EP3827238A1

EP3827238A1 - Method and device for identifying the occurrence of a defect in a pipeline by means of estimation

Info

Publication number: EP3827238A1
Application number: EP19786462.2A
Authority: EP
Inventors: Michael Hornacek; Daniel SCHALL
Original assignee: Siemens Energy Global GmbH and Co KG
Current assignee: Siemens Energy Global GmbH and Co KG
Priority date: 2018-09-27
Filing date: 2019-09-24
Publication date: 2021-06-02
Also published as: CA3114323A1; US20210356350A1; WO2020064741A1; EP3628993A1

Abstract

The invention relates to identifying the occurrence of a defect (11) in a pipeline (10) by means of estimation, wherein at least one first indicator (40) is identified by a first detection means (20) assigned to a first detection location (30), from which indicator a first estimation value regarding the occurrence of the defect (11) in the pipeline (10) is determined by means of a first estimation function, and at least one second indicator (41-43) is identified by at least one second detection means (21-23) assigned to a second detection location (31-33), from which indicator at least one second estimation value regarding the occurrence of the defect (11) in the pipeline (10) is determined by means of at least one second estimation function, and an overall estimation value is determined from the first estimation value and the at least one second estimation value by means of an overall estimation function by taking into account the respective positions of the first detection location (30) and of the second detection location (31-33), from which overall estimation value the occurrence of the defect (11) is estimated.

Description

Method and device for determining the occurrence of a line fault by means of an estimate

The invention relates to a method and a device for determining the occurrence of a fault in a line by means of estimation.

Lines for the transport of fluid media such as liquids or gases are often laid underground, for example to protect the line from unfavorable external influences and to improve the visual appearance for residents, especially with very long lines such as pipe lines. However, this type of installation also poses a number of problems. Unwanted disturbances in the form of leaks and overflows of water, oil or gas often occur with pipes, which can lead to extensive environmental damage, which can be difficult or impossible to remedy at times, but almost always too difficult and very expensive Remedial measures leads.

The publication Kishawy, Hossam A., and Hossam A. Gabbar, "Review of pipeline integrity management practices.", International Journal of Pressure Vessels and Piping

87.7 (2010), pages 373-380, describes various methods currently used for the detection and prevention of leaks. It shows that prevention begins with the appropriate installation of the pipeline, since even minor damage to the line such as dents can result from small cracks. In the state of the art, the following systems can be used to detect line faults during line operation:

1. Supervisory Control and Data Acquisition (SCADA): for monitoring the flow or pressure of the medium in the line, with appropriate sensors being arranged in pump stations on the line. SCADA is generally understood to be the monitoring and control of technical processes using a computer system. 2. Measuring lines, for example made of glass fibers, which can be used as a sensor element to detect the smallest vibrations and to detect movements in the vicinity of the line.

3. Inserting devices into the line (so-called “smart pigs”) and flowing through with the medium in order to obtain, for example, corrosion data on the line.

4. Flying the line and determining the depth of the cover ("depth of cover") or vegetative changes over the line, as well as the use of specially trained detection dogs to detect leaks.

The first two variants represent continuous leak detection systems, which allow continuous monitoring of leaks on lines. A distinction can be made between external and internal systems. Internal systems can be, for example, fiber optic sensors, acoustic sensors, sensor hoses and video surveillance. External systems can, for example, use a pressure point analysis or the mass balance method (difference between mass inflow and mass loss), or on statistical systems, "Real-Time Transient Model" (RTTM-) or "Extended Real-Time Transient Model" (E- RTTM) based systems.

Variants 3 and 4 are non-continuous leak detection systems, which are only carried out when required and do not allow permanent monitoring of leaks.

In the publication Guerriero Marco Et Al: "Bayesian data fusion for pipeline leak detection", 2016 19th International Conference on Information Fusion (FUSION), ISIF, July 5, 2016, pages 278-285, XP032935023, as well as in the

US 2017/076563 Al (Guerriero M [US] Et Al) March 16, 2017, discloses a method for determining the occurrence of a line fault by means of an estimate. It is performed propose to combine a treasure algorithm for detection and a treasure algorithm with one another using a dynamic Bayesian network (DBN).

Two heterogeneous systems are combined, namely "fiber optic distributed acoustic sensing" (DAS) and "internal leak detection (ILD)" technology. ILD technology typically has low sensitivity and poor localization. DAS technology, in turn, typically has a relatively high sensitivity, but a high false alarm rate. The combination of ILD and DAS technology is intended to achieve high sensitivity with a low, high false alarm rate.

