You are on page 1of 11

PETROLEUM SOCIETY OF CIM AND CANMET

PAPER NO. 95-09

TOWARDS OPTIMIZING GAS CONDENSATE RESERVOIRS


F .B. Thomas

X.L. Zhou
D.B. Bennion
D.W. Bennion
Hycal Energy ResearchLaboratoriesLtd

ABSTRACT

2.
3.

In the last year the authors have fielded many


questions from companies, both international and domestic,
concerning gas condensate reservoirs. It appears that gas
condensates are becoming more important throughout the
world Many international petroleum societies are beginning to
have conferences specifically oriented to gas condensate
reservoirs and discussing all parameters germane to such
systems. In light of this increased interest, the authors have

madea short list of questionswhich are mostoften asked.


Indeed; these questions point to two specific areas which govern
the production and future exploitation plans for gas condensate
systems. These two areas are characterization and retrograde
condensate influences on relative permeability.

It has beenfound that the characterizationof the gas


condensatefluid\' can be strongly influenced by two main
factors:
1.

2.

Any degree of contamination by a free liquid phase insitu;


Hold-up of the retrograde condensate in the formation
resulting in excessiveproducing GOR's.

Care must be taken when sampling gas condensate


wells in order to produce representativerecombinedfluids. In
order to gain an appropriate evaluation of the gas condensate
reservoironemustbe able to adequatelycharacterizethefluids
in-situ. Experimental and theoretical work performed on
evaluatingretrogradecondensateeffectshaspointed to thefact
that the influence of retrograde condensateis much more
deleteriousin tighter formations and higher interfacialftuw.
Theability to identify the influenceof retrogradeliquid on gas
phaseproduction rates is a difficult taskand data are provided
herein which comparethe retrogradecondensateeffectsatlwo
levels of interfacial tension and as a function of rock
permeability.
It has been found that in a review of four gas
condensatereservoirs, one of which included afractured system,

therewasa couplingof a multiplicityoffactorsincluding:


1

Interfacial tensioneffects

Viscosity ratio
The healing of fractures with its concomitant effect on
absoluJe permeability

In order to adequatelyforecast such systems, a


simulator must incorporate theseeffects.

SAMPLING CONDENSATE RESERVOIRS


Condensatereservoirsare inherently more difficult to
characterizecorrectly. The literature showsmany differences
betweengas condensatereservoirsand dry gas reservoirs(I-6).
One questionoften askedis during and after sampling. Figure
1 providesa fairly typical GOR versustotal flow rate response
from a gas condensatereservoir. One seesthat, at very low
flow rates, one has a high producing GOR and, beyond the
certainminimum value in GOR, the trend is againupwards. It
is easyto identify why this occurs,but sometimes,when faced
with the possibility of havingextrasamplingrunsand spending
more time in the field, the generationof a plot such as Figure
1 is not easy.
In the same plot one comparesthe responsewhich
would nonnally be seen for an oil reservoir. With the oil
reservoir,the samplingtechniqueis fairly easyto specify. All
one must do is try to produce the well in the domain low
enough so that a constant GOR is produced. Since the
behaviouris asymptoticas a function of decreasingtotal flow
rate from the well, it is easyto identify what production level
oneneedsto apply for taking the gasand liquid samples.Such
is not the casewith gascondensatereservoirshowever. At low
flow rates,asshownin Figure I, one may be producingenough
liquid in the wellborethat, unlessthe flow rate is high enough,
the liquid hold-up will increaseand slugging may result. In
such a case,the GOR, dependingupon at which interval the
sampleis taken,may fluctuateand result in an excessivelyhigh
GOR. The increasedGOR to the left of the vertical line is due
to the low flow rate not providing enoughlift to transportthe
liquids in the wellbore.
By contrast,onemay be inducing liquid dropoutin the
reservoirat high flow ratesto the right of the vertical line in
Figure I. In such a case, as the pressuredrops below the

