Systems for monitoring equipment

Abstract

The present invention relates to a system for monitoring equipment, the system comprising: an infrared camera; a processing unit; and an output unit. The infrared camera is configured to acquire a plurality of infrared images of the equipment, wherein the plurality of infrared images comprises a first infrared image and a second infrared image, and the second infrared image is acquired in a time period after the first infrared image. The processing unit is configured to determine the pixel with the highest temperature in the first infrared image, and to determine the pixel with the highest temperature in the second infrared image. The processing unit is configured to determine a first number of pixels in the first infrared image and a second number of pixels in the second infrared image. The processing unit is configured to determine a temperature amplitude. The processing unit is configured to determine a rate of change of temperature. The processing unit is configured to determine that a hot spot exists in the equipment and / or is being formed in the equipment. The output unit is configured to output an indication of a fault in the equipment based on determining that a hot spot exists and / or is being formed.

Description

Technical Field

[0001] The invention relates to a system for monitoring a device such as a switching device or a motor and to a method for monitoring a device. Background Art

[0002] Infrared (IR) images can be used to identify technical problems within electrical equipment (e.g. switchgear) that cause components to overheat. However, while humans are very capable of identifying hot spots, automated systems require specific configuration to correctly identify problems. A common approach is to identify areas of concern for each phase. These areas can then be compared. If one area is significantly hotter than the others, a fault has occurred.

[0003] This approach has several disadvantages. First, the phases must be manually identified in the IR image for each model and rating of switchgear. This is a time-consuming, expensive and error-prone process, as the exact location of the phases in the infrared image depends on many switchgear-specific parameters (panel size, current rating, internal structure) as well as camera-specific parameters (field of view, resolution and optics manufacturing tolerances). Secondly, any thermal activity outside the defined area of ​​interest will be ignored, so any errors will reduce the efficiency of the algorithm and any developing problems in different components will not be detected. Thirdly, modifications to the switchgear during field service missions may affect the identification of the phases in the IR image, and / or any movement or change in the camera field of view alignment may result in a change in the location of the area and lead to incorrect operation.

[0004] There is a need to address these issues. Summary of the invention

[0005] Therefore, it would be advantageous to have an improved system for processing infrared images of electrical equipment into actionable information.

[0006] The objects of the invention are solved by the subject matter of the independent claims, wherein further embodiments are incorporated in the dependent claims. It should be noted that the system is described in some aspects with respect to a switchgear, but is also useful in other electrical systems that may suffer from overheated components, such as motors. These different devices, systems or apparatus are collectively referred to as apparatus hereinafter.

[0007] In one aspect, a system for monitoring a device is provided, the system comprising:

[0008] - Infrared camera;

[0009] - a processing unit; and

[0010] - output unit;

[0011] The infrared camera is configured to acquire a plurality of infrared images of the device, wherein the plurality of infrared images includes a first infrared image and a second infrared image, the second infrared image being acquired at a time period after the first infrared image. The processing unit is configured to determine a pixel having a highest temperature in the first infrared image, and to determine a pixel having a highest temperature in the second infrared image. The processing unit is configured to determine a first number of pixels in the first infrared image whose temperature is within a threshold temperature of a highest temperature in the first infrared image, and to determine a second number of pixels in the second infrared image whose temperature is within a threshold temperature of a highest temperature in the second infrared image. The processing unit is configured to determine a temperature amplitude, including utilizing a highest temperature in the second infrared image and / or utilizing a highest temperature in a different infrared image of the plurality of infrared images that was acquired at a different time than the second infrared image. The processing unit is configured to determine a rate of change of temperature, including utilizing a highest temperature in the second infrared image and utilizing a highest temperature in a different infrared image, and utilizing a duration between the acquisition of the second infrared image and the acquisition of the different infrared image. The processing unit is configured to determine that a hot spot exists in the device and / or that a hot spot is forming in the device, including utilizing a comparison between a first number of pixels and a second number of pixels, and utilizing a temperature amplitude and a rate of change of temperature. The output unit is configured to output an indication of a fault in the device based on determining that a hot spot exists and / or is forming.

[0012] Thus, if a hotspot is forming, the hottest area of ​​the device begins to get hotter than expected. As the hotspot forms, it gets hotter than other areas of the device, and the number of pixels within the temperature range of the hottest pixel actually decreases. At the same time, based on the temperature of the hottest pixel and the rate of change of temperature, it is concluded that a hotspot can be determined to be forming. In this way, for sufficiently high maximum temperatures, a reduction in the number of pixels associated with the rate of change of temperature can determine with high confidence that a hotspot is forming, and that remedial actions, such as shutting down the device or reducing power can be completed before the hotspot causes damage to the device.

