阅读说明：本技术 用于识别具有运行设备的设施的电晕放电的方法和组件 (Method and assembly for detecting corona discharges in installations having operating devices ) 是由 J·A·比克鲍尔 O·克勒 于 2019-08-21 设计创作，主要内容包括：用于识别具有运行设备的设施的电晕放电的方法和相应的组件,其中-借助于交通工具(16)沿着具有运行设备(2、3)的设施(17)引导传感器组件(18),以及-第一相机(20)被用于传感器组件(18)以获取UV辐射,该第一相机具有用于阻挡日光的日光过滤器(21),以及-借助于传感器组件(18)记录设施(17)的图像,其中-借助于评估装置(22)为相机(20)的图像标记三维位置,以及-借助于评估装置(22),在每个单独图像中检测可能的电晕放电(5至15),并且基于相应的三维位置将可能的电晕放电变换到单个三维空间中,并且-借助于评估装置(22),创建关于可能的电晕放电的频率的空间统计,基于该空间统计,实际的电晕放电(5至8、10、13至15)被识别为与噪声相比地点固定且更频繁出现。(Method and corresponding assembly for identifying corona discharges of a facility with operating equipment, wherein-a sensor assembly (18) is guided along the facility (17) with operating equipment (2, 3) by means of a vehicle (16), and-a first camera (20) is used for the sensor assembly (18) for acquiring UV radiation, which first camera has a daylight filter (21) for blocking daylight, and-images of the facility (17) are recorded by means of the sensor assembly (18), wherein-a three-dimensional position is marked for the images of the camera (20) by means of an evaluation device (22), and-possible corona discharges (5 to 15) are detected in each individual image by means of the evaluation device (22) and transformed into a single three-dimensional space on the basis of the respective three-dimensional position, and-by means of the evaluation device (22), spatial statistics are created about the frequency of possible corona discharges, on the basis of which the actual corona discharges (5 to 8, 10, 13 to 15) are identified as occurring stationary and more frequently than the noise.)

1. A method for detecting a corona discharge (5 to 15) of a plant (17) having an operating device (2, 3),

-guiding a sensor assembly (18) by means of a vehicle (16) along the installation (17) with the operating devices (2, 3), and

-a first camera (20) is used for the sensor assembly (18) to acquire UV radiation, the first camera having a daylight filter (21) for blocking daylight, and

-recording a plurality of images of the facility (17) by means of the sensor assembly (18),

it is characterized in that the preparation method is characterized in that,

-marking a three-dimensional position for said image of said camera (20) by means of an evaluation device (22), and

-detecting possible corona discharges (5 to 15) in each individual image by means of the evaluation device (22) and transforming the possible corona discharges into a single three-dimensional space based on the respective three-dimensional positions, and

-creating, by means of the evaluation device (22), spatial statistics regarding the frequency of the possible corona discharges, on the basis of which the actual corona discharges (5 to 8, 10, 13 to 15) are identified as stationary and occurring more frequently than noise.

2. Method according to claim 1, characterized in that a second camera (19) is additionally used for the sensor assembly (18) to acquire visible light, wherein images of the installation (17) are recorded with the camera and a three-dimensional position is marked for the images.

3. Method according to claim 1 or 2, characterized in that the status of the operating device (2, 3) and/or the maintenance of the operating device (2, 3) is verified, the actual corona discharge (5 to 8, 10, 13 to 15) being identified at the location of such operating device.

4. Method according to any of the preceding claims, characterized in that information about the operating devices (2, 3) at the location of a possible corona discharge (5 to 15) and/or the location of an actual corona discharge (5 to 8, 10, 13 to 15) is provided by means of a geographical information system.

5. The method according to any one of the preceding claims, characterized in that the spatial statistics are formed by assigning to each possible corona discharge (5 to 15) in a three-dimensional space a plurality of entries in a quantized three-dimensional counter state grid, wherein a plurality of lines of sight (25) is determined in the three-dimensional space based on a viewing direction of the sensor assembly (18), wherein frequencies are filled into the counter state grid in case a plurality of lines of sight from temporally offset individual images of the sensor assembly (18) at the same three-dimensional position intersect.

