Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
METHOD AND APPARATUS FOR OPERATING A PIPELINE
Document Type and Number:
WIPO Patent Application WO/2009/022148
Kind Code:
A1
Abstract:
A method of operating a pipeline (2) comprising the steps of: detecting (20) a leak (12) in the pipeline (2); reducing (22) the flow rate into the pipeline over a reduction period of time to a reduced rate; and determining (26) an estimate of the location of the leak (12) from data relating to the fluid flow in the pipeline (2) determined (24) during said reduction period of time.

Inventors:
AAMO OLE MORTEN (NO)
Application Number:
PCT/GB2008/002768
Publication Date:
February 19, 2009
Filing Date:
August 15, 2008
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
NTNU TECHNOLOGY TRANSFER AS (NO)
SAMUELS ADRIAN (GB)
AAMO OLE MORTEN (NO)
International Classes:
F17D5/02; G01M3/28
Domestic Patent References:
WO2000072750A12000-12-07
Foreign References:
US20060287837A12006-12-21
DE3720345A11988-12-29
DE4333095A11995-03-30
EP0105229A21984-04-11
Other References:
AAMO, OLE, M. ET AL.: "Observer design using boundary injections for pipeline monitoring and leak detection", 5 April 2006, IFAC, INT. SYMP. ON ADVANCED CONTROL OF CHEMICAL PROCESSES, GRAMADO, BRAZIL, XP002508469
RAHIMANN W. ET AL.: "Circle criterion based nonlinear observer design for leak detection in pipelines", 30 May 2007, XP002508470
LIU, MING ET AL.: "Fast leak detection and location of gas pipelines based on an adaptive particle filter", INT. J. APPL. MATH. COMPUT. SCI., vol. 15, no. 4, 2005, pages 541 - 550, XP002508467
WATANABE K. ET AL.: "A comparison of location methods for pinholes in pipelines", 10 May 1994 (1994-05-10), XP002508468, Retrieved from the Internet [retrieved on 20081216]
WATANABE K. ET AL.: "Pinhole location in a pipeline", 1991, XP002508471
Attorney, Agent or Firm:
FRANK B. DEHN & CO. (10 Salisbury Square, London EC4Y 8JD, GB)
Download PDF:
Claims:

Claims

1. A method of operating a pipeline comprising the steps of:

- detecting a leak in the pipeline;

- reducing the flow rate into the pipeline over a reduction period of time to a reduced rate; and

- determining an estimate of the location of the leak from data relating to the fluid flow in the pipeline determined during said reduction period of time.

2. A method as claimed in claim 1 wherein said data relating to the fluid flow in the pipeline comprises one or more of: pipeline inlet flow rate, pipeline inlet pressure, pipeline outlet flow rate and pipeline outlet pressure.

3. A method as claimed in claim 1 or 2 further comprising making an initial estimate of the location of the leak prior to the reduction in the flow.

4. A method as claimed in claim 3 comprising refining said initial estimate as the flow is being reduced.

5. A method as claimed in any preceding claim comprising initiating the flow reduction automatically upon detection of a leak.

6. A method as claimed in any of claims 1 to 4 comprising initiating the flow reduction manually upon detection of a leak.

7. A method as claimed in any preceding claim comprising triggering an alert in response to detection of a leak.

8. A method as claimed in any preceding claim comprising reducing said flow rate as a monotonically-decreasing function during said reduction period.

9. A method as claimed in any preceding claim comprising reducing said flow rate substantially linearly over time.

10. A method as claimed in any preceding claim comprising reducing said flow rate during said reduction period by at least 50%; more preferably by at least 80%; or to zero.

11. A method as claimed in any preceding claim further comprising determining an estimate of the size of the leak from said data.

12. A method as claimed in claim 11 wherein said reduced rate is dependent on the estimated size of the leak.

13. A method as claimed in any preceding claim further comprising displaying the estimated location of the leak.

14. A method as claimed in any preceding claim further comprising storing the estimated location of the leak in a non- volatile medium.

15. A method as claimed in any preceding claim further comprising uploading the estimated location of the leak to a networked server.

16. A method as claimed in any preceding claim wherein the step of detecting a leak in the pipeline comprises comparing one or more parameters of the fluid flow in the pipeline with the value(s) of said parameter(s) predicted by a computer simulation of the fluid flow.