[0013] In this way, the change in the number of pixels can be calculated continuously for two temporally adjacent infrared images, and these images are also used to determine the maximum temperature and the rate of change of temperature, all of which is used to determine whether a hot spot exists. However, in order to maximize the difference in the number of pixels, an initial infrared image can be acquired and the number of pixels calculated and compared to the number of pixels in a series of subsequently acquired infrared images. However, for later infrared images in the series, the temporally adjacent images can be used to determine both the maximum temperature and the temperature rate, which, when used together with the ongoing change in the number of pixels, enables a hot spot to be identified as being formed.

[0014] In an example, the different infrared image is a first infrared image.

[0015] In an example, the different infrared image is not the first infrared image, and the different infrared image is acquired at a time between the first infrared image and the second infrared image, or the second infrared image is acquired at a time between the first infrared image and the different infrared image.

[0016] In an example, a time period between acquisition of the first infrared image and the second infrared image is greater than a time period between acquisition of the second infrared image and the different infrared image.

[0017] In an example, the threshold temperature is a fixed temperature offset.

[0018] In an example, the comparison between the first number of pixels and the second number of pixels includes determining a pixel count difference equal to subtracting the second number of pixels from the first number of pixels.

[0019] In an example, determining that a hotspot exists in the device and / or that a hotspot is developing in the device includes comparing a pixel count difference to a threshold pixel count number.

[0020] In an example, the threshold pixel count number is fixed.

[0021] In an example, the threshold pixel count number is a function of a maximum temperature in the first infrared image and / or the second infrared image.

[0022] In an example, determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes analysis of one or more of: a shape of at least one area within determined pixels in the first infrared image and the second infrared image; a size of at least one area within determined pixels in the first infrared image and the second infrared image; a location of at least one area within determined pixels in the first infrared image and the second infrared image.

[0023] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes determining that an area in the second infrared image is smaller than a corresponding area in the first infrared image.

[0024] In other words, for the example of a switchgear having, for example, three current-carrying components that should be equally hot (such as circuit breakers of the same type), then there should typically be three separate hot regions within the threshold temperature range of the hottest pixel. However, when one of these components fails and becomes hot, then when the threshold temperature range is correctly chosen, then the maximum temperatures of the other components will be outside the maximum temperature of the failed component (within the temperature range of the highest temperature). The result is that only one region has a temperature within the highest temperature, rather than three in this example, and this can be used to determine that the component with that highest temperature is faulty.

[0025] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes determining that a number of regions in the second infrared image is less than a number of regions in the first infrared image.

[0026] In an example, determining that a hotspot exists in a device and / or that a hotspot is forming in a device includes utilizing a machine learning algorithm.

[0027] In an example, multiple infrared images are acquired at different times, wherein a processing unit is configured to: determine pixels in each of the multiple infrared images that are associated with a highest temperature in each of the multiple infrared images, wherein the processing unit is configured to: determine a threshold number of pixels in each of the multiple infrared images that are associated with a temperature within a threshold temperature of the highest temperature in each of the multiple infrared images, and wherein determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes: determining a rate of change of the number of threshold pixels over time.

[0028] In a second aspect, a method for monitoring a device is provided, the method comprising:

[0029] Acquiring a plurality of infrared images of the device by an infrared camera, the plurality of infrared images comprising a first infrared image and a second infrared image, the second infrared image being acquired in a time period after the first infrared image;

[0030] determining, by the processing unit, a pixel having a highest temperature in the first infrared image and a pixel having a highest temperature in the second infrared image;

[0031] determining, by the processing unit, a first number of pixels in the first infrared image whose temperature is within a threshold temperature of a maximum temperature of the first infrared image, and determining a second number of pixels in the second infrared image whose temperature is within a threshold temperature of a maximum temperature of the second infrared image;

[0032] determining, by the processing unit, the temperature amplitude, including utilizing a maximum temperature in the second infrared image and / or utilizing a maximum temperature in a different infrared image of the plurality of infrared images acquired at a different time than the second infrared image;

[0033] determining, by the processing unit, a rate of change of temperature, including utilizing a maximum temperature in the second infrared image and utilizing a maximum temperature in the different infrared image, and utilizing a duration between acquisition of the second infrared image and acquisition of the different infrared image;

[0034] Determining, by the processing unit, that a hotspot exists in the device and / or that a hotspot is forming in the device includes utilizing:

[0035] a comparison between the first number of pixels and the second number of pixels; and

[0036] The temperature amplitude and the rate of change of temperature; and

[0037] The output unit is configured to output an indication of a fault in the device based on determining that a hot spot exists and / or is forming.

