalwww.sciencedomain.org

PVT Fluid Sampling, Characterization and GasCondensate Reservoir Modeling

Julius U. Akpabio1,2*, Sunday O. Isehunwa2 and Oluwatoyin O. Akinsete2

1Department of Chemical and Petroleum Engineering, University of Uyo, Uyo, Nigeria.

2 Department of Petroleum Engineering, University of Ibadan, Ibadan, Nigeria.

Authors’ contributions

This work was carried out in collaboration between all authors. Author JUA designed the study, wrote

the protocol, and wrote the first draft of the manuscript. Author OOA confirmed the cited literatures,and author SOI reviewed the simulation process, analysis and results. All authors read and approvedthe final manuscript.

Article Information

DOI: 10.9734/AIR/2015/16000Editor(s):

(1) Pradip K. Bhowmik, Department of Chemistry, University of Nevada Las Vegas, Las Vegas, USA.Reviewers:

(1) Anonymous, Suez Canal University, Egypt.(2) Anonymous, Inha University, South Korea.

(3) Anonymous, Ladoke Akintola University of Technology (LAUTECH), Nigeria.Complete Peer review History: http://sciencedomain.org/review-history/10630

Received 31st

December 2014 Accepted 9

th June 2015

Published 23rd

August 2015

ABSTRACT

When reservoir pressure decreases in gas condensate reservoirs, there is a compositional changewhich makes the system difficult to handle. This type of system requires an Equation of State (EOS)to ensure proper fluid characterization so that the Pressure Volume Temperature (PVT) behavior ofthe reservoir fluid can be well understood. High quality and accurate PVT data will help reservoirengineers to predict the behavior of reservoir fluids and facilitate simulation studies. The aim of thisstudy is to determine what to do on reservoir fluid before carrying out reservoir modeling.

PVT data were obtained from a reservoir fluid in the Niger Delta which was sampled followingstandard procedures. Then the laboratory experiments were critically examined to ensure accuracy,consistency and validity before PVT analysis. Finally, the results from the PVT experiments wereimported into PVT software and subsequently in a reservoir simulator for simulation studies. Theseprocesses generate the EOS model for reservoir modeling of gas condensate reservoirs.The mass balance test, Hoffman plot and CCE/CVD (Constant Composition Expansion andConstant Volume Depletion) comparison plots were used to validate PVT data. From these tests,

Original Research Article

7/21/2019 Akp a Bio 552014 Air 16000


7/21/2019 Akp a Bio 552014 Air 16000


Akpabio et al.; AIR, 5(5): 1-11, 2015; Article no.AIR.16000

3

EOS can be modeled with the following generalprocedure:

a) EOS model is built with an EOS correlationusing available composition at reference

pressure, temperature and depth.b) To ensure that the GOR and density

match, the molecular weight or specificgravity of the pseudo component ischanged by 5-10% to get close to thedensities.

c) Binary Interaction Parameters (BIP) areused to match saturation pressure whilethe Pseudo components are split.

d) Laboratory data is entered and deviationbetween the calculated and experimentalvalues is checked.

e) The BIPs and critical properties of pseudocomponents are regressed if a large

deviation is noticed.f) EOS model is lumped for use in simulation

model, after linear regression to reduce thesimulation time.

EOS correlation is used with accurate PVTcharacterization to develop gas condensatereservoir models. When developing gascondensate reservoirs, a major challenge is thephenomenon called “condensate banking.” Thisphenomenon occurs when condensate drops outnear wellbore region as pressure drops belowthe dew point pressure, causing condensate todrop and form a ring-like structure whichultimately reduces the well deliverability [7].

There is need for accurate sampling of thereservoir fluids to achieve a good EOS model.Laboratory experiments must be performedcorrectly to develop accurate EOS models whenthe sample is a good representative of thereservoir fluids [6]. PVT models are used togenerate mathematical algorithms expressed asEquation of State [8]. Some practical limitationsin obtaining gas condensate PVT data includeaccurate laboratory PVT data and compositionalchanges due to pressure drop below the dewpoint and uncertainties associated with smallliquid volumes [9]. The profitability in the

development of any gas condensate reservoirdepends on four factors: field location and size,local markets for separated gas and condensate,phase behavior of reservoir fluid and tax regime.

