METHOD AND ARRANGEMENT FOR DETECTING CORONA DISCHARGES IN A PLANT WITH OPERATING EQUIPMENT

Description

[0001] The invention relates to a method according to the preamble of claim 1 and an arrangement according to the preamble of claim 13.

[0002] European patent application 17161027.2 of 15.03.2017, published as EP 3376213 A1, entitled "Method and arrangement for condition monitoring of a plant with operating resources" discloses a method for condition monitoring of a plant with operating resources, in which overview data are acquired by means of a first vehicle with an overview sensor arrangement for optically detecting the plant, and By means of an evaluation device, the operating equipment is recognized in the overview data and the positions of the operating equipment are determined taking into account the position of the first vehicle, characterized in that detailed images of the operating equipment are generated by means of a second vehicle with a detail camera that is aimed at the respective positions of the operating equipment. For example, only a single aircraft such as a drone or a helicopter is used to detect masts and insulators when flying over an overhead line using the overview camera, to determine the position of the insulators and then to acquire high-resolution images of the insulators using the detail camera. This allows defective insulators to be detected easily and reliably.

[0003] Corona discharges on elements of overhead lines or other high-voltage infrastructure are a well-known yet undesirable physical phenomenon, described in detail on Wikipedia (permanent link: https: / / de.wikipedia.org / w / index.php?title=Koronaentladung&ol did=173331289). In combination with the nitrogen content of the ambient air, corona discharges can produce acid products that attack the surfaces of high-voltage fittings. Corona discharges also have other undesirable side effects, such as interference with radio frequency bands. To prevent corona discharges, protective devices are installed on components such as e.g. Protective fittings, so-called corona rings, are attached to the insulators. Corona rings are described, for example, on Wikipedia (permanent link: https: / / de.wikipedia.org / w / index.php?title=Koronaring&oldid=1 71645402).

[0004] Corona discharges (especially under the influence of daylight) are usually not visible to the human eye.

[0005] Therefore, cameras for the ultraviolet (UV) frequency range are used, such as the DAYCOR product from OFIL (known from the website http: / / www.ofilsystems.com / products). Such cameras are equipped with an image intensifier that can make individual light quanta visible. A daylight blocking filter is also installed to minimize the influence of daylight. Various preprocessing steps are known in the state of the art for converting the raw signal into an image that is easier for a human observer to perceive. This is done, for example, by overlaying an image in the UV spectrum of light on an image of an installation such as an overhead line for rail vehicles or an overhead power line in the visible spectrum of light. The evaluation of whether displayed discharges represent a relevant corona is the responsibility of a human evaluation technician.

[0006] Video recordings of overhead lines made with a corona camera as described above often reveal a certain amount of background noise in the recorded signals. While numerous discharges are recorded in a single image, these are no longer visible in subsequent images. Few methods are known from the state of the art for automatically evaluating such images, especially with regard to the location-based detection of a corona effect in a moving recording system.

[0007] A well-known publication is "An Automatic Coronadischarge Detection System for Railways Based on Solar-blind Ultraviolet Detection" by Li et al., Current Optics and Photonics, Vol. 1, No. 3, June 2017, pp. 196-202. In the approach described – tailored to rail vehicles – the linear motion of the detected corona points is detected directly in the image space, applying the concept of a so-called "Hough Transformation" (in 2D). This exploits the fact that a corona discharge appears in several consecutive images as a bright point moving along a straight line as the rail vehicle moves past it on the straight railway tracks. The mathematical procedure of the Hough Transformation is explained, for example, on Wikipedia (permanent link: https: / / de.wikipedia.org / w / index.php?title=Hough-Transformation&oldid=165672024).Another system for detecting corona discharges in rail vehicles is the DayCor Rail System from OFILSYSTEMS, known from the website http: / / www.ofilsystems.com / products.

[0008] Furthermore, the publication by Nguyen van Nhan et al: "Automatic autonomous vision-based power line inspection: A review of current status and the potential role of deep learning", International Journal of Electrical Power & Energy Systems, Jordan Hill, Oxford, UK, Vol. 99, 9 January 2018, pages 107-120, ISSN: 0142-0615, DOI: 10.1016 / J.IJEPES.2017.12.016 and the publication by Lan Chen et al: "The Ultraviolet Detection of Corona Discharge in Power Transmission Lines", Energy and Power Engineering, Vol. 05, No. 04, 1 January 2013 (2013-01-01), pages 1298-1302, USDOI: 10.4236 / epe.2013.54B246 Methods and systems for detecting corona discharges are known.

