You are on page 1of 17

NOTICE: This is the author's version of a work that was accepted for publication in Chemi-

cal Engineering Journal. Changes resulting from the publishing process, such as peer re-
view, editing, corrections, structural formatting, and other quality control mechanisms, may
not be reflected in this document. Changes may have been made to this work since it was
submitted for publication. A definitive version was subsequently publication in Chemical
Engineering Journal, Volume 81, Issue 1-3, Pages 41-51 (January 2001), DOI:
10.1016/S1385-8947(00)00194-7

Page 1
Comparison of isothermal and non-isothermal pipeline gas
flow models

Andrzej J. Osiadacz, Maciej Chaczykowski, Warsaw University of Technology, Institute of Heat-


ing and Ventilation, Nowowiejska 20, 00-653 Warsaw, Poland tel (48-22) 660-53-11, fax (48-22)
825-29-92, E-mail: osiadacz@hetman.iis.pw.edu.pl, mchacz@hetman.iis.pw.edu.pl

Summary
The transient flow of gas in pipes can be adequately described by a one-dimensional approach.
Basic equations describing the transient flow of gas in pipes are derived from an equation of mo-
tion (or momentum), an equation of continuity, equation of energy and state equation. In much of
the literature, either an isothermal or an adiabatic approach is adopted. For the case of slow tran-
sients caused by fluctuations in demand, it is assumed that the gas in the pipe has sufficient time
to reach thermal equilibrium with its constant - temperature surroundings. Similarly, when rapid
transients were under consideration, it was assumed that the pressure changes occurred instanta-
neously, allowing no time for heat transfer to take place between the gas in the pipe and the sur-
roundings. For many dynamic gas applications this assumption that a process has a constant tem-
perature or is adiabatic is not valid. In this case, the temperature of the gas is a function of dis-
tance and is calculated using a mathematical model, which includes the energy equation. In the
paper comparison of different (isothermal and non-isothermal) models is presented. Practical ex-
amples have been used to emphasize differences between models.

Keywords: gas flow, mathematical modelling, simulation, gas networks.

Notation:

A - cross-section area of the pipe, m2


cp - specific heat at constant pressure, J/(kgK),
cV - specific heat at constant volume, J/(kgK),
D - pipe diameter, mm,
f - Fanning friction coefficient, -,
g - the net body force per unit mass (the acceleration of gravity) m/s2
k - pipe roughness, mm
kL - heat transfer coefficient, W/(mK),
L - pipeline length, m,
M -mass flow, kg/s,
p(x) - pressure at x, Pa
q - the heat addition per unit mass per unit time, W/kg
Qn - flow (under standard conditions) m3/h,
R - specific gas constant, J/(kgK),
t - time, s
T - gas temperature, K
Tsoil - soil temperature, K,
w - flow velocity, m/s
x- spatial coordinate, m,
Z - compressibility factor, -.

Page 2
Greek symbols
α - the angle between the horizon and the direction x,
λ - thermal conductivity coefficient of gas, W/(mK),
µ - viscosity of natural gas, N s/m2,
ρ - density of gas, kg/m3

Note:
Flow rate Qn is shown in the standard conditions of 273.15 K, 0.1 MPa

1. Introduction
It is a well established fact that flow in gas pipelines is unsteady. Conditions are always changing
with time, no matter how small some of the changes may be. When modelling systems, however,
it is sometimes convenient to make the simplifying assumption that flow is steady. Under many
conditions, this assumption produces adequate engineering results. On the other hand, there are
many situations where an assumption of steady flow and its attendant ramifications produce un-
acceptable results. Dynamic models are just a particular class of a differential equation model in
which time derivatives are present.
During transport of gas in pipelines the gas stream loses a part of its initial energy due to friction-
al resistance which results in a loss of pressure. This is compensated for by compressors installed
in compressor stations.
Compression of the gas has the undesired side effect of heating the gas. The gas may have to be
cooled to prevent damage to the main transmission pipeline. If the cooler is installed heat from
the gas is passed to the air in a force draught heat exchanger in which one or more fans operate,
depending on the number of compressors in service.
Cooling of the gas is desirable because it improves the efficiency of the overall compression pro-
cess. As always it is a matter of balancing capital and maintenance costs against operating costs.