A fundamental tool for planning the developmentof a field and evaluation of field productionperformances is reservoir simulation studies.Part of the requirement for any integratedreservoir studies is the reservoir fluid PVT model

[8].

1.1 Fluid Characterization and GasCondensate Fluids

One way of obtaining representative reservoirfluid, is by sampling the fluid just after thecompletion of the drilling, since pressure is lesslikely to drop below the dew point therebycreating a two phase in the reservoir [10-13]. Amono-phasic condition should be maintainedduring sampling and transfer to laboratory foranalysis. In order to realize this objective, thesample drawdown pressure should be controlledand kept as close as possible to the reservoirpressure and above the dew point [14,15].

From the well stream composition of a reservoirfluid (which is obtained from the recombinationprocess), it is possible to know the type ofreservoir fluid. The major characteristic feature ofa gas condensate fluid is the Gas GOR (Gas-OilRatio). The condensate fluid can be furtherclassified into four categories: Lean, medium,rich and very rich condensate [16] as shown inTable 1. As the isothermal condition of the

reservoir fluid approaches the critical point, in thephase envelope, the richness of the fluid isincreased. Fig. 1 shows the lean and rich gascondensate phase envelopes [17].

Rich and lean gas condensate phase envelopescan be differentiated by Fig 1a and 1b andknown by the size of heptane plus as well as thepercentage of liquid dropout. When pressuredeclines at reservoir temperature, a rich gascondensate forms more liquid dropout than alean gas. The constant compositional changes inthe gas condensate reservoir, makes it acomplex system, requiring compositionalsimulation to be able to model the phasebehavior of the fluid and evaluate the recoveryprocesses properly.

Table 1. Classification of Gas Condensate Fluids [16]

Lean Medium Rich Very rich

CGR(STB/MMSCF) <50 50 - 125 125 - 250 >250

7/21/2019 Akp a Bio 552014 Air 16000



4

Fig. 1a. and 1b. Phase envelopes of rich and lean gas condensate fluids [17]

2. METHODOLOGY

Fluid samples for this study were obtained fromthe Niger Delta region of Nigeria. The record ofthe sampling process and information about thetemperature and pressures were used toascertain the suitability of the condition forobtaining representative reservoir fluid samples.Table 2 shows the compositions of the gascondensate fluid and other properties of the fluid.Standard sampling procedures will be highlightedand laboratory analysis performed on the fluidbefore characterization and EOS generation.

2.1 Sampling Procedures for GasCondensate Fluid

2.1.1 Subsurface samzpling

The following steps are the specific procedure forsubsurface sampling:

i. Condition the well to insure that a single-phase, representative fluid is flowing at the

productive intervalii. Either shut in the well or allow it to

continue flowing at a very low rateiii. Run pressure and temperature surveys to

determine fluid levels and pressuresiv. Select the sampling point and run the

bottom hole fluid sampler to depthv. Actuate the sampler and retrieve the

sample.

vi. Repeat the sampling operation to obtainduplicate samples (preferably threesamples should be retrieved).

vii. Perform a quality check on the samples atthe surface

viii. Transfer the samples to a storagecontainer for transport to the laboratory

A minimum of triplicate samples should becollected. This is to permit comparison of samplecompositions and properties and to have backupsamples in case of leakage during transit fromfield to laboratory. For gas-condensate samples,the optimum procedure is to ship the entiresubsurface sampling tool section to thelaboratory, to minimize the possibility of leakage[13,18].

2.1.2 Surface sampling

The following steps are the specific proceduresfor surface sampling [19]:

i. Condition the well to insure that a single-phase representative fluid is flowing intothe wellbore

ii. Maintain the final conditioning flow rateiii. Accurately measure and record the GORiv. Sample the gas and oil streams at the

primary or first stage separator and atseparator pressure.

v. Accurately record sample data and tag forshipment to laboratory.

