.

SPE
SPE 23442

Production and Pressure Decline Curves for Wet Gas Sands

With Closed Outer Boundaries
S. Mohaghegh and T, Ertekin, Pennsylvania State U,
SPE Members

Copyright 1S91, society of Petroleum Engineers, Inc.

This pepar wes prepared for preaantetlon et the SPE Eastern Regionel Maellng held In Lexington, Kentucky, October 22-25, 1991.

Thie paper was selected for presentation by an SPE Program Committ- following review of Information contained In an abstract submitted by the author(s). Contents of the paper,
as presented, have not been reviewed by the SNety of Petroleum Englnaara end are aub]ect to corraotlon by Ihe author(a). Tha material, as presented, does not rmcessarily reflect
eny position of the society of Petroleum Englneere, Ita officere, or membere, Pepera presented at SPE meatinge are subject 10publication review by Editorial Commillaes of Iho Society
of Petroleum Englnaara. Permieeionto copy is restricted !Gan abstract of not more Ihan 300 words. Illustrations may not tM copied. The abstract should contai,l conspicuous acknowladgmant
of where and by whom the papw la presented. Write Librarian, SPE, P.O. Sox SWWS, Richardson, TX 7S0834S36 U.5A. Telex, 730989 SPEDAL.

industry and for more than 40 years in hydrogeology. Applica-

ABSTRACT tion of this method to gas reservoirs has been investigated by
several scientists including Carte#, and Fetkovich4, In gas reser-
voirs as well as other petroleum reservoirs the focus of type
curve analysis is on single-phase flow. Application of the type
A family of pressure and production decline cuwcs am gen- curves in analyzing wet-gas sands is the subject of this paper.
erated for wet-gas sands with closed outer boundaries. Wet-gas Wet-gas sarrds are gas reservoirs in which simultaneous flow of
sands are characterized as gas reservoirs which produce substan- water and gas phases takes place. These reservoirs may produce
tial amounts of water together with ~as. Production of water substantial amounts of water with gas production. Modeling of
introduces complications when practicing engineers use decline the flow phenomena in such reservoirs requires the consideration
xwves designed for gas resemoirs in which gas is the only of two separate flow equations. Combining gas and water flow
Rowing phase. This usually translates to over estimation of the equations into a single compact form is the first step toward the
production performance of the reservoir, generation of the decline curves. This equation can then be
transformed into a dimensionless form and necess~
dimensionless groups need to be identified, The identified
Irr this paper a series of pressure and production decline dimensionless groups then can be used in the construction of the
curves which accounts for the water production in wet-gas sands type curves,
is presented. These decline cumes provide a simple way of
extracting valuable information from avail~bleda!a, using graphi-
cal methods and simple calculations, The prcposed decline The type curves presented in this work are getieratcd for a
wves arc generated for a radiid system with closed outer boun- radiul system with closed outer boundary itmJ with a fully
dary with one centrally loeuted well which fully penetrates the penetrating well at the cxxrter, Constant pressure and constant
formation. The application of he proposed decline curves is tlow rate cases at the inner boundary arc invcstigitted, [n the
illustrated through a series of examples, case of constant pressure inner boundwy production type curves
are generated for different mtios of wellbore to initial formation
pressure, and different reservoir extents. Within a certuin
In the case of constant pressure inner boundary, production wellbore to inithd formation pressure mtio, type curves are gen-
decline curves are generiited for different ratios of sandface pres- erated for different initial pressures. When constant flow rate
sure and initial formation pressure, and in the case of constant inner boundary is specified, pressure type curves are labeled for
flow rate inner boundary, pressure decline curves are generated both different initial fommtiort pressures and different rescwoir
for different initial formation press;res, The proposed decline extents.
cu~es covers a broad spectrum of reservoir extents and have
proven to be unique for a wide range of fluid and formation
parameters, 7trispaper should be considered iis ii conlinuittion of a pre-
vio~s paper presented by (lie iiuthors$, In this puper a change in
th( formulation of the problem has been introduced through
INTRODUCTION wh):h iI previous limitation which wtts restricting the usc of the
type curves to low pressure reservoirs is successfully removed,