2. Basic equations
The transient flow of gas in pipes can be adequately described by a one dimensional approach.
Basic equations describing the transient flow of gas in pipes are derived from an equation of mo-
tion (or momentum), an equation of continuity, equation of energy and state equation [7]. In prac-
tice the form of the mathematical models varies with the assumptions made as regards the condi-
tions of operation of the networks. The simplified models are obtained by neglecting some terms
in the basic model as a result of a quantitative estimation of the particular elements of the equa-
tion for some given conditions of operation of the network [3], [8].
2.1 Conservation of mass: continuity equation
Generally, the continuity equation is expressed in the form:
∂ ρw ∂ρ
b g
− = (1)
∂x ∂t
where: w - flow velocity,
ρ - density of gas
Substituting M = ρ w A, we have:
1 ∂M ∂ρ
− = (2)
A ∂x ∂t

Page 3
where: A - cross-section area of the pipe,
M - mass flow
2.2 Newton's second law of motion: momentum equation
According to [2] the basic form of momentum equation can be expressed in the form

∂ ρw ∂ ρw
b g c h
2
∂p 2 fρw 2
− − − gρ sin α = + (3)
∂x D ∂t ∂x
where:
f - Fanning friction coefficient,
g - the net body force per unit mass (the acceleration of gravity)
α - the angle between the horizon and the direction x.
∂ ρw 2 fρw 2
b g ∂
The constituent factors , , ρg sin α andc h ρw 2 define the gas inertia, hydraulic
∂t D ∂x
friction force, force of gravity and the flowing gas dynamic pressure respectively.
2.3 State equation
An equation of state for a gas relates the variables p, ρ, and T. The type of equation which is
commonly used in the natural gas industry is [6], [11]:
p
= ZRT (4)
ρ
where the deviation from the ideal gas law is absorbed in the compressibility factor Z, which is a
function of p and T.
2.4 Conservation of energy
The basic form of energy equation, according to [10] is the following

∂ LMb IJ OP + ∂ LMbρwAdxgFG c T + p + w IJ OP
gFGH
w2 2
qρAdx = ρAdx cv T + + gz + gz (5)
∂t K Q ∂x N H ρ 2
v
N 2 KQ
where: q - the heat addition per unit mass per unit time,
T - gas temperature,
cV - specific heat at constant volume.
Before going over to analysis of transient conditions, the simple case of steady state conditions
will be considered.

3. Steady state non-isothermal model


The temperature of gas, as a function of length of the pipe is calculated using the heat balance
equation, assuming that heat transfer process is quasi steady state, expressed by the following
equation
c p MdT = − k L (T − Tsoil )dx (6)

where: cp - specific heat at constant pressure, J/(kgK),


M -mass flow, kg/s,
kL - heat transfer coefficient, W/(mK),
T - gas temperature, K,
Tsoil - soil temperature, K.
The form of equation (6) results from the following transformations of the equation (11). Since
under steady state conditions
Page 4
∂ LMb IJ OP = 0
∂t
ρAdx cv T +
2
gFGH
w2
+ gz
KQ
N
we can write the energy equation describing the flow of gas throw the horizontal pipe in the form
∂ LMb IJ OP
qρAdx =
∂x
p w2
ρwAdx cv T + +
ρ 2
gFGH KQ
N
The quantitative analysis of energy equation (paragraph 4.1) has shown that under steady state
conditions term (III) (spatial derivative of the kinetic energy) can be neglected in comparison
with other convective terms. Substituting the enthalpy and neglecting term (V) (heat conduction
through the gas along the pipeline) we finally obtain the energy equation in the form of the heat
balance equation. Equation (6) can be written in the form:
dT k
= − L dx
(T − Tsoil ) cp M
By integrating between T(0) (Tx=0) and T(x), x∈(0,L] we get:
T( x) x
dT k
z
T ( 0)
(T − Tsoil )
= − L dx
cp M 0 z
Finally:
T ( x ) = Tsoil + T (0 ) − Tsoil e − βx
b g (7)
where: β=kL/(cpM)

