Documente Academic
Documente Profesional
Documente Cultură
Zhenhua Rui, Quoqing Han, He Zhang, Sai Wang, Hui Pu, Kegang Ling
PII: S1875-5100(17)30349-9
DOI: 10.1016/j.jngse.2017.08.025
Reference: JNGSE 2293
Please cite this article as: Rui, Z., Han, Q., Zhang, H., Wang, S., Pu, H., Ling, K., A new model to
evaluate two leak points in a gas pipeline, Journal of Natural Gas Science & Engineering (2017), doi:
10.1016/j.jngse.2017.08.025.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Zhenhua Rui, Independent Project Analysis, Inc., Ashburn, Virginia, United States
He Zhang, University of North Dakota, Grand Forks, North Dakota, United States
PT
Sai Wang, University of North Dakota, Grand Forks, North Dakota, United States
Hui Pu, University of North Dakota, Grand Forks, North Dakota, United States
RI
Kegang Ling*, University of North Dakota, Grand Forks, North Dakota, United States
SC
*Corresponding Author: Kegang Ling
U
Keywords: Gas Pipeline, Leak Detection, Pipeline Flow Modeling
AN
Abstract
M
Natural gas is a clean fossil energy and an important sector in the energy consumption chart. Because of its reliability
D
and low carbon-hydrogen ratio, the demand for natural gas increases steadily to replace coal and wood to better protect
TE
environment. To accommodate the ever rising in natural gas production and transportation, more gas pipelines are
being constructed. Meanwhile, the existing gas pipelines are aging inevitably. One of the critical needs in natural gas
flow assurance is detecting and locating pipeline leak in a timely manner. A reliable and timely detection of the
EP
leakage of gas pipeline can not only reduce the loss of hydrocarbon, but also limit the damage to facilities, possible
loss of life, and the extent of environmental pollution. Two or more leakage points in a pipeline were observed in the
C
field. Physical methods and mathematical models were employed to detect pipeline leakages. However, literature
AC
review indicates that no mathematical model has been developed to detect multiple leaks in the same pipeline. This
study focused on the detection of two leak points in a pipeline. Multi-flowrate tests are proposed to evaluate the
locations and sizes of leakages in two leak points. The new mathematical model is crucial when no physical inspection
is available. The proposed model can be used to monitor possible leak in real-time because flowrate and pressure that
are utilized to estimate multiple leaks are monitored in real-time and are available almost simultaneously. Therefore,
the new method provides a practical, quick, and low computational cost approach to detect multiple leakages. The
1
ACCEPTED MANUSCRIPT
proposed method is important because existing mathematical models assumed single leak in a pipeline, which limits
their applications because the detection will be misleading if there is more than one leakage in the same pipeline. The
proposed model can differentiate single-leakage scenario from multi-leakage scenario based on multi-rate tests. The
identification is critical because it guides the leakage detection to the right direction.
PT
Introduction
RI
Natural gas supplies almost one-fourth of all energy used in the world (Rui et al, 2017a). Due to the increasingly
environmental protection concern and natural gas’s efficiency, natural gas consumption increases steadily and expect
SC
to exceed the oil because more gas fields are discovered and produced comparing with oil fields (Rui et al, 2017c).