However, DAS technology is not always available, and it cannot be retrofitted at great expense or economically.

It is an object of the invention to improve the reliability of a detection of an impurity compared to the prior art with simple means.

The object is achieved by a method of the type mentioned at the outset, in that a first indicator is determined by a first detection means assigned to the line using at least one first detection parameter, from which a first estimate with regard to the occurrence of the line fault occurs by means of a first estimation function and a first detection location is determined, and at least one second indicator is determined by at least one locally detecting second detection means assigned to a second detection location using at least one second detection parameter, from which at least one is determined by means of at least one second estimation function second estimated value with regard to the occurrence of the fault in the line is determined, the first detection parameter and the at least one second detection parameter being selected such that a first operating mode is configured for the first detection means and / or the at least one second detection means, and from which a first estimated value and the at least one second estimated value is determined by means of an overall estimation function, taking into account the respective position of the first detection location and the respective second detection location, from which the occurrence of the fault location is estimated, and if a threshold value is exceeded, the overall estimation function Estimation for the occurrence of the fault based on the total estimated value, a second operating mode for the first detection means and / or the at least one second detection means is determined, which by means of the at least one first detection parameter and the at least one second iten acquisition parameter is configured, and the total estimated value is determined again on the basis of the second operating mode and the occurrence of the fault location is re-estimated therefrom.

The result of this is that a fault can be determined in an incremental split method by firstly estimating the event in a first operating mode and, upon detection, carrying out a further, specifically adapted estimation in a second operating mode. This can improve the reliability and accuracy of a detection. In other words, the reliability of a detection is improved by means of a two-stage method. The two-stage process thus comprises a first stage and a second stage for the estimation. Here, a simple second detection means can be used, since only the occurrence of the fault location is estimated and no localization has to be carried out during operation, since the detection location is known by the assembly or installation position of the second detection means.

The detection means have at least two operating modes, for example a “monitoring mode”, in which an error event occurs over a longer period of time by means of statistical monitoring and evaluation of the system in order to suppress short-term faults and to make the system more robust and reliable, and an “error mode” in which a new estimate is carried out by means of an optimized parameterization, which permits an improved statement with regard to the type and position of the detected fault location.

The respective detection locations are determined and known in advance by the installation locations of the sensors or their geographic areas, which are covered by the respective sensor.

An estimator to form an estimate, too

Estimation statistics, or estimators for short, are used in mathematical statistics to determine an estimated value based on existing empirical data from a sample and thereby to obtain information about unknown parameters of a population. Estimation functions are the basis for calculating point estimates and for determining confidence intervals using range estimators and are used as test statistics in hypothesis tests. They are special sample functions and can be determined using estimation methods, for example the least squares estimate, the maximum likelihood estimate or the moment method.

Furthermore, it is expedient if in the second operating mode an additional third, that is to say still in the first operating mode unused detection means is used to improve the estimate, for example, with regard to accuracy and / or susceptibility to failure.

The number of detection means used can thus be higher in the second operating mode than in the first operating mode.

Consequently, a third indicator can be determined by a third detection means assigned to the line using at least one third detection parameter, from which a third estimate is determined by means of a third estimation function with regard to the occurrence of the line fault and a third detection location, and in addition to this the first estimate and the at least one second estimate, the third estimate is determined by means of the overall estimate of the total estimate, taking into account the respective position of the first, second and third detection location, from which the occurrence of the fault point is estimated.

The invention is also achieved by a device of the type mentioned at the outset, comprising at least one first detection means assigned to a first detection location, which is set up to determine a first estimate as at least a first indicator by means of a first estimation function, and at least one second detection location assigned second detection means, which is set up to determine at least one second estimate as at least one second indicator by means of at least one second estimation function, and an analysis device with a processor and a memory, which is set up to determine estimated values , wherein the first and second detection means are connected to the analysis device, and wherein the device is configured to carry out the method according to one of the preceding claims.

In a further development of the invention, it is provided that the total estimated value based on the first operating mode and the total estimated value based on the second operating mode are correlated with one another.

As a result, a resulting overall estimated value of the method can be determined, which further improves the reliability and accuracy.