For non-isothermal steady state flow, pressure at specific point of the pipe can be expressed by
the following equation [1]:
p( x ) = − KM 2
2
b p(0)g (8)

where: p(0) - pressure at x=0, Pa,


M - mass flow, kg/s,
K - coefficient defined by the equation:

K=
LM FG
ZR 4f
Tsoil x +
T(0) − Tsoil T(0) − Tsoil − βx
− e − 2 T(0) − T(x)
IJ b gOP (9)
2
A D N H β β K Q
where: x - spatial coordinate, m,
f - Fanning friction coefficient,
Z - compressibility factor,
R - specific gas constant, J/(kgK),
A - cross section area of the pipe, m2,
Specific heat at constant pressure can vary strongly with temperature of the gas. In the gas trans-
portation systems the temperature ranges are modest and the value of specific heats may be as-
sumed constant [4]. Such assumption, however, cannot be applied with regard to the compressi-
bility factor. Its value varies significantly with temperature and pressure of the gas and separate
calculations are carried out for every discretization section of the pipeline.

3.1 Steady state simulation


Differences between two models (isothermal and non-isothermal) were analyzed using a part of
existing real gas system, Yamal – West Europe. This gas transportation system (Fig. 1) consists
of five compressor stations, installed on the Polish terrain. At each compressor station there are

Page 5
installed between 2 and 3 centrifugal compressors, driven by gas turbines. For the purpose of our
investigation, one pipe between two compressor stations was taken.
German Belarus
border compressor border
station
5 4 3 2 1

139.0 km 177.0 122.0 117.0 126.5 2.5

Fig. 1. Structure of gas transportation system.

Calculations were carried out for the following parameters:


pipe diameter D = 1422 mm, pipe wall thickness 19.2 mm,
pipeline length L=122 km,
pressure at x=0 (discharge pressure) p1 = 8.4 MPa,
temperature at x=0 (discharge temperature) Tx=0=42.5 °C,
density of natural gas (under standard conditions) ρn = 0.7156 kg/m3,
viscosity of natural gas µ=0.135×10-4 N s/m2
flow (under standard conditions) Qn=2019950 m3/h,
soil temperature Tsoil=12 °C,
heat transfer coefficient kL=25 W/(mK),
pipe roughness k=0.03 mm.
Fanning friction coefficient and compressibility factor were calculated for every discretization
section of the pipeline using Nikuradse [1] and SGERG 88 [5] equations respectively. Results of
the investigations are shown in Fig. 2 and Fig. 3.
45

o
T = CONST, To = 12 C
40 o
T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C
35

30

25
T ( C)
o

20

15

10

0
0 12.2 24.4 36.6 48.8 61 73.2 85.4 97.6 109.8 122
x (km)

Fig. 2 Temperature profile along the pipeline for a) T=const, To=12 oC, b) T≠const, To=42.5 oC,
c)T≠const, To=30 oC.

Page 6
8.5

o
T = CONST, To = 12 C
o
8.4 T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C

8.3

8.2
p (Mpa)

8.1

7.9

7.8
0 12.2 24.4 36.6 48.8 61 73.2 85.4 97.6 109.8 122
x (km)

Fig. 3 Change of pressure along the pipeline for a) T=const, To=12 oC,b) T≠const, To=42.5 oC, c)
T≠const, To=30 oC.
Maximum pressure difference between isothermal and non-isothermal flow:
pisotherm x = J − pn −isotherm 7.894 − 7.879
δ max = x= J
⋅ 100% = ⋅ 100 = 019
. %
pisotherm x = J 7.894
Maximum pressure difference between non-isothermal flow (without cooling system) and non-
isothermal flow (with cooling system):
pn −isotherm x = J − pn −iso −c 7.894 − 7.885
δ max = x= J
⋅ 100% = ⋅ 100 = 011
. %
pn −isotherm x = J 7.894

Discharge temperature Capacity Capacity difference


o
C m3/h m3/h %

42.5 (without cooling system) 2019949.7 - -

30.0 (with cooling system) 2032010.2 12 060.5 + 0.6

12.0 (soil temperature) 2049998.4 30 048.7 + 1.5

Table 1 Comparison of capacity of the pipeline for above analyzed cases.