Anderson and Driscoll (2000) estimated that natural gas consumption will increase 50% within the next 20 years. More
U
gas pipelines are being constructed to transport natural gas from fields to gas plants and then to refinery and end users
AN
(Rui et al, 2017b). Leak detection is important in the hydraulic test of new pipeline to examine the pipeline integrity
before commissioning (Rui and Wang, 2013). The risk of leak occurred in the existing gas pipelines increases because
M
of the aging of pipeline material, corrosion, erosion, and pressurization caused by partial blockage. To reduce the
damage to the facility and life and limit the pollution to environment caused by pipeline gas leak, it is necessary to
D
develop an accurate method to detect the leak in a timely manner. Method to detect pipeline leak can be classified to
TE
two categories: physical inspection and mathematical model. The advantage of physical method is its accuracy and
reliability. Physical method has the limitations of production shut-down, high cost/long time to run detection, and
EP
difficulty operation in harsh environment. The mathematical simulation can evaluate leakage quickly without
interference with pipeline operations but with high uncertainty due to the fact that real conditions deviate from the
C
assumptions in the mathematical model and/or inaccurate measurement of flow parameters such as pressure, flowrate,
and temperature. Usually, single leak point in a gas pipeline was commonly found in the pipeline operation. Two and
AC
more leakage points in a pipeline have been observed in the field although their percentages are low. The physical
methods have been developed to detect single leak and multiple leaks in a pipeline. Different physical methods that
have been applied in the field were discussed by Van der Marel and Sluyter (1984), Erickson and Twaite (1996),
Sadovnychiy et al. (2005), Elliott et al. (2008), Naranjo and Neethling (2011), Eisler (2011), Rui, et al. (2011a, 2011b,
2012a, 2012b), and Vralstad et al. (2011). The advantages and disadvantages of these technologies were also
discussed. Many mathematical models have been proposed to detect single leak in a pipeline. Huber (1981) detected
2
ACCEPTED MANUSCRIPT
possible hydrocarbon liquids such as ethylene, ethane, and propane in Cochin pipeline system using a computer-based
pipeline simulator. With simulation model, the pressure, temperature, density, and flow profiles of entire pipeline were
determined. Cesar (1986) linearized the partial differential equation and used numerical method to solve the linear
parabolic partial differential equation to model the gas transient flow in a pipe. Gas leak can be detected by the analysis
of transient flow behavior. Employing the principles of mass balance, momentum conservation, energy conservation,
PT
and equation of state (EOS). Massinon (1988) presented a real-time transient hydraulic model to detect leak in a liquid
pipeline system. Mactaggart (1989) utilized single volume balance transient model, and compensated volume balance
RI
to detect sour gas leak. He believed the combination of different methods ensures accurate modelling and acceptable
leak detection performance on gas as well as liquids pipelines. Scott et al (1998) applied back-pressure technique to
SC
evaluate leaks in deepwater gas flowline assuming fully rough flow. Zhou et al (2000) improves the accuracy of leak
detection model by including the kinetic energy term and solving the partial differential equation numerically. Geiger
U
(2001) described the basic principles and application of a model to monitor the pipeline operation for the presence of
AN
leaks. The model used measured flowrate, pressure, temperature to calculate the static and dynamic behavior of
pipelines by a computer online and in real time. The one-dimensional continuity and momentum equation is used in the
model. Bai et al. (2004) discussed the pipeline leak detection using a fluid transient model coupled with the extended
M
Kalman filter. The extended Kalman filter has a time variant track and is suit for the stochastic process with Gaussian
D
noise. The hydraulic transient model was lumped to a discrete-time/discrete-space by the characteristic method. In
Reddy’s model (2006), a fully nonlinear second-order accurate finite difference method was incorporated into
TE
simulation models for state estimation and leak detection. Wang and Carroll (2007) used a transient model to analyze
real-time data to detect leaks in gas and liquid pipelines. Gajbhiye and Kam (2008) detected leak in subsea pipeline
EP
under fixed pressure boundaries by a mechanism model. They utilized the change in inlet total flowrate and the change
in outlet total flowrate as two possible leak detection indicators to monitor possible pipeline leak in real-time. The
C
effect of different parameters on the mechanistic leak detection modeling is investigated accounting for gas
AC
compressibility, back pressure of the system, and the pressure drop across the system. Li et al. (2009) found that the
leak location error depends on friction factor from their research on pressure gradient. They proposed backpropagation
neural network to forecast friction factor. A water pipe with a length of 48.775m and a diameter of 53mm has been set
up in the hydraulics laboratory to verify their method. They claimed that their method is efficient to locate the leak in
long-distance pipeline. Hauge et al. (2009) indicated that given inlet/outlet velocity, pressure, and temperature leak
location and size can be estimated using an adaptive Luenberger-type estimator. Bustnes (2011) employed a
3
ACCEPTED MANUSCRIPT
commercial real-time transient model to detect leak in oil pipeline. The model required no prior knowledge of the
software calculation method. But the accuracy of leak detection is low and has high uncertainty under transient
operation conditions. Harriott (2011) applied gas dynamic principles to develop a leak detection system for long gas
pipeline of arbitrary connectivity that contain point resistances to flow. The leaks were identified from the differences
between simulated and measured flow. The simulated flow was obtained through solving the governing equations by a
PT
novel finite element method with boundary conditions. Civan and Balda (2013) located leak in liquid pipelines
considering the transient flow in pipeline. Mass balance was applied in their model. Michkova and Strelnikova (2013)
RI
presented a mathematical model to better describe the transient processes of fluid motion in a pipeline. The model
calculated pressure distribution along the pipeline using flowrate, density, viscosity, and pipeline diameter as input.