In a development of the invention, it is provided that the first operating mode provides detection by means of the first indicator and the at least one second indicator, in which the at least one first detection parameter and the at least one second detection parameter are selected such that the reliability of the detection is optimized.

The result of this is that the reliability of the method is improved. The reliability can be described, for example, by using statistical variables when viewing the indicators over time. Furthermore, a corresponding adaptation of the sampling frequency can advantageously be used when the indicators are detected by the detection means.

In a further development of the invention, it is provided that the first operating mode provides detection by means of the first indicator and the at least one second indicator, in which the at least one first detection parameter and the at least one second detection parameter are selected such that the current consumption of the detection is optimized. This can be done, for example, by appropriately adapting the sampling intervals when the indicators are recorded by the detection system. means can be achieved, or also by a targeted connection of more sensitive sensors or electronic amplifiers in the detection means.

It is thereby achieved that the operating properties are improved or the operating costs of the method are reduced. The operating costs can be, for example, required resources with regard to data acquisition, data storage, statistical data evaluation or also current or power consumption of a corresponding device.

In a further development of the invention, it is provided that the second operating mode provides for detection using the first indicator and the at least one second indicator, in which the at least one first detection parameter and the at least one second detection parameter are selected such that the accuracy the acquisition is optimized.

It is thereby achieved that the accuracy of the method is improved. As previously stated, operating properties can be improved or operating costs of the method can be reduced.

By combining the aforementioned first with the second operating modes, an improvement of the method is achieved, which allows an improved, reliable and at the same time accurate detection of faults.

It is advantageous if the line is an oil, gas or water line, since such lines are already used in the prior art via a number of sensors for monitoring the line, and therefore a simple way to implement the merge the combined, two-stage evaluation of several indicators can take place, especially in the case of already existing lines and associated sensor systems. In a development of the invention, it is provided that when the threshold value of the estimate for the occurrence of the fault location is exceeded, a hypothesis for the type and / or position of the fault location is determined, which hypothesis when determining the at least one first detection parameter and the minimum a second acquisition parameter is used.

The result of this is that a configuration during the detection of the indicators in the second operating mode verifies the hypothesis and thus the fault location is estimated in an efficient manner.

It is also advantageous if the fault point is a leak in the line. This allows a specially constructed, as well as existing sensors from lines to be used in the arrangement.

Furthermore, it is advantageous if the fault location is influenced by an event outside the line that is detected. In other words, the detected Störstel le for example, the line surrounding soil, mounting ge ge or connecting elements such as screws, which are attached to the line or connect the line with other parts, relate. This allows preventive faults to be recorded before damage occurs directly in the line, for example due to the line being flushed out due to a temporary water flow caused by a severe storm or a landslide.

In a development of the invention, the position of the fault in or on the line is estimated from the total estimated value. The total estimated value is determined from the previously determined estimated values, in each of which the position of the fault location can also be determined. Only through a joint assessment of the individual estimates of the first stage with the respective positions of the common fault location can the accuracy of the position determination of the estimates of the first stage overall in the second stage of the second stage process can be improved. It can be an estimate of the position of the fault in the line, as well as in the adjacent area around the line, as previously described, for example, the earth surrounding the line, be determined.

In a development of the invention, the first estimate is determined by means of the first estimate and / or the at least one second estimate is determined by means of the at least one second estimate by an analysis device. The respective detection devices and their development can be simplified by calculating the respective estimated values in the analysis device.

In a development of the invention, the first detection location of the first detection means and the second detection location of the at least one second detection means are adjacent to one another and to the line, preferably within a distance of 40 m, more preferably within 20 m and particularly preferably within 10 m. By appropriate positioning of the respective detection means can be achieved that the determination of the position of jamming points is more accurate.

In a development of the invention, the first acquisition means and / or the at least one second acquisition means are formed by performing a visual inspection of the line, preferably in the second operating mode.

This ensures that, for example, the type and quality of the vegetation can be analyzed in a simple manner via a buried pipe, in particular if the visual inspection is carried out using air, for example with a drone or a helicopter. The visual inspection is preferably carried out with the aid of an imaging sensor in the optical and / or in the infrared range while at the same time determining the position of the inspection locations.

It is advantageous if a more detailed visual inspection is carried out in the second operating mode, for example through a new detection using an air-based detection device. The estimate from the first operating mode can be used as a basis, for example to limit the location for carrying out the detection in the second operating mode and to carry it out at a lower altitude in order to collect more precise and more detailed data, which are then processed further. Thus, parameters for the second operating mode can also include, for example, the flight altitude, the recording path or the spectral sensitivity (such as UV, IR or optical spectrum) of the sensor of the air-based detection means.