Fig.4 shows that for non-isothermal flow, the pressure drop along the pipeline is greater than in
the isothermal case. It is resulted by the decrease of the density of the gas, allowing the smaller
mass of gas to be transported at the specific velocity.

Page 7
7.95
7.916
7.903
7.894
7.9 7.879
7.875
7.859
7.85

7.8 T ≠ C O N ST,
T o = 42.5 o C
p (MPa)

7.75

7.7

7.656 T = C ON ST,

7.65 7.639 T o = 12 o C

7.6

7.55
2 3 4 5
COMPRESSOR STATION #

Fig. 4. Comparison of suction pressures for the whole transportation system.


On the Fig. 5 difference in HP used by compressor stations is presented. We can see significant
difference in energy consumption by the drivers of the compressors of each case. It means that
the cost of running of the transportation system is a function of discharge temperature. The low-
est costs correspond to the isothermal flow.

1.740

1.729

1.720

1.700 1.698
HP (MW)

1.680

1.660 1.657

1.640
12.0 42.5 30.0

T0 (oC)

Fig. 5. HP as a function of discharge temperature for the whole transportation system.

It is clear that decreasing the discharge temperature of the gas, the effectiveness of the transporta-
tion process can be significantly increased.
Page 8
4. Transient non-isothermal model
With respect to the conveyance of gas in pipelines, there are two technically relevant extreme cas-
es of pipe flow:
– flow without heat exchange with the ground outside: adiabatic and more especially, isentropic
flow,
– flow with heat exchange with the ground outside, which is regarded as being a heat storage unit
of infinite capacity with constant temperature T0: isothermal flow.
The flow processes of greatest concern here are those in which temperature equalization with the
ground outside cannot take place. The temperature profile is a function of pipeline distance. In this
case the transient, non-isothermal flow of gas in a horizontal pipe ( ρg sin α = 0,

ρwAgzdx = 0 ) is described by the system of equations 2÷5.
b g
∂x
Two contradictory constraints are imposed on the above equations. The requirement is that on the
one hand the description of the phenomenon is accurate, and on the other there is sufficient sim-
plification to allow solving of this model by reasonable computation means. As a rule simplified
models are sought which present a reasonable compromise between the accuracy of the descrip-
tion and the costs. The simplified models are obtained by neglecting some terms in the basic (ac-
curate) equation. This results from the quantitative estimation of particular elements of the equa-
tion, carried for some given conditions of operation of the pipeline. This means that the model of
transient flow used for simulation should be fitted to the given conditions of operation of the pipe.
A necessary condition for proper selection of the model is therefore the earlier analysis of these
conditions. Estimation of the particular terms of the equation (5) for given operating conditions
and a given geometry of the pipe is given below.

4.1 Energy equation simplifications


Assuming that heat transfer is limited only to conduction through a walled tube and the gas along
a pipeline the following equation can be written
∂ ∂T FG IJ
qρAdx = A λ dx − k L T − Tsoil dx b g (10)
∂x ∂x H K
where: λ - thermal conductivity coefficient of gas, W/(mK),
kL - heat transfer coefficient, W/(mK).
Combining equations (5) and (10), the final version of the equation can be put into the form
∂ ∂ ρwAp FG ∂ ρAw 3
FG IJIJ b ∂ FG ∂ ∂T IJ
ρwAcv Tdx +
b g dx + dx + ρwAgzdx -
g λA dx +
∂x442443 ∂x
1 ρ H ∂x
H K2K ∂x44244
1 H 3 1 ∂x442∂44 x 3 K
I
1 44 2 44 3 1 4 4 2 44 3 IV
II III V
(11)
∂ ∂ F ρAw I ∂ 2
+ k bT − T gdx + b ρAc Tdx g + G dxJ + b ρAgzdx g = 0
L
144244 3 1 ∂t 4
soil
4244 3 1 ∂t H 2
v
K ∂t
VI
VII
44244 3 14243 IX
VIII