SC
The leak can be located based on the pressure drop surge. The model can minimize the false leak detection by
excluding "external" disturbances in steady-state pumping. Geiger and Vogt (2014) describes a new leak detection
U
methodology which uses pattern recognition techniques to combine two or more internal methods seamlessly into one
AN
scheme hence improving performance, robustness and applicability. Their approach is a generalization of the extended
real-time transient model. Although numerous mathematical models exist to evaluate pipeline leak, to the best of our
knowledge, no model is available to detect two or more leakage points in a pipeline. To fill this gap, we propose a new
M
model using multi-rate tests to identify two leakages in the same gas pipeline. By conducting multi-rate tests, the
D
locations and sizes of leakages can be evaluated. The new method makes the evaluation of multiple leak in the same
pipeline possible. It is crucial when physical inspection is unavailable. Furthermore, existing mathematical models
TE
assumed single leak in the pipeline, which limits their applications because the detection will be misleading if there are
more than one leakage in the same pipeline. The proposed model can identify single-leakage scenario from multi-
EP
leakage scenario based on multi-rate tests. The identification is critical because it guides the leak detection to the right
direction.
C
AC
Model Development
To detect leak, the gas pipeline flow equations is used to evaluate the flow parameters such as pressure and flow rate
4
ACCEPTED MANUSCRIPT
When the pipeline diameter is smaller than 16 inches, Weymouth equation (Weymouth, 1912) is used to calculate gas
flow rate in pipelines. If pipeline diameter is larger than 16 inches, Panhandle A and B equations are appropriate for
PT
q=
3.23Tsc (p 2
inlet − e s poutlet
2
D5 )
.. .................................................................................................................... (1)
fγ g T z Le
RI
psc
SC
where
U
(e s − 1) L
Le = .. ......................................................................................................................................................... (2)
AN
s
0.0375γ g ∆z
M
s= ....................................................................................................................................................... (3)
Tz
D
TE
where
Chen’s (1979) correlation can be used to calculate Fanning friction factor. Then friction factor can be obtained.
C
AC
where relative roughness, eD, is the ratio of the absolute roughness to the pipe internal diameter
5
ACCEPTED MANUSCRIPT
ε
eD = ....................................................................................................................................................................... (6)
D
and Reynolds number, NRe, can be expressed as
Duρ
N Re = .. ............................................................................................................................................................ (7)
µ
PT
Panhandle A and B equations are
RI
( )
SC
1.07881 0.5394
D 2.6182 T pinlet
2
− poutlet
2
q = 435.87 0.4604 b .. ....................................................................................... (8)
γg pb T zL
U
AN
and
( )
1.02 0.510
M
T pinlet
2
− poutlet
2
q = 737D 2.530 b ,....................................................................................................... (9)
T zLγ g
0.961
pb
D
TE
respectively.
EP
To locate the gas leak and evaluate the leak size for two leaks in the same pipeline, four flowrate tests are required to
obtain flow parameters to solve the governing equations. Following assumptions are made in the leak detection
analysis:
6
ACCEPTED MANUSCRIPT
• Flow rate and pressure are measured at the inlet and outlet of pipeline.
The gas leak rate, which is the sum of gas leak rates at leak points 1 and 2, is the difference between inlet and outlet
gas rates. In other words, although individual leak rate at each leak point is unknown, the total leak rate can be
calculated with measured inlet and outlet rates. Accordingly, the pressure at leak points 1 and 2 are unknown since leak
points are unknown, but the total pressure drop across the pipeline is the difference between inlet and outlet pressures.