Analyzing petroleum reservoir; with the aid of type curves FORMULATION

is a practical method available to petroleum engineers, Type
curve analysis has been used for more than 20 years in the oil
In a previous paped the methodology used in combining
References and figures at end of paper. gas and wa!er transport equations inks a single expression was

261
Iutlined in detail. A summary of this methodology which also quation 9 can be written as:
Ictails the implemented changes to the previous approach are
Irescntedin this section.
(~+k)V2p+k1Vh(~)vw
s
Starting with the gas and water equations which represent
he two-phase flow in wet-gas sands, one can write:
&Vln(#)vp=Oq~f (11)
w

(1)
[5)=+ [%1 n the linearization of the atmve expression it is assumed that:

k
($s. (12)
v,
[1
#.vr
w
.$
-1IIw
(2) =ap
B@

L
=bp (13)
B.
(3)
vhere a and b are constants. Therefore, Equation 11 can be
mitten as:
(4)

Multiplying Equation 1 by B, and Equation 2 by B. and

Issuming that porosity is not a function of pressure, and then jy recalling that:
~ddingthe two equations one obtains:

Vln(#)= Vln(ap)=Vlnp=~Vp (15)

W?VI+B4*)V1=
I

If one defines the total system mobility as:

(5; h=k,+k (16

Zquation 14 can be written as:

After some mathematical manipulations, Equatio:~5 can be writ-
:en as:
V2P+:(VP)2=+R (17:

:v[(t)vl+B.v[ (=)vl=o~~ 6)Equation 17 can be rewritten in a more compact form as:

where total system compressibility, c,, is defined as:
(18:

Finidly, when Equation 18 is written in one-dimensional radia

mordinates gives:
Expanding the left hilnd side of Equation 6, one obtains:
(19

Equation 19 is similar to Equation 18 of Reference 5, Th(

which con be also expressed as: ~iffwence is that in Equation 19 pressure squared approach ha!
xen used instead of the pressure itpprowh,

(AB+L)V2P+?+V $%+
I I
Trimsforming Equation 19 into dimc,lsionless form yields:

ilA~ aApD
.la
1
r~ a rD D a rr2 = -
a ID
(2(

Using !he following identity, where

Vlnu=$Vu (lo)

262
 .

Production Type Curves

Ah= R2-P2 (22)
(pi*-PM*)%
Using Ihe formation and fluid data shown in Table 1, gas
and water production decline curves were constructed. Figures 1
2.637Xl@~ L
b= (23) and 2 show gas production type curves for ~ = 0,1 and initial
~ q rW2
formation pressure of 2000 psi and 6000 psi, respectively. The
nd corresponding water production type curves are shown in Figures
, 3 and 4. The proposed type curves were extensively tested
against a wide range of fluid and formation properties. Figures 5
qwBB
1
~. through 8 show the agreements obtained in some of these
(24) verification tests, Several other verification tests were conducted
[ 2rrklqh(p~-pti2) W
and in every case high quality matches were achieved, Gas and
water production type curves were generated and verified for
To construct the type curves for gas and water phases it is
eeessary io identify dimensionless gas production and water other ratios of ~* and different initial formation pressures and
reduction rates explicitly. These dimensionless groups can be are available from the authors upon request. The relative per-
ientified as follows: meability characteristics that was used in the construction of the
proposed type curves are shown in set A of Figure 9. The
~w Z T
14.24X1(+ uniqueness of the proposed type curves with respect to the rela-
(25) tive permeability characteristics will be discus~ later.
= Ach(p; -pd2) Following example demonstrates use of the gas and water pro-
duction type curves,
282.S3qWXBW P
(26) .Example #1
=~h(p/-pti2)
A simulation run was conducted to generate gas and water
I order to enerate the type cuwes which.characterize the pres- production data. The formation and fluid data used for this run
Jre drop, f imensionless presure drop group must be identified.
?ris can be achieved by substituting Equation 25 into Equation are shown in Table 2, Gas production type curve for ~ = 0,1
2 which in turn yields: and initial formation pressure of 4tXKlpsi were used to perform
typr: curve matching from gas flow rate versus time data. In Fig-
~h(~2-p2) ure 10 a match point with coordinates
A~= (27)
14.24X1($
~ Z T
[~=0.05 ,~= 2000(S)
],@uN, md [qt. lXIO,t=0.9]hUn@
The constants irr Equations 23, 25, 26 and 27 imply the use
f the field units for all of the entries in these equations. is identified, Using the information provided in Table 2 and
relative permeability characteristics of set A in Figure 9 follow-
ing values were obtained:
DISCUSSION OF RESULTS
~ = 0,022 Cp z, = 0.86
kr~= 0,1024 kW=0.12%
A one-dimensional, two-phase, radial numerical simulatof
was used to generate the rate-time and pressure-time data neces-
wry for the construction of the type curves. In the case, of pro- Substituting above values and information from Table 2 in Equa-
duction type cuwes the rate-time data generated by the smurlator tion 25 gives:
was used to calculate the dimensionless production rate and the
dimensionless time groups, These calculated values of the 0.05= ~ 424)( lx07 ) ( 530) (@@=
k ( 0.1024) ( 85.6)( 38342-383,42)
dimensionless groups were then plotted on Iog-lcg scale. It was
found that for different wellbore to initial formation pressure
ratios it was necessary to generate separate production type Jr
curves, Thus, production type curves were generated for a broad k = 22.39 md
spectrum of ~ ratios, Itwas also observed that within a given
wellbore to initial formation pressure ratio type curves are unique Substituting above calculated permeability in Equation 23, poros-
for different initial formation pressures, ity of the reservoir is calculated as 15,7%.

