IPC2000-257
MODELING OF BURIED NATURAL GAS PIPELINE DECOMPRESSION
X. L. Zhou
Greenpipe industries Ltd.
1600, 715-5th Ave. S.W. Calgary, AB T2P 2X6
R. G. Moore
University of Calgary
2500 University Dr. N.W. Calgary, AB T2N 1N4
ABSTRACT
A pseudo-homogeneous multi-component equilibrium
model is developed to study fluid flow behavior and heat
transfer during decompression of a buried natural gas pipeline.
The model includes non-isentropic effects of heat transfer and
viscous dissipation. The heat transfer involves transient heat
conduction through the pipe wall and the soil. A numerical
scheme is developed to efficiently solve the resulting
conservation equations simultaneously with the heat transfer
equation. The study shows that the inclusion of transient heat
transfer is necessary for accurate modeling the decompression
of natural gas pipelines. The temperature reduction on the pipe
wall due to fluid expansion can be crucial for the proper design
of pipelines to prevent brittle behavior of the pipe steel.
INTRODUCTION
Gas pipeline decompression behavior is a subject that is
particularly interested in the petroleum and chemical process
industries. The risk assessment of consequences of an
accidental release of flammable, toxic or volatile fluid due to
pipeline breaks or ruptures requires an accurate prediction of
mass efflux and thermal-hydrodynamic behaviors in the
pipeline. Most gas transmission pipelines in service today are
buried below ground. Although many precautions are taken in
the operation of pipelines, failures can and do occur. Most
common causes of pipeline failure are accidental damage,
ground movement, or corrosion.
Numerous incidents have been reported that have resulted
in loss of life and extensive property damage. In 1976 a road
Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
G. G. King
Greenpipe Industries Ltd.
1600, 715-5th Ave. S.W. Calgary, AB T2P 2X6
grader ruptured a 20-inch transmission pipeline in Cartwright,
Louisiana; natural gas escaped and ignited within seconds. In
1980 a pipeline transporting naphtha through Long Beach,
California, ruptured as a result of corrosion and overpressure.
Escaping fuel under high pressure blew a hole through the
paved street over the line and sprayed 6 m into the air. When it
ignited, the naphtha caused a fire that reached 22 m high. In
1985 natural gas at a pressure of 6.8 MPa ruptured a 30-inch
transmission line in Beaumont, Kentucky. The rupture, caused
by corrosion, tore out 9 m of the pipe; blew apart 5 m of a 36-
inch casing pipe that surrounded the carrier; blasted an opening
across Kentucky Sate Highway 60; and cut out a crater 30 m
long, 12 m wide'341. Recently in February 1999, a pipeline was
ruptured and exploded near Rocky Mountain House, Alberta.
The explosion sent up a mushroom cloud that was visible for
100 km. About 100 m of pipe at the site of the rupture had to be
replaced. The cause of incident was determined to be stress
corrosion cracking [1I] .
Mathematical models are useful in understanding the
characteristics of the rapid transient flow that occurs when
pipelines rupture. Numerical simulation can provide detailed
data for analyzing the consequences of pipeline bursts and
subsequent mechanical performance as they decompress. There
is extensive literature on pipeline decompression and various
types of experimentation. The majority of published work is
related to the design of nuclear powered steam generators
motivated by the analysis of loss of coolant accident (LOCA)
scenario. The focus has been on single component
(water/steam) systems during blowdown of either vessels or
relatively short lengths of pipeline ( M H o m ^ m . ^ m
Although the fundamental conservation equations are the same,
Copyright 2000 by ASME
 the constitutive closures can be very different between these
models for steam and those applicable to decompression of
long petroleum pipelines. Thus, it is insufficient to directly
apply the steam solutions to facilities involved in the
production and transportation of petroleum fluids, where multi-
component hydrocarbon mixtures and high ratio of length to
diameter are common. Nevertheless, such fundamental research
has laid the path for applications in the petroleum and chemical
process industries.
Relatively less attention has been given to studies that are
relevant to pipeline decompression using petroleum fluids.
Over last 25 years, only a few published papers have addressed
specifically the problems of decompression in long pipeline
carrying petroleum fluids. The proposed theoretical models
range from a simple semi-empirical model[33] through the
homogeneous equilibrium model (HEM) M t ' ^ ' W P O M C T
[28][32][31] t 0 t h e two.fluid m odel [3Flf241 . All models consider
pipeline decompression as one-dimensional flow. The
empirical model is generally considered to be accurate for
specific experimental setups that have similar parameters
throughout the system. The HEM is a step more advanced and
generalized for fluid flow in pipelines where thermodynamic
equilibrium is approximately achieved. There are many
successes in applying this model to pipeline decompressions of
high-pressure natural gas system where dispersed flow results
when condensation occurs. The two-fluid model is the most
general flow model for two-phase flow in pipes. It can include
mathematical representations for various decompression
phenomena that occur during the decompression process, such
as thermal-hydrodynamic non-equilibrium, phase slip, different
phase pressure and temperature etc. The major difficulty for
this type of model is the mathematical closure of the model.
All models mentioned above do not couple the hydraulic
simulation with the transient heat transfer from surrounding.
Many of them assume constant wall temperature to simulate the
heat conduction. The duration of typical long pipeline
decompression is in an order of hours. The thermal capacitance
of pipe and heat transfer from surrounding can have appreciable
effect on the decompression behavior of hydraulic simulation.
In this paper, a transient pseudo-homogeneous multi-
component equilibrium model is developed to study flow
behavior coupled with heat transfer for a buried natural gas
pipeline during decompression. A numerical scheme is
developed that efficiently solves the resulting conservation
equations simultaneously.
DECOMPRESSION PROCESS
Depending on initial conditions, various characteristics can
be observed during the decompression process. Figure 1
