8.1 Physical Properties of The Gas: 8.1.1 Pressure and Volume
8.1 Physical Properties of The Gas: 8.1.1 Pressure and Volume
8.1 Physical Properties of The Gas: 8.1.1 Pressure and Volume
8. Gas
There is no qualitative difference between the oil and gas flow. However, since the
gas is lighter and less viscous than the oil, the effects are negligible for oil prove
important for gas flow, and vice versa.
8.1 Physical properties of the gas
8.1.1 Pressure and volume
The relationship between pressure and volume can be expressed by the general
equation of state
nzRT pV
(8-1)
ased on !an der "aals equation or other extended state equations, one can in
principle express the # factor as a function of reduced pressure and temperature$
pressure % critical pressure and temperature % critical temperature. &or gas mixtures, it
is common to estimate the so'called pseudo critical pressure and temperature, based
on averaging the critical pressure and temperature of the gas components, or simply
from the specific gravity of the composition.
) , (
pR pR
T p z z
(8-2)
p
pR
: pseudo reduced pressure
T
pR
: pseudo reduced temperature
&igure 8.1 illustrates (8')* by the +tanding',at# correlation.
&igure 8.) gives the correlation between pseudo critical pressure and temperature
and specific gravity
y formation factor, we understand the ratio between the actual volume and the
volume of the same quantity of gas at atmospheric standard conditions. &rom
equation (8.1* follows
o o
o o
o g
z
z
T
T
p
p
V
V
B
(8-3)
-ompressibility means relative volume change with pressure change. -ompressibility
at constant temperature can be estimated from the general equation of state. .f the #
factor can be considered constant, compressibility corresponds to the inverse of
pressure
p p
z
dp
d
z
p
dp
dV
V
c
T T
g
1 1
,
_
(8-4)
The result implies that gas compressibility decreases rapidly when the pressure
increases. "e have all experienced that bicycle tires become harder when we pump
them up.
Prod.karakterist-
09.01.2014
skr/221/HA
2
Figur2 8.1 z-factor diagram, by Yarborough & Hall /19!/ "#t"$d"d "%uatio$ of stat"
8.1.2 Density
/ensity equals the ratio between mass and volume. 0as density depends on
molecular weight$ 1, pressure and temperature. &rom the state equation (8.1* we
can express the relationship
zRT
pM
V
nM
V
m
(8-)
.t is worth noting that even if the gas density is less than the oil density, the gas
density, at reservoir conditions it can still be large compared what we associate with
gas. (2s a rule of thumb, the gas density in the reservoir approximately equals the
reservoir pressure in bar$ thus at pressure 133 bar, the density will be of the order of
133 4g%m5*
Prod.karakterist-
09.01.2014
skr/221/HA
3
Figur" 8.2 &s"udo critical 'r"ssur" a$d t"m'"ratur"
8.1.5 !iscosity
The viscosity is much less gas than for oil, so that the viscous resistance of flow is
correspondingly less. The viscosity of natural gas can reasonably accurately
estimated by the correlations of -arr',obayashi'urrows % 1678 %, reproduced below.
&igure 8.5 gives the viscosity at atmospheric conditions, as a function of the average
molecular weight (specific gravity 9 )8.6:* &igure 8.8 gives viscosity increase at
higher temperatures and pressures.
Prod.karakterist-
09.01.2014
skr/221/HA
4
&igure 8.5 !iscosity for natural gas at atmospheric pressure
&igure 8.8 !iscosity change as a function of pressure
8.1.8 Temperature changes with pressure
The gas will release, or absorb, heat when the volume is changed. .n the reservoir,
the heat released is absorbed by the roc4, usually without much temperature change.
so that the reservoir temperature may in practice be considered constant ..