dewpoint, the liquid will drop out and begin to collect in the
nearwellbore region. In so doing, producedhydrocarbonswill
containlessliquid than they shouldand thereforethe GOR will
be high. Thus,when oneasksthe question,"At what producing
rate should a well be sampled?",the only way that one can
adequatelyrespondis if one alreadyknows the characterof the
fluid and the dynamics of the production well. Since this
informationis not available,samplinggascondensate
reservoirs
in an optimal mannercan sometimesbe non-linearand include
sometrial and error.
How does it influence the overall characterizationif
one is producing in other than the optimal zone of the
production scenario? Figure 2 shows a typical response
obtainedwhen recombiningseparatorgasand liquid from a gas
condensatewell. In this case,the asteriskcorrespondsto the
separatorGOR which was observedin the field during the
sampling. As is often the case,the relationship in Figure 1
was not scopedduring sampling, and thereforeFigure 2 may
correspondto producing the well in the region to the right of
the vertical line in Figure 1. Figure 2 indicates that the
saturationpressureis higher than the reservoir pressureand
thereforesomethingis wrong. For systemslike this wherethe
deviation between the observed saturation pressureof the
recombinationandthe reservoirpressureis lessthan or equalto
1000 psi, one can normally trust that a manipulation of the
GOR will result in satisfying the necessarycondition for the
recombination. The necessarycondition is that the saturation
pressureof the recombinedsamplemust be equalto or lessthan
the reservoirpressureat reservoirtemperature. By decreasing
the GOR by adding additional separator liquid, one will
suppressthe dewpoint down to a level which meets the
reservoirpressurecriteria. Should the real dewpoint pressure
be lower than the reservoirpressure,one really hasno way of
knowing that unlessthe owner was preparedto return to the
field and measurethe GOR dependencyon the flow rate of the
system. One may also observesituationswhere,by increasing
the GOR slightly, the saturationpressuremay convergequickly
to the reservoirpressure. This is an option but, in light of the
fact that the gasphaseis more mobile thanthe liquid phase,this
is usually not the procedurewhich is followed.
Nevertheless, with the response as in Figure 2, the
modified GOR results in a saturation pressure equal to the
reservoir pressure and the recombined fluid properties can then
be measured including such things as overall composition,
constant composition expansion data as well as constant volume
depletion properties.
Figure 2 provides the best case scenario in terms of the
credence which can lend to the resulting recombination. The
modified GOR fluid still possessesall of the characteristics
which are consiStent with expectation with a slight modification

for GOR which in and of itself is also logical in tenns of the


higher gasmobility asmentionedearlier. A more difficult case
would be that shown 1n Figure 3 where there is a substantial
deviationbetweenthe saturationpressureat the separatorGOR
and the reservoir pressure. In the experienceof the authors,
often when the saturation pressureof the system shows a
positivedeviationfrom the reservoirpressureof morethan 1500
psi, it is usually due to the fact that somethinghasoccurredin
the liquid phase. Work has beenperfonned by the authorsin
the past, wherein a very slight change in the heavy end
characterof a gas condensatesystemcan result in significant
changesin the saturationpressure.This againis muchdifferent
than for a conventionaloil system. For example,if one hasa
slight mixing of two different oil zonesin the reservoir, the
influence of the two may be fairly insignificant on the
saturationpressureproperties. This is due to the fact that the
oil is already a bubblepoint system and, therefore the
contaminantswhich alsohavea tendencyto remain in the liquid
phase may modify the bubblepoint behaviour, but the
componentsthemselvesare both consistentwith a bubblepoint
response. On the other hand, if one were to have a
contaminationdue to a free oil phasein-situ, which exhibits a
bubblepoint response,then the dewpoint behaviour of the
dominantgascondensate
phasecanbe radically different. This
is consistentsinceone must increasethe pressureof the system
to a high enoughlevel in order to changethe behaviourof the
contaminantcomponentsfrom a naturalbubblepointresponseto
a dewpointresponse.This is tantamountto having to vaporize
all of the heaviercomponentsand in orderto do so a significant
deviation in pressure from the reservoir pressure is often
encountered.This canthereforebe usedasa criterion by which
onecanjudge if the deviationfrom the reservoirpressureis due
to a liquid phasecontaminantor simply a perturbationin the
level of producingGOR.
The ways of remedyingthe deviant behavioursuchas
that in Figure 3 is also more challenging. Obviously, oncethe
contaminating liquid components have entered into the
condensatecomponents,it is extremelydifficult, if not totally
impossible,to separatethe two. Therefore,if that is the only
separatorliquid obtained, which is a combination of gas
condensatecomponentsand oil components,then one must
either be preparedto re-sampleor reducethe GOR to a low
enoughvalue to simply meet the necessarycondition of the
saturationpressurebeing equalto the reservoirpressure. In so
doing however,one will only exacerbatethe characterization
problems. Two reasonsare responsible. The first is that by
decreasingthe GOR, one will only emphasizethe liquid phase,
and therefore if the liquid phase is the one which is
contaminated,onewill be emphasizingthe contamination.This
will serveto make the subsequenttesting with this fluid less
representative
of the bulk of the reservoir,assumingthat the oil
phaseis only a very small portion of the overall reservoirand