By integrating the above equation between x = 0 and x = L (where L is the length of the pipe) for
the following parameters of the system:
pipe diameter D = 1422 mm, pipe wall thickness 19.2 mm,
pipeline length L=122 km,
pressure at x=0 (discharge pressure) p1 = 8.4 MPa,
suction pressure p2 = 7.88 MPa,

Page 9
temperature at x=0 (discharge temperature) Tx=0=42.5 °C,
temperature at x=L (suction temperature) Tx=L=13 °C,
density (under standard conditions) ρn = 0.7156 kg/m3,
flow (under standard conditions) Qn=2019950 m3/h,
soil temperature Tsoil=12 °C,
heat transfer coefficient kL=25 W/(mK),
thermal conductivity coefficient of gas λave=3.4×10-2 W/(mK),
specific gas constant R = 518.8 J/(kgK),
specific heat at constant pressure cp = 2.278×103 J/(kgK),
specific heat at constant volume cv = 1.759×103 J/(kgK),
we get the following values for each term of the equation


L

z ρwAcv T dx ≈ ρ n Qn cv T
b g c −T h=
∂x x= L x=0
(I) 0 → O(107)
2019950
= 0.7156 ⋅ ⋅ 1.759 ⋅ 10 3 ⋅ 286 − 3155
b . = −2.08 ⋅ 10 7 W g
3600

∂ ρwAp FG IJ
L

z dx ≈ ρ n Qn ZR T c −T h
∂x ρ H K x= L x=0
(II) 0 → O(106)
2019950
= 0.7156 ⋅ ⋅ 0.97 ⋅ 518.8 ⋅ 286 − 315.5 = −6.0 ⋅ 10 6 W
b g
3600

L
∂ ρAw 3 FG
1
dx ≈ ρ n Qn w 2
IJ e j=
(III)
z 0
∂x 2 2H K x= L
− w2
x =0
→ O(102)
1 2019950
=
2
⋅ 0.7156 ⋅
3600
c
⋅ 4.96 2 − 513
. 2 = −3.4 ⋅ 10 2 W h


L

ρwAgz dx ≈ ρ n Qn g z x = L − z x =0
z b g c h
(IV) 0 ∂x

Assuming that the difference of levels is e.g. 100 m, we get


2019950
ρ n Qn g z x = L − z x =0 = 0.7156 ⋅
c h ⋅ 9.81 ⋅ 100 = 3.9 ⋅ 10 5 W → O(105)
3600
L ∂ F ∂T I O
L
∆T − ∆T
(V) − z M G λA J Pdx ≈ − λ ave A
x= L x =0

N∂x H ∂x K Q
0
∆x
from the steady state analysis we have
∆T 4.8
= K/m
∆x x =0 6100

Page 10
∆T 0.2
= K/m
∆x x= L 6100

∆T − ∆T π ⋅ 13836
. 2
4.8 − 0.2
− λ ave A x= L x=0
= −3.4 ⋅ 10 −2 ⋅ ⋅ = 3.9 ⋅ 10 −5 W → O(10-5)
∆x 4 6100
L

(VI) ∫ k L (T − Tsoil ) dx ≈ k L L∆ts


0

Substituting the mean temperature difference between the gas and the soil ∆Ts=15.75 K we
have
k L L∆Ts = 25 ⋅ 1.22 ⋅ 10 5 ⋅ 15.75 = 4.8 ⋅ 10 7 W → O(107)

∂ ∆T
L

(VII) z ρAcv T dx ≈ ρAcv L


b g
0
∂t ∆t

Let the increment of the temperature be 5 K and be achieved within ∆t = 1h. Substituting the
mean value of gas density along the pipeline ρ = 64.27 ⋅ kg / m 3 we get
∆T π ⋅ 1.3836 2 5
ρ Acv L = 64.27 ⋅ ⋅ 1759
. ⋅ 10 3 ⋅ 122
. ⋅ 10 5 ⋅ = 2.88 ⋅ 10 7 W → O(107)
∆t 4 3600
∂ ρAw 2 FG Lw ∂ ρAw Lwρ n ∆Qn
IJ b g
L