In the case of pressure decline curves, pressure-time data
generated by the simulator was used to caI.ulate the dimension-
less pressure drop and the dimensionless time groups. These
dimensionless values were then plotted on log-log scale. Pres-
sure decline type curves are generated for different initial forma-
tion pressures,
To demonstrate the compatibility of the corresponding water
production type curve to that of the gas, a second type curve
matching for this example was also performed this time using the
water production type curve with the water flow rate versus time
data, As it can be seen in Figure 11 a match point with coordi-
nates

In both cases type curves are constructed for different ~=o,2, tD=2m ~YHWC
and q. = 700, t = 0,9 MtPld
drainage areas. In the constmction of type curves all of the pres- [ 1. [ 1
sure and saturation dependent variables are evaluated at initial
conditions,
263
s used. It should be noted that in both cases of gas and water
mxluction type curves the value of dimensionless time and real
ime in match points arc the same, This is essentially true since
Ksth type cumes share the same abscissa. This observation
hould be used as a guide line in obtaining a match point.
Following the same procedure as discused before and using
Equation 26 the rmeability and porosity of the formation are
xdculated to be r 3.48 md and 16.4%, respectively. The actual
Ares used in the simulation run were 23.5 md for the permea-
bility and 16,2% for porosity.
As it was mentioned earlier, existing type curves for gas
eservoti assume flow of single phase. This assumption creates
;ignificsnt limitations on the use of these type curves when ini-
,ial water saturation is higher than the formations critical water
mturation. Use of single-phase type curves, when analyzing a
eservoir which is producing under two-phase flow conditions,
will result in the over estimation of the reservoirs performance
when all reservoirs parameters are known, On the other hand,
lse of the single-phase type curves will lead to under estimation
]f reservoir properties if type curve matching is performed. Fol-
owing example focuses on demonstrating these shortcomings.
Example #2
Using the same reservoir as Example 1, a single-phase type
:urve3 was used to find the permeability and porosity of the for-
nation. Type curves in Reference 3 are generated for different
dues of L, which is defined as follows:
(28)
For this example the value of L is calculated to be 0.5S. Figure
12 shows the type curves matching performed using the single-
>hase type curve generated for k = 0.05. l%e match point has
he following coordinates:
[ 1 1
qDR= 0,033,IDR= 3WKJT@Jmc and [ q = 1X1O,t= O.01~mnd
Using Reference 3, permeability of the formation is calculated to
x 3.47 md, Using the calculated permeability value, one can
:alculate the formation porosity as 1.5%. Comparison of these
ralues to actual input values of 23.5 md and 16,2%, highlights
:he inadequacy of using single-phase type curves in analyzing the
iata generated flom a gas reservoir under two-phase flow condi-
tions, If performance prediction of the reservoir is the goal then
Jse of single-phase type curves will result in overestimation of
;he reservoir capabilities, For the above match point, if the input
values for permeability and porosity were used the single-phase
!ype curves would predict a gas flow rate of 380 MMSCFD