6. Method according to any of the preceding claims, characterized in that the spatial statistics are formed by detecting possible corona discharges (5 to 15) in the images and projecting the possible corona discharges into temporally subsequent images, respectively, wherein the probability of identifying one actual corona discharge (5 to 8, 10, 13 to 15) is increased in case of matching a possible corona discharge identified in a subsequent image.

7. Method according to any of the preceding claims, characterized in that the spatial statistics are created taking into account a previously known three-dimensional model of the installation (17), such that in the three-dimensional space the search space for corona discharges (5 to 15) is limited to the vicinity of the surroundings of the installation (17).

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

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

10. The method according to any one of the preceding claims, characterized in that the evaluation device (22) is arranged in the vehicle.

11. The method according to any one of claims 1 to 9, characterized in that the evaluation means is arranged as a central server.

12. The method according to any of the preceding claims, characterized in that the three-dimensional position is separately found based on "global positioning system" (GPS) data.

13. An assembly for identifying corona discharges of a facility having an operating device, having:

a vehicle (16) designed to guide a sensor assembly (18) along a facility (17) with operating devices (2, 3), wherein

The sensor assembly (18) has a first camera (20) for capturing UV radiation, and the first camera (20) has a daylight filter (21) for blocking daylight, and

wherein the sensor assembly (18) is designed to record a plurality of images of the installation (17), an

Wherein the vehicle (16) has a positioning device (23),

it is characterized in that the preparation method is characterized in that,

an evaluation device (22) is designed to:

marking the image of the camera (20) with a three-dimensional position, and

detecting possible corona discharges (5 to 15) in each individual image and transforming the possible corona discharges into a single three-dimensional space based on the respective three-dimensional positions, and

spatial statistics are created about the frequency of the possible corona discharges (5 to 15), on the basis of which the actual corona discharges (5 to 8, 10, 13 to 15) can be identified as occurring stationary and more frequently than the noise.

14. Assembly according to claim 13, characterized in that the sensor assembly has a second camera (19) for acquiring visible light, wherein by means of the positioning device (23) a three-dimensional position can be marked for the image of the installation (17) recorded with the camera.

15. The assembly according to any one of claims 13 to 14, characterized in that the vehicle is an airplane, a helicopter or a drone (16).

Technical Field

The invention relates to a method according to the preamble of claim 1 and an assembly according to the preamble of claim 13.

Background

European patent application 17161027.2, which was not published to date, 2017, 03, 15, entitled "method and assembly for condition monitoring of a facility having operating equipment", discloses a method for condition monitoring of a facility having operating equipment, in which overview data are acquired by means of a first vehicle having an overview sensor assembly for optically detecting the facility, and the operating equipment is identified in the overview data by means of an evaluation device, and the position of the operating equipment is determined taking into account the position of the first vehicle,

it is characterized in that the preparation method is characterized in that,

by means of a second vehicle having a detail camera, a detail image of the operating device is generated, which detail camera is directed at a corresponding position of the operating device. For example, only one aircraft, such as a drone or a helicopter, is used in order to identify poles and insulators by means of overview cameras when flying over an overhead line, to determine the position of the insulators and then to obtain high-resolution images of the insulators by means of detail cameras. In this way, a damaged insulator can be identified simply and reliably.

Corona discharge at overhead wires or other high voltage infrastructure is a known but undesirable physical phenomenon described in detail, for example, in wikipedia (permanent link: https:// de. wikipedia. org/w/index. phitlite. koronaentaeld & oldid. 173331289). In combination with the nitrogen content of the ambient air, acidic products can be formed by corona discharge which attack the surface of the high-voltage fitting. In addition to this, corona discharge has other undesirable side effects, such as interference with the radio frequency band. In order to prevent corona discharges, protective fittings, so-called corona rings, are mounted on the components (e.g. insulators). For example, corona rings are described in wikipedia (permanent link: https:// de. wikipedia. org/w/index. phitlite. korronaring & oldid. 171645402).