(VIII) z 0
∂t 2
dx ≈
H 2 ∂t

2 K ∆t

Let the increment of the load of the pipeline at x=L be 0.5×Qn and be achieved within ∆t=1h.
We then get
Lwρ n ∆Qn 1.22 ⋅ 10 5 ⋅ 5.04 ⋅ 0.7156 1009975
= ⋅ = 618
. ⋅ 10 7 W → O(107)
2 ∆t 2 3600
∂ ∂z
L

(IX) z ρgzA dx ≈ ρgAL = 0 (for each gas pipeline)


b g
0
∂t ∂t
Based on above given analysis we can write the simplified form of the energy equation

∂ LMb IJ OP + ∂ LMbρwAdxgFG u + p IJ OP
k L Tsoil − T dx =
b g ∂t
ρAdx u +
w2
2
gFGH K Q ∂x N H ρKQ
N
The character of the results cannot be generalized. This can only be the starting point, which al-
lows the forwarding of the hypothesis that in the case when the selected parameters do not change
rapidly, transient non-isothermal flow through the horizontal pipe can be represented be the set of
equations in the form

Page 11
R|− ∂ bρwg = ∂ρ
|| ∂p∂x 2 fρw∂t ∂ bρwg ∂ cρw h
2 2

|− ∂x − D = ∂t + ∂x
S| ∂ L F w IJ OP + ∂ LMbρwAdxgFG u + p IJ OP
2
k bT − T gdx = Mb ρAdx gG u +
|| ∂t N H 2 K Q ∂x N H ρKQ
L soil

|| ρp = ZRT
T
4.2 Transient simulation
The investigations were carried out using the method of lines [9] for the following values of pa-
rameters
compressor compressor
station 1 station 2

D(1422×19.2 mm)

122.0 km

Fig. 6 Structure of gas transportation system.

pipe diameter D = 1422 mm, pipe wall thickness 19.2 mm,


pipeline length L=122 km,
temperature at x=0 (discharge temperature) Tx=0=42.5 °C,
density (under standard conditions) ρn = 0.7156 kg/m3,
soil temperature Tsoil=12 °C,
heat transfer coefficient kL=25 W/(mK),
thermal conductivity coefficient of gas λave=3.4×10-2 W/(mK),
specific gas constant R = 518.8 J/(kgK),
specific heat at constant pressure cp = 2.278×103 J/(kgK),
specific heat at constant volume cv = 1.759×103 J/(kgK),
p(0, t ) = const = 8.4 MPa UV boundary conditions
QN ( L, t ) = f (t ) W
The function QN ( L, t ) = f (t ) is shown in Figure 5. For given range of variation of the load flow
behavior in the pipeline is approximated accurately by the rough-pipe flow law [1]. This law is
represented by the horizontal lines in the Moody diagram of friction factor, with each line corre-
sponding to a specific value of the relative roughness of the pipe.

Page 12
4000

3500

3000
Qn x10 (m /h)

2500
3
3

2000

1500

1000

500
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
t (h)

Fig. 7 Change of flow with time (x=L) – (boundary condition).


Results of calculation are presented in Figs. 8, 9, 10, 11, 12.
8

7.8

7.6

7.4
p (Mpa)

7.2

o
T = CONST, To = 12 C
6.8 o
T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C

6.6
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
t (h)

Fig. 8 Change of pressure at x=L.


Maximum pressure difference between non-isothermal flow (without cooling system) and iso-
thermal flow:
pisotherm x = J − pn −isotherm 7.145 − 7.067
δ max = x= J
⋅ 100% = ⋅ 100 = 109
. %
pisotherm x = J 7.145

Page 13
Maximum pressure difference between non-isothermal flow (without cooling system) and non-
isothermal flow (with cooling system):
pisotherm x = J − pn −isotherm 7.145 − 7.117
δ max = x= J
⋅ 100% = ⋅ 100 = 0.39 %
pisotherm x = J 7.145
21

19

17

15
T ( C)

13
o

11

o
T = CONST, To = 12 C
o
7 T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C

5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
t (h)

Fig. 9 Change of temperature at x=L.