instead of simulated value of 67.7 MMSCFD, at t=O,Olday.
Pressure Type Curves
To construct the pressure type curves, formation and fluid
propeflies shown in Table 3 were used. Again pressure versus
time data were generated using the numerical modelf. Dimen-
sionless ressure drop and dimensionless time groups were calcu-
lated anJ Iottul on log-log scale. Relative permeability charac-
teristics of set A in Figure 9 was used in the construction of the
proposed pressure drop type curves. Figurw 13 and 1,4 show
ths dimensionless pressure drop curves for initial fonnatlon pres-
sures of 20(KIpsi and 6000 psi, respectively. In Figures 15 and
16 results of some of the verification tests are displayed. Similar
type curves have been constructed for other values of initial pres-
sure and are available from the authors upon rquest. As it is
shown in Reference 5 for production decline curves and demon-
strated here for pressure type cumes, the pro~ed type cumes
have been found to exhibit unique characteristics for different
sets of dative permeability curves. It should be noted that for a
given real time wet-gas sands which share the same properties
with the exception of relative permeability characteristics, will be
displayed at different locations on the type curve. These shifts
caused by the relative permeability characteristics arc shown in
Figure 17.
Application of the proposed pressure diop type curve is
demonstrated in the following example,
Example #3
Using the data shown in Table 4, a simulation run was con-
ducted and a pressure drop versus time data was generated,
Using the pressure drop type curve constructed for initial forma-
tion pressure of 4000 psi, a type curve matching was performed.
Figure 18 displays the match point as:
Using infornwtion in Table 4 and set A of Figure 9 following
values are obtained:
Be= 0,0235 Cp Z+=0.865
q= 0.3951 & = 0.0123
Using Equation 27 permeability of formation is calculated to be
4.56 md. Using the calculated permeability in Equation 23 one
can calculate the porosity as 9.07%. Actual input perrneabdity
and porosity values were 4.41 md and 9.3%, respectively and
agree very well with the values derived from type curve match-
ing.
During the construction of the presure type curves for
different initial formation pressures, it was observed that all of
these type cutwes were showing qualitatively the same behavior
and they could be collapsed into a single pressure type curve. It
was found that this can be done with a proper shifting in the
dimensionless time group. Pressure type curves for initial forma-
tion pressure of 2000 psi was chosen to provide the base for this
purpose. Figure 19 shows the amount of shift in dimensionless
time needed for any initial formation pressure. These shifts can
also be represented using the two following simple relationships
for two different pressure ranges:
Imrt =3.8574X10-Spi+ 0.882852 for R c 2350 psig (29)
and
1~~, = 4,3574X
10A pi-0.00507 for p,> 2350 psig (30)
Following example will demonstrate the use of the shifting
functions described in Equations 29 and 30,
Example #4
In order to demonstrate the use of pressure ty~ curves
developed for initial formation pressu~ of 2000 psi, in analyzing
the data generated from a reservoir wuh dtfferent initial pressure,
a simulation run was designed with initial pressure of 5183 psi.
Formation and fluid properties for this run arc shown in Table 5.
 .
SPE 23442 S. MOHAGHEGH ANDT.ERTEKIN 5