In general, corona discharges (especially under the influence of sunlight) are not visible to the human eye. Thus, a camera for the Ultraviolet (UV) frequency range is used, for example DAYCOR, a product of OFIL corporation (known from the website http:// www.ofilsystems.com/products). Such cameras are equipped with an image intensifier that can make a single light quantum visible. In addition, a sunlight blocking filter is installed to minimize the influence of sunlight. A number of preprocessing steps are known in the prior art by which the raw signal is converted into an image that is more easily understood by a human observer. This is done, for example, by: images of a facility (e.g., an air wire or overhead line for a railway vehicle) in the visible spectrum of light are superimposed with images in the UV spectrum of light. The human assessment engineer is responsible for assessing whether the displayed discharge represents a relevant corona.

From the imaging of overhead lines with a corona camera as described previously, some background noise is typically observed in the recorded signal. Although a plurality of discharges have been recorded in the individual images, these discharges are no longer visible in the temporally delayed individual images. There are few methods for automatically evaluating such images in the prior art, in particular in the context of a location-dependent detection of corona effects in the case of a moving recording system.

It is known that the publication "An Automatic Corona-discharge Detection System for Railwalls Based Solar-blue Ultraviolet Detection" by Li et al, Current Optics and Photonics, volume I, phase 3, 6 months 2017, page 196-. In the described solution, the linear movement of the detected corona point is detected directly in the image space, in a manner tailored to the rail vehicle, wherein the concept of the so-called "Hough transform" (in 2D) is applied. Here, the following facts are utilized: as the rail vehicle moves past it on a linear track, corona discharges in multiple successive images appear as bright spots moving along a straight line. The mathematical method of the Hough transform is described in wikipedia (permanent link: https:// de. wikipedia. org/w/index. phitlite. Hough-Transformation & old. 165672024). Another system for identifying corona discharges in the case of rail vehicles is the DayCor rail system from OFILSYSTEMS, which is known from the website http:// www.ofilsystems.com/products.

Disclosure of Invention

It is an object of the invention to provide a method with which a corona discharge at a facility can be determined automatically and reliably.

The invention achieves this object by a method according to claim 1.

The facility may be, for example, an electrical facility, such as an air wire or overhead line. The operating device of an electrical installation in the sense of the present invention can be, for example, an insulator or a current-carrying cable.

The ultraviolet light is recorded by the first camera. Typically, the wavelength of light visible to humans is between 380nm and 780nm, and the wavelength of UV radiation is typically between 10nm and 380nm (permanent link: https:// de. wikipedia. org/w/index. ptititle. elektromagnetics _ Spektrum & oldi. 178702023).

Preferably, so-called weak UV radiation with a wavelength between 230nm and 380nm can be detected by the first camera. Even more preferably, weak UV radiation in the so-called solar-blind wavelength range (i.e. having a wavelength between 240nm and 280 nm) can be detected by the second camera.

The daylight filter of the first camera functions to block visible light for the camera and the corona discharge appears in front of an otherwise substantially black background, e.g. as a bright light spot or light source. Thus, the daylight filter is a daylight blocking filter that blocks all other daylight wavelengths. Within the scope of the invention, an unobstructed bright light spot is first acquired as a possible corona discharge. Preferably, a wavelength range of 240nm to 280nm should be acquired with the second camera, since in this range the UV radiation of the sun ("part of the sunlight") is filtered out by the earth's ozone layer. Therefore, everything measured in this wavelength range (assuming that the ozone layer is functioning properly) originates from the earth. It can therefore be assumed that the corresponding signals are artificial. However, since only few photons can be received in a limited wavelength range, an image intensifier may also be provided together with a daylight filter.