in-situ. Rememberthat, although the reservoir was almost


totally dominatedby the gas condensatephase,the production
of the oil phasewas enoughto result in approximately 100kg
of solids producedper 1000barrelsof liquid phaserecovered.
Although in a laboratory perspectivethis mass percentageis
extremelysmall,from an operatingperspectiveit wasmorethan
an annoyance.

that the bulk of the reservoir is a gas condensate system. By


decreasing the GOR to a level where the Pial is equal to the
reservoir pressure, the doubt will increase as to whether the data
produced from any studies will then be meaningful.

Secondly,asoneseessomeliquid phasecontamination
which requires a distinct drop in the GOR for the
recombination, one will often see a response where the
behaviour will change from a dewpoint to a bubblepoint.
Therefore,the implications of this decreasein GOR are very
serious.
Moreover, the approach to be taken for the
exploitation of such a reservoirmay also becomeobscuredin
that, ratherthan being a gascondensatesystem,it may become
a light oil system,therebybringing otherquestionsinto the fray
such as, "Would this be a better candidatefor a gas injection
process or waterflooding instead of primary depletion or
possibly a gascycling project?".

Another system recently analyzed showed a very waxy


condensate phase containing components as high as Cso'
Because of the very high temperature and pressure of this
reservoir, these components could actually exist in the gas phase
at reservoir conditions and, upon depletion and liquid phase
dropout, the system exhibited vapour liquid solid behaviour. In
this case, die cloud point temperature was exactly the same as
the melting point of the solid components which were
centrifuged out of the liquid phase at the cloud point
temperature. In other words, it was a totally reversible
phenomenon normally associated with nothing but a pure wax
system. That was a fairly unique fluid but. in that case. there
was vapour liquid solid equilibrium with a system which
exhibited a dewpoint pressure 200 psi less than the reservoir
pressure and no signs of contamination. The differentiating
feature in that case was that the molecular weight of the solid
phase was on the order of 450 and the solid was reversible.
For these reasons it is more credible that it was associated with
a gas condensatesystem with no free oil contamination effects.

One other behaviour often associatedwith a gas


condensatesystem which shows liquid phase contamination
problems is that there may be some solid phase instability.
Typically, therearevery few solids which will precipitatefrom
a gas condensatesystem. Rarely one sees diamondoid
precipitationand, althoughmore frequentlythe precipitationof
sulphur-containingcompoundsmight be a problem, one can
usuallytracethoseto the compositionand the conditionsof the
productionstring. The situations,however,showingdark high
molecularweight solids beingproducedin the productionwells
are usually representativeof solid phase componentsbeing
precipitated from the phase which is contaminatingthe gas
condensate
componentsand is not an inherentproblemwith the
light gascondensate
componentsthemselves.It is usuallya rule
of thumb that if the molecularweight of the producedsolids is
greaterthan 500 - 700 andthe solubility of thosesolids is high
in an aromatic solvent, such as toluene, then there is a good
chancethat the solids originatewith a separateoil phasein-situ,
andnot only is therea problem with the characterizationof the
gas condensatephasebut the oil is also causing production
problemsdue to solid componentsincompatibilities.

In summary,therefore,when characterizingthe fluid


obtainedfrom samplinga gas condensatereservoir, one needs
to be awareof the following practices:
There will be an optimal producing flow rate for the
well in question. There is little possibility of being
able to predict a priori what the appropriateflow rate
is for samplingthe well. This is due to the fact that
the separator response will be a function of the
interfacial tension, the viscosity ratio, the reservoir
rock characteristics,the tubing size and type as well
the depletion parametersof the gas condensateitself.
The best possiblething to do if one needsto obtain a
representative characterization without having to
repeatedlyreturn to the well, if the fluids are not
representative,is to perform a GOR versusflow rate
sequencewhile sampling so that the best possible
samplescan be taken.