In a development of the invention, the first acquisition means and / or the at least one second acquisition means are formed by carrying out a soil analysis of the surroundings of the line. It is thereby achieved that, for example, the contamination of the soil by the medium which is carried in the line can be analyzed in a simple manner above, next to or below a buried line, in particular by means of a moisture sensor.

In a development of the invention, the first detection means is formed by an analysis device of weather data. Weather data can be forecast data, current or previous weather influences. This data can either be determined manually or determined by weather simulations by a computer-aided weather service. Local weather data are determined at a specific location, such as hail, floods, storms or hot spots.

In a further development of the invention, the first detection means is formed by an analysis device of environmental data. The environmental data can be analysis data that originate from or are derived from satellite-based imaging sensors, for example maps of bodies of water, burned-off areas, flooding or subsidence. Environmental data can be data from satellite-based data services such as Copernicus (http: // www. Copernicus. Eu / main / Services), and can include data on the atmosphere, climate change, marine environment or land data, for example.

In a further development of the invention, the first indicator and the at least one second indicator represent independent physical measurement variables, preferably pressure, movement and temperature. A movement can be detected, for example, by motion sensors or acceleration sensors.

In a development of the invention, the first indicator and the at least one second indicator represent a change in physical properties of the line, preferably material properties or properties relating to the aging of material. The corresponding properties or their changes can be stored in a model or a database.

In a development of the invention, the overall estimator is a Bayesian estimator. In this way, an improved two-stage estimate can be implemented in a simple manner, and the line and its indicators can be modeled particularly clearly by using a Bayesian network.

In mathematical statistics, a Bayes estimator is an estimator that takes into account any previous knowledge of a parameter to be estimated in addition to the observed data. In accordance with the procedure of Bavarian statistics, this prior knowledge is modeled by a distribution for the parameter, the a priori distribution. Bayes' theorem gives the conditional distribution of the parameter under the observation data, the a posteriori distribution. In order to obtain a clear estimate from this, positional measures of the a posteriori distribution, such as expected value, mode or median, are used as so-called Bayes estimators. Because the a posteriori expected value is the most important and is the most frequently used estimator in practice, some authors also refer to it as the Bayes estimator.

A Bayes estimator is generally defined as the value that minimizes the expected value of a loss function under the a posteriori distribution. For a quadratic loss function, the a posteriori expected value is then obtained as the estimator.

The invention is explained below with reference to an embodiment shown in the beige closed drawings. In the drawings:

1 is a schematic representation of a device for estimating the occurrence of a fault on a line,

Fig. 2 is a schematic representation of a Bayesian network for use in a method according to the invention.

Fig. 1 shows schematically an embodiment of a Anord voltage with a line 10 which is provided for the long-distance transport of oil. The figure shows the arrangement for estimating the occurrence of a fault point 11 according to the invention, for example a leak due to a crack in the line 10, the line 10 being introduced into the ground 12 and covered with soil and also vegetation 13, for example a meadow, covered the floor 12.

The device comprises a first detection means 20 in the form of a line sensor with a measuring line 25 which runs adjacent to the line 10. By means of the measuring line 25, a fault point 11 can be determined, for example, by detecting a change in the resistance of the measuring line 25 in the ground, and the location of the resistance change as a first detection point 30.

The detection of the change in resistance can, for example, be the first indicator 40 by the first detection means 20 by measuring reflections of a high-frequency signal fed into the measuring line 25, which occur due to the action of the leakage at the fault point 11 of the line 10 on the first detection location 30.

The first detection means 20 is connected to an analysis device 50 for wired or wireless transmission of the first indicator 40 and the first detection location 30.

The analysis device 50 is set up to determine a first estimate from the first indicator 40 by means of a first estimate. Alternatively, the first estimate can also be determined by the first detection means 20 and the first estimate can be transmitted to the analysis device 50 for further use.

Furthermore, the device comprises a second detection means 21 in the form of a camera, which is set up to carry out a visual inspection of the line 10. The camera is an imaging sensor that is sensitive, for example, in the optical and infrared range. For example, the camera can be moved along line 10 with a drone or a helicopter and record the vegetation 13 or the surroundings directly above line 10. Changes in the vegetation 14 at a second detection location 31 can be combined in the form of camera recordings with position data from a GPS system, recorded as a second indicator 41 and transmitted to the analysis device 50.