Profile of temperature at x=L is caused by load variations and by distributed velocity of the gas
along the pipeline.
8.6

o
T = CONST, To = 12 C
8.4 o
T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C
8.2

8
p (MPa)

7.8

7.6

7.4

7.2

7
0 12200 24400 36600 48800 61000 73200 85400 97600 109800 122000
x (m)

Fig. 10. Change of pressure along the pipeline for t=12 h.

Page 14
1.3
o
T = CONST, To = 12 C
o
T ≠ CONST, To = 42.5 C
o
1.25 T ≠ CONST, To = 30 C

1.2

1.15
ε

1.1

1.05

1
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
t (h)

Fig. 11 Change of compressor ratio at compressor station No. 2.

Maximum compressor ratio difference between non-isothermal flow (without cooling system) and
isothermal flow:
ε − ε n −isotherm . − 126
118 .
∆ε max = isotherm ⋅ 100% = ⋅ 100 = 6.78 %
ε isotherm 118
.
Maximum compressor ratio difference between non-isothermal flow (with cooling system):
and isothermal flow:
ε − ε n −isotherm 1074
. − 1078
.
∆ε max = isotherm ⋅ 100% = ⋅ 100 = 0.37 %
ε isotherm 1074
.

Page 15
1.4

o
T = CONST, To = 12 C
o
1.2 T ≠ CONST, To = 42.5 C
o
T ≠ CONST, To = 30 C

0.8
Ne (MW)

0.6

0.4

0.2

0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
t (h)

Fig. 12 Changes of HP at compressor station No. 2 for the isentropic compression.

Maximum HP difference necessary to keep constant discharge pressure at compressor station No.
2 for non-isothermal flow (without cooling system) and isothermal flow:
N − N n −isotherm 0.666 − 1255
.
∆N max = isotherm ⋅ 100% = ⋅ 100 = 88.43 %
N isotherm 0.666
Maximum HP difference necessary to keep constant discharge pressure at compressor station No.
2. for non-isothermal flow (with cooling system) and isothermal flow:
N − N n −isotherm 0.666 − 0.807
∆N max = isotherm ⋅ 100% = ⋅ 100 = 2117
. %
N isotherm 0.666

5. Conclusions
It is clear that cooling of the gas improves the efficiency of the overall compression process. There
exists a significant difference in the pressure profile along the pipeline between isothermal and
non isothermal process. This difference increases when the quantity of gas increases. This shows
that in the case when gas temperature does not stabilize the use of an isothermal model leads to
significant errors. The problem of choosing the correct model is a function of network structure
and network complexity.

6. Bibliography
[1] The American Gas Association 1990. Distribution – System Design. A.G.A., Arlington.
[2] Daneshyar, H. 1976. One-dimensional Compressible Flow. Pergamon Press, New York.
[3] Goldwater, M. H. and Fincham, A. E. 1981. Modelling of Dynamical Systems (ed. H. Ni-
cholson). Peter Peregrinus, Stevenage.
[4] Holman, J. P. 1980. Thermodynamics. McGraw-Hill Book Company, New York.
[5] ISO 12213–3 Natural gas – Calculation of Compression factor – Part 3: Calculation using
physical properties.

Page 16
[6] Kralik, J. et al. 1988. Dynamic Modelling of Large Scale Networks With Application to Gas
Distribution. Elsevier, New York.
[7] Osiadacz, A. J. 1989. Simulation and Analysis of Gas Networks. Gulf Publishing Company,
Huston.
[8] Osiadacz, A. J. 1996. Different Transient Models - Limitations, Advantages And Disad-
vantage. PSIG The 28th Annual Meeting, San Francisco.
[9] Schiesser, W. E. 1991. The Numerical Methods of Lines. Academic Press, London.
[10] Shapiro, A. H. 1954. The Dynamics and Thermodynamics of Compressible Fluid Flow. The
Ronald Press Company, New York.
[11] Wylie, E. B., Streeter V. L. 1978. Fluid Transients. McGraw – Hill, New York.

Page 17