17. A method as claimed in any preceding claim wherein the step of determining an estimate of the location of the leak comprises comparing one or more parameters of the fluid flow in the pipeline during the reduction period with the value(s) of said parameter(s) predicted by a computer simulation of the fluid flow.

18. A method as claimed in claim 16 or 17 wherein the computer simulation uses a model based on equations derived from the physical principles of conservation of mass and momentum of fluid flowing through the pipeline.

19. A method as claimed in claim 18 comprising using the measured parameter(s) of the fluid flow to update the internal state of the model.

20. A method as claimed in claim 19 wherein the model is designed so that said updating causes the model's state to converge towards said measured parameter(s).

21. A method as claimed in any of claims 18 to 20 wherein the model incorporates functions of flow velocity, flow pressure and flow density, each expressed over both time and distance along the pipeline.

22. A method as claimed in any of claims 18 to 21 wherein the model takes consideration of frictional losses in the pipeline.

23. A method as claimed in claim 22 wherein the model includes a friction coefficient that depends on pipeline relative roughness and fluid viscosity.

24. A method as claimed in claim 23 wherein the model includes an unknown constant that accounts for uncertainty in the friction coefficient.

25. A method as claimed in any of claims 18 to 24 comprising modelling said leak as a loss of fluid at a specific point along the pipeline's length.

26. A method as claimed in any of claims 18 to 25 comprising modelling said leak as having a flow rate that is dependent on one or more selected from the group comprising: time; ambient pressure on the exterior of the pipeline; fluid density inside the pipeline; fluid pressure inside the pipeline; and discharge coefficient of the leak.

27. A method as claimed in any of claims 18 to 26 wherein the model is arranged to estimate the discharge coefficient of said leak.

28. A control apparatus for operating a pipeline comprising means for detecting a leak in the pipeline and means for determining an estimate of the location of the leak from data relating to the fluid flow in the pipeline determined after a reduction in flow rate in the pipeline has been initiated.

29. A control apparatus as claimed in claim 28 further comprising means for triggering an alert in response to detection of a leak.

30. A control apparatus as claimed in claim 28 or 29 further comprising means for generating a signal to an inlet valve to reduce the flow rate in the pipeline in response to detection of a leak.

31. A control apparatus as claimed in any of claims 28 to 30 comprising means for displaying said estimate of the location of the leak.

32. A control apparatus as claimed in any of claims 28 to 31 comprising means for storing said estimate of the location of the leak in a non- volatile medium.

33. A control apparatus as claimed in any of claims 28 to 32 comprising means for uploading said estimate of the location of the leak to a networked server.

34. A system for conveying fluid comprising: a pipe; an inlet valve controlling the rate of flow of fluid into the pipe; a flow rate detector; a pressure detector; and a computer-implemented model arranged to calculate the location of a leak along the pipe using data gathered as fluid flow in the pipe is being reduced by closing said inlet valve.

35. A system as claimed in claim 34 wherein the computer-implemented model is further arranged to detect a leak in said pipe.

36. A system as claimed in claim 35 wherein the computer-implemented model is arranged to generate a signal to said inlet valve to reduce the flow rate in the pipe in response to detection of a leak.

37. A system as claimed in claim 35 or 36 further comprising an alarm, the computer-implemented model being arranged to sound said alarm on detection of a leak.

38. A system as claimed in claim any of claims 34 to 37 further comprising a display for displaying said calculated location of a leak.

39. A system as claimed in any of claims 34 to 38 further comprising a non- volatile storage medium for storing said calculated location of a leak.

40. Computer software adapted, when executed on computing means, to calculate an estimate of the location of a leak in a pipeline for carrying fluid, using data on the fluid flowing in the pipeline as said flow is being reduced.

41. Computer software as claimed in claim 40 further adapted to detect a leak in said pipeline.

42. Computer software as claimed in claim 41 further adapted to trigger an alert in response to detection of a leak.

43. Computer software as claimed in claim 41 or 42 further adapted to generate a signal to an inlet valve to reduce the flow rate in the pipeline in response to detection of a leak.