In the last two years,the authorshaveencounteredtwo


gas condensatesystems which exhibited solid precipitation
problems. One had the responseshown in Figure 4 where
solids precipitatedat high temperatureas long as the pressure
was high whereasat lower pressuresno solids were fonned,
even at low temperatures. A responsesuch as this would
usually be representativeof a liquid phasecontaminationof the
gas condensatesince the solids in this case were pressuredominatedand thereforecorrespondedto an asphaltenetype of
deposit. Sincethe vapourpressuresof aspbalteneareextremely
low, it is highly unlikely that thesecomponentsare produced
from the gas phasein-situ and thereforethis is a telltale sign
that the incompatibility wascausedby a contaminatingoil phase

't

In recombiningthe separatorgasand liquid to achieve


a representativereservoir fluid one must be very
cautiousabout how the GOR is being adjusted. The
approachtaken by the authors is that as long as the
downwardadjustmentin the GOR doesnot changethe
behaviour of the fluid from a dewpoint to a

bubblepointthen the affect of that adjustmenton the


fluid phasebehaviourwill not be significant. Indeed,
if therearesignificant changesin the liquid dropoutby
doing this GOR modification then equation of state
techniquescan be used to conductsensitivity teststo
seehow the different characterizationwill impact the
long-term reservoir forecasting including the
economicsof the system.
3

of the capillary pressure equation, which can be viewed


intuitively as equation 1, wherein in order to sweepthrough a
pore in a two phasescenario,the differential pressuremust be
equalto or exceedthe capillary pressure.
AP ~ p

The saturation pressure deviation from reservoir


pressureshould not usually be in excessof 1000psi.
If it is, then one may expectto seesomeliquid phase
contamination and therefore the resulting
recombinationmay bearno resemblanceto the actual
fluid which is in-situ in the reservoir. Many times
with a system which exhibits liquid phase
contamination,a changefrom dewpointto bubblepoint
is evoked by decreasing the GOR to meet the

/U",.

T19"oo

(1)

Extrapolations of this thinking would indicate,


therefore,that for gas condensatesystemswhich exhibit high
interfacialtensionswherethe pore throatsarevery smaIl,which
may correspondeither to low permeability rock or higher
permeabilityrocksbut with very largecoordinationnumber,the
successof flowing the liquid from the rock, once it has
condensed,
will be limited. In suchcasesvaporization(leangas
cycling) or injection of 1FTreducingagents(CO2)may be the
only option to enhancethe performance.On the other hand if
the equilibrium gas and liquid exhibit low interfacial tension,
then the liquid may flow freely from most of the pore throats
in the rock and very little retrograde condensaterelative
permeabilityreductionwill be observed.

One needsto be very vigilant and careful in systems


that are like this.
If there are solid instabilities then this is usually a
strong indicator that a liquid phasecontaminationis
occurring in-situ, particularly if the molecularweight
of the solid componentsis high and the solid phase
response is more sensitive to pressure than to
temperatureand is irreversible.

RETROGRADE CONDENSATE EFFECTS ON RELATIVE


PERMEABILITY

Conversely,wherethe pore throat diametersaremuch


larger,eventhough the interfacial tensionmay be high, it may
be easyto overcomethe capillary forceswhich are keepingthe
liquid in thosepore throats. Therefore,for larger pore throat
syStemswhich may correspondto higher permeability rocks
and/orsmaIlcoordinationnumbers,the interfacialtensioneffect
may not be very important.

Retrograde condensation results in a number of


problems. The most obvious and serious of these is lost
productive capacity due to accumulation of liquid in the
reservoir. This hastwo facets: the first is associatedwith not
being able to producethe higher value liquid componentsand
secondly,the increasedliquid saturationresultsin reducedgas
flow rates. Thesefactorswork in concertand the more serious
the liquid dropout, the greater the reduction in gas relative
permeability.

