ModelingHighlyCompressibleFlow I
ModelingHighlyCompressibleFlow I
Jeffrey A. Olsen
Senior Development Engineer
Applied Flow Technology Corp.
Woodland Park, Colorado
ABSTRACT Subscripts
A highly accurate compressible flow 1 = Location 1 in pipe
solution engine has been combined with an advanced 2 = Location 2 in pipe
graphical interface in AFT Arrow for Windows. AFT i = junction at which solution is sought
Arrow models the flow of real gases, including j = junctions with pipes connecting to
steam, in complex network piping systems, and also junction i
offers heat transfer and system energy balance
capabilities. A discussion of the governing equations
is given, as well as a summary of the solution method
for network pipe systems. An overview of sonic INTRODUCTION
choking in gas piping systems is also given. Several Compressible flow modeling in pipe
applications are discussed. networks is a challenging technical application
because of the coupled nature of the governing
NOMENCLATURE equations and the large number of parameters
a = sonic speed involved. For single pipe systems engineers can
A = cross-sectional flow area of a pipe frequently perform a rough compressible flow
D = diameter of a pipe calculation for prediction of pressure drop. However,
f = friction factor even slight increases in complexity, including
FTo = Parameter in Equation 9 branched or looped systems, results in extreme
Ff = Parameter in Equation 9 analytical difficulties. If real gas effects are included
h = enthalpy, static (which is typical of steam and natural gas systems)
ho = enthalpy, stagnation the complications compound further.
L = length of a pipe Because of the analytical difficulties
m& = mass flow rate involved, engineers will typically overdesign the
M = Mach Number piping system using seat-of-the-pants
P = pressure approximations. One serious drawback of
Po = pressure, stagnation overdesigning systems is the increased cost to build
R = gas constant and operate the system.
T = temperature, static To accurately calculate the pressure drop,
To = temperature, stagnation flow rates and temperatures in gas systems, a coupled
V = velocity solution method that simultaneously solves all of the
x = length relevant equations is required. In addition, a piping
Z = compressibility factor network in which flow splits are unknown further
γ = specific heat ratio requires a matrix method to balance flow and energy
ρ = density in the network.
Still another formidable challenge is that of
sonic choking. Sonic choking occurs when the gas
* Member ASME
cannot flow through a pipe or flow restriction
without accelerating past its sonic speed. In such
cases the flow “chokes”, which means that a shock Equation of State: P = ZρRT (4)
wave is generated and changes in downstream
conditions cannot produce any additional flow V
Mach Number: M= (5)
through the system. γZRT
To assist engineers in designing gas piping
systems which encounter the previously mentioned In order to obtain a system mass and energy
difficulties, AFT Arrow for Windows, a compressible balance, the following two equations must be
flow modeling software product for pipe networks, satisfied for all branching nodes:
was developed. AFT Arrow addresses all of the
relevant phenomena in gas systems including sonic Balance Mass at Branches:
choking.
n
∑ m& ij = 0
AFT Arrow incorporates these modeling
capabilities within an advanced Windows graphical (6)
interface (Figure 1). All modeling is performed with j =1
simple drag-and-drop methods, offering a short
learning curve and increased ability to identify Balance Energy at Branches:
modeling errors.
n
∑ ⎛ ⎞
1
m& ij ⎜ hij + Vij 2 ⎟ = 0 (7)
FUNDAMENTAL EQUATIONS OF ⎝ 2 ⎠
j =1
COMPRESSIBLE FLOW
There are five equations that govern
compressible flow in piping systems. For each The Length March Method
individual pipe these are: AFT Arrow implements three independent
methods to solve these equations, of which two will
dρ dV be discussed here. For systems that do not approach
Mass: + =0 (1) sonic velocity, the Length March Method (LMM) is
ρ V
optimal. The LMM breaks each pipe into a certain
1 f number of sections and solves the governing
Momentum: dP + ρV 2 dx + ρVdV = 0 (2) equations over each section.
2 D
⎛ 1 ⎞
Energy: m& d ⎜ h + V 2 ⎟ = q (3)
⎝ 2 ⎠
Applying calculus and after some algebraic
2
manipulation, the following equation can obtained Integration yields:
which is based on a fixed length step between
computing sections. M22 To,2 Z γ
ln − FTo ln − FTo ln 2 − FTo ln 2
M 12 To,1 Z1 γ1
dPo γM 2 ⎛ fdx dTo dZ dγ ⎞ x 2 = x1 +
=− ⎜ + + + ⎟ Ff
f
Po 2 ⎝ D To Z γ ⎠ D
(9)
Integration yields:
By selecting an increase in Mach Number,
⎡ γM 2 ⎛ f ⎞⎤ say by increments of 0.01 from M1 to M2, the
2 + ln Z 2 + ln γ 2 ⎟ ⎥
To,
Po,2 = Po,1 exp ⎢ − ⎜ ( x 2 − x1 ) + ln distance to x2 can be computed that is required to
⎢ 2 ⎜D T Z γ 1 ⎟⎠ ⎥
⎣ ⎝ o ,1 1 ⎦ obtain this change in Mach Number. Again, heavy
(8) iteration is required because there are multiple
unknowns at section 2 in this equation which are
Conditions at section 1 are known and the converged upon by applying equations 1-5
goal is to find conditions at section 2 which satisfies repeatedly.