For example, a conventional computer device with a corresponding data memory can be used as evaluation device. From the separate images of the two cameras, all the acquired possible corona discharges are projected into a three-dimensional (3D) space. Based on this projection, statistics of the identified bright spots can be derived in order to distinguish the actual corona discharge from the random noise of the image. The actual corona discharge always appears repeatedly at the same point in a chronological sequence of individual images, with respect to the three-dimensional space. This makes it possible to distinguish them from random noise signals, which may only occur for a short time (in a few individual images or even in only one individual image) and which may occur at random locations in a series of images. Depending on the camera system used for the sensor assembly, a suitable threshold value for the frequency of the identified bright spots may be determined beforehand by testing or calibration measurements with artificially generated corona discharges, so that above this threshold value for the frequency may result from the actual corona discharge. This makes it possible to automatically distinguish an unstable discharge phenomenon from a true discharge. Since repeated discharge effects can be distinguished from spontaneous discharge effects, the frequency of false alarms or of incorrectly detected discharges is particularly greatly reduced.

A decisive advantage of the method according to the invention is that, in the automatic detection of the corona discharge, a connection is established with respect to the three-dimensional geometry. Thus, for example, when flying over from above with an aircraft, the exact position can be identified and stored for each corona discharge. This allows for accurate assessment and optionally for accurately located maintenance or repair of the identified damage. This solution enables the use of an aircraft, since in the case of mobile airborne platforms (such as helicopters or drones), the solutions of Li et al known in the art cannot be used, since the motion of the aircraft itself (in contrast to rail vehicles) does not generally follow a predetermined or pre-limited (linear) form in the individual images.

The invention allows the use of a freely moving camera, compared to an evaluation with a still camera. This makes it possible to localize the discharge in three-dimensional space, while, for example, using a still camera and without further prior knowledge about the recorded scene, only a limitation of the 3D points along the line of sight of the respective camera can be achieved.

In a preferred embodiment of the method according to the invention, a positioning device is also used for the sensor device. With the positioning device, the three-dimensional position of the corresponding possible corona discharge or the three-dimensional position of the bright spot in the UV image can be found. For example, GPS signals may be evaluated. For example, each individual image is assigned a time stamp and an accurate position of the vehicle, from which an accurate position of a possible corona discharge can then be calculated in three-dimensional space based on the viewing direction or line of sight of the sensor assembly. For example, a position sensor may be used to determine the viewing direction. Preferably, position sensors may be used which determine the orientation based on so-called "inertial measurement units" (IMUs) and/or "inertial navigation systems" (INS). Such location sensors are known, for example, from the website of Vectornav corporation (https:// www.vectornav.com/support/library/imu-and-ins).

Thus, within the scope of the present invention, a possible or actual three-dimensional position of the corona discharge already refers to a calculation result obtained from the position of the vehicle and the perspective of the camera. Thus, the sensor assembly enables accurate determination of the position and orientation of the UV camera, optionally in combination with an evaluation device.

In a preferred embodiment of the method according to the invention, for the sensor assembly, a second camera is also used for acquiring visible light, wherein an image of the installation is recorded with the second camera and a three-dimensional position is marked for the image. This has the advantage that a human observer can see an image (e.g. the image shown in fig. 1) and optionally a manual verification or plausibility check can be performed with respect to the identified corona discharge. The second camera records a conventional daylight image, relative to which the bright spots can be placed in a spatial relationship or superimposed.

In a further development of the aforementioned embodiment, the photogrammetry can be carried out by means of an evaluation device on the basis of the image of the second camera. Here, multiple images of the same section of the facility may be obtained as the second camera moves along the facility. Alternatively or additionally, other cameras for visible light may also be used, so that multiple images of the facility can be obtained at any time. Using photogrammetry, the spatial position and/or three-dimensional shape of the facility can be inferred from the images of the facility. The principle of photogrammetry is known, for example, from the wikipedia (permanent link: https:// de.