Unfortunately, other factors cloud the issue. One of


the factorswhich complicatesthis developmentis the mobility
effect. If one wereto analyzethe viscosity ratio betweenmost
equilibrium gasandcondensate
systems,the viscosityof the gas
would normally be at least 10 - 20 times lower than that of the
condensate
phase.Due to the inherentnatureof the lessviscous
phaseto flow more readily, it will tend to take the path of least
resistanceand will preferentially flow through the larger pore
throats. Gardnef1>has shown correlations wherein for
gas/Iiquid flows, exponential viscous finger growth is often
seen. Thereforein light of the 1FTcriterion (equation I), one
may automaticallyconcludethat the largerporethroat sizeswill
contribute to much easierproduction from a reservoir which
exhibits liquid condensateeffects. This is usually true.
However,eventhoughthe fluids may havethe capacityto flow

For a systemwhich is single phaseinitially, relative


permeabilityeffectsareabsent.Relativepermeabilityshouldbe
viewed as a dependentvariable determined by three other
parameters or influences. These general influences are
associated with:

I.
2.
3.

JDtmaci8/

Onecanobservefrom Equation I that asthe interfacial


tensiondecreases,
the capillary pressuredecreases.Conversely
as the radius of the pore throat which containsthe retrograde
liquid decreases,
the capillary pressureholding the liquid in the
pore increases. Therefore, to be able to produce retrograde
liquid from small porethroats,onemust haveeithera very high
differential pressuredriving force or low interfacial tension.

necessarycondition of P181 equalling reservoir pressure.

4.

interfacial tension effects


viscosity ratio
pore size distribution

By definition. 1FTeffectsare only involved whentwo


phasesarepresent. The interfacialtensionis importantbecause
4

throughmost of the rock wherethe pore throatsare larger,the


gas may preferentially "choose" only the largest pore throats
and may bypassthe rest. Thereforeeven though the gas may
havea low enough1FTto potentially sweepall pore throats,it
may, due to mobility effects, only contact those of larger
dimension. Thereforea compromisewill be reachedbetween
interfacial tension and mobility.
If the system is
mobility-dominated.then the only way to effectively reducethe
liquid saturationmay beto vaporizecomponentsfrom the liquid
into the flowing gas phase. Whetherthe free liquid will flow
or improved recovery will have to rely on extraction effects
dependson the compromisereachedbetweenIFf andmobility
in the presenceof the actual porousmedia.

In this caseone seesthe expectedeffect of a reduction in gas


relative permeability very severefor the lowest penneability
systemand leastseverefor the highestpenneability core. The
techniquesfor measuringretrogradecondensatethemselvesare
very difficult from a laboratory perspective since all of the
influenceswhich were mentionedearlier are at play.
Figure 6 showsan exampleof the influence of liquid
saturationon the relative permeability. In this test which was
perfonnedby the authorsfor a very extremecondition reservoir
fluid one seesthe drastic impact of small liquid saturationson
the gas productivity as well as the influence of the interfacial
tension. In this case,the interfacial tensionsof the fluids were
measuredusing the machinevision pendantdrop techniqueas
shown in Figure 6. This technique has been describedby
Rotenberg(').Using this technique,interfacialtensionsdown to
10-2dyne/cmcan be measuredquickly and consistently.

In order to assessthe optimal way to producea gas


condensate,one must be preparedto perform the testing in the
presenceof the porousmedia. If the actualreservoirrock is not
used, one may have the appropriate viscosity ratio and
interfacial tension, but the conclusions drawn may be
inappropriatein light of the fact that the compromisebetween
mobility and1FTis not the one which would be consistentwith
the reservoirrock.

With the interfacial tension at a level of 0.2 dyne/cm,


one seesa slightly lower critical condensatesaturationthan at
the 2 dyne/cmsystem. Moreover,the influence of the wr on
the endpoint saturationsand relative penneabilities is also as
intuitively expected. The influence of the critical condensate
saturationmay be very importantdependingupon the degreeof
retrogradecondensationeffect. For example,if the maximum
retrograde condensation effect is lower than the critical
condensatesaturation then the only way in which the gas
productivity can be remediedis by extracting the components.
Thomaset a1(IO)
defmedextraction effects for systemssuch as
this. For retrogradecondensatelevels higher than the critical
condensatesaturationone may be able to consistentlymobilize
and produce the condensatesaturation above the critical
saturation. This therefore points to the importanceof being
able to accurately define the liquid dropout as well as the
critical condensatesaturation.