44. Computer software as claimed in any of claims 40 to 43 further adapted to render said estimate of the location of the leak in tangible form.

45. Computer software as claimed in any of claims 40 to 44 further adapted to store said estimate of the location of the leak in a non- volatile medium.

46. Computer software as claimed in any of claims 40 to 45 further adapted to upload said estimate of the location of the leak to a networked server.

47. A computer software product on a data carrier adapted, when executed on computing means, to calculate an estimate of the location of a leak in a pipeline for carrying fluid, using data on the fluid flowing in the pipeline as said flow is being reduced.

48. A computer software product as claimed in claim 47 further adapted to detect a leak in said pipeline.

49. A computer software product as claimed in claim 48 further adapted to trigger an alert in response to detection of a leak.

50. A computer software product as claimed in claim 48 or 49 further adapted to generate a signal to an inlet valve to reduce the flow rate in the pipeline in response to detection of a leak.

51. A computer software product as claimed in any of claims 47 to 50 further adapted to render said estimate of the location of the leak in tangible form.

52. A computer software product as claimed in any of claims 47 to 51 further adapted to store said estimate of the location of the leak in a non- volatile medium.

53. A computer software product as claimed in any of claims 47 to 52 further adapted to upload said estimate of the location of the leak to a networked server.

Description:

Method and Apparatus for Operating a Pipeline

This invention relates to methods and apparatus for operating and analysing pipelines for transporting fluids, in particular the handling and location of leaks.

Leaks in pipelines for carrying fluids such as natural gas, oil or water are a growing problem as more pipelines are laid and existing pipelines age. Such leaks can have both an economic impact, in terms of loss of valuable fluid and downtime for repairs, and an environmental impact. It is therefore important to be able to detect a leak as soon as possible after it appears. To aid in rapidly repairing the leak, it is also important to be able to estimate the leak's location along the pipe as accurately as possible.

It is known to detect leaks by direct observation of cracks or leaking fluid using, for example, remote cameras or gas detectors. It is also known to detect leaks indirectly, using physical measurements of the fluid flow through the pipe. Some leak detection methods perform statistical analysis on physical measurements to detect leak conditions (so-called "black box" methods) while other approaches utilise mathematical models based on physical principles. A simple example of a physical approach is to employ the principle of conservation of mass to detect a leak by noting a reduced mass of fluid flowing out of the pipe compared with the mass of fluid flowing into the pipe over a period of time.

Another model-based approach for detecting a leak is presented in "Observer Design using Boundary Injections for Pipeline Monitoring and Leak Detection" by Ole Morten Aamo et al (Department of Engineering Cybernetics, Norwegian University of Science and Technology). A software-based model of a pipe is constructed having the ability to model a single point leak in the pipe. The size and location of the leak are treated as unknown constants. Measurements from the real pipe are

used with the model to detect and estimate the location of a leak once one has occurred.

This approach for leak detection, however, suffers from a number of limitations especially in the accuracy of its estimates of the size and location of any leak.

It is an object of the present invention to improve on known systems for detecting and dealing with leaks.

From a first aspect the present invention provides a method of operating a pipeline comprising the steps of: detecting a leak in the pipeline; reducing the flow rate into the pipeline over a reduction period of time to a reduced rate; and determining an estimate of the location of the leak from data relating to the fluid flow in the pipeline determined during said reduction period of time.

Thus it will be seen by those skilled in the art that in accordance with the present invention, once a leak in a pipeline has been detected, data used to establish the location of the leak can be gathered whilst the flow is being reduced. This is advantageous as it means flow reduction does not need to be delayed until an accurate estimate of the leak's location has been determined. Rather flow rate into the pipeline may be reduced immediately after a leak is detected, thereby minimising the negative effects of the leak.

When viewed from a second aspect the invention provides control apparatus for operating a pipeline comprising means for detecting a leak in the pipeline and means for determining an estimate of the location of the leak from data relating to the fluid flow in the pipeline determined after a reduction in flow rate in the pipeline has been initiated.

When viewed from a third aspect the invention provides a system for conveying fluid comprising: a pipe; an inlet valve controlling the rate of flow of fluid into the pipe; a flow rate detector; a pressure detector; and a computer-implemented model arranged to calculate the location of a leak along the pipe using data gathered as fluid flow in the pipe is being reduced by closing said inlet valve.