PT
In this study, dimensionless analysis is introduced to locate leak point and estimate leak rate. In the dimensionless
RI
Dimensionless leak location: It is the ratio of distance between leak locale and pipeline inlet to pipeline length, which
is expressed as
SC
Lleak
Lleak , D =
U
.. ......................................................................................................................................................... (10)
L
AN
Accordingly, dimensionless leak locations 1 and 2 is defined as
M
Lleak 1
Lleak 1,D = ........................................................................................................................................................ (11)
D
L
TE
and
EP
Lleak 2
Lleak 2,D = .. ..................................................................................................................................................... (12)
L
C
AC
Dimensionless gas leak rate: It is the ratio of gas leak rate to gas rate at inlet of pipeline, which is expressed as
qleak
qleak , D = .. ......................................................................................................................................................... (13)
qinlet
7
ACCEPTED MANUSCRIPT
Dimensionless pressure drop: It is defined as the ratio of pressure drop through pipeline under gas leak conditions to
∆pleak
∆p D = .. ..................................................................................................................................................... (14)
∆p no leak
PT
The detection procedure is illustrated through following example. Data used in the example are listed in Table 1.
RI
During the multi-rate tests, pipeline inlet and outlet flow rates, pressures, and temperatures are measured. The
measured data are utilized to determine leak points and sizes by following steps below:
SC
1) Run four rate tests 1, 2, 3, and 4; record pipeline inlet and outlet temperatures, flow rates, and pressures
2) Calculate total leak rates (the difference between inlet and outlet rates) and pressure drops (the difference
U
between inlet and outlet pressures) for all four rate tests
AN
3) For all four rate tests, calculate pressure drop in pipeline assuming no leak in the pipeline using Weymouth or
Panhandle A and B Equations (Equation (1), (8) or (9)). It should be noted that the pressure drop without gas
leak is maximum compared with any gas leak case, then calculate dimensionless gas leak rates and
M
dimensionless pressure drops for all four rate tests using Equations (13) and (14)
D
4) Now for the first flowrate test, assume a leak rate of q1_leak 1 for leak point 1, then leak rate at leak point 2 is the
5) Calculate pressure drops that are corresponding to different leak locations of leak points 1 and 2 using Equation
(1), (8) or (9), also calculate dimensionless leak locations, dimensionless gas leak rates, and dimensionless
EP
pressure drops using Equations (10) through (14), plot type curves on 3-dimension with dimensionless leak
location 1 (x-axis) - dimensionless leak location 2 (y-axis) - dimensionless pressure drop (z-axis) by using
C
different leak locations of 1 and 2 as Fig. 1, which constructs type curve plane
AC
6) Connecting the intersection points between dimensionless pressure drop plane obtained in step 3 and type curve
8) Repeating steps 4 through 7 to obtain line C’D’ for the second rate test as shown in Fig. 2,
8
ACCEPTED MANUSCRIPT
9) Projecting the intersection point (point E) of lines A’B’ and C’D’ on y-axis gives point F, which is
dimensionless leak location 1; Projecting point E on x-axis gives point G, which is dimensionless leak location
2 as shown in Fig. 2
10) With points F and G, leak locations 1 and 2 can be calculated with Equations (11) and (12)
11) Using the third and fourth rate tests, repeat steps 1 through 10 to obtain another set of leak locations 1 and 2
PT
12) If one or two of leak locations obtained from step 10 is/are different from those obtained in step 11, use average
of leak location 1 obtained from steps 10 and 11 as the new leak location 1 and average of leak location 2
RI
obtained from steps 10 and 11 as the new leak location 2 to back calculate the leak rates for all four rate tests
SC
13) Use the leak rates calculated in step 12 to repeat steps 4 through 12 until the estimated two leak locations
converge
U
14) Two leak locations obtained in step 13 are the solution for leak locations
AN
15) Use leak locations obtained in step 13 to back calculate leak rates at two locations using Equation (1), (8) or (9).