The analysis device 50 is set up to determine a second estimated value from the second indicator 41 by means of a second estimation function.

In this exemplary embodiment, the device also includes a third detection means 22 with a soil moisture sensor 27, which is set up to detect the relative humidity of the soil at a third detection location 32 as a third indicator 42. In other words, A soil analysis of the area around the line 10 is carried out by the moisture sensor 27.

The third detection means 22 is connected to an analysis device 50 in a wired or wireless manner in order to transmit the third indicator 42 and the third detection location 32.

The analysis device 50 is set up to determine an overall estimated value from the third indicator 42 by means of an overall estimation function.

The device also comprises a fourth detection means 23 in the form of a computer-aided weather service, which is set up to forecast local weather data at a fourth detection location 33 or to determine current or already occurring weather influences and to transmit them to the analysis device 50 as a fourth indicator 43. Local weather influences can include hail, floods, storms or hot spots.

The analysis device 50 is set up to determine a fourth estimated value from the fourth indicator 43 by means of a fourth estimation function.

In this example, the device comprises a fifth detection means 24 in the form of a database, in which properties about the aging of physical material properties of sections of the line 10 form a fifth indicator. A fifth detection location for the position of the corresponding section of line 10 is additionally assigned and stored in the database.

The analysis device 50 forms a SCADA system and is set up there to determine a fifth estimated value from the fifth indicator by means of a fifth estimation function.

It is clear that the detection locations 30-33 of the respective detection means 20-24 are already known and can be static. In these cases, the positions of the respective Detection locations 30-33 in the analysis device 50 are stored in a memory and are assigned to the respective detection means 20-24 for further use, the detection means then only transmitting the respective indicators 40-43 to the analysis device 50.

The analysis device 50 is also set up to take a total estimated value from the first estimated value, the second estimated value, the total estimated value, the fourth estimated value and the fifth estimated value by means of an overall estimation function, for example using a Bayesian estimator to determine the respective location of the first detection location 30, the second detection location 31, the third detection location 32, the fourth detection location 33 and the fifth detection location, from which total estimated value the occurrence of the fault location 11 or its position can be estimated.

In one exemplary embodiment of the method, a first indicator 40 is determined by a first detection means 20 assigned to the line using at least one first detection parameter, from which a first estimate with regard to the occurrence of the fault point 11 of line 10 and a is generated by means of a first estimation function first detection location 30 determined.

Furthermore, at least one second indicator 41-43 is determined using at least one second detection means 21-23, each assigned to a second detection location 31-33, using at least one second detection parameter, from which at least a second detection function is used by means of at least one second estimation function Estimated value with regard to the occurrence of the fault point 11 of the line 10 is determined.

The first detection parameter and the at least one second detection parameter are selected such that a first operating mode for the first detection means 20 and / or the at least one second detection means 21-23 is configured.

Furthermore, an overall estimated value is determined from the first estimated value and the at least one second estimated value by means of an overall estimation function, taking into account the respective position of the first detection location 30 and the respective second detection location 31-33, from which the occurrence of the fault point 11 is estimated .

On the basis of the total estimated value, a second operating mode for the first detection means 20 and / or the at least one second detection means 21-23 is determined, which is configured by means of the at least one first detection parameter and the at least one second detection parameter.

The total estimated value is determined again on the basis of the second operating mode and the occurrence of the fault location 11 is again estimated therefrom.

The two-stage estimation according to the invention enables indicators that are independent of one another to be combined with one another and a precise estimate of defects to be made possible, the indicators being able to be independent physical measurement variables, such as pressure, movement and temperature. Consequently, the detection means 20-23 can be a pressure sensor, an acceleration sensor, a temperature sensor or the like, it being clear that corresponding sensor evaluation electronics must also be provided in order to detect the sensor values as corresponding indicators.

The first operating mode provides for detection by means of the first indicator 40 and the at least one second indicator 41-43, in which the at least one first detection parameter and the at least one second detection parameter are selected such that, for example, the reliability or the current consumption the acquisition is optimized. The second operating mode provides for detection by means of the first indicator 40 and the at least one second indicator 41-43, in which the at least one first detection parameter and the at least one second detection parameter are selected such that, for example, the accuracy of the detection is optimized .