7/21/2019 Akp a Bio 552014 Air 16000



5

2.2 Analysis and Tuning of PVT Data

The results from the laboratory experimentswhich are the CCE, CVD, viscosity and separatortests were inputted into the PVTsim software.

The Fluid is characterized by delumping andlumping of the plus fractions and assigning ofindividual properties (e.g. Tc, Pc, accentric factor,Mw etc.) to these components using an equationof state (EOS). The EOS parameters were tunedto match experiments’ PVT data of the CCE andCVD tests with the simulation results. Thelumping and delumping of the C7 plus fractionwas necessary to reduce the number ofcomponents used in the EOS calculations alsoreduce compositional model computing time [20].

2.3 Basic Experiments before FluidCharacterization

Constant Composition Expansion (CCE) andConstant Volume Depletion Experiments (CVD)are the two basic experiments carried out on gascondensate fluids before characterization. CCEexperiment is performed in a high-pressure gas

condensate windowed cell. A part of therecombined sample was changed to the cell andexpanded thermally to the reservoir temperatureof 176.6ºF. This experiment is started at apressure much higher than the reservoir

pressure and reduced stepwise until the dewpoint pressure is observed and recorded. Otherparameters recorded in this test are the deviationfactor, the compressibility factor, the liquiddropout, gas density and relative gas volumes asseen in Table 3.

CVD experiment is carried out on the fluid atreservoir temperature of 176.6ºFand dew pointpressure which was determined by the CCEexperiment. The experiment involves a series ofpressure expansions and constant pressuredisplacements to maintain the sample in a

constant volume that was equal to the volume ofsample at dew point pressure. The process isrepeated until an abandonment pressure, whichwas 524 Psia. The well stream was pumped fromthe PVT cell into a pre-weighed flask submergedin liquid nitrogen and condensed. The condensed

Table 2. Wellstream compositions of fluid A

Component Gas mol % (yi) Liquid mol % (xi) Reservoir fluid mol % (zi)

N2 0.15 0 0.14CO2 0.18 0 0.18H2S 0 0 0CI 87.98 0 87.26C2 5.29 0.1 5.25

C3 2.83 0.07 2.81i-C4 0.68 0.06 0.67n-C4 0.91 0.13 0.9i-C5 0.41 0.21 0.41n-C5 0.31 0.25 0.31C6 0.56 1.73 0.57C7+ 0.7 97.45 1.5

TOTAL 100 100 100

Liquid density (IB/FT3) 50.067Liquid MW 156.37Gas gravity (air = 1) 0.802GOR (scf/sepbbl) 82918.7

Table 3. Constant volume depletion test-produced well stream properties at 176.6ºF

Steps Press.

(Psia)

Gas Den.

(g/cc)

Gas

gravity

Gas Z

factor

Gas FVF Gas

Visc.(cp)

2-Z

factor

Retrog

radeliq. (%)

Cum.

Prod.fluid

DewPoint.

4191 0.238 0.721 0.862 0.00369 0.0276 0.862 0 0

1 3799 0.187 0.702 0.831 0.00392 0.0251 0.824 0.27 4.922 3099 0.153 0.678 0.794 0.0046 0.0214 0.786 1.1 13.363 2388 0.118 0.664 0.79 0.00594 0.0183 0.768 1.93 28.654 1610 0.078 0.657 0.83 0.00925 0.0158 0.746 2.65 46.985 978 0.046 0.656 0.905 0.0166 0.0142 0.729 2.61 64.046 524 0.025 0.662 0.962 0.0325 0.0134 0.708 2.25 74.05

7/21/2019 Akp a Bio 552014 Air 16000



6

gas phase is then gradually allowed to return toambient temperature. The gas evolves and theresidual condensate are collected separately,weighed and analyzed. The quantities that arerecorded during this experiment are liquid

dropout, cumulative produced fluid; gas densityand gas z-factor, see Table 4. Phase andvolumetric behavior of mixtures using any of theEOS models can be predicted by obtaining suchproperties as the critical properties (Tc, Pc) andaccentric factor, ɷ for each component in themixture.