[0009] The object of the invention is to provide a method by which corona discharges in systems can be detected automatically and reliably.

[0010] The invention solves this problem with a method according to claim 1.

[0011] A system can, for example, be an electrical system such as an overhead line or an overhead power line. Equipment of an electrical system within the meaning of the invention can, for example, be insulators or current-carrying cables.

[0012] The first camera captures UV light. Human-visible light is typically measured in wavelengths between 380 nm and 780 nm, while UV radiation is typically measured in wavelengths between 10 nm and 380 nm (permanent link: https: / / de.wikipedia.org / w / index.php?title=Elektromagnetische s_Spektrum&oldid=178702023).

[0013] Preferably, the first camera can detect so-called weak UV radiation with a wavelength between 230 nm and 380 nm. Even more preferably, the second camera can detect weak UV radiation in the so-called solar-blind wavelength range, i.e., with a wavelength between 240 nm and 280 nm.

[0014] The daylight filter of the first camera blocks visible light for this camera, causing corona discharges to appear, for example, as bright points of light or light sources against an otherwise essentially black background. The daylight filter is therefore a daylight blocking filter that blocks all other daylight wavelengths. Within the scope of the invention, the bright points of light that are not blocked are initially recorded as possible corona discharges. Preferably, a wavelength range of 240–280 nm should be recorded with the second camera, since in this range the sun's UV radiation (a portion of "daylight") is filtered out by the Earth's ozone layer. Thus, everything measured in this wavelength range (assuming a functioning ozone layer) has somehow originated on Earth. It can therefore be assumed that corresponding signals are man-made.However, since only very few photons are received in the limited wavelength range, an image intensifier can be provided together with the daylight filter.

[0015] A conventional computer with appropriate data storage can be used as an evaluation device, for example. From the individual images of the two cameras, all possible corona discharges detected are projected into three-dimensional (3D) space. Using this projection, the statistics of the detected bright points can be determined in order to distinguish actual corona discharges from random noise in the image recordings. Actual corona discharges will repeatedly appear in a temporal series of individual images at the same location in relation to the three-dimensional space. This makes them distinguishable from random noise signals, which may only appear for a short time (in a small number of individual images or even just a single image) and may appear at random locations in a series of images. Depending on the camera system used for the sensor arrangement, calibration measurements orFor tests with artificially induced corona discharges, a suitable threshold for the frequency of detected bright spots must be set in advance, so that an actual corona discharge can be assumed above the threshold. This allows volatile discharge phenomena to be automatically distinguished from genuine discharges. In particular, the frequency of false alarms or false positive discharge detections is significantly reduced, as recurring discharge effects can be separated from spontaneous discharge effects.

[0016] A key advantage of the method according to the invention is that a reference to a three-dimensional geometry is established during the automatic detection of corona discharges. This allows, for example, an exact position for each corona discharge to be detected and stored during an overflight with an aircraft. This allows precise evaluation and, if necessary, precise location-based maintenance or repair of detected damage. This approach enables the use of aircraft, since the prior art solution by Li et al. cannot be used for a mobile airborne platform such as a helicopter or drone, because the aircraft's own movement usually does not follow a predetermined or pre-restricted (linear) shape in the individual images (unlike rail vehicles).

[0017] Compared to analyses using stationary cameras, the invention allows the use of freely moving cameras. This enables the localization of discharges in three-dimensional space, whereas with stationary cameras and without any prior knowledge of the recorded scene, only a restriction to 3D points along the line of sight of the respective camera is possible.

[0018] According to the claimed method according to the invention, a position-determining device is additionally used for the sensor device. The position-determining device can be used to determine the three-dimensional position of each possible corona discharge or bright spot in the UV image. GPS signals, for example, can be evaluated. For example, a timestamp and a precise position of the vehicle are assigned to each individual image, from which a precise position of possible corona discharges can later be calculated in three-dimensional space based on the viewing direction or the line of sight of the sensor arrangement. A position sensor, for example, can be used to determine the viewing direction. Preferably, a position sensor can be used that determines the orientation using the so-called "Inertial Measurement Unit (IMU)" and / or an "Inertial Navigation System (INS)".Such position sensors are known, for example, from the website of the company Vectornav (https: / / www.vectornav.com / support / library / imu-and-ins).

[0019] Within the scope of the present invention, the three-dimensional position of a possible or actual corona discharge already represents a calculation result derived from the position of the vehicle and the viewing angle of the cameras. The sensor arrangement thus allows, if necessary in conjunction with the evaluation device, the position and orientation of the UV camera to be precisely determined.