the above equation. Obviously, extensive iteration is The MMM is well suited to pipes which
required because there are multiple unknowns at have sonic choking, as it is able to accurately follow
section 2 in this equation which are converged upon the rapid acceleration at the end of the pipe and
by applying Equations 1-5 repeatedly. rapidly changing conditions. An example of such an
The logarithmic terms that involve γ and Z application will be given in later section.
account for real gas effects, and have been observed
by AFT to be significant for many real gas systems,
including steam systems. AFT is not aware of any NETWORK SOLUTIONS
other published derivations which include these To solve Equations 6 and 7 for all junctions
terms and thus account for the real gas effects with in a network, a network solution method is required.
complete accuracy. AFT Arrow uses a modified Netwon-Raphson
method to determine the junction pressures and pipe
The Mach March Method flow rates which satisfy the balance equations.
An alternative solution method is optimal
for systems that have sonic choking. This method,
the Mach March Method (MMM), uses a variable SONIC CHOKING
length step based on how fast the Mach Number
Sonic choking is the phenomenon in gas
changes. It takes solution steps over equal Mach
piping systems where the flow conditions reach a
Number increments rather than length increments.
Mach Number of 1, and cannot accept any further
Applying calculus and algebra, the
increase in flow rate. In such cases a lowering of
following equation can be obtained for the MMM:
downstream pressure will not produce any additional
flow through the system.
dM 2 dTo fdx dZ dγ A convenient equation can be derived (Saad
= FTo + Ff + FTo + FTo
M2 To D Z γ (1993)) which relates all relevant parameters which
affect sonic choking. This equation is based on the
where: continuity equation. The equation shown here is that
from Saad (1993), which is for an ideal gas, but
(1 + γM 2 )⎛⎜⎝1 + γ 2− 1 M 2 ⎞⎟⎠
corrected for real gas effects.
FTo = m& = A
Po
f (M) (10)
1− M 2 To
⎛ γ −1 2⎞
γM 2 ⎜ 1 +
⎝ 2
M ⎟
⎠ γ ⎛ γ −1 2⎞ [
−( γ +1) 2( γ −1) ]
Ff = f (M) = M ⎜1 + M ⎟
ZR ⎝ 2 ⎠
1− M 2
3
Flow chokes at exit 1.0
into atmosphere
0.8
Mach 0.6
Number
Flow chokes at 0.4
restriction in pipe
0.2
0.0
0 20 40 60 80 100
Pipe Length (feet)
4
J1 J2
100
P1>
80
Static 60
Figure 4a. 500 psi steam system example model. Temperature
(deg. F)
40
500 20
400 0
0 20 40 60 80 100
Stagnation 300 Pipe Length (feet)
Pressure
(psia) Figure 4d. Static temperature variation in a 100
200
ft. sonically choked pipe predicted by
100 AFT Arrow.
0 1500
0 20 40 60 80 100
Pipe Length (feet)
1000
Figure 4b. Stagnation pressure variation in a 100 Velocity
foot sonically choked pipe predicted (feet/sec)
by AFT Arrow. 500
500
0
400 0 20 40 60 80 100
Static 300 Pipe Length (feet)
Pressure Figure 4e. Velocity variation in a 100 foot
(psia) 200
sonically choked pipe predicted by
100 AFT Arrow.
0
0 20 40 60 80 100
Pipe Length (feet)
Figure 4c. Static pressure variation in a 100 foot
sonically choked pipe predicted by
AFT Arrow.
5
J4 J5 J6
Po=196.35 psia Po=195.89 psia Po=14.70 psia
DP=0.3401 psid DP=159.8662 psid DP=19.8766 psid
P4> P5>
M# Out=0.1059 M# Out=0.9947
P3^ Po In=196.01 psia Po In=36.03 psia
M# Out=0.1056 Po Out=195.89 psia Po Out=34.58 psia
Po In=196.61 psia L=10 feet L=10 feet
Po Out=196.35 psia Dh=16.876 inches Dh=16.876 inches
L=21 feet
Dh=16.876 inches
P2^
M# Out=0.9950
Po In=277.33 psia
Po Out=269.40 psia
J1 L=2 feet
Dh=6.065 inches Second choke point
Po=300.00 psia
DP=0.0000 psid at sonic orifice (area
ratio = 0.18)
J2
P1> Po=297.41 psia
M# Out=0.7020 DP=20.0725 psid
Po In=300.00 psia
Po Out=297.41 psia
L=1 feet
Dh=6.065 inches
200
Stagnation
Pressure
(psia) 150
MULTIPLE CHOKING POINTS
100
It should be noted that sonic choking can
50
exist in any multiple of the above geometries and at
multiple locations in the same pipe system. AFT 0
0 10 20 30 40 50
Arrow’s solution method incorporates this reality. Flow Length (feet)
The model shown in Figure 5 is that of a “triple Figure 6. Stagnation pressure drop in Triple
choke steam flow”, where sonic choking occurs in Choke Steam Flow model predicted by
three places, as indicated in Figure 5. The pressure AFT Arrow.
drop in the system is as shown in Figure 6.