One of the assumptions in the proposed method is that there are two leak points in a pipeline. In pipeline operation,
leak number is unknown before the detection is applied. Before mathematical model is used to detect leak, it is
TE
necessary to develop an approach to identify leak number in a pipeline. Many simulation models assumed single leak
point and will give wrong results if there are two or more leak points occur. To guide the leak detection to the right
EP
At least two rate tests are required to differentiate single leak from multiple leak. For each rate test, assuming single
leak scenario and plotting 3-dimension with dimensionless leak location 1 (x-axis) - dimensionless gas leak rate (y-axis)
- dimensionless pressure drop (z-axis) can determine a leak location (Ling et. al, 2015). For the case of single leak
point occurring in a pipeline, the fact is that only one leak location exists in the pipeline. Therefore, any flowrate test
should give the same leak location. If multi-rate tests give different leak locations, it should not be single leak in a
9
ACCEPTED MANUSCRIPT
pipeline. In other words, there are two or more leak points in the pipeline. If two or more rate tests result in a same leak
PT
Based on the aforementioned identification, if two or more leak points in the same pipeline exist, the procedure to
determine the leak location and size for two-leak scenario can be used to differentiate two-leak scenario from other
RI
leak scenarios. For the case of two leak points occurring in a pipeline, the fact is that there are two leak locations exist
in the pipeline. Therefore, any flowrate tests should give the same two leak locations. According to the two-leak
SC
location detection procedure discussed above, the four rate tests should give the same two leak locations at the end of
calculation, i.e., two leak locations obtained from the first and second rate tests should be the same as those obtained
U
from the third and fourth rate tests. This means the calculation in the aforementioned detection procedure should
AN
converge if there are two leaks take place, Otherwise, there are more than two leak points in the same pipeline.
To evaluate the accuracy of the proposed model, numerical simulation was used to detect leak locations and estimate
leak rates assuming two leak points occur in a pipeline. Input data are given in Table 2. Pressure and flow rate at the
TE
inlet and outlet of pipeline are measured to be used for the leak detection using both approaches. The locations of leak
points 1 and 2 detected by the proposed method are at 355513 ft and 555609 ft from the inlet of pipeline, respectively.
EP
The leak rates at leak points 1 and 2 calculated by the proposed method are 120006 ft3/hour and 79994 ft3/hour,
respectively. Meanwhile, the locations of leak points 1 and 2 detected by the numerical simulation are at 355497 ft and
C
555596 ft from the inlet of pipeline, respectively. The leak rates at leak points 1 and 2 calculated by the numerical
AC
simulation are 120023 ft3/hour and 79977 ft3/hour, respectively. The comparison indicates that leak detection by the
proposed method is close to that by numerical simulation (Table 3). This suggests that the proposed method is reliable.
The difference is due to the different gas properties caused by average temperature and pressure. Therefore, the
proposed method can narrow down the possible pipeline interval involving gas leak that will be examined by physical
inspection.
10
ACCEPTED MANUSCRIPT
The proposed methods require accurate measurement of temperature, pressure, and rate. Commonly, A high gas leak
rate is easier to be detected comparing with a low gas leak case. The noises always exist in the flow parameter
measurement. The resolutions of instruments are critical to detection capability. It is important to detect small leak in
PT
early stage while avoiding false alarm. Although high resolution pressure gauge and metering equipment that can
provide high quality data are expensive, the total cost is relatively low comparing to that of the physical inspection
RI
method and/or loss of hydrocarbon, damage to facilities, possible loss of life, and environmental pollution. One should
note that in the real world there is no perfect pipeline with uniform geometry and uniform gas properties throughout a
SC
long gas pipeline. Change of pipeline geometry resulting from deformation, scaling, solid deposition, chemical
precipitation, corrosion, and erosion affect the accuracy of the detection. Inaccurate measurements of temperature,
U
pressure, and flow rate, inaccurate estimation of gas properties, and possible liquid condensed in pipeline introduce
AN
errors into the leak estimation Usually, the distributions of these errors are randomly distributed. Therefore, multi-rate
test can be employed to eliminate the randomly distributed errors (or noises) and amplify the leak points (or signals)
because the leak locations remain the same for different flow rate tests. Statistically, the more number of multi-rate test
M
is, the lower the uncertainty is in leak detection considering the fact that leak locations are fixed and results from
D
The proposed method can detect and evaluate the two leak points in a pipeline. However, pipeline networks in
operation can be very complicated and the leak point can be one, two, or more in the same pipeline or different
EP
pipelines within the systems. Future work should focus on expanding the application of proposed method to more
complicated scenarios such as multiple leaks in same or different pipelines in pipeline networks. A gas transportation
C
network can be decomposed into numerous single-pipeline and parallel-pipeline sharing junctions and/or nodes. The
AC
analysis of the leak in the basic units is a critical step to determine which pipeline/s contain leak/s. Another direction of
future work should be the validation of model by experiment or field data. Through the experimental tests, one can
evaluate the effects of inaccurate measurements of temperature, pressure, and flow rate, change of pipeline inner
diameter due to scaling, corrosion, and erosion, liquid condensed in pipeline, and inaccurate estimation of gas
11
ACCEPTED MANUSCRIPT
Conclusions
• A new method was presented to locate two leak points and estimate leak sizes in a pipeline. It fills the gap in the
PT
mathematical models because existing models assumed single leak point in a pipeline.