[0020] In a preferred embodiment of the method according to the invention, a second camera for detecting visible light is additionally used for the sensor arrangement, with the second camera taking images of the system and labeling them with a three-dimensional position. This has the advantage that a human observer can see images (such as e.g. which in Fig. 1 The second camera can view the images shown) and, if necessary, perform a manual review or plausibility check regarding detected corona discharges. The second camera takes ordinary daylight images, to which the bright light points can be placed in a spatial relationship or superimposed.

[0021] In a further development of the aforementioned embodiment, photogrammetry can be performed using the evaluation device based on the images recorded by the second camera. As the second camera moves along the system, multiple images of the same section of the system can be obtained. Alternatively or additionally, additional visible-light cameras can be used to acquire multiple images of the system at any time. Photogrammetry can be used to determine the spatial position and / or three-dimensional shape of the system from the images recorded. The principle of photogrammetry is known, for example, from Wikipedia (permanent link: https: / / de.wikipedia.org / w / index.php?title=Photogrammetrie&ol did=179451745).

[0022] In a preferred embodiment of the method according to the invention, the condition of equipment at whose locations actual corona discharges are detected is checked and / or repaired. This is advantageous because an optimal operating condition of the system can be quickly and safely restored, preventing further damage caused by corona discharges, further radio interference, or even system failures.

[0023] In a further preferred embodiment of the method according to the invention, information on operating equipment at the locations of potential and / or actual corona discharges is provided using a geoinformation system. This is advantageous because the plant operator often has information on the operating status and any known wear and tear or maintenance requirements. This information can be linked to statistics, for example, to detect corona discharges even more reliably.

[0024] In a further preferred embodiment of the method according to the invention, the spatial statistics are formed by assigning a number of entries in a quantized three-dimensional counter reading grid to each possible corona discharge in three-dimensional space. Visual rays in three-dimensional space are defined based on a viewing direction of the sensor array. In the event of an intersection of several visual rays from several temporally offset individual images of the sensor array at the same three-dimensional position, a cluster is entered in the counter reading grid. This concept uses, for example, a so-called "Hough Space," i.e., an extension of Hough's approach described above, related to a 3D space. This embodiment is advantageous because it is reliable.

[0025] The quantized three-dimensional counter reading grid can be understood, for example, as a subdivision of three-dimensional space into many equal-sized cubes. If the position of a corona discharge falls within a specific cube, the counter for that cube is incremented by 1. A person skilled in the art can use a calibration measurement to find a suitable size for the cubes so that the supposed corona discharges appear sufficiently frequently in a specific cube. The quantized three-dimensional counter reading grid is therefore used in the manner of a histogram.

[0026] In a further preferred embodiment of the method according to the invention, the spatial statistics are formed by detecting possible corona discharges in each of the image recordings and projecting them into subsequent image recordings. If a match with a possible corona discharge detected in the subsequent image recordings is found, an increased probability of an actual corona discharge is assumed. If further possible corona discharges are present that cannot be associated with already known bright light points, further possible corona discharges are detected along the corresponding lines of sight. Compared to the previous embodiment with a counter reading grid, a reduction in the memory requirement in the evaluation device is to be expected, which is advantageous.

[0027] In a further preferred embodiment of the method according to the invention, the spatial statistics are created taking into account a previously known three-dimensional model of the plant, so that in the three-dimensional model, a search space for corona discharges is restricted to the immediate vicinity of the plant. For example, in three-dimensional space, only possible corona discharges that are located in the vicinity of the plant, such as an overhead power line, are recorded. Although the corona discharges do not occur directly on physical objects that are measured using conventional methods (lines, insulators, etc.), they are to be expected in the vicinity (e.g. at a distance of a few cm to m) of such objects. The respective backprojections along the lines of sight can thus be correlated with corresponding 3D objects. Instead of the entire line of sight, further processing can be reduced to an area around these objects.This is an advantage because the computational effort required by the evaluation device is significantly reduced. In this embodiment, the evaluation device can find a result more quickly at a given computing speed, or can be used with significantly reduced computing power and thus costs.

[0028] In a further preferred embodiment of the method according to the invention, an aircraft is used as the vehicle. An aircraft in the sense of the invention is any object that is not only two-dimensionally positioned on the Earth's surface. e.g. a road or railway track, but also at altitudes above the earth's surface. The use of an aircraft is a great advantage because facilities can be inspected quickly and reliably by flying over them, even when there are no roads or similar facilities available for inspection from the ground. This is particularly advantageous for overhead line inspections.