2.4 Compositional Consistency

The consistency of the fluid composition can bedetermined by evaluating the PVT data. Onemajor consistency test is the summation of thecomposition to ensure they add up to 100%. In

modern PVT reports, inconsistencies incomposition are usually very small and are seenin unnoticeable figures instead of errors [21]. Apart from the summation of the composition,other techniques for consistency checks are, themass balance plot, Hoffman plot and Buckleyplot. While the mass balance plot is aquantitative method, the Hoffman and Buckleyplots are qualitative methods of assessing PVTdata.

2.4.1 Mass balance test

This test is used to assess the feed compositionand the separator vapour and liquid compositionfor consistency. The basis for the test is themass balance criteria of the component. Onemole of fluid of composition z is considered at acertain temperature and pressure (T, P); it canbe split into liquid and vapour of L moles and Vmoles of compositions respectively [22]. An Lmole of liquid has the compositions x1, x2, … xn,and a V mole of vapour has compositions of y1,y2, …yn. Then

)1(1V L

and

)2(niii Z Vy Lx

)3(1111

n

i

i

n

i

i

n

i

i z y x

Therefore the K-value of the ith

component isexpressed as:

)4(ii

x y K

Hence, the expression for the xi and yi are asfollows:

Dividing through by Z

)6( F z yV z x L iiii

Dividing through by V will give

)7(V F z y z xV L iiii

This translates to

This is a straight line equation obtained by

plotting ii z y against ii z x to generate an

intercept, F/V and a negative slope (L/V); thisnegative slope is equivalent to the measuredGOR. Table 2 is used to generate this massbalance plot. Deviation from the straight line canbe seen as mass balance inconsistency. The R

2

value must tend towards unity (Fig. 2). This plotis sometimes used to identify discrepancies inthe reported compositions. The reciprocal of theslope may be used to compute GOR andthereafter compare with the measured GOR. Theconversion from mole to barrels is necessary

when the values of the liquid density andmolecular weight are provided. The feedcompositions used in this study yielded a goodresult in terms of consistency. The separatorliquid and vapour can be mathematicallyrecombined to obtain the well streamcomposition. From the graph, the value of theslope was 0.0083 which gives a GOR of 82201.8scf/sep. bbl. upon conversion of the reciprocal ofthe slope. The difference between the calculatedand measured GOR is 0.86%. The measuredGOR is 82918.7 scf/sep. bbl. This result showsthat the value of the liquid molecular weight anddensity are close to the reported GOR. If this

difference is large, then the values of thereported liquid molecular weights and densitiesare inaccurate. This will make the tuning bylinear regression of the equation of state difficult.

The mass balance of the gas composition can beexpressed by Equation (9).

)9(0088.10083.0 x y

)8( V F z L z i iii

) 5(iii Fz Vy Lx

7/21/2019 Akp a Bio 552014 Air 16000



7

And R2 is 0.9997

Using some conversion factors, the slope of the mass balance plot can be expressed as GOR.

)10(6145.5*067.50*37.156

1*5.379*0083.01

3

3 bbl ft

ft lb

lblbmol

lbmol scf

Ibmol lbmol GOR

stb scf GOR /8.82201

Fig. 2. Mass balance plot of recombined fluid sample

Fig. 3. Hoffman plot for the fluid A

y = -0.0083x + 1.0088

R² = 0.9997

0.0

0.20.4

0.6

0.8

1.0

1.2

0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0

y i / z i

xi/zi

HC Component

7/21/2019 Akp a Bio 552014 Air 16000



8

2.4.2 The hoffman plots

A method for checking the recombined separatorsample was proposed by Hoffman et al. [23].This following expression was used to obtain the

Hoffman plot based on the correlation of Kvalues.