illustrates the characteristics of decompression wave
propagation. When a rupture occurs, the fluid escapes from the
Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
pipe and accelerates the flow to the local sonic velocity, at
which the first decompression wave propagates through the
undisturbed fluid. A series of pressure waves forms inside the
pipe and travel upstream from the break location. The actual
wave propagation velocity is the local sonic velocity minus the
fluid flow velocity. If the pressure in the pipe is high enough to
maintain the mass discharge, the flow will be choked at the
rupture location with a steep pressure decrease. The choked
flow prevents pressure waves generated downstream of the
pipeline break from travelling upstream inside the pipeline.
Initially, the viscous dissipation and wall heat transfer do
not have enough time to take effect. As the decompression
wave propagates along the pipeline, it causes the fluid to
experience an incremental expansion resulting in a slight lower
pressure and temperature. The pressure at closed end is
unchanged until the first wave arrives. The decreasing pressure
and temperature rapidly reaches the saturation conditions of the
fluid, which is given as the saturation characteristic line
(Figure 1). Owing to the non-equilibrium effects during rapid
decompression, the equilibrium condensation of fluid may not
take place at this line. Further expansion results in a super-
saturation state of the fluid, which reaches the maximum along
the particle path at the condensation front line. Time ti and t2
corresponding to location x, and x 2 represent the degree of fluid
supersaturation of the decompression system. The positive
slope of the particle path indicates that the fluid flows in a
direction opposite to that of decompression waves. These
waves result in expansions of fluid and produce a continuous
decrease of local sonic velocity and increase of fluid flow
velocity.
The decompression then enters into a quasi-steady state
flow, which, for a long pipeline, is the primary period of mass
discharge. This period has the greatest impact and practical
importance on the consequence analysis of the decompression
process. The flow is still choked at the open end of die pipe.
Two-phase flow starts at the choked plane and quickly reaches
the whole pipeline along the condensation front line (Figure 1).
The local sonic velocity of fluid decreases significantly when
the mixture crosses the phase boundary. The cooling effect of
fluid expansion results in large temperature difference between
the pipe wall and flowing fluid leading to intensive heat
transfer on the interface, which lowers the temperature of the
pipe wall. For a buried pipeline, the surrounding soil acts as a
protective layer that can also affect the heat transfer behavior
appreciably. As the cooling rate reduces due to decreasing
pressure, the effects of heat generation, heat transfers from
surrounding and capacitance of pipe steel gradually balance and
then dominate the process, resulting in slowly recovery of the
fluid temperature.
When the pressure at the pipe exit drops to the atmospheric
pressure, the fluid flow from die pipe ceases to be choked. Only
a small amount of fluid remains in the pipe at this time. The
viscous dissipation dominates the rest of the mass discharging
 process. The fluid exit pressure remains constant, which close the entire system. In this paper, the BWRS equation of
becomes a boundary condition for the hydraulic flow. The fluid state is employed.
flow velocity gradually reduces to zero to end the
decompression process. The closure of the model requires the evaluation of sonic
velocity, friction factor and mixture properties etc. The viscous
dissipation is given as [8]:
MODEL FORMULATION dp
f = 2i /f
P v Iv I (6)
For a long natural gas pipeline, the decompression duration dx d
is in an order of hours. The effect of heat transfer from the
surrounding soil is significant due to the large temperature For a two-phase system, the friction factor f is calculated
difference between the fluid and the pipe wall, thus, must be using the Colebrook correlation [I3] with the mixture fluid
taken into account for a complete description of this process. In properties. The heat conduction through the pipe wall into the
this paper, the hydrodynamic model is coupled with the heat fluid system is:
transfer model as described below.
Q= 4-{TP-T) (7)
Decompression Flow Model
A long pipeline has a large length to diameter ratio. It is It should be noted that all thermodynamic properties,
reasonable to assume the hydraulic flow to be one-dimensional. including mass density and enthalpy, used in the model
For a natural gas system, the potential liquid fraction is small formulation are the homogeneous mixture properties when the
and the flow velocity is high. The flowing fluid is considered at fluid system is in the two-phase region.
thermodynamic equilibrium for each phase and mixture,
implying the fluid system is single-phase flow or two-phase Heat Transfer Model
dispersed flow with no phase slip and equal pressure and For the decompression of a long pipeline, the time scale is
temperature for each phase. For an equilibrium system, the in an order of hours. The heat transfer will be significant
two-phase interfacial mass transfer is considered instantaneous. enough to affect the decompression behavior of the fluid in the
The overall mixture composition is assumed to be constant pipe. Due to the rapid flowing fluid during decompression, the
along the pipeline. The conservation equations can be given as: convection dominates the process. The temperature gradient in
Mass balance: axial direction is much smaller than that in the radial direction,
thus, the thermal gradients and consequently heat transfer
Dp d\ . within the pipe in axial direction is negligible compared to the
Z- + P = 0
Dt ^ dx
(1) radial gradients.
Figure 2 shows the cross-sectional two-dimensional heat
Momentum balance: transfer domain of a buried pipeline. By using the steady state
shape factor1301, the domain can be mapped into an equivalent
D\ dP dp one-dimensional heat transfer problem. The steady state heat
p + + f Jpgsm0=O (2)
Dt ox ox transfer in the radial direction of a pipe can be considered as an
infinite horizontal circular channel in semi-infinite solid, which
Energy balance: is given as:
q = S k(f -Ta) (8)
Dh DP _ dpf
p O-v+ vpg sin# = 0 (3)
Dt Dt dx where the shape factor S is:

and an equation of state: 2M

S =- (9)
P=f ( p j ) (4) cosh" ^

and the thermodynamic relation:

For a circular pipe in a cylindrical domain, the shape factor is:
h= f(P,T) (5)
In
The set of conservation equations (Eq.1-3) has 4 S = D> d (10)
-1
unknowns with 3 equations. The equation of state (Eq.4) cosh
provides the 4th equation, but it also introduces a 5th unknown, 2dD
T. The thermodynamic relation (Eq.5) is the 5th equation to

Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 The burial depth of a pipeline can then be mapped into the For one-dimensional heat transfer at the radial direction,
cylindrical soil domain by equating shape factors, which gives: the third kind boundary condition is considered on the internal
interface between the fluid and pipe, which is given as:
D = 2y + ^4y2-d2 (11)
Consequently, the solution is transformed into a one-
= U(TS-T) (17)
' dr
dimensional heat transfer problem. The validity of this process
will be demonstrated in the results of this simulation for buried For two-phase flow system, the heat transfer coefficient U
pipes. Thus, the heat conduction surrounding the pipe can be should account for two-phase flow. In this work, the mixture
expressed as one-dimensional in the radial direction: properties are used to calculate U. For the external boundary, a
constant temperature is used:
8TS _ 1 8 dZ
Psci kr (12)
dt r dr dr s
\r=Rr = *
Tn (18)