[0029] In a further preferred embodiment of the method according to the invention, an airplane, a helicopter or a drone is used as the aircraft.

[0030] In a further preferred embodiment of the method according to the invention, the evaluation device is provided in the vehicle. This is advantageous because the image data from the sensor array can be evaluated during the flight. If actual corona discharges are detected, the coordinates (e.g., GPS data) can be transmitted to the system operator either after the inspection of the system or even immediately via radio data communication.

[0031] In a further preferred embodiment of the method according to the invention, the evaluation device is provided as a central server. This is advantageous because it saves weight and space for an evaluation device on board the vehicle, thus reducing costs. This is particularly advantageous for aircraft. The image data captured by the two cameras can, for example, be compressed and saved in a data storage device. The data storage device is read out after the inspection. Alternatively, the image data can be transmitted immediately to the central server via wireless data communication. The central server can, for example, be implemented as a "cloud application."

[0032] In another preferred embodiment of the method according to the invention, the three-dimensional position is determined using Global Positioning System (GPS) data. This is an advantage because GPS is proven, reliable, and accurate.

[0033] The invention further aims to provide an arrangement with which corona discharges in systems can be detected automatically and reliably.

[0034] The invention solves this problem by an arrangement according to claim 13. Preferred embodiments emerge from claims 13 to 15.

[0035] In a further preferred embodiment of the arrangement according to the invention, the evaluation device is provided in the vehicle.

[0036] In a further preferred embodiment of the arrangement according to the invention, the evaluation device is provided as a central server.

[0037] The same advantages arise for the arrangement according to the invention and its embodiments as explained above for the method according to the invention. It will be clear to those skilled in the art that the individual embodiments described for the method according to the invention, in particular the configurations of the image analysis steps performed by the evaluation device, can also be implemented in the arrangement according to the invention and can be freely combined.

[0038] To better explain the invention, the following schematic representations show Figure 1 shows a first overlay of a single image in the visible light spectrum with a single image in the UV spectrum, and Figure 2 shows a second overlay after the first overlay according to Figure 1 located superposition of a single image in the visible light spectrum with a single image in the UV spectrum, and Figure 3 shows an embodiment of the invention.

[0039] The Figures 1 and 2show published, known images. A mast 1 with cables 3 and operating equipment such as insulators 2 is shown. Numerous white or bright spots are clearly visible, which were detected by the UV camera as UV sources and superimposed on the visible image. These bright spots are possibly corona discharges. Figure 2 The same mast 1 is shown a little later, with no bright spots visible anymore.

[0040] The Figure 3shows a preferred embodiment of the invention. It depicts an installation 17, an overhead line with a first mast 1 and a second mast 4, insulators 2, and cables 3. Cables 3 are present on the masts 1, 4. A drone 16 is used as the vehicle, which guides a sensor arrangement 18 along the installation. The sensor arrangement 18 has a second camera 19 for detecting visible light and a first camera 20 for detecting UV radiation. The first camera 20 additionally has a daylight filter 21 for blocking daylight.

[0041] The sensor array 18 captures images of the system 17 and processes them on board the drone 16 using an evaluation device 22. The images captured by the two cameras 19, 20 are marked with a three-dimensional position, which is determined using a GPS satellite and a positioning device 23. The positioning device 23 assigns a timestamp and a precise position of the drone 16 to each individual image, from which the precise position of possible corona discharges 5-15 can later be calculated in three-dimensional space based on the viewing direction or line of sight 25 of the sensor array 18.

[0042] The evaluation device 22 detects possible corona discharges 5-15 in each individual image of the second camera 20 as bright spots. Based on the respective three-dimensional position of the bright spot, it is transferred into a single three-dimensional space. This allows the information from a time series of images to be analyzed. The evaluation device can create spatial statistics on the frequency of possible corona discharges. Based on the spatial statistics, actual corona discharges 5-8, 10, 13-15 can be identified as stationary and occurring more frequently compared to noise. In the example shown, the corona discharges 5-8, 10, 13-15, which are located in the immediate vicinity of the system 17, were identified in this way. This meansthat actual corona discharges are imaged more frequently in a temporal series of images than random noise, which repeatedly occurs at different locations and is not recognizable at one location for a longer period of time over several individual images.

[0043] A geographic information system (not shown) can provide information on assets 2,3 and the positions of possible 5-15 and / or actual 5-8,10,13-15 corona discharges.