"hen gas flows through a no##le, there is little time to transfer heat and the heat
Prod.karakterist-
09.01.2014
skr/221/HA
,
_
(8-")
k
i
i
p
p
1
,
_
(8-8)
k : adiabatic exponent, for ideal gases: k = c
p
/c
v
c
p
: specific heat capacity at constant pressure
c
v
: specific heat capacity at constant volume
i
: initial density
T
i
: initial temperature
8.2 ($flo)
8.2.1 ($flo) '"rforma$c"
The &orcheimer equation (<'5* combines viscous friction and =turbulence>. +olved
for stationary, radial flow of gas, gives the pressure profile
( )
2
2 2
2 2
1 1
2
#$
g
e
o
o o
g
g
e
g
o
o
e
q
r r h T
z T p
q
r
r
kh T
z T p
r p p
,
_
(8-10)
%it& r
e
>>r
w
,
2 2 2
g g k w e
q A q A p p
+
(8-11)
%&ere
kh
r r
T
z T p
A
w e g
o
o
k
) / #$(
w
o
o o
g
r h T
z T p
A
2 2
2
'
+ (8-12)
Here g B is the formation factor (8'5* at, the average flow conditions. y comparing
(8'1)*, with the flow characteristics of liquid (<'8*, we see that the relations are
analogous. The difference is that the liquid volume changes so little with pressure
that we usually can consider formation factor constant. &or gas, volume changes
may be significant .
,or - (urdere .ruk.ar&ete$ a( (8-12) som ti#$/rmi$' ti# (8-10), ka$ (i samme$#ik$e a((ik
dersom (i .ruker tr)kk (ed )tre 're$se, i stedet *or et i ut'a$'spu$ktet uk0e$t
'0e$$oms$ittstr)kk:. 1ette 'ir a$s#a' *or *ei#mar'i$e$, utr)kt som *raks0o$
To use (8'1)* as approximation to (8'13*, we may use pressure at the outer
boundary, instead of an initially un4nown pressure$
( ) ( )( ) ( )
w R R w R w r
2
w
2
R
p p p 2 p p p p p p + This implies the error margin
Prod.karakterist-
09.01.2014
skr/221/HA
"
,
_
+
R
w
R
w R
p
p
"
2
"
p 2
p p
"
&or example$ if the reservoir pressure is )33 bar and the well pressure 163, the error
neglecting gas expansion becomes$ @ 3.7. (1'163%)33* @ 3.3)7, ie ).7A
8.2.2 +ac,'r"ssur" "%uatio$
2n alternative way of expressing the inflow of a gas well is by the empirical
;bac4pressure equation; introduced by the B.+. ureau of 1ines in 16)6
n
w R g
p p # q ) (
2 2
(8-
13)
The bac4pressure equation (8'15* has two parameters$ n and -. These must be
determined from multi'rate production test.
The flow characteristics based on &orchheimer equation also involves two
parameters that can be determined from the production test$ : A
k
o' A
. . ut these
are also connected to the reservoir and fluid parameters. The bac4pressure equation
is purely empirical no such correlations.
8.- Flo) i$ 'roductio$ tubi$g
8.-.1 Flo) "%uatio$
Pressure drop %it& statio$ar) *#o% i$ pipes is 2ua$ti*ied as .e*ore
0
2
1
2
+ + + d$ v
d
% d$ g vdv dp
$
(8-14)
%&ere
A
m
v
o'
d
A
m
dv
2
1
2t stationary conditions, the mass flow. m
, will be constant all along the pipe, while
the speed varies with pressure. .t is therefore convenient to express (8'18* in mass
terms
0
8 1!
2
2
2 4 2
2
+ + d$
d
m %
d$ g d
d
m
dp
$
(8-1)
8.-.2 &r"ssur" i$ static gas colum$
.f we have a closed gas well, #ero flowrate, the second and fourth terms of (8'17*
become #ero. .f we express the gas density at the general state equation, we get
Prod.karakterist-
09.01.2014
skr/221/HA
8
0
+ d$
zRT
g
p
dp
$
(8-1!)