180
Sonic choking in multiple locations in a
160
flow path can be understood as follows. Recall that
Sonic Fow Rate (lbm/sec)
140
the sonic flow rate will have a unique value given
120
upstream conditions and the piping losses and
100
thermal environment. No matter what happens at the
80
exit boundary condition, the flow rate will not
60
increase or change in any way. However, by
40
changing the upstream conditions the sonic flow rate
20
can be changed. This is because the conditions are
being changed on the same side as the shock wave 0
0 100 200 300 400 500
and can thus influence the fluid behavior. Figure 7 Supply Pressure (psia)
graphs the supply pressure vs. sonic flow rate for a
series of conditions for this system. Figure 7. Sonic flow rate vs. supply pressure in
Triple Choke Steam Flow model.
6
J4 J5
P4>
J6
P5>
P3^
J3
P2^
Sonic flow rate determined
J1 J2
by system up to expansion
P1>
Figure 8. Triple Choke Steam Flow model pseudo system for determining shock
wave pressure drops.
7
J21 ^
J27 ^
P15^
J15
P26^
J26 P10^
J37 J38 J39 J7 ^
<P14 P8> P9>
P25^
P2^ P12^
J32 J30
J4 J41 J5 J31
<P27
P6> <P7
P3v
P24v
Figure 9. AFT Arrow model of 800 psi steam system flowing in a cogeneration piping system.
Some examples of various pipe systems choking a numerous sonic orifices as it is being
modeled with AFT Arrow are shown in Figures 9-11. delivered to a 100 psi receiver location.
In Figure 9, a steam distribution system for a
cogeneration facility is shown. This system involved
determination of delivery pressures at multiple CONCLUSIONS
locations around the facility. Operating pressures AFT Arrow offers a highly accurate solution
were 800 psi. method for real gas compressible flow modeling,
A second application is that shown in Figure including steam and natural gas systems. Previous
10 of a the fuel supply to a Rolls Royce Avon gas real world applications have demonstrated its ability
turbine. In this application a gaseous hydrocarbon is to model a wide range of gas piping systems. The
supplied around a manifolding system, with the graphical interface implementation within Microsoft
objective being a determination of the delivery Windows offers ease-of-use benefits and numerous
temperatures to ascertain the thermal distribution. customizing features making AFT Arrow an excellent
Unbalanced thermal distribution significantly impacts tool for power system engineers.
gas turbine operations and expected lifetime.
A third application is shown in Figure 11. REFERENCES
This system is a feedwater heater vent system, and
Saad, M.A., Compressible Fluid Flow, 2nd Edition,
involves 500 psi steam that experiences sonic
Prentice-Hall, Englewood Cliffs, NJ, 1993.
8
J12
P11^
J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 J11
<P10 <P9 <P8 <P7 <P6 <P5 <P4 <P3 <P2 <P1
P12^
J13 J14 J15 J16 J17 J18 J19 J20 J21 J22 J23
<P22 <P21 <P20 <P19 <P18 <P17 <P16 <P15 <P14 <P13
P33^
J24 J25 J26 J27 J28 J29 J30 J31 J32 J33 J34
<P32 <P31 <P30 <P29 <P28 <P27 <P26 <P25 <P24 <P23
P45^
J35 J36 J37 J38 J39 J40 J41 J42 J43 J44 J45
<P43 <P42 <P41 <P40 <P39 <P38 <P37 <P36 <P35 <P34
P44^
J46 J92
<P91
P89v
J47 J48 J49 J50 J51 J52 J53 J54 J55 J56 J57
<P55 <P54 <P53 <P52 <P51 <P50 <P49 <P48 <P47 <P46
P56v
J58 J59 J60 J61 J62 J63 J64 J65 J66 J67 J68
<P66 <P65 <P64 <P63 <P62 <P61 <P60 <P59 <P58 <P57
P77v
J69 J70 J71 J72 J73 J74 J75 J76 J77 J78 J79
<P76 <P75 <P74 <P73 <P72 <P71 <P70 <P69 <P68 <P67
P88v
J80 J81 J82 J83 J84 J85 J86 J87 J88 J89 J90
<P87 <P86 <P85 <P84 <P83 <P82 <P81 <P80 <P79 <P78
P90v
J91
Figure 10. AFT Arrow model of fuel manifolding system for a 140 psi Rolls Royce
Avon gas turbine.
J34 J13
P40>
P22^
J24 J20 J15
P28> <P20
P27> <P26 P18> <P17
J22 J23 J21 J11 J4 J12
<P24 P25> <P15 P16>
Figure 11. AFT Arrow model of 500 psi feedwater heater steam vent system.