When viewed from a fourth aspect the invention provides computer software and a computer software product on a data carrier, both adapted, when executed on computing means, to calculate an estimate of the location of a leak in a pipeline for carrying fluid, using data on the fluid flowing in the pipeline as said flow is being reduced. Said software is preferably arranged to render said estimate of the location of the leak in tangible form.

The flow rate may be reduced to zero after detecting a leak - i.e. the pipe could be shut off completely. Equally however it may be undesirable to stop flow completely. For example some flow may be required to enable precise localisation of the leak once an engineer or repair pig arrives at the leak site. Preferably the flow rate is reduced by at least 50%, more preferably at least 80% during the reduction period.

Preferably the flow rate is reduced as a monotonically-decreasing function over the reduction period. More preferably, it is reduced substantially linearly.

In general the rate at which the flow rate can be safely reduced will be dependent on the nature of the pipeline. Moreover it is envisaged that the flow rate could be reduced over a longer period than the theoretical minimum in order to give more data for the leak locating algorithm, recognising that that this is still an advantage over prior art proposals in which the position of the leak is only calculated at full flow. For example in some embodiments, it takes under one minute to reach the reduced level; in other embodiments, it takes between one and five minutes; in yet others, more than five minutes.

In some embodiments the flow reduction is initiated automatically upon detection of a leak. For example in such embodiments the control apparatus of the second aspect of the invention would be configured to generate a signal to an inlet valve to reduce the flow rate in the pipeline in response to detection of a leak. In other embodiments a user intervention is required. For example the detection of a leak might trigger an alert to allow the user to decide whether to reduce the flow and, possibly, by how much. As well as an estimate of the location of the leak, an estimate of the size of the leak may also be determined, hi such embodiments, the flow reduction could be dependent on the estimated size of the leak.

In some preferred embodiment an initial estimate of the location of the leak is made prior to the reduction in the flow. This can then be refined as the flow is being reduced.

Preferably the step of detecting a leak in the pipeline comprises comparing one or more parameters of the fluid flow in the pipeline with the value(s) of said parameter(s) predicted by a computer simulation of the fluid flow. For example a leak state could be defined as an over-threshold divergence between the measured and simulated parameters.

Preferably the step of determining an estimate of the location of the leak comprises comparing one or more parameters of the fluid flow in the pipeline during the reduction period with the value(s) of said parameter(s) predicted by a computer simulation of the fluid flow. As the measured and simulated parameters converge again, the location of the leak is given by the simulation.

Preferably the computer simulation uses a model based on equations derived from the physical principles of conservation of mass and momentum, of fluid flowing through the pipeline. Preferably the measured parameter(s) of the flow is/are used to update the internal state of the model. The model is preferably designed so that such updating causes the model's state to converge towards the measured parameter(s) pertaining to the real pipeline.

Some preferred embodiments use established flow rate simulators, such as those of the commercially-available OLGA (RTM) flow rate simulators sold by SPT Group AS. The Applicant has found advantageously that several commercially available flow-rate simulators can be configured to operate in accordance with the invention.

Preferably measurements of one or more of: inlet flow rate, inlet pressure, outlet flow rate and outlet pressure for the pipeline are carried out. Measurements can also be carried out at other points along the pipeline.

Preferably the data relating to the fluid flow in the pipeline from which the estimate of the location of the leak is determined comprises, or is derived from, one or more such measurements.

The preferred mathematical model of the pipeline incorporates functions of flow velocity, pressure and density, each expressed over both time and distance along the pipeline. The speed of sound in the fluid may feature as a constant term.

Preferred embodiments of the model take consideration of frictional losses in the pipeline. This may be achieved by including a friction coefficient, which may depend on pipe relative roughness and fluid viscosity. Preferably the model also includes an unknown constant that accounts for uncertainty in the friction coefficient.

In preferred embodiments the model is arranged to model a leak in the pipeline.