The total estimated value based on the first operating mode and the total estimated value based on the second operating mode can be correlated with one another.

The method according to the invention can be more precise if the detection locations 30-33 of the respective detection means 20-24 are located adjacent to one another and to the line 10, preferably within a distance 16 of 40 m, more preferably within 20 m and particularly preferably within 10 m.

The distance 16 is preferably the diameter of an imaginary circle, in which both the position of the leak and the detection locations 30-33 of the respective detection means 20-24 lie.

A degree of coverage 17 is due to the depth of the line below the covering surface of the earth, that is, the vertical distance from the top of the pipeline to the surface of the ground.

Furthermore, the method according to the invention is suitable for detecting disturbance points 11 which are influenced by an event outside the line (10), such as, for example, flushing the line 10 through a flowing water.

2 shows an exemplary embodiment for describing a modeling of the method according to the invention using a Bayesian network.

The concept of the invention provides for heterogeneous information to be combined by a uniform model. This can be used to support an operator of an oil pipeline in his conclusions and decisions.

Bayesian networks provide a formal basis for the integration of information and inference of events.

A Bayesian network is a graphical model that can be described as a directed acyclic graph, where nodes represent discrete random variables and directional arrows represent dependencies between variables in terms of conditional probabilities.

A Bayesian network makes it easy to distribute expert knowledge and empirical data in models with Bayesian networks. Bayesian networks are also robust against erroneous, missing and scattered data. Furthermore, Bayesian networks allow semantic interpretation from your own data, which enables a trustworthy and secure relationship to be used with the data used.

The model of FIG. 2 can answer queries regarding the "health" or the tightness state of a section of a pipeline, which can be described by the node HEALTH 100.

A section can be defined very flexibly, for example as the distance between two pumping stations, in order to achieve a sufficient limitation of information and alarms. In general, the tightness of a section of a pipeline can be a result of an analysis of SCADA events, earth observation data and events that are triggered by external activities, for example a nearby construction site.

In connection with oil, gas and water, a SCADA system with a status node 110 can monitor the flow and pressure values of a line and trigger corresponding alarms when a specified limit value is exceeded. The SCADA In a similar form, the system can be the analysis device 50 and comprise the detection means 20-24 of FIG. 1.

Leak detection systems on a statistical basis each subject a statistical test to previously determined values for indicators 40-43. General statistical variables can be formed from a change in pressure over time or using the mass balance method. The so-called mortgage test is a frequently used method.

If leak detection systems on a statistical basis, which carry out an evaluation of a SCADA system, determine a leak with a high safety probability, the likelihood of a declining tightness condition of a pipeline section is often very high, especially for creeping or emerging leaks . Thus, the SCADA node 110 in the model of FIG. 2 is directly related to the HEALTH node 100.

Earth observation data can be used to determine the vegetation status of the soil covering the pipeline. The so-called "Normalized Difference Vegetation In

dex "(NDVI) can be used, which can be detected, for example, by airborne visual observation using a camera. Individual camera images can be assigned current position data of the camera from a GPS system to enable automated evaluation and geo-referenced electronic map of the NDVI The NDVI is a simple and widely used vegetation health indicator.A geographically restricted area with an NVDI grid can be calculated using a combination of aerial photographs of optical and near infrared bands (NIR) On the basis of two overlapping layers of NDVI maps, which come from different observations, a map with NDVI changes, for example, can be calculated pixel-by-pixel, which are divided into polygons with regard to areas of minimum and maximum change can. Numerous attributes can be assigned to these polygons, such as the area in square meters, which in turn can be used for filtering.

While an NDVI value or a change in the NDVI in the Bayes model of FIG. 2 can detect a death of the vegetation, which may be caused by a leak in the line, a single NDVI value or a change in the NVDI could become undesirable False alarms ("false positives") lead, whereby the sole NDVI value is rather unsuitable for leak detection. Therefore, in FIG. 2, an NDVI node 120 is connected to the SCADA node 110 and to an ASSET Node 150.

A SCADA alarm would exacerbate a problem that was detected by an NDVI change and provide further information about the detected leak. Furthermore, the ASSET node 150 would also exacerbate the problem by adding information regarding the corrosion status of the coating of the pipeline, as further explained below.