• The procedure to differentiate two-leak scenario from single-leak or more-than-two-leak scenario in a pipeline
RI
was proposed and is invaluable to avoid mis-detection. The identification is critical because it guides the
SC
• To reduce the leak detection uncertainty, more flowrate tests should be conducted. Theoretically, every rate test
should lead to the same two leak locations. Considering the possibility of inaccurate measurement or irregular
U
pipeline, more rate tests could eliminate the “noise” and amplify the “signal” or real leak location.
AN
• Any deviation from the assumptions can cause the detection by the mathematical model different from the real
leak locations and sizes. Therefore, the model should be used to narrow down the possible leak locations and
M
sizes only. Other methods should be used to verify the real leak locations and sizes.
D
Acknowledgment
TE
The authors are grateful to The Petroleum Engineering Department at the University of North Dakota. This research is
EP
supported in part by the North Dakota EPSCoR Program under award number EPS-0814442.
C
AC
Nomenclature
12
ACCEPTED MANUSCRIPT
D = pipe diameter
d2 = choke diameter
eD = relative roughness
f = friction factor
PT
ff = Fanning friction factor
RI
L = pipe length
SC
Lleak 1 = leak location 1 (measured from inlet of pipeline to leak locale 1)
U
Lleak 2 = leak location 2 (measured from inlet of pipeline to leak locale 2)
AN
Lleak 2,D = dimensionless leak location 2
MW = molecular weight
M
pb = base pressure
TE
13
ACCEPTED MANUSCRIPT
Tb = base temperature
PT
Tup = upstream temperature
RI
z = gas compressibility
SC
∆pD = dimensionless pressure drop
U
∆pno leak, = pressure drop through pipeline without gas leak
AN
∆z = outlet elevation minus inlet elevation (note that ∆z is positive when outlet is higher than inlet)
ε = absolute roughness
µ = gas viscosity
D
ρ = gas density
TE
EP
Appendix A
C
Ap pr
z= .. ................................................................................................................................................................. (A-1)
Y
14
ACCEPTED MANUSCRIPT
and
2
0.06125 1
A= exp − 1.21 − .. .......................................................................................................................... (A-3)
T pr T pr
PT
14.76 9.76 4.58
B= − 2 + 3 ........................................................................................................................................... (A-4)
T pr T pr T pr
RI
90.7 242.2 42.4
C= − 2 + 3 .. ........................................................................................................................................ (A-5)
SC
T pr T pr T pr
U
2.82
D = 2.18 + .. ...................................................................................................................................................... (A-6)
T pr
AN
where
M
Given gas z-factor, gas density can be calculated by real gas law
EP
MW p
ρ = 2.7 .. ................................................................................................................................................... (A-7)
28.96 zT
C
Lee-Gonzalez-Eakin (1970) correlation is used to calculate viscosity providing gas z-factor and density.