)11(log 1 ioi F A A K

Where

)12(log

11

11 scci

cibi

bii P P

T T

T T F

A1 and Ao are slope and intercept respectively fora semi-log plot of Ki versus Fi. This plot showsthat the trend of the graph is linear forcomponents C2 through C6 for all pressures asseen in Fig. 3. However, it bends slightlydownwards for heavier components at lowpressures.

2.4.3 Buckley plot

The Buckley plot is expressed by the semi logplot of K-value versus the square of the criticaltemperature (Fig. 4). Usually, the heavier

component deviates downwards away from thestraight line.

3. RESULTS AND DISCUSSION

The specific gravity of Heptane plus is generatedfrom the reservoir fluid equation for expressingthe relationship between specific gravity andmolecular weight before characterization of thefluid [24].

)13(9418.652855.0 1299.0 MiC f i

Where, Cf is the specific gravity correlation

characteristic factor. This factor is adjusted toobtain equality between calculated andexperimental specific gravity of Heptane plus.The value of Cf is between 0.27 and 0.31.

Table 4. CCE experiment at 176.6ºF

Steps Pressure Rel. vol. Retro. liq. % Density FVF Z Comp. (1/psia)

1 7062 0.769 0 0.309 0.00285 1.119 6.15E-052 6559 0.793 0 0.3 0.00294 1.072 7.19E-053 6056 0.822 0 0.289 0.00305 1.025 8.36E-054 5553 0.855 0 0.278 0.00318 0.979 9.89E-055 5051 0.899 0 0.265 0.00334 0.936 9.96E-05Res. Press 4868 0.915 0 0.26 0.0034 0.918 1.16E-047 4548 0.949 0 0.25 0.00353 0.89 1.50E-04

PD 4191 1 0 0.238 0.00371 0.8629 3543 1.118 0.5710 3041 1.247 1.2211 2741 1.36 1.612 2440 1.512 2.0813 2139 1.718 2.3814 1837 1.987 2.7115 1536 2.391 2.8416 1215 2.896 2.79

Fig. 4. Buckley plot for the reservoir fluid components

y = -0.5779x + 4.0214

R² = 0.9036

-3.0

-2.0

-1.0

0.0

1.0

2.0

3.0

0.0 2.0 4.0 6.0 8.0 10.0

L o g K

Tc2 /105

Non-HC Component HC component Linear (HC component)

7/21/2019 Akp a Bio 552014 Air 16000



9

Non-linear regression is used to determine thecorrelation constants to fit the experimental datafrom the True Boiling Point (TBP) and Mi is themolecular weight of heptane plus. This equationwas originally generated based on a comparison

of experimental and calculated specific gravity of68 samples.

3.1 CCE and CVD Comparison Plots

Two major experiments for gas condensate fluidsfor determination of the reservoir phase behaviorare Constant Composition Expansion (CCE) andConstant Volume Depletion (CVD) [25]. Theseexperiments help in tuning the equation of statemodels for modeling reservoirs. Informationobtained from these experiments are importedinto PVT softwares to generate the actualreservoir fluid properties. Important information is

obtained by plotting the CVD and CCE together.The dew point pressure of the reservoir was4191 psi. and the plot shows the consistency oftge two experiments. The plot has an identicaldew point while the liquid dropout of the CCE isless than that of the CVD. The CCE test gives amaximum liquid dropout at a lower pressure thanthe CVD test. At the beginning of the retrogradeprocess, the curves are identical but the CVD willshow greater liquid dropout than the CCE asmore liquid condenses. This happens becausewhen retrograde condensation starts, thecomposition of the gas phase becomes leanerthan the liquid phase. During CVD experiments,

the composition of gas will be richer than theoriginal composition during the CCE experimentwhen the leaner gas is removed. More liquiddropout is experienced because the total CVDfluid composition is richer. This is the reason whythe CCE and CVD curves will initially track each

other before the CVD curve lies beneath theCCE curve. (Fig. 5).