This may be modelled as a loss of fluid at a point somewhere along the pipe's length. The loss is preferably modelled as having a flow rate that is a function of time. Additionally or alternatively the flow is rate is dependent on one or more selected from the group comprising: ambient pressure on the exterior of the pipeline; fluid density inside the pipeline; fluid pressure inside the pipeline; and discharge coefficient of the leak. These parameters may each be taken to be variable or known or unknown constants. In preferred embodiments, the model is arranged to estimate

the discharge coefficient at the leak site. Preferably the model is configured to give an estimate of the size of the leak as well as its location.

Preferably the method comprises the step of displaying the estimated location of the leak. Additionally or alternatively preferably the estimated location is stored in a non- volatile medium or uploaded to a networked server.

Throughout this specification, the terms "pipe" and "pipeline" are used interchangeably to denote a simple pipe or network of pipes, having one or more inlet openings and one or more outlet openings. The pipes may be of any suitable material, such as plastic, metal or concrete, and are not limited to any particular size or shape. They may include components such as valves and seals.

Certain preferred embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 is a schematic view of elements involved in operating a pipeline in accordance with the invention;

Figure 2 is a flow diagram showing a method embodying the invention; Figure 3 is a further flow diagram showing another method embodying the invention; and

Figure 4 is a schematic view of a pipeline and observer system in accordance with the invention.

Figure 1 shows a pipeline 2 having a start point 4, at which fluid can enter, and an end point 6, at which fluid can exit the pipeline. Fluid can therefore flow from entry to exit in the direction indicated by the arrow 8. An inlet valve 10 determines the rate of flow of fluid into the pipeline. In the event of a leak 12, fluid exits the pipeline unintentionally, upstream of the end point 6. An alarm 14 is provided to alert an operator to the leak. A terminal 16 is provided to show details of the size and location of leak 12 to allow an engineer to investigate and fix it. This is described in more detail below.

A method embodying the invention will now be described with reference to Fig. 2. In this method periodic measurements of various parameters of fluid flow in the pipeline 2 are made such as inlet flow rate, inlet pressure etc. These are used as input data to a modified OLGA (TM) pipeline flow simulator model. Measurements of the resulting parameters such as outlet flow rate and pressure are compared with the corresponding parameters determined by the simulator. In ordinary circumstances the simulator gives closely similar results to the measured parameters. In the event of a leak occurring however, the simulator will no longer be a good representation of the physical system and therefore the simulated and measured parameters will begin to diverge.

As indicated in the first step 20 in Fig. 2, once this divergence between the simulator and the physical system exceeds a predetermined threshold, it is flagged as indicating the detection of a leak. The alarm 14 is therefore sounded to alert the pipeline operator to the occurrence of the leak, allowing corrective action to be taken. Alternatively such action could be taken automatically.

In the second step 22 a shutdown of fluid flow into the pipeline 2 is initiated in response to the leak. This could be a complete or partial shutdown. Even after shutdown has been initiated by the progressive closure of the inlet valve 10, in the next step 24 measurements of the reducing inlet pressure and flow rate of the fluid at the start point 4 and the outlet pressure and flow rate at the end point 6 continue to be taken. However the model is now altered to include a leak with initially unknown location and discharge coefficient. As the measurements of the outlet pressure and flow rate are compared to the simulated parameters generated by the model based on the inlet parameters, the estimated leak parameters are updated using the measured parameters so that the simulation converges onto the new physical state of the pipeline.

Once the model has converged to the real system, the leak location and discharge coefficient (indicative of the leak size) estimated in the model are given in the final

step 26. This information is displayed on a terminal 16 so that the leak can be investigated and fixed. Of course in addition or instead the information can be uploaded to a web server to allow remote access - e.g. by an engineer or pipeline operator.

Thus the measurements used to establish the location and size of the leak are taken while the flow rate is being reduced which means that the shutdown can be commenced as soon as the leak is detected.

Figure 3 illustrates a second embodiment similar to that described with reference to Figure 2. However, in the embodiment of Figure 3, a further step 30 is taken of returning an initial estimate of the leak parameters based on the fluid flow measurements taken after a leak has been detected but before shutdown 20 of the pipeline is initiated. This early estimate of the size and location of the leak, which may be displayed on a terminal 16 and/or uploaded to a server, could be useful for example in allocating responsibility for fixing the leak to the correct local engineer. Subsequent fluid flow measurements 24 taken during the shutdown can then be used in a final step 26 to produce an improved estimate of the location of the leak 12 as previously.