y integrating (8'1<* along the pipe, we get the relationship between static tubing
head pressure and bottom pressure
T R z
& g
w
'h
$
e
p
p
(8-1")
& : #e$'t& a#o$' tu.i$'
M : 'as mo# %ei'&t : g
M 9" . 28
g
: spesi*ic 'ra(it)
Comparison with pressure in liquid filled well
+eries development of exponential gives$
((( ) * $ ) 2 $ $ " e
* 2 $
+ + + +
Bsing the ) first
terms (8'1:* can be re'written
& g p
T R z
& g
o p & g
T R z
p
p p
$ w w
$ 2
w $
w
w 'h
,
_
The result corresponds the liquid gradient, with 3o
2
( )3 indicating the order'of'
magnitude of the deviation. Thus the approximatio will be acceptable when the
exponent$ " T R z & g
$
<<
8.3.3 Pressure in producing gas well
Bnder normal flow conditions in pipes of constant diameter, we can neglect
acceleration. y expressing the gas density at the general state equation, the flow
equation becomes
+ d$
p
" zRT
d
m % ,
p
zRT
g
dp
- 2
2
$
1
]
1
,
_
+
,
_
+
(8-18)
Bsually, parameter groups in square brac4ets ;C D; vary so little along the production
pipe that we can consider them as constants. .ntegration of (8'18* then gives
2
2 2
2
2
2
2
1
8
m e
T R z
d g
%
p e p
T R z
& g
$
w
T R z
& g
'h
$ $
1
1
]
1
,
_
,
_
,
_
1
]
1
(8-19)
The relationship above express a simple relationship between pressure and flow
rate$
2 2
,
_
,
_
w
g
% g
w
'h
p
q
A A
p
p
(8-20)
Prod.karakterist-
09.01.2014
skr/221/HA
9
2
1
]
1
T R z
& Mg
e A
$
g
( ) ( )
g
$
g
o
g
$
%
A
T
T p z
d g
%
A
M
T R z
d g
%
A
,
_
,
_
1
8
1
8
2
0
0
2
2
2
2
2long a production pipe, the parameter groups 2g and 2f can usually considered
constants. This means that we can estimate these from production data, by plotting$
( )
2
/
w 'h
p p
(ersus$
( )
2
/
w g
p q
The plot should define a straight line. The intersection with the y'axis provides
parameter$ 2g, while the slope provides parameters and 2f.
&igure 8'< compares shut'in pressure along the production pipe, calculated by (8'
1:*, and production pressure by (8')3*. Production is set relatively high$ )3 +m5 % s,
which here gives the flow speed between 1) and 18 m % s. +ince the speed increases
towards the top, the pressure drops profile curves slightly.
Figur" 8-. &r"ssur" 'rofil"s i$ gas )"ll
8.-.! Horiso$tal" r/yr
Prod.karakterist-
09.01.2014
skr/221/HA
10
,or et &oriso$ta#t r4)r, *a##er det statiske #eddet .ort: g
$
= +. 5ed - #a g
$
'- mot $u## , ka$ (i
(ise (med #6Hopita#6s re'e#) at *riks0o$sparametere$ i (8-20) '-r as)mptotisk mot
2
0
4
,
_
o
o
g
%
T
p
Rd
& z T %M
A
$
7ik$i$' (8-20) *ore$k#es da ti# de$ k#assiske 8e)mout&#ik$i$'e$ *or tr)kktapet i&oriso$ta#e
r4)r
( )
. 0
2
2
0
0
2
4
,
_
,
_
,
_
& T M z %
Rd p p
p
T
q
i .
g
(8-21)
8.-.* 0yst"ma$alys"
5ed kom.i$as0o$e$ a( (8-11) o' (8-20), ka$ (i uttr)kke samme$&e$'e$ me##om tr)kk o'
str4m$i$'
( )
2 2 2
g % g k g e g 'h
q A A q A A p A p +
(8-22)
1ette k(a$ti*iserer tr)kktap i 'assproduks0o$ss)stemet, p- 'ru$$ a( str4m$i$'.
9)stemparametre$e: A
g
/ A
k
/ A
/ A
%
, er k$)ttet ti# parametere o' .es#ut$i$'s(aria.#er '0e$$om
re#as0o$e$e ut#edet o(e$*or.