Thus, the pipeline and surrounding soil can be divided into a

number of isothermal slices that are coupled with the
NUMERICAL SOLUTION TECHNIQUE
hydrodynamic model.
Numerous numerical methods can be found in
Boundary Conditions
computational fluid dynamics. Methods based on
For hydrodynamic flow model, the boundary conditions characteristics are useful when the systems are hyperbolic and
can be determined from the characteristic lines for the set of the compatibility and characteristic equations can be derived
conservation equations. There are 3 real eigenvalues that can be easily. The pros and cons of various numerical methods can be
obtained at choking plane[231[271: found in many textbooks[1][6]. In this work, the semi-implicit
finite difference method is used to solve the coupling set of
4,2,3 = v , v a (13) hydrodynamic and heat transfer equations (Eq.1-5, and 12). In
this method, the primary variable in each equation is treated
The method of characteristics indicates that at a boundary, implicitly. Other variables in the equation are treated explicitly.
where one characteristic propagates into the domain, then the With the central difference for the numerical discretization, a
value of one dependent flow-field variable must be specified at tri-band matrix can be formulated for each equation in the
that boundary. If one characteristic line propagates out of the following form:
domain, the value of another dependent flow-field variable
must be allowed to float at the boundary. The streamline plays A* u = S (19)
the same role as the characteristic direction1-11. Thus, at the exit
where
plane, one boundary condition has to be specified. When the
exit pressure at x=L is higher than the external pressure, the "C, A 0 0 0 "I
fluid flow will be choked. The sonic velocity is used as a B2 c2 0 0 "2 s2
boundary condition and the other variables allow floating. A= 0 0 ,s = (20)
When the exit pressure is equal to the exterior pressure, the
0 0 A,-, u
n-l S.-1
choking condition no longer holds and the condition is replaced
by a constant exit pressure (sub-sonic flow): 0 0 0 "n Sn

V which can be solved efficiently using Thomas algorithm[S].

[ L=* (Pexit>Pa) (14)

RESULTS AND DISCUSSION

and the closed end at x=0:
v (15)
u=0
At the pipe wall, the heat addition to the fluid system is
determined from Eq.(7). The sonic velocity is calculated along
the isentropic thermodynamic path using numerical
differentiation due to two-phase flow.
2 Ap
a = (16)
Ap
Validation
The model was validated using pipe burst test data[25]. The
pipe outside diameter is 1219 mm with wall-thickness 13.71
mm. The pipe was buried below surface estimated to be 1 m.
The test section was 57.72 m long. Initial natural gas
temperature was measured at -5C, and pressure at 8143 kPa.
The test gas contained primarily methane (85.71%), ethane
(7.94%) and propane (4.27%). Pressure transducers were
placed along the pipeline. The experimental results show clear

Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 two-phase discontinuity in the pressure range 6,000-6,140 kPa.
The time for such a short length of pipe to decompress is less
than one second. The heat transfer was not measured during the
experiment.
The calculation was made using 100 spatial grid points
along the pipeline. An equilibrium fluid property table was
generated using the BWRS equation of state. The property
tables include quantities such as mass density, sonic velocity,
Cp, Cv, enthalpy as functions of pressure and temperature. The
total calculation time for this system was of an order of
minutes.
Figure 3 shows the predicted pressure profile in
comparison with the experimental data. Good agreement is
evident including the two-phase effect at the pressure around
6,000 kPa. The discrepancy arises at the later time probably due
to the assumption of full-bore break without fracture
propagation, while the experimental ductile fracture has a
different configuration of exit geometry.
Long Pipeline Simulation Result
The model validated above is used to predict the
decompression behavior of a long buried pipeline. The
simulation is performed on the same kind of pipe as in the field
test except the length extended to 35 km long. The initial
pressure is 20,786 kPa and temperature 0C. 700 grid blocks
are used for the simulation. Figure 4 shows the pressure
profiles at different time during the decompression. It clearly
exhibits the choked effect and pressure wave reflection from
the intact end. Figure 5 depicts the fluid velocity profiles at the
corresponding times. The fluid exit pressure and temperature
are plotted in Figure 6 and velocity in Figure 7. Within seconds
after starting the decompression, the fluid pressure drops to
approximately 5,000 kPa and temperature reduces to -60C.
The fluid flow is choked at a velocity between 230 m/s and 300
m/s. The continuous expansion of fluid decreases the pressure,
and the temperature is cooled down and approaches -100C.
At time 1,500 s, the pressure lowers to about 500 kPa and the
cooling rate of expansion fluid is reduced, the viscous heat
generation and heat transfer effects begin to dominate the
process. This is indicated by the recovering of the exit fluid
temperature. The fluid flow remains choked until the
simulation time reaches about 2,500 s. At this time, the flow
becomes sub sonic with continuing decrease of flow velocity
and increase rate of fluid temperature recovery at a constant
exterior pressure.
Figure 8 and 9 present the radial temperature profiles in the
pipe and the surrounding soil at the mid point and exit point of
the pipeline. The first 2 points of linear profile in each line are
temperatures within the pipe wall. The temperature gradient
within the pipe degrades over times from sharp increase at 1 s
to flat at 3,000 s. The thermal penetration to the surrounding
soil is approximately 20 cm. This suggests that the boundary
condition at r=Ro in Eq. (18) is valid and has no appreciable
Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use
impact on the heat transfer calculation. This also implies that
the domain transfonnation in Eq. (11) is valid. Heat flow from
the ground surface to pipe during this period is therefore
negligible and the ground heat flux does not influence final
pipe wall temperature.
Assuming the average temperature drop of 50C for the
pipe and 10 cm thermal penetration to the surrounding soil
during the decompression, the amount of heat transfer to the
gas is well over 2xl0 9 kJ. Thus, The inclusion of transient heat
transfer is necessary for to accurately model the decompression
of natural gas pipelines. The temperature reduction on the pipe
wall due to fluid expansion can be crucial for die proper design
of pipeline to prevent brittle behavior of the pipe steel.
CONCLUSIONS
A decompression flow model is formulated and coupled
with a one-dimensional heat transfer model in the surrounding
soil. The thermal capacitance of the pipe steel and soil are
accounted for in the simulation. The flow model is validated
with a pipeline burst test data that results in good agreements
(Figure 3).
For a long pipeline, the decompression can last over an
hour. The effect of heat transfer is significant and should be
considered in the modeling process (Figure 6).
The thermal penetration to cool down the surrounding soil
is less than 20 cm during the decompression (Figure 9). The
effect of die surface heat flux is negligible for a buried pipeline
during the decompression.
ACKNOWLEDGEMENTS
The authors acknowledge the support from Greenpipe
Industries Ltd. during the course of this work.
NOMENCLATURE
a sonic velocity (m/s)
A tri-band matrix
cp heat capacity (J/kg.K)
d pipe inside diameter (m)
D soil domain external diameter (m)
f Fanning friction factor
g gravity (m/s2)
h enthalpy (J/kg)
k thermal conductivity (W/m.K)
Pf friction pressure (Pa)
P pressure (Pa)
Pa external pressure (Pa)
Pexit fluid exit pressure (Pa)
Q rate of heat transfer per unit volume of fluid (J/m3.s.K)
 r radial coordinate 11. Ewart, S.; "Stress Corrosion: Cause of Blast"; Calgary
r0 pipe inside radius (m) Herald, Feb. 19,1999
R soil domain radius (m) 12. Fairuzov, Y.V.; "Blowdown of Pipelines Carrying Flashing
S shape factor Liquids"; AIChEJ, Vol.44, No.2, Feb. 1998
S solution vector 13. Govier,G.W. and Aziz, K.; "The Flow of Complex
t time (s) Mixtures in Pipes"; Rrieger Pub., Florida, 1982
T fluid temperature (K) 14. Groves, T.K., Bishnoi, P.R. and Wallbridge, J.M.E.;
Ta ground surface temperature (K) "Decompression Wave Velocities in Natural Gases in Pipe
Tp pipe inside temperature (K) Lines"; Can. J of Chem. Eng., Vol.56, Dec. 1978
Ts soil temperature (K) 15. Hancox, W.T., Ferch, R.L., Liu, W.S. and Nieman, R.E.;
u variable vector "One-dimensional Models for Transient Gas-Liquid Flows
U pipe-fluid interface heat transfer coefficient (W/m2.K) in Ducts"; Int. J of Multiphase Flow, Vol. 6, No.l, 1980
v velocity (m/s) 16. Henry, E.E. and Fauske, H.K.; "The Two-Phase Critical
x coordinate Flow of One-Component Mixtures in Nozzles, Orifices,
y depth of cover (m) and Short Tubes"; J of Heat Transfer, Trans, of ASME,
May 1971
Greek symbol 17. Hopke, S.; "Interaction between Gas Decompression and
p fluid mass density (kg/m3) Propagating Ductile Fractures in Gas Pipelines"; Technical
ps soil mass density (kg/m3) Service Report, EPR.29PS. 75, Exxon Prod. Res. Co., Mar.
9 pipeline inclination () 1975
X eigenvalues 18. Ishii, M.; "Thermo-Fluid Dynamic Theory of Two-Phase
Flow"; Eyrolles, 1975
19. Jones, O.C. Jr.; "Toward a Unified Approach for Thermal
Non-equilibrium in Gas-Liquid Systems"; Nuc. Eng. &
REFERENCES Des., Vol.69,1982
1. Anderson, J.D. Jr. ; "Computational Fluid Dynamics: The 20. King, G.G.; "Decompression of Gas Pipelines During
Basics with Applications"; McGraw-Hill, Inc, 1995 Longitudinal Ductile Fractures"; J. of Energy Resources
2. Ardron, K.H. and Furness, R.A.; "A Study of the Critical Tech., Trans, of the ASME, Vol.101,1979
Flow Models Used in Reactor Blowdown Analysis"; blue. 21. King, G.G.; "Geothermal Design of Buried Pipelines";
Eng. & Des., Vol.39,1976 Presented at Energy Tech. Conf & Exhib., Pet. Div. of
3. Bendicksen, K.H, Maines, D, Moe, R. and Nuland, S.; ASME, New Orleans, La., Feb. 3-7,1980
"The Dynamic Two-Fluid Model OLGA: Theory and 22. Latrobe, A. and Grand, D.; "On Some Difficulties in
Application"; SPE19451, SPEProd. Eng., May 1991 Numerical Modeling of Two-Phase Flows"; Proc. of CSNI
4. Botros, K.K., Jungowski, W.M. and Weiss, M.H.; "Models Specialists Meeting, Toronto, Vol.1, Aug. 3-4,1976
and Methods of Simulating Gas Pipeline Blowdown"; Can. 23. Lyczkowski, R.W., Gidaspow, D., Solbrig, C.W. and
JChem. Eng., Vol.67, Aug. 1989 Hughes, E.D.; "Characteristics and Stability Analyses of
5. Boure, J.A.; "Mathematical Modeling of Two-Phase Transient One-dimensional Two-Phase Flow Equations
Flows, A Review of Its Bases and Problems"; Proc. of the and Their Finite Difference Approximations"; Nuc. Sci. &
CSNI Specialists Meeting, Toronto, Vol.1, Aug. 3-4,1976 Eng., Vol.66,1978
6. Burden, R.L. and Faires, J.D.; "Numerical Analysis, Fourth 24. Mucharam, L. and Adewumi, M.A.; "Study of Continuity
Edition"; PWS-KENT, Boston, 1989 and Pressure Waves in Condensing Two-Phase Flow in a
7. Chen, J.R., Richardson, S.M. and Saville, G.; "Modeling of Pipeline"; AIChE Symp. Series, Heat Transfer, Vol.87,
Two-Phase Blowdown from Pipelines, II. A Simplified No.283, Minneapolis 1991
Numerical Method for Multi-Component Mixtures"; 25. National Energy Board; "Final Report on the Test Program
Chem. Eng. Set, Vol.50, No.13,1995 at the Northern Alberta Burst Test Facility", Public Doc.,
8. Davies, J.T.; "Turbulence Phenomena, An Introduction to National Energy Board, GH-3-97, Aug 10, 1981
the Eddy Transfer of Momentum, Mass, and Heat, 26. Norris, H.L. and Puis, R.C.; "Single-Phase or Multiphase
Particularly at Interfaces"; Academic Press, New York, Blowdown of Vessels or Pipelines"; SPE 26565,1993
1972 27. Picard, D.J and Bishnoi, P.R.; "The Importance of Real-
9. Dobran, F.; "Non-equilibrium Modeling of Two-Phase Fluid Behavior and Non-isentropic Effects in Modeling
Critical Flows in Tubes"; J of Heat Transfer, Trans, of Decompression Characteristics of Pipeline Fluids for
ASME, Vol.109, Aug. 1987 Application in Ductile Fracture Propagation Analysis";
10. Edwards, A.R. and O'Brien T.P.; "Studies of Phenomena Can. J of Chem. Eng., Vol.66, Feb. 1988
Connected with the Depressurization of Water Reactor"; J.
of Brit. Nuc. En. Soc., Vol.9, 1970

Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 28. Richardson, S.M. and Saville, G.; "Blowdown of
Pipelines"; SPE 23070,1991
29. Richter, H.J.; "Separated Two-Phase Flow Model:
Application to Critical Two-Phase Flow"; Int. J of
Multiphase Flow, Vol.9, No.5, 1983
30. Rohsenow, W.M., Hartnett, J.P. and Ganic, E.N.;
"Handbook of Heat Transfer Fundamentals, 2nd Ed.";
McGraw-Hill, 1985
31. Sand, I.O., Sjoen, K. and Bakke, J.R.; "Modeling of
Release of Gas from High-Pressure Pipelines"; Int. J for
Numerical Methods in Fluids, Vol.23,1996
32. Seynhaeve, J.M., Lombre, R., Ducrocq, J. and Bolle, L.;
"Physical Modeling of Rapid Transients in Long Pipes, in
Case of Vaporization: An Efficient Tool for Safety
Management"; Process Safety Progress, Vol.13, No.2,
Apr. 1994
33. Tam, V.H.Y. and Higgins, R.B.; "Simple Transient
Release Rate Models for Releases of Pressurized Liquid
Petroleum Gas from Pipelines"; J of Hazardous Materials,
Vol.25, 1990
34. TRB (Transportation Research Board); "Pipelines and
Public Safety", Trans. Res. Board, National Research
Council, Special Report 219,1988
35. Wallis, G.; "Critical Two-Phase Flow"; Int. J of
Multiphase Flow, Vol.6, 1980