In a preferred embodiment of the method according to the invention, the state of the operating device at the location of which the actual corona discharge is detected is verified and/or the operating device is repaired. This is advantageous because the optimum operating state of the installation can be restored quickly and reliably, which avoids further damage due to corona discharges, further radio interference or even an installation failure.

In a further preferred embodiment of the method according to the invention, information about the operating device at the location of the possible and/or actual corona discharge is provided by means of a geographical information system. This is an advantage, since information about the operating state and optionally known wear or maintenance requirements is usually present at the operator of the facility. For example, such information may be correlated with statistics to more reliably identify corona discharges.

In a further preferred embodiment of the method according to the invention, the spatial statistics are formed by assigning a plurality of entries in a quantized three-dimensional counter state grid for each possible corona discharge in a three-dimensional space, wherein a plurality of lines of sight are determined in the three-dimensional space on the basis of the viewing direction of the sensor assembly, wherein the frequencies are populated into the counter state grid in the event of the intersection of a plurality of lines of sight from a plurality of individual images of the sensor assembly at the same three-dimensional position, which are offset in time. This concept uses, for example, the so-called "Hough Space", i.e. an extension of the previously described Hough scheme with respect to 3D Space. This embodiment is advantageous because it is reliable. Here, a quantized three-dimensional counter state grid is to be understood as, for example, subdividing a three-dimensional space into a number of equally sized cubes. If the location of the corona discharge falls within a particular cube, the counter for that cube is incremented by 1. A person skilled in the art can find a suitable cube size by means of calibration measurements so that the assumed corona discharge occurs sufficiently frequently in a particular cube. Thus, the quantized three-dimensional counter state grid is used in a histogram manner.

In a further preferred embodiment of the method according to the invention, the spatial statistics are formed by detecting possible corona discharges in the images and projecting the possible corona discharges into temporally subsequent images, wherein the probability of an actual corona discharge is assumed to increase if it matches a possible corona discharge identified in a subsequent image. If there are possible corona discharges that cannot be associated with a known bright light spot, further possible corona discharges are acquired along the respective line of sight. Advantageously, a reduced memory requirement in the evaluation device can be expected compared to the previously described embodiment with a counter status grid.

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 installation, so that in the three-dimensional model the search space for the corona discharge is limited to the vicinity of the surroundings of the installation. For example, only possible corona discharges located near a facility (e.g., overhead electrical line) are acquired in three-dimensional space. Although corona discharge does not occur directly at physical objects (wires, insulators, etc.) measured using conventional methods, it is still expected to occur in the vicinity of these objects (e.g., in distances of several centimeters to several meters). Thus, a respective back projection along the line of sight may be associated with the corresponding 3D object. Instead of the entire line of sight, further processing may be reduced to the area around these objects. This is an advantage since the computational cost of the evaluation device is greatly reduced. In this embodiment, the evaluation device can find the result faster at a given calculation speed, or can be used with significantly reduced calculation power and thus at significantly reduced cost.

In a further preferred embodiment of the method according to the invention, an aircraft is used as vehicle. An aircraft in the sense of the present invention is any object that travels not only in two dimensions on the earth's surface (e.g. on roads or rails), but also in altitude above the earth's surface. The use of an aircraft is very advantageous because the facility can be quickly and reliably inspected by flying over even when there is no road or the like on the ground for inspection. This is particularly advantageous for overhead line inspection.

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

In a further preferred embodiment of the method according to the invention, the evaluation device is arranged in the vehicle. This is an advantage since the evaluation of the image data of the sensor assembly can already be performed during flight. If an actual corona discharge is recognized, the coordinates (e.g., GPS data) can be transmitted by data communication to the operator of the facility by radio after the end of the inspection of the facility or even immediately.