It should be noted at this point that in light of the


couplednatureof interfacialtension,mobility andthe pore size
distribution of the porous media that laboratory testing must
adequatelyrepresenteachof thesethree factors. If any portion
of laboratory analysis does not adequatelyrepresentthese
parameters,then the credencewhich one can lend to the
conclusionsdrawn from a moreroutine laboratorystudymay be
minimal. That is, if viscosity ratio is usedas a referencefor
atmosphericrelative permeability testing, but the interfacial
tensionof the fluids is not matched.then onehasa priori biased
the conclusions in favour of the system being mobility
dominated.
One of the standard tests which is often applied to
detennine the response of a gas condensate system is the
constant volume depletion test wherein the liquid phase as a
function of pressure is measured. On this basis one can
detennine if the liquid is of high enough volumetric proportion
to cause a problem. Usually, those fluids which exhibit less
than I % retrograde condensation show very little tendency to
reduce gas ~. However, in some cases, even at low liquid
dropouts, the liquid tends to migrate into the production
wellbore and result in reductions in gas penneability. The
evaluation via CVD, however, introduces one to how serious the
problem may be and initiates the overall evaluation on a
laboratory scale.

The liquid dropout is fairly easy to do since it is a


routinephasebehaviourexperiment.However,to determinethe
critical condensatesaturation is much more difficult. The
critical condensatesaturation is defmed as the condensate
saturationbelow which the condensate
will not be ableto move
or to be movedandabovewhich the condensatewill be mobile.
Therefore,to measurethe critical condensatesaturationstrictly
one must deplete the dewpoint fluid in the representative
reservoirrock and then subsequentlyflow equilibrium vapour
throughthe specimen.Dependingupon the pore volume of the
system,the dead volumes of the experimentalapparatusand
inherentnatureof the liquid phasebeing droppedout, this can
be an extremelydifficult procedure.If the condensateis mobile
but is trappedin the accessories
of the experimentalapparatus,
one may have too high of critical condensatesaturation. One
may then possiblyget involved in a gas cycling schemewhen,
in fact, the critical condensatesaturationis lower andwould not

Following the perfonnance of a constant volume


depletion test one often will quantify the influence of the
retrogradecondensateon the gas relative penneability. Figure
5 providesan exampleof somedata obtainedby the authors.

necessitatea cycling operation. Because of the inherent


difficulties in measuringthis, somelaboratorieshavejudged the
residual condensatesaturationto be the same as the critical
condensatesaturation. This, however,is not the case.

This brings us to another point wherein the


measurement of retrograde condensate effects can be very
sensitive. This is mainly the case where the differential
pressure of the system is fairly high. For a large differential
pressure, the thermodynamic behaviour may be coupled in with
the fluid flow behaviour. Figure 9 shows a comparison between
an equilibrium assumption versus anon-equilibrium assumption.
For the equilibrium assumption following each saturation level
using the analog methane binary, the core was depressurized.
The gas was collected and the gas composition was used to
interpolate the gas phase saturation. The pressure effect,
however, resulted in a compression of that gas phase over what
would normally be present if there had not been a substantial
differential pressure. It should be noted that in this case the
differential pressure was as high as 300/0of the backpressure
setting. Therefore, by including a compression effect of the
vapour phase only, the non-equilibrium values were generated
in Figure 9. Compared to that, are the values calculated from
the equation of state based on the produced gas composition
upon depressurizing the core. One can see that although they
are different there is a reasonable comparison between the two.
Knowing this may be particularly important for some of the gas
condensate relative permeability testing that may have to be
done very close to the saturation pressure corresponding to very
low interfacial tensions. For example, if an 1FT required was
in the order of 0.1 dyne/cm which meant that the operating
pressure had to be within 50 psi of the saturation pressure, then
the differential pressure would result in a pressure higher than
the saturation pressure. Therefore, knowing the comparison and
the reasonableresponsebetween the equation of state calculated
values and the non-equilibrium assumption is an important
parameter to note for those performing the experimentation.