Claims

1. Method for detecting corona discharges (5-15) of a system (17) comprising equipment (2, 3), in which - a sensor assembly (18) is moved by means of a vehicle (16) along the system (17) comprising equipment (2, 3), and - a first camera (20) for registering UV radiation with a daylight filter (21) for blocking out daylight is used for the sensor assembly (18), and - images of the system (17) are recorded by means of the sensor assembly (18), characterized in that - by means of an evaluation device (22), the images from the camera (20) with a three-dimensional position, wherein - by means of the evaluation device (22), possible corona discharges (5-15) in each individual image are detected and transferred into a single three-dimensional space with the aid of the respective three-dimensional position, and in that - by means of the evaluation device (22), a spatial statistic relating to the frequency of the possible corona discharges is compiled, with the aid of which actual corona discharges (5-8, 10, 13-15) are detected as stationary and occurring more frequently in comparison with noise, wherein the sensor assembly (18) additionally uses a position determination device (23), wherein the position determination device determines the three-dimensional position of the respective possible corona discharge (5-15) in the UV image.

2. Method according to Claim 1, characterized in that a second camera (19) for registering visible light is additionally used for the sensor assembly (18), images of the system (17) being recorded with the camera and labelled with a three-dimensional position.

3. Method according to Claim 1 or 2, characterized in that verification of the status and / or repairing of equipment (2, 3), at the positions of which actual corona discharges (5-8, 10, 13-15) are detected, is carried out.

4. Method according to one of the preceding claims, characterized in that information relating to equipment (2, 3) at the positions of possible (5-15) and / or actual (5-8, 10, 13-15) corona discharges is provided by means of a geoinformation system.

5. Method according to one of the preceding claims, characterized in that the spatial statistic is formed by assigning a number of entries in a quantized three-dimensional counter reading array to each possible corona discharge (5-15) in the three-dimensional space, visual lines (25) in the three-dimensional space being established with the aid of a viewing direction of the sensor assembly (18), a cluster being entered in the counter reading array in the case of intersection of a plurality of visual lines from a plurality of time-offset individual images of the sensor assembly (18) at the same three-dimensional position.

6. Method according to one of the preceding claims, characterized in that the spatial statistic is formed by detecting respectively possible corona discharges (5-15) in the images and projecting them into chronologically subsequent images, an increased probability of an actual corona discharge (5-8, 10, 13-15) being assumed in the event of a match with a possible corona discharge detected in the subsequent images.

7. Method according to one of the preceding claims, characterized in that the spatial statistic is compiled while taking into account a previously known three-dimensional model of the system (17), so that a search space for corona discharges (5-15) in the three-dimensional space is restricted to the close proximity of the system (17).

8. Method according to one of the preceding claims, characterized in that an aircraft (16) is used as the vehicle.

9. Method according to Claim 8, characterized in that an aeroplane, a helicopter or a drone (16) is used as the aircraft.

10. Method according to one of the preceding claims, characterized in that the evaluation device (22) is provided in the vehicle.

11. Method according to one of Claims 1 to 9, characterized in that the evaluation device is provided as a central server.

12. Method according to one of the preceding claims, characterized in that the three-dimensional position is respectively determined with the aid of "Global Positioning System" (GPS) data.

13. Assembly for detecting corona discharges of a system comprising equipment, having a vehicle (16) which is configured to move a sensor assembly (18) along a system (17) comprising equipment (2, 3), the sensor assembly (18) comprising a first camera (20) for registering UV radiation, and the first camera (20) comprising a daylight filter (21) for blocking out daylight, and the sensor assembly (18) being configured to record images of the system (17), and the vehicle (16) comprising a position determination device (23), wherein an evaluation device (22) is configured to label the images from the camera (20) with a three-dimensional position, and to detect possible corona discharges (5-15) in each individual image and transfer them into a single three-dimensional space with the aid of the respective three-dimensional position, and to compile a spatial statistic relating to the frequency of the possible corona discharges (5-15), with the aid of which actual corona discharges (5-8, 10, 13-15) are detectable as stationary and occurring more frequently in comparison with noise, wherein the position determination device (23) is suitable for determining the three-dimensional position of the respective possible corona discharge (5-15) in the UV image.

14. Assembly according to Claim 13, characterized in that the sensor assembly (18) comprises a second camera (19) for registering for registering visible light, images of the system (17), recorded by means of the camera, being labellable with a three-dimensional position by means of the position determination device (23).

15. Assembly according to one of Claims 13 to 14, characterized in that the vehicle is an aeroplane, a helicopter or a drone (16).