AC
( )
µ = 10−4 K exp Xρ Y .. .............................................................................................................................................. (A-8)
15
ACCEPTED MANUSCRIPT
986.4
X = 3.448 + + 0.01009M w .. ..................................................................................................................... (A-10)
T
PT
where
MW = molecular weight
RI
SC
References
U
Anderson, R., and Driscoll, D. 2000. Pathways for Enhanced Integrity, Reliability and Deliverability. Report prepared
Bustnes, T. E., Rousselet, M., and Berland, S. 2011. Leak Detection Performance of a Commercial Real Time
Transient Model for Troll Oil Pipeline. Pipeline Simulation Interest Group. 2011, January 1
D
Cesar, A. L. 1986. An Efficient Program for Transient Flow Simulation in Natural Gas Pipelines, Pipeline Simulation
TE
Chen, N.H.: “An Explicit Equation for Friction factor in Pipe,” Ind. Eng. Chem. Fund., 18: 296, 1979.
EP
Civan, F., and Balda, K. V. 2013. Application of Mass Balance and Transient Flow Modeling for Leak Detection in
Eisler, B. 2011. Leak Detection Systems and Challenges for Arctic Subsea Pipelines. Offshore Technology
AC
Elliott, J., Fletcher, R., and Wrigglesworth, M. 2008. Seeking the Hidden Threat: Applications of a New Approach in
Erickson, D., & Twaite, D. (1996, January 1). Pipeline Integrity Monitoring System for Leak Detection, Control, and
16
ACCEPTED MANUSCRIPT
Gajbhiye, R. N., and Kam, S. I. 2008. Leak Detection in Subsea Pipeline: A Mechanistic Modeling Approach With
Geiger, G., and Vogt, D. 2014. A Combined Leak Detection Method Using Pattern Recognition Techniques. Pipeline
Geiger, G., Werner, T., Matko, D., and Dunker, H. 2001. Model-Based Leak Detection And Localization. Offshore
PT
Mediterranean Conference.
Gonzalez, M.H. Eakin, B.E. and Lee, A.L. 1970. Viscosity of Natural Gases. New York: Institute of Gas Technology
RI
(Chicago) Monograph on API Research Project 65
Hall, K.R. and Yarborough, L. 1973. A New Equation of State for Z-Factor Calculations. Oil & gas Journal, June 18,
SC
1973, 82
Harriott, G. M. 2011. Gas Pipeline Simulation: Leak Detection. Pipeline Simulation Interest Group.
U
Hauge, E., Aamo, O. M., and Godhavn, J.-M. 2009. Model-Based Monitoring and Leak Detection in Oil and Gas
AN
Pipelines. Society of Petroleum Engineers. doi:10.2118/114218-PA, 2009, September 1
Huber, D. W. 1981. Real-time Transient Model For Batch Tracking, Line Balance And Leak Detection. Petroleum
Li, J., Liu, W., Su, Z., and Cui, L. 2009. A New Method For Location Leak In Long-distance Pipeline And Study On
D
Ling, K., Han, G., Ni, X., Xu, C., He, J., and Pei, P. 2015. A New Method for Leak Detection in Gas Pipelines, SPE
TE
Massinon, R. V. J. 1988. A Real Time Transient Hydraulic Model For Leak Detection And Sa Tch Tracking On A
EP
Mactaggart, R. H. 1989. A Sour Gas Leak Detection System Implementation. Pipeline Simulation Interest Group.
C
1989, January 1
AC
Michkova, D. E., and Strelnikova, S. A. 2013. Mathematical Modeling of Wave Processes in Leak Detection Systems
Naranjo, E. and Neethling, G.A. 2011. Best Practices in the Allocation, Commissioning, and Maintenance of
Ultrasonic Gas-Leak Detectors, SPE Projects, Facilities & Construction, December 2011, pp. 205-210
Reddy, H. P., Narasimhan, S., and Bhallamudi, S. M. 2006. Simulation and State Estimation of Transient Flow in Gas
Pipeline Networks Using a Transfer Function Model, Ind. Eng. Chem. Res. 45, 3853-3863
17
ACCEPTED MANUSCRIPT
Rui, Z., Metz, P. A., Reynolds, D., Chen, G., and Zhou, X., 2011a. Historical pipeline construction cost analysis.
Rui, Z., Metz, P. A., Reynolds, D., Chen, G., and Zhou, X., 2011b. Regression models estimate pipeline construction
Rui, Z., Metz, P. A., and Chen, G., 2012a. An analysis of inaccuracy in pipeline construction cost estimation.
PT
International journal of oil, gas and Coal Technology. 5(1), 29-46.