As it can be seen irr Figure 20 following match point from

pi=20CKItypc CUIVC is obtained:
NOMENCLATURE

B= formation volume factor, $#-

Using information provided in Table 5 and set A of Figure c= comprcssibility, psi-l
9 following values are derived: k= permeability, md
Pv= 0,0274 Cp q = 0.980 h= reservoir thickness, ft
p= pressure, psi
km=0.3951 ~ = 0.0123 q= flow rate, MMSCF/day
radius, ft
since the shift only takes place in abscissa, no adjustment is :: saturation, fraction
needed for dimensionless pressure drop. Using Equation 27 per- t= time, day
meability is calculated to be 9,1 md. Using the relationship T= temperature, R,
given in Equation 30, the value of dimensionless time is adjusted z . gas compressibility factor, fraction
as follows:
Greek
rmm= ( 4.3574x10A) ( 5183.0) -0.00507 = 2,2534 k = mobility, md/cp
~. viscosity, cp
$ = porosity, fraction

Subscripts
Using the new value of ~ and the calculated permeability, poros-
ity is calculated as 11.65% from Equation 23. Actual input D= dixmc;onless properties
e=
values for_permeability and porosity were 9.45 rrd and 12.85%, g=
respectively. gas properties
Sc = standard condition
1= total properties

SUMMARY w= water properties

Wf= Wellboreproperties

Major achievements of this study can be summarized as fol-

lows:
). Gas and water transpost equations were combined
into a single expression in the form of a pressure
squared expression. This expression was
transformed into a dimensionless form and necessary
dimensionless groups were identified.
2, Using the identified dimensionless production rate
and dimensionless time groups, water and gas pro-
duction type curves were construct for different
wellbore to initial formation pressure ratios for a
given initial formation pressure.
3. Using the dimensionless pressure drop and dimen-
sionless time groups identified in this study, pressure
type curves were constructed for different values of
imtial formation pressure.
4, Extensive tests were conducted to ensure the unique-
ness of the proposed type curves within a wide range
of fluid and formation properties.
It is envisa cd that the proposed type curves could serve as
t practicing engineers in performance prediction
a practical toolor
as well as formation characterization of the wet.gas sands.
These type curves will be useful in avoiding the overestimation 6.
of the reservoir capabilities which could be caused by the usc of
single-phase decline models when simultaneous production of
gas and wateris encountered.
,.
REFERENCES
1, Gnngarten, A, C. :Type Curve Analysis: What It Can and
Cannot Do, JPT, (Jan, 1987), 11-13.
2. Carter, R. D. :Type Curves for Finite Radial and Linear
Gas Flow Systems: Constant-Terminal-pressure Case,
SPEJ, (Oct. 1985),7 19-728,
3, Carter, A. D, :Characteristic Behavior of Finite Radial and
Linear Gas Flow Systems - Constant Terminal Pressure
Case, SPE/DOE Low Permeability Symposium, (May 27-
29, 1981), Denver, Colorado.
4, Fetkovich, M. J. :Decline Curve Analysis Using Type
Curves, 48th Annual Fall Meeting of SPE, (Sept. 30- Oct.
3, 1973), Las Vegas, Nevada,
5. Mohaghegh, S., Ertckin, T, :Production Decline Curves for
Low Pressure Gas Rescsvoirs Undergoing Simultaneous
Water Productions, SPE 21269, SPE Eastern Regional
Meeting, (Oct. 31- Nov. 2, 1990), Columbus, Ohio.
Kin% G. R. :Numerical Simulation of the Simultaneous
Flo@of Methane and Water Through Dual Porosity coal
Seams: Ph.D. Dissertation, The Pennsylvania State Univer-
sity (1985).

26s
 Table 3. Formmbo snd nuld data wed
in tbe KwBtsuetloo of tbe PsW8MN
type curves fw wet*ae mods.

Speci6ed Gee RoW Rue fhfSCFD) 10.0

Well Redius (ft) 0.5

Tabk L Formatim and tluld dsta used Pwmesbiiily (red) 1.0

10 the ametructbsr of gaehmter
productbn type curvee for wet.gae esnds. Pcloaity (%) 6

Fmmaticm ThiCkOtsS (ft) 20

Snilid Psessurc (psi) 2000
Ges Grtvity (eir=l ,0) 0.6
Wellbore Presure (psi) 200
Reservoir Temperature @R) 530
Well Redius (ft) 0.5
Waler Viscosity (cp) 1.0
PermeabUi~ (red) 1.0