In a further preferred embodiment of the method according to the invention, the evaluation device is provided as a central server. This is an advantage, since weight and space for the evaluation device can be saved on the vehicle, thus saving costs. This is advantageous in particular for aircraft. The image data acquired by the two cameras may be compressed, for example, and stored in a data memory. After the end of the check, a read of the data memory is performed. Alternatively, the image data may be sent immediately to the central server by radio by means of data communication. The central server may for example be designed as a "cloud application".

In a further preferred embodiment of the method according to the invention, the three-dimensional position is determined on the basis of "global positioning system" (GPS) data. This is advantageous because GPS is long-tested, reliable and accurate.

It is a further object of the invention to provide an assembly with which a corona discharge at a facility can be determined automatically and reliably.

The invention achieves this object by means of an assembly as claimed in claim 13. Preferred embodiments can be derived from claims 13 to 15.

In a further preferred embodiment of the assembly according to the invention, the evaluation device is arranged in the vehicle.

In a further preferred embodiment of the assembly according to the invention, the evaluation means is provided as a central server.

With the assembly according to the invention and its embodiments, the same advantages as previously set forth for the method according to the invention are correspondingly obtained. It is clear to the person skilled in the art that the embodiments described for the method according to the invention, in particular the design of the image analysis step carried out by means of the evaluation device, can also be implemented in the assembly according to the invention and can be freely combined.

Drawings

To better explain the invention, the invention is shown in schematic form,

fig. 1 shows a first overlay of an individual image in the visible spectrum and an individual image in the UV spectrum; and is

Fig. 2 shows a second superimposition of the individual images in the visible spectrum and the individual images in the UV spectrum, which are located temporally after the first superimposition according to fig. 1; and is

Fig. 3 shows an embodiment of the invention.

Detailed Description

Fig. 1 and 2 show known images that have been disclosed. A utility pole 1 is shown with a cable 3 and running equipment (e.g., insulators) 2. Many white or bright spots can be clearly seen, which are acquired as UV source by means of a UV camera and superimposed with the visible image. These bright spots may be corona discharges. The same pole 1 is shown later in fig. 2, where the bright spots are no longer recognizable.

Fig. 3 shows a preferred embodiment of the invention. A facility 17, an overhead line with a first pole 1 and a second pole 4, an insulator 2, and a cable 3 are shown. Cables 3 are present at poles 1, 4. Using the drone 16 as a vehicle, the drone 16 guides the sensor assembly 18 along the facility. The sensor assembly 18 includes a second camera 19 for acquiring visible light and a first camera 20 for acquiring UV radiation. The first camera 20 also has a daylight filter 21 for blocking daylight. Images of the facility 17 are recorded with the sensor assembly 18 and processed with the evaluation device 22 on the drone 16. The images of the two cameras 19, 20 are marked with three-dimensional positions, which are determined by means of GPS satellites and a positioning device 23. With the positioning device 23, each individual image is assigned a time stamp and an exact position of the drone 16, from which the exact position of the possible corona discharges 5 to 15 can then be calculated in three-dimensional space based on the viewing direction or line of sight 25 of the sensor assembly 18.

The evaluation device 22 identifies possible corona discharges 5 to 15 in each individual image of the second camera 20 as bright spots. The bright spots are transformed into a single three-dimensional space based on their respective three-dimensional positions. This enables analysis of information from a time series of images. The evaluation means may create spatial statistics related to the frequency of possible corona discharges. Based on spatial statistics, the actual corona discharges 5 to 8, 10, 13 to 15 can be identified as stationary in place and occurring more frequently than noise. In the example shown, the corona discharges 5 to 8, 10, 13 to 15 in the immediate vicinity of the installation 17 have been identified in this way. That is, the actual corona discharge is imaged more frequently in a chronological sequence of images than random noise, which always recurs elsewhere and cannot be recognized here at one location for a longer time on individual images.

A geographical information system (not shown) may provide information about the location of the operating devices 2, 3 and possibly the corona discharges 5 to 15 and/or the location of the actual corona discharges 5 to 8, 10, 13 to 15.