By a residual condensatesaturation,one definesthis


value by filling the specimencompletely with a bubblepoint
equilibriwn phaseproducedfrom the depletionexperiment.By
filling the core completelywith that phase,onehas filled all of
the pore throats and poreswith the liquid. Upon injecting the
gas, even though it is an equilibrium gas, the pore throats,
which are so small that the Iff betweenthe gasand the liquid
precludeentry, will still be filled with the condensate.In other
words, it will only be a judge of which poresthe equilibrium
gashasbeenableto sweep. For the pore sizedistribution given
in Figure 7 at a level of 1FT which correspondsto 20 microns
and the obviously artificial pore size distribution containedin
Figure 7 (for explanationpurposes),the residual condensate
saturationwill be associatedwith all of the poreslower than 20
microns. In such a case,the critical condensatesaturation,if
measuredas a residual condensatesaturation,could be very
high; the residualcondensatesaturationwill be an indicationof
the saturationof the condensate
remainingin the corethat could
not be swept.
If one compares that to the critical condensate
saturation,whenthe condensate
saturationbuilds up to a certain
value,thenthe fIrst sign of productionof that condensate
would
be calledthe critical condensatesaturation. This doesnot mean
that all of the condensatewill move from the core but only
representsthe saturationat which the condensatehas become
mobile. Necessarily,therefore,the critical condensate
saturation
is going to be governedby the gas condensateresponsein the
largestpore throatswhereasthe residualcondensatesaturation
is going to be governed by the gas liquid interaction in the
smallestporethroats. The critical condensate
saturationwill be
the most optimistic value andthe residualcondensate
saturation
will be the most pessimisticvalue. Otherwisestated,the critical
condensatesaturationwill indicate the lowest saturationthat
needsto be built up before the condensatein the largestpore
throatswill be mobilized. This, however.doesnot meanthat
the condensatein the lower pore throatswill be mobilized and,
in fact, will not be. Indeed,as the condensatephasebeginsto
migrate throughout the rock one may result in the net effect
associatedwith the residual condensatesaturationbeing more
representativethan the so-calledcritical condensatesaturation.
At a level of 1FT the smaller pore throats will have the
possibility of imbibing the liquid phaseand thereforeoncethe
smallerporethroatsare filled with the condensate
theremay be
no possibility other than through gas cycling or through 1FT
manipulationof accessingthosepore throatsandmitigating the
influence of retrograde condensation on the relative
permeabilities.

SIMULATION APPLICATIONS
The ability to producerepresentative
experimentaldata
is inherentlyimportant. Nevertheless,this importancewould be
diminishedsignificantly if there was no ability to couplethese
datainto making forecastestimatesfor gascondensate
reservoir
performance.The constantvolume depletiondataarenormally
readily importedinto a simulationmodel. The ability, however,
to couple the fluid phase behaviour with the fluid flow
characteristics is more challenging from a simulation
perspectiveand one must be very awareof someof the factors
which needto be presentin a simulator. A quick review of
theseparameters,as mentionedearlier, are:

2.

3.
6

Influence of interfacial tension on relative


penneabilities.
The inclusion of viscosity/IFT compromise in the
relative penneabilities.
An adequatemeansof representingthe liquid dropout.

4.

The influence of pore volume compressibility as a


function of pressure.
For fractured systems, the mechanism whereby
fracturesheal as the reservoir is depletedand the net
overburdenpressureincreases.

The authors have worked in the last year with a system


which required all of these parameters. For instance, upon
discovery and initial depletion. the behaviour was above the
dewpoint and corresponded mainly to a simple gas
decompression problem. Once the dewpoint pressure was
reached and liquid phase dropout commenced then this had an
influence, based upon the retrograde condensate effect. Not
only did that occur but simultaneously with the decrease in
pressure the 1FT began to increase. This increasing 1FT
resulted in more drastic relative permeability decreases. In a
reverse synergistic effect, the decreasing pressure and increasing
1FT also resulted in decreasing absolute permeability due to the
fractured nature of some of the systems which the authors have
modelled. This calls into play, therefore, an influence in the
absolute permeability of the model along with an ability to
interpolate between relati ve permeabiIities corresponding to high
and low 1FT regimes. The authors found it necessary to have
the simulator used (Computer Modelling Group compositional
simulator GEM) modified to include all of these effects and
every one of these effects was required in order to get an
adequate response of the field data.