Rui, Z., Metz, P. A., Chen, G., Zhou, X., and Wang, X., 2012b. Regressions allow development of compressor cost
RI
estimation models. Oil & Gas Journal, 110(1a), 110-115.
Rui, Z. and Wang, X., 2013. A comprehensive analysis of natural gas distribution pipeline incidents. International
SC
journal of oil, gas and Coal Technology. 6(5), 528-548.
Rui, Z., Peng, F., Chang, H., Ling, K., Chen, G., and Zhou, X., 2017a, Investigation into the performance of oil and
U
gas projects, Journal of Natural Gas Science and Engineering, 38, 12-20.
AN
Rui, Z. Li, C., Peng, P., Ling, K., Chen, G., Zhou, X., and Chang, H., 2017b. Development of Industry performance
metrics for offshore oil and gas project, Journal of Natural Gas Science and Engineering, 39, 44-53.
Rui, Z., Lu, J., Zhang, Z., Guo, R., Ling, K., Zhang, R., Patil, S., 2017c. A quantitative of oil and gas reservoir
M
evaluation system for development, Journal of Natural Gas Science and Engineering, 42, 31-39.
D
Sadovnychiy, S., Bulgakov, I., and Valadez, J. 2005. System for Remote Detection of Pipeline Leakage. Society of
Scott, S. L. and Yi. J. H. 1998. Detection of Critical Flow Leaks in Deepwater Gas Flowlines. Society of Petroleum
Van der Marel, M., and Sluyter, E. A. 1984. Leak Detection Survey of A 36 Inch Diameter 78 Mile Long Submarine
Vrålstad, T., Melbye, A. G., Carlsen, I. M., and Llewelyn, D. 2011. Comparison of Leak-Detection Technologies for
AC
Wang, S. and Carroll, J. J. 2007. Leak Detection for Gas and Liquid Pipelines by Online Modeling. Society of
Weymouth, T.R. 1912. Problems in Natural Gas Engineering. Trans. ASME, Vol. 34, 185
18
ACCEPTED MANUSCRIPT
Zhou, J. and Adewumi, M. A. 2000. Simulation of Transients in Natural Gas Pipe Lines Using Hybrid TVD Schemes.
PT
Table 1. Input data for the two leak points detection
RI
Pipeline length 1,000,000 ft
Pipeline diameter 10 in.
Inlet pressure 1500 psia
SC
o
Pipeline inlet temperature 580 R
Outlet pressure 1250 psia
o
Pipeline outlet temperature 540 R
Gas specific gravity 0.7
U
Gas rate at inlet 1,000,000 scf/h
Pressure at standard condition 14.7 psia
AN
o
Temperature at standard condition 520 R
Absolute roughness of pipe 0.0005 in.
Elevation difference (=Outlet elevation-Inlet elevation) 500 ft
M
Table 2. Input data for the two leak points detection using both the
proposed model and numerical simulation
D
o
Temperature at standard condition 520 R
Absolute roughness of pipe 0.0005 in.
AC
Table 3. Comparison of leak detection results by the proposed model and numerical simulation
Proposed method Numerical simulation Difference
Leak point 1 location (measured from
355513 ft 355497 ft 16 ft
pipeline inlet)
Leak point 2 location (measured from
555609 ft 555596 ft 13 ft
pipeline inlet)
3 3 3
Leak rate at leak point 1 120006 ft /hour 120023 ft /hour -17 ft /hour
3 3
Leak rate at leak point 2 79994 ft /hour 79977 ft /hour 17 ft3/hour
19
ACCEPTED MANUSCRIPT
PT
A
B’
RI
A’
U SC
AN
Fig. 1. Plot of dimensionless pressure drop-dimensionless leak location 1- dimensionless leak location 2 for the first rate test
M
D
D
TE
C
B’
EP
F
D’
E
C’
A’
G
C
AC
Fig. 2. Plot of dimensionless pressure drop-dimensionless leak location 1- dimensionless leak location 2 for the second rate
test
20
ACCEPTED MANUSCRIPT
Highlights
The model can detect pipeline leak in real-time using flow parameters.
PT
RI
U SC
AN
M
D
TE
C EP
AC