Fix Location Pressure Profiles

Condensation Front

Test@14m
Test@18im
Model @14m
P(X3,t)
Model @l&5m

Distance X

Figure 1: Characteristics of Pipeline Decompression Figure 3: Pressure Profile Comparison

0 5000 10000 15000 20000 25000 30000 35000

Length (m)

Figure 2: Heat Transfer Domain Transformation Figure 4: Decompression Pressure Profiles of Long
Pipeline

Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 u
8- -40

_ _ _ v//

I!
-1000
-2000
1000 -2500
-20(10 -3000
-400)

0 5000 10000 15000 20000 25000 30000 35000 0 58 0.68

Length (m) Radius (m)

Figure 5: Decompression Velocity Profiles of Long Figure 8: Soil Temperature Profiles at Mid Point of Long
Pipeline Pipeline during Decompression

20000

10000

I "l
0 500 1000 1500 2000 2500 3000 3500 4000
Time (s)

Figure 6: Exit Pressure and Temperature Profiles of Long Figure 9: Soil Temperature Profiles at Exit Point of Long
Pipeline during Decompression Pipeline during Decompression

Figure 7: Exit Velocity Profile of Long Pipeline during

Decompression

Downloaded From: https://proceedings.asmedigitalcollection.asme.org/ on 10/25/2017 Terms of Use: http://www.asme.org/about-asme/terms-of-use