In the embodiments of Figures 2 and 3, the steps 20 and 26 of detecting a leak and estimating its size and location are implemented with the aid of an "observer". In control theory, an observer is a mathematical model that uses plant inputs and outputs to produce an estimate of the plant's state, typically for the purpose of providing feedback control to the plant.

Figure 4 further illustrates the connections between the physical system 34, comprising the pipeline 2 and suitable measuring equipment, and a Luenberger-type observer 36, suitable for use with the embodiments of Figures 2 and 3 above. The observer 36 comprises a mathematical model of the pipeline 2, implemented on computer processing apparatus. For the purposes of the model, the flow rate at the start point of the pipeline 4 and the fluid pressure at the end point of the pipeline 6

are regarded as controlled inputs 38 to the physical system 34. The flow rate at the end point of the pipeline 6 and the fluid pressure at the start point 4 are, by contrast, regarded as process measurements 40. Measurements of all of these variables are sent (38, 40) to the observer 36. These measurements may be sampled from the physical system periodically; for example, every 4 seconds.

The observer 36 uses these measurements to update its internal state model so as, ideally, to converge on the true state of the physical system 34. The observer can, therefore, determine an estimate of the state of the physical pipeline. In particular, it can determine when the pipeline may be in a leak state, and output 42 this information along with estimates of the size and location of the leak as described above.

A detailed construction of the mathematical observer model now follows.

Fluid flow in a pipeline 2 obeys the mass conservation equation and the momentum conservation equation (ignoring friction for the time being), for (x, t) e (0, L) x (0, ∞) , where x is distance along the pipeline 2 from start point 4 to end point 6, t is time, u is flow velocity, p is pressure and p is density.

For a liquid flow, the relation between pressure and density can, following A.O. Nieckele et al, "Transient pig motion through gas and liquid pipelines", Journal of Energy Resources Technology, vol. 123, No. 4 (2001), be modelled as where p re f is reference density at reference pressure p re / and c is the speed of sound. It can be assumed that c is sufficiently large to ensure p > 0. For gas flow, (3) can

simply be substituted with the ideal gas law. The following calculations, however, assume liquid flow, in which temperature can be ignored.

Defining k = c 2 p ref - p ref and substituting equation (3) into equations (1) and (2) gives

and The boundary conditions are and For the purposes of constructing a Luenberger observer, M 0 (0 and p L (t) are viewed as inputs to the process, with the remaining boundary conditions p Q (t) = p(0, t) and u L (0 = u(L,t) being viewed as process measurements.

A Luenberger-type observer constructed from equations (4) and (5), and boundary injections

and has favourable convergence properties for |& 0 | < 1 and \k L < 1 when compared with a plain copy of the pipeline; i.e. M(0, 0 = M 0 (0 m & P(L, 0 = P L (0 •

Adding friction to the model equations (4) and (5) gives the mass balance equation and momentum conservation equation where D is the pipe diameter and δ is an unknown constant to account for uncertainty in friction coefficient/, which, in turn, is defined by

in which ε I is the pipe's relative roughness, Re rf is the Reynolds number defined as and μ is the fluid viscosity.

The observer now becomes

and which incorporates an estimate δ of δ , with boundary conditions (8) and (9).

An heuristic parameter update law is given by where K δ is a strictly positive constant, and where and

Adding a leak to the model equations (10) and (11) with δ = 0 gives the mass balance equation and the momentum conservation equation where A is t pipe's cross-sectional area and f t (x,t) represents the leak. Assuming a point leak occurring at t = t, , define / ; (x, t) to be where w / and are the size and position of the leak respectively, δ denotes the Dirac distribution and H denotes the Heaviside step function. The size of the leak is modelled as the time-dependent equation where C y is a discharge coefficient and p amb is the ambient pressure of the exterior of the pipe at the leak. While p amb is assumed to be known, C v is an unknown constant to be estimated by the leak detection system.

The observer now becomes

and incorporating estimates of the leak discharge coefficient C 1 , and position x, . Heuristic parameter update laws are given by

and where λ and φ x are as given in (17) and (18), and k C , K x and γ axe strictly- positive constants.