The degree of coverage, the so-called “Depth of Co

ver "(DOC) or layer thickness, is the vertical distance from the top of the pipeline to the surface of the ground and is taken into account in the model of FIG. 2 by node 130. At the time of their construction, pipelines must meet minimum legal requirements which can also regulate the necessary cover, but this cover can change over time, for example through excavation work, erosion, cultivation, construction activities, flooding, subsidence or other environmental influences or human influences.

A pipeline reference model and a terrain model, which was determined, for example, from aerial photographs, can be used to determine DOC in meters with respect to the terrain model. For a certain point, DOC is the difference between see the terrain model and the pipeline reference model at this point, as known from WO 2017/174426 Al.

Another node 140 can be the weather (WEATHER). Weather data can be forecast data, current or previous weather influences. These data can either be determined manually or determined by weather simulations of a computer-aided weather service. Local weather data are determined at a specific location, such as hail, floods, storms or hot spots.

ASSET node 150 contains all of the information related to the physical line, such as information in the form of a geo-referenced pipeline reference model, the material properties of the line, and potential corrosion parameters of the pipeline or portions thereof. The corrosion parameters can be recorded, for example, using various measurement techniques within the line, as well as reliability and survival models and analyzes.

Node 150 connects nodes 110 (SCADA) and nodes 120 (NDVI). An alarm from the SCADA system can thus be additionally supported by data from the ASSET node (150) by means of a correspondingly high probability of occurrence, for example of age-related corrosion.

Another node 160 can take events (EVENTS) into account, such as excavation work, cultivation, construction activities, floods or earthquakes. The processes can be recorded and classified, for example, by using glass fiber sensors. Processes usually have a major impact on the "health" of the pipeline, which is shown in node 100, provided that DOC node 130 is also considered critical. An operator must carry out an evaluation to determine whether corrective measures are among the given ones Circumstances are appropriate. Knots 130 and 140 are preferably with the knot

ten 160 (EVENTS) connected.

After the Bayesian model has been trained and the corresponding data stored by a sensor network, for example in the form of an Internet of Things (IoT), an automatic and ongoing evaluation of the "health" of the pipeline can take place, which is an inference referred to as.

For example, node 100 (HEALTH) can include the following values: OK, DETORIATED (deteriorated),

SEVERE DAMAGE (severe damage).

A query to the SCADA system, for example the analysis device 50 of FIG. 1, could then be: cpquery (

evidence

NDVI CHA the following parameters according to FIGS. 1 and 2 define the query as an example:

"Pipeline_health_bn" is a Bayesian network model of the Pipe line 10 with the health status of the HEALTH node 100.

"Event" defines a specific health status of the HEALTH node 100, which is determined by the condition of an "evidence" parameter. The evidence parameter can be a combination of several states of different nodes, such as the "ALARM_PRESSURE_LOSS" state of the SCADA node 110 and the "LARGE" state of the NVDI node 120.

In this case, the result would be that the pipeline health status is SEVERE_DAMAGE, supported by a message from the SCADA system, which has given an alarm to a pressure sensor of the SCADA system about a pressure loss, and the vegetation in the relevant area has undergone a major change of the NDVI parameter, which by means of Aerial photography and a subsequent computer-aided evaluation was determined.

Reference symbol list:

10 line

11 Impairment point, leak

12 bottom

13, 14 vegetation, meadow

15 Leakage or accumulation of the medium

16 distance

17 Degree of coverage

20-24 means of registration

25 measuring line

27 humidity sensor

30-33 Location of the means of detection

40-43 indicator

50 evaluation device

100 State of the line, HEALTH

110 SCADA system

120 Vegetation status, normalized difference vegetation

Index, NDVI

130 State of Coverage, Depth of Cover, DOC

140 WEATHER, WEATHER

150 physical state, properties of the line,

ASSET

160 events, EVENTS

Claims

Expectations

1. A method for determining the occurrence of a fault (11) in a line (10) by means of an estimate,

characterized in that a first indicator (40) is determined by a first detection means (20) assigned to the line using at least one first detection parameter, from which a first estimate with respect to the occurrence of the line fault (11) occurs by means of a first estimation function (10) and a first detection location (30) is determined,

and at least one second indicator (41-43) is determined by at least one locally detecting second detection means (21-23) assigned to a second detection location (31-33) using at least one second detection parameter, from which at least one is used second