Poroeiry (%) 6

Formedon Tbkhss (ft) 20

e3seGfaviLy (eiA.0) 0.6

Rewrvois Tcmpaeture @) 530 Tabk 4, Formation id 8Ukt d8ta used

h Rxwsspk #3,
Wuer Viscoeiry (q) 1.0

Speci6wf Gse Pbw Rate (MSCFD) 14.39

Well Rutiw (ft) 0.5

Fawwdon ~CkSSCSS (ti)

I 17.4

(3s5 Gravity (sir=l.0)

I 0.6

Reurvoir Tcnspmrure ~)

Water Viscosiry (cp)

I 530

1.0
I
WeUtme PressureQA) 383.4

WeU Redius (ft) C.5

Mid Ge Semetion (%)

I
30
I
Famaion TM-s (fr)
I 85.6
I Table $. Formation met tluid data wed
b Gravity (sir=l .0)
I 0.6
I h Ehmpk $4,

Ruervoir Tempcrerurc ~R)

I 530
I
~
Speci5ed Gas FbW R@e (MSCPD) 21,7S

WeU Ra&ue (ft)

Forrneth fhic&I-S (h)

I 0.5

39,45

au Grevity (*1 .0) 0.6

Wster Viscoeiry
(Cp)
I 1.0

206
 Figure
1. Gas production type
curves for ~~ = 0.1 and p,
=ZOOI
ptit.

Figure 2. Gas productiontype curvesfw ~ = 0.1 and n = ~ PJf#.

i!

,-o
$
,:

Dimt?nsionless Time

Figurs 3. Water prmiuction t~ curvss fw ~. O.I and p, .2013 pdl.

(n
Q
m

.
 SPE 23442 :
.

Set A Set B
Figsuc7. Afpfdim Of waurpmduaiontypecurves10a rcscwoirwishfollowing Chamcteristix

5-=S0%.
P.=l@Opsi.w=lfXI pSi. k=13.7 d +=.6% h=352 fL

*,
,,
87
1 *
;0s
,.
.3
PI
.,
.,
u *
.

Set C Set D
~ Figure
9. Ftelativc
pcnncahili!y
chamcwristics used in analyzing the Pcrfomance of k proposed
s
i5 type Curws

D-MuAollrcssmm?

Fipe8. A@iasim ofwabcrpmduaiwsYFcavvcs wamsahr Wm fLltlLwingChaIxmisti:s

Sm=SO%,P,=@M psi. x=fKKt F&. k-40.O md. 0=13.5%.h=73.5 fL

 5442

Figure 10, Type curve matching using gas production type curve for Example #l,

Figure 11, Type curve matching using water production type curve for Example #l,

lga ,

q
107 D= .-:

w
0,1

0,01
%R

106 1111 I
P- mmmm m I 0.001

Fiwme 12, Twe

. . curve matching using Carters) type cuwe for Example #2.

270
 Dimendohess Time

Figure 13. Rcssure IYIK cutves rOfp,EZo@I

PW.

F@R 14, Prcwre typs curves fnf p, = MM pif.

Figure IS, Appllc&IIon of prewwe Iype cwvcs to a rcwvok with following chsractctistics:

S,, .S0%, A=1783 Pd, qxE7MXl SC.PD, kmo, I ml, 4u8%, hm49 fl,
 Dimeaa{oheas I{me

Figure 16, Application of prcmIre IypO cosvcs to a resrmoir with following chuactcnstics

S,, -60%, PI-5872@, e,m=29300 SCFD, k=23.1 rm!, 0=1 1,3%, h=44 fl.

9 Set D
i

Figure 17, Sensitivity of the pmposcd IYIM curves 10 the relative pcrmcahllity curves given In Figure 9,

Flgusw 18. Type curve rrulchlng ucln~ prmwe Iyp+ curve for Example #3,

2?2
,,*
SPE 23442

I 1
I

1 ! 1
2000 4000
Pressure (psi)

Figun 19. Shift in tfJwhenpressuretype curve for pi =20C@

Psis is USC4! as the base

3
l!
I
Jxmmdon&41 mu

Figure20, Type curvematchingusingpressuretypecurvefor Example#4,

279