Retrogradecondensateeffects on relative permeability


have been shown. These effects becomemore serious
with increasinginterfacial tension and decreasingrock
penneability. This effect is determinedby the interplayof
interfacial tension, pore size distribution, viscosity ratio
and, to somedegree,connatewater saturation.

4,

Detennination of retrograde condensation effects on


relative penneability is a challenging measurementto
make experimentally and some proceduralconcernsare
identified with possiblesolutions.

ACKNOWLEDGEMENTS
The authors would like to express appreciation to the
personnelat Hycal EnergyResearchLaboratoriesLtd. for their
time and effort in producing the experimentalresults for this
paper.

REFERENCES

Dake, L.P.: "Fundamentalsof Reservoir Engineering",


Elsevier Publishing,Oxford (1978).

Only in using a simulator with this much detail and


basedupon detailed experimental data can one hope to close the
gap between the initial characterization and the initial hopes
associated with discovery and the ultimate dollar value which
can be realized from such reservoirs.

2.

craft, B.C. and Hawkins, M.F.: "Applied Petroleum


Reservoir Engineering", Prentice-Hall Inc., Englewood
Cliffs, NJ, (1959) Chapter2.

3,

Gilchrist, R.E. and Adams, J.E.: "How to Best Utilize


PVT Reports", Petroleum Engineer International (July
1993).

Moses, P.L.:
"Engineering Applications of Phase
Behaviour of Crude Oil and CondensateSystems",JPT
(July 1986)pp. 715-723.

s,

Strong,J., Thomas,F.B., Bennion,D.B.: "ReservoirFluid


Samplingand RecombinationTechniquesFor Laboratory
Experiments", paperpresentedat the CIM 1993 Annual
TechnicalConferencein Calgary, May 9-12, 1993.

CONCLUSIONS
Characterizationof the gas condensateis often a very
challengingendeavour.Many times the separatorGOR's
will be too high dependingupon whetherthe well is being
producedat too high or too Iowa flow rate.
2.

Techniquesfor modifying GOR difficulties are routine


whereasif there is any contaminationof the produced
liquid phasetheseproblemsrequireresampling. A rule of
thumb for saturation pressuredeviation from reservoir
pressureis defined. Usually, if a GOR manipulation is
requiredthe saturationpressureis normally lessthan 1000
psi higher than the reservoirpressure. If there is greater
than 1000 psi deviation from reservoir pressurethen
normally this is associated with a liquid phase
contaminationproblem.

6.

7.

'f

McCain,W.D., and Alexander,R.A.: "SamplingGas


CondensateWells", Society of PetroleumEngineers
Reservoir
Engineering(Aug. 1992)pp. 358-362.
Gardner,J.W." Ypma, J.G.J.: "An Investigationof Phase
Behaviour- MacroscopicBypassingIntersectionin CO2
Flooding", SPE 10686, 1992 SPE DOE Symposiumon
EnhancedOil Recovery,Tulsa, OK., April 4-7, 1992.

9.

Rotenberg, Y., Boruvka, L., Neumann, A.W.:


"Determination of Surface Tension and Contact Angle
from the Shapes of Axisymmetric Fluid Interfaces",
Journal of Colloid and InterfaceScience,Vol. 93, No.1
(May 1983).
Bennion,D.B., Thomas,F.B.: "RecentImprovementsin
Experimental and Analytical Techniques for the
Determination of Relative Permeability Data from
UnsteadyStateFlow Experiments",presentedat the SPE
10thTechnicalConferenceand Exposition held in Port of
Spain,Trinidad, June26-28, 1991.

10. Thomas, F.B., Holowach, N., Zhou, X.L., Bennion. D.B.:


"Optim izing Production from Gas CondensateReservoirs"
CIM Paper 94-04, 1994.

I
I
Ni
Ii
,
~

~I
I~.
~

'.

tit

!~

\..

.
--

tl

III

I'
..7

],

~f

il
a

;
I

il~
I

\.
"
'"
~.G

..

1-

.1
s~

~I
t

-!J
II

I
I
~-.:_~~-~

=IQ~i
0000-

1111

irlil

I
.1
.

c!

.~
I~

I
~

il
~I
~
i
4
~

tlt~l~

I
I

I
I
I
I
~,,~

-IIIIIJII;.

"
.J

-I

I!

I
J

i
;1
~t
~I
~
I

You might also like