3.2 Linear Regression of Data

After the CCE and CVD data are obtained, thereis need to use non-linear regression to adjust thepredicted EOS characterization in order to obtainan acceptable match. Some parameters need tobe selected based on experience because thetuning process is usually a trial and errorprocess. Out of the several non-linear leastsquares regression methods being implementedand tested, the rotational discrimination methodis recommended. To obtain a good matchbetween the simulated and experimental data aset of BIPs need to be adjusted. These BIPs helpto predict the capabilities of the EOS used inVolume Liquid Equilibrium (VLE) calculations of

reservoir fluids .A reasonable match could beachieved with an error of±(5-10%) then the EOScan be imported into a simulation model such asECLIPSE100, 300, Dynamo/Mores which willnow function as a compositional model.

3.3 Gas Condensate Reservoir Modeling

A compositional simulator tracks eachcomponent of the oil and gas in the reservoirfrom the least component to the largest (C1,C2……Cn). By doing this it models fluid near thecritical point in the phase envelope, wherechanges in the pressure and temperature of thecompositional system can result in very differentphase behavior. In this system, oil and gasphases are represented in multi-componentmixtures while composition and time arerepresented by EOS.

Fig. 5. Comparison plots of CVD and CCE experiments

0

1000

2000

3000

4000

5000

6000

7000

8000

0 1 2 3

P r

e s s u r e ,

P s i

Liquid dropout %

CCE Liquid dropout %

CVD Liquid dropout %

7/21/2019 Akp a Bio 552014 Air 16000


7/21/2019 Akp a Bio 552014 Air 16000



11

Peer-review history:The peer review history for this paper can be accessed here:

http://sciencedomain.org/review-history/10630

17. Akpabio JU, Udofia EE, Ogbu MP. PVTfluid characterization and consistencycheck for retrograde condensate reservoirmodeling SPE 172359 prepared forpresentation at the SPE Nigeria Annual

International Conference and Exhibitionheld in Lagos, Nigeria, 5-7 August; 2014.

18. Strong J. Thomas FB. Bennion DB.Reservoir fluid sampling andrecommendation techniques for laboratoryexperiments. Paper CIM 9354 presented atthe annual Technical Conference inCalgary, 9-12 May; 1993.

19. International Human ResourcesDevelopment Corporation (IHRDC) andInternational Petroleum IndustryMultimedia System (IPIMS); 2013. Avaliable:http://www.ihrdc.com

20. PVT sim Software, Calsep; 2012.

Avaliable:http://www.pvtsim.com21. Imo-Jack O, Emelle C. An analytical

approach to consistency checks ofexperimental PVT data. Paper SPE167560 presented at the Nigerian AnnualInternational Conference and Exhibition

held in Lagos, Nigeria, 30 July–1 August;2013.

22. Bengherbia M, Tiab D. Gas-condensatewell performance using compositionalsimulator: A case study. Paper SPE75531 presented at the SPE GasTechnology Symposium, Calgary,Canada, 30 April–2 May; 2002.

23. Hoffman AE, Crump JS, Hocott CR.Equilibrium constants for a gas-condensate system. PetroleumTransactions, AIME. 1953;198:1-10.

24. Soreide I. Improved phase behaviorprediction of petroleum reservoir fluidsfrom a cubic equation of state. Dr. Techn.Theses Norwegian Institute of Technology,Department of Petroleum Engineering;1989.

25. Trengove RD, Hann JH, Skates JR. Theimpact of PVT data quality on hydrocarbonrecovery predictions SPE 22988 preparedfor presentation at the SPE Asia PacificConference held in Perth Western Australia, 4-7 November; 1991.

_________________________________________________________________________________© 2015 Akpabio et al.; This is an Open Access article distributed under the terms of the Creative Commons Attribution License( http://creativecommons.org/licenses/by/4.0 ), which permits unrestricted use, distribution, and reproduction in any medium,

provided the original work is properly cited.

akp a bio 552014 air 16000

Documents