8.! Flo) through 1al1"s a$d r"strictio$s
"hen gas flows through the valves and restrictions, will usually acceleration forces
dominate. The flow equation from (8'18*, then becomes
0 + vdv dp
(8-24)
y solving this, we can predict the flow capacity and pressure drop
8.!.1 ($com'r"ssibl" a''ro#imatio$
"hen the pressure drop across a valve or restriction is less than 7A (p
c
% p
i
E 3.67*,
corresponding to 1ach$ 1a F3.)7, we can neglect compressibility and calculate as
for incompressible fluid
8.!.2 2om'r"ssibl", subcritical flo)
"hen the pressure drop is greater than 7A, we should ta4e into account that the gas
will expand. +ince the speed is high and distance through the valve short, there is
little time to heat conduction. "e can therefore assume adiabatic flow, so that
equation (8')8* becomes
Prod.karakterist-
09.01.2014
skr/221/HA
11
0
1
,
_
+ vdv
p
p
dp
k
i
i
(8-2)
:) i$te'rati$' *rom upstream pressure: p
i
a$d (e#ocit):
v
i
;0, to out#et pressure a$d (e#ocit):
p
c
/v
c
, %e 'et
,
_
,
_
k
k
i
c
i
i
c
p
p p
k
k
v
1
1
1
2
(8-2!)
<=pressed .) mass *#o% u$its
,
_
,
_
,
_
,
_
,
_
,
_
k
k
i
c
k
i
c
i i
i c
k
k
i
c
k
i
c
i i c c c c
p
p
p
p
RT z
M
k
k
p A
p
p
p
p
p
k
k
A A v m
1 2
1 2
1
1
2
1
1
2
(8-2")
>eatio$s (8-2!), (8-2") are (a#id *or pressured rop #ess t&at 0? : p
c
/p
i
>+(-/ e2ui(a#e$t#):
M01"(
8.!.- 2ritical flo)
The pressure difference across a no##le accelerates the fluid. This means that when
the downstream pressure falls, both speed and throughput increases. However,
since the gas is compressible, its volume also expands when the pressure falls.
"hen the downstream pressure drops a lot, more and more pressure energy is
needed accelerate expansion, without contributing to greater throughput. This means
that by decreasing the downstream pressure, the flow approach an upper limit. This
is what we call$ critical flow
2cceleration due to fluid expansion is included in (8'):* above. &igure 8': shows the
estimated throughput of falling downstream pressure
Prod.karakterist-
09.01.2014
skr/221/HA
12
Figur 8- 3ass flo) through orific", calculat"d by diff"r"$t a''ro#imatio$s
The figure shows the formulas for incompressible and compressible fluid flow, as
previously claimed, provides almost identical results as long as the pressure drop
across the no##le is relatively small (F7A*.
8.4.4 as !alve characteristics
The results above show that we must distinguish between subcritical and critical flow.
&or gas wells, it is common to express the output in standard volumemes. This is
directly related to the mass flow
o
g
g
m
q
"ncompressi#le flow
.f the pressure drop is small, the volume flow through the no##le change so little that
we can often ignore this and approximate pressure drop as for incompressible flow$.
2
i
v - ( + p p
.
y expressing the density of state equation, we get the mass flow through the no##le
,
_
i
i c c
p
p
"
zRT
M 2
p A vA m
Prod.karakterist-
09.01.2014
skr/221/HA
13
2s the limit of this approximation, we use 7A pressure drop 2- ( + p p
i .
>
$u#critical flow
The equation for compressible flow (8'):* provide a maximum at
1
2
1
k
k
i
p
p k
_
+
,
@&is correspo$ds to approximately at 73A pressure drop. &or larger pressure reduction
over the no##le, the formula predicts reduced throughput. This is illogical and not in
accordance with reality. 2fter the maximum according to (8'):*, further pressure
reduction does not affect the mass flow.
-ritical flow
The speed through the valve becomes critical when the pressure difference across
greater than critical
1
2
1
k
k
i
p
p k
_
+
,
-orresponding speed through the no##le, we will call$ critical speed. y putting (8')8*
into (8')<*, we get
M
kRT
M
RT
k
k
v
c i
c
A
A
1
2
+
(8-
29)
-ritical mass flow can we predict by putting the critical pressure ratio into (8'):*
1
1
1
2
+
,
_
k
k
i i
i c
k
k
RT z
M
p A m
(8-30)
.f upstream pressure$ p
i
rises, it will not affect the critical speed. However, the gas
density dependence of pressure, so that a larger upstream pressure and greater
critical mass flow in accordance with (8'53*.