Estimating function at least one second estimated value with regard to the occurrence of the fault point (11) of the line (10) is determined,

wherein the first acquisition parameter and the at least one second acquisition parameter are selected such that a first operating mode for the first acquisition means (20) and / or the at least one second acquisition means (21-23) is configured,

and from the first estimated value and the at least one second estimated value by means of an overall estimation function, an overall estimated value is determined taking into account the respective position of the first detection location (30) and the respective second detection location (31-33), from which the occurrence of the Fault point (11) is estimated,

and if a threshold value of the estimate for the occurrence of the defect (11) is exceeded on the basis of the total estimated value, a second operating mode for the first detection determination means (20) and / or the at least one second acquisition means (21-23) is determined, which is configured by means of the at least one first acquisition parameter and the at least one second acquisition parameter,

and the total estimated value is determined again on the basis of the second operating mode and the occurrence of the fault location (11) is estimated again therefrom.

2. The method according to the preceding claim, wherein the number of detection means (20-23) used in the second operating mode is higher than in the first operating mode.

3. The method according to any one of the preceding claims, wherein the first operating mode provides detection by means of the first indicator (40) and the at least one second indicator (41-43), in which the at least one first detection parameter and the at least one second detection parameter meter is selected so that the reliability of the detection is optimized.

4. The method according to any one of the preceding claims, wherein the first operating mode provides for detection by means of the first indicator (40) and the at least one second indicator (41-43), in which the at least one first detection parameter and the at least one second detection parameter meter is selected so that the current consumption of the detection is optimized.

5. The method according to any one of the preceding claims, wherein the second operating mode provides detection by means of the first indicator (40) and the at least one second indicator (41-43), in which the at least one first detection parameter and the at least one second detection parameter is selected so that the accuracy of the detection is optimized.

6. The method according to claim 1, wherein the total estimated value based on the first operating mode and the total estimated value based on the second operating mode are correlated with one another.

7. The method according to any one of the preceding claims, wherein when the threshold value of the estimate for the occurrence of the fault location (11) is exceeded, a hypothesis for the type and / or location of the fault location (11) is determined, which in determining the at least one first Detection parameters and the at least one second detection parameter is used.

8. The method according to any one of the preceding claims, wherein the defect (11) is a leak in the line (10).

9. The method according to any one of the preceding claims, wherein the fault location (11) is influenced by an event outside the line (10).

10. The method according to any one of the preceding claims, wherein the position of the fault point (11) in or on the line (10) is estimated from the total estimated value.

11. The method according to any one of the preceding claims, wherein the first estimate is determined by means of the first estimate and / or the at least one second estimate is determined by means of the at least one second estimate by an analysis device (50).

12. The method according to any one of the preceding claims, wherein the first detection location (30) of the first detection means (20) and the second detection location (31-33) of the at least one second detection means (21-23) within a distance (16) from 40 m, preferably within 20 m and particularly preferably within 10 m adjacent to each other and to the line (10).

13. The method according to any one of the preceding claims, wherein the first detection means (20) and / or the at least one second detection means (21-23) by performing a visual inspection of the line (10) with the aid of an imaging sensor (21) in optical and / or in the infrared range, preferably in the second operating mode.

14. The method according to any one of the preceding claims, wherein the first detection means (20) and / or the at least one second detection means (21-23) are formed by performing a soil analysis of the surroundings of the line (10), preferably by a moisture sensor (22 ).

15. The method according to any one of the preceding claims, wherein the first detection means (20) is formed by an analysis device (23) of weather data or environmental data.

16. The method according to any one of the preceding claims, wherein the first indicator (40) and the at least one second indicator (41-43) represent mutually independent physical measurement variables, preferably pressure, movement and temperature.

17. The method according to any one of the preceding claims, wherein the first indicator (40) and the at least one second indicator (41-43) represent a change in physical properties of the line (10), preferably material properties or properties about the aging of material .

18. The method according to any one of the preceding claims, wherein the overall estimator is a Bayesian estimator.

19. Apparatus for estimating the occurrence of a fault point (11) in a line (10),

comprising at least one first detection means (20) which is assigned to a first detection location (30) and which is set up for this purpose by means of a first estimation function To determine the estimated value as at least one first indicator (40),

and at least one second detection means (21-23) assigned to a second detection location (31-33), which is set up to determine at least one second estimate as at least one second indicator (41-43) by means of at least one second estimation function ,

and an analysis device (50) with a processor and a memory, which is set up to determine estimated values,

said first and second detection means (20-23) being connected to the analysis device (50), and

characterized in that the device is set up to carry out the method according to one of the preceding claims.