&igure 8.8 illustrates the gas valve characteristics, calculated from (8'):* and (8'53
Prod.karakterist-
09.01.2014
skr/221/HA
14
Figur" 8.8 Gas 1al1" charact"ristics
8.* Futur" 'roductio$
8e %i## i'$ore &ere possi.#e i$*#u= o* %ater a$d assumi$' t&at t&e reser(oir produces %it&
'as e=pa$sio$. 8e ca$ *orma##) 2ua$ti*) t&e e=pa$sio$ %it& compressi.i#it) e2uatio$
R
T
R
dp
dV
V
c
1
(8-
31)
c : compressi.i#it)
V
R
: 'as (o#ume o* t&e reser(oir (co$sta$t)
T
R
: reser(oir temperature (co$sta$t)
3p : pressure reductio$
dV : (o#ume c&a$'e
Bompressi.i#it) o* t&e 'as %as deri(ed pre(ious#) (8-). C$serted i$ (8-38) 'i(es t&e
re#atio$s&ip .et%ee$ pressure drop a$d t&e c&a$'e i$ 'as (o#ume
,
_
p
z
d
z
p
V dV
R (8-
32)
Prod.karakterist-
09.01.2014
skr/221/HA
1
C* t&e pore (o#ume i$ t&e reser(oir is co$sta$t, t&e c&a$'e is: d5, e2ua# to t&e (o#ume
produced. Ct is co$(e$ie$t to e=press t&e productio$ (o#ume at sta$dard co$ditio$s (%e cou#d
possi.#) use mass produced, or mo#es). 8e ca$ e=press t&e productio$ rate o* (o#ume c&a$'e
o(er time, di(ided .) t&e *ormatio$ *actor
,
_
,
_
,
_
z
p
d'
d
T p
T
V
p
z
d'
d
z
p
T p
T
V
d'
dV
B
q
R
o
o
R
R
o
o
R
g
g
2
1
(8-
33)
Bom.i$i$' (8-33) %it& t&e pre(ious#) deri(ed s)stem re#atio$ (8-22), %e ca$ predict
productio$ pro*i#es. @&e com.i$atio$ ca$ .e do$e a$a#)tica##), .ut o*te$ do$e $umerica##).
@&is makes it possi.#e to predict &o% (arious impro(eme$ts o* t&e productio$ s)stem %i##
a**ect productio$ rates a$d ear$i$'s.
Here %e %i## #imit our o.0ecti(e to predict t&e re#atio$s&ip .et%ee$ cumu#ati(e productio$
a$d reser(oir pressure. :) i$te'rati$' (8-40), t&is ca$ .e e=pressed as
,
_
z
p
z
p
T p
T
V d' q 4
i
i
R
o
o
R
'
'
g p
i
(8-
34)
D*te$ %e %i## a#so e=press t&e reser(oir (o#ume, &o% muc& 'as is co$tai$ed i$ t&e p#ace
4
T p
z T p
4 B V
o
i
i R
o
gi R
(8-
3)
:) putti$' (8-3) i$to (8-34), %e o.tai$ t&e c#assica# mass .a#a$ce *or t&e 'as reser(oirs
p
i
i
i
i
p
i
i
4
z
p
4 z
p
4
4
z
p
z
p
,
_
,
_
1
1
(8-
3!)
@&e resu#t mea$s t&at t&e re#atio$s&ip .et%ee$ pressure a$d +-*actor: p/z,, *a## i$ proportio$
to t&e cumu#ati(e productio$. (8e &a(e pre(ious#) s&o%$ t&at i* a c#osed reser(oir co$tai$i$'
sma## compressi.#e oi#, t&e pressure %i## *a## proportio$a##) %it& cumu#ati(e productio$.)
C* t&e measured p/z p#ots as a strai'&t #i$e a'ai$st cumu#ati(e productio$, t&is i$dicates t&at
productio$ is dri(e$ .) t&e e=pa$sio$. 8e ca$ t&e$ predict t&e 'as co$te$t i$ t&e reser(oir .)
e=trapo#ati$' it to +ero pressure.
Prod.karakterist-
09.01.2014
skr/221/HA
1!
Figur" 8.94 &roductio$ history for th" Frigg fi"ld
Prod.karakterist-
09.01.2014
skr/221/HA