[0038] The above aspects and examples will be apparent from and elucidated with reference to the embodiments described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Exemplary embodiments will be described below with reference to the following drawings:

[0040] Figure 1 shows a detailed workflow related to an example of a process for determining a fault in a device such as a switchgear;

[0041] Figure 2 The number of pixels in the image showing, for example, a normal (healthy) condition and a fault condition;

[0042] Figure 3 showing changes in the number of pixels within a threshold of a maximum temperature in an image, for example, for a normal (healthy) condition and a fault condition; and

[0043] Figure 4 A representation of a range of maximum temperatures relative to different operating scenarios is shown for normal (healthy) conditions and fault conditions. DETAILED DESCRIPTION

[0044] Figures 1 to 4 The invention relates to a system for monitoring equipment and a method for monitoring equipment.

[0045] In an example, a system for monitoring a device includes an infrared camera, a processing unit, and an output unit. The infrared camera is configured to acquire a plurality of infrared images of the device, wherein the plurality of infrared images include a first infrared image and a second infrared image, the second infrared image being acquired at a time period after the first infrared image. The processing unit is configured to determine a pixel having a highest temperature in the first infrared image, and to determine a pixel having a highest temperature in the second infrared image. The processing unit is configured to determine a first number of pixels in the first infrared image whose temperature is within a threshold temperature of a highest temperature in the first infrared image, and to determine a second number of pixels in the second infrared image whose temperature is within a threshold temperature of a highest temperature in the second infrared image. The processing unit is configured to determine a temperature amplitude, including utilizing a highest temperature in the second infrared image and / or utilizing a highest temperature in a different infrared image among the plurality of infrared images acquired at a different time than the second infrared image. The processing unit is configured to determine a rate of change of temperature, including utilizing a highest temperature in the second infrared image and utilizing a highest temperature in a different infrared image, and utilizing a duration between the acquisition of the second infrared image and the acquisition of the different infrared image. The processing unit is configured to determine that a hotspot exists in the device and / or that a hotspot is forming in the device, including using:

[0046] a comparison between the first number of pixels and the second number of pixels, and

[0047] The temperature amplitude and the rate of change of temperature; and

[0048] The output unit is configured to output an indication of a fault in the device based on determining that a hot spot exists and / or is forming.

[0049] In an example, the temperature amplitude is the highest temperature in the second infrared image.

[0050] In the example, the temperature amplitude is the highest temperature in the different infrared images.

[0051] In an example, the temperature amplitude is an average of a maximum temperature in the second infrared image and a maximum temperature in a different infrared image.

[0052] According to an example, the different infrared image is a first infrared image.

[0053] According to an example, the different infrared image is acquired at a time between the first infrared image and the second infrared image, or the second infrared image is acquired at a time between the first infrared image and the different infrared image.

[0054] According to an example, the time period between the acquisition of the first infrared image and the second infrared image is greater than the time period between the acquisition of the second infrared image and the different infrared image.

[0055] According to an example, the threshold temperature is a fixed temperature offset.

[0056] According to an example, the comparison between the first number of pixels and the second number of pixels comprises determining a pixel count difference equal to subtracting the second number of pixels from the first number of pixels.

[0057] According to an example, determining that a hotspot exists in the device and / or that a hotspot is developing in the device includes comparing a pixel count difference to a threshold pixel count number.

[0058] According to an example, the threshold pixel count number is fixed.

[0059] According to an example, the threshold pixel count number is a function of a maximum temperature in the first infrared image and / or the second infrared image.

[0060] According to an example, determining that a hotspot exists in a device and / or that a hotspot is forming in a device includes analyzing one or more of: a shape of at least one area within determined pixels in a first infrared image and a second infrared image; a size of at least one area within determined pixels in a first infrared image and a second infrared image; a position of at least one area within determined pixels in a first infrared image and a second infrared image.

[0061] According to an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device comprises determining that an area in the second infrared image is smaller than a corresponding area in the first infrared image.

[0062] According to an example, determining that a hot spot is present in the device and / or that a hot spot is forming in the device comprises determining that a number of regions in the second infrared image is smaller than a number of regions in the first infrared image.

[0063] In an example, determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes determining that a number of regions in the second image is one.

[0064] In an example, a device includes two or more components of the same type, which are loaded with substantially the same current; wherein determining that a hot spot exists in the device and / or a hot spot is being formed in the device includes one or more of the following: a shape of a first area within a determined pixel in a second infrared image is different from a shape of a second area within a determined pixel in the second infrared image; a size of the first area within a determined pixel in the second infrared image is different from a size of the second area within a determined pixel in the second infrared image; a position of the area within the determined pixel in the second infrared image is independent of the position of the two or more components; and the number of areas within the determined pixel in the second infrared image is less than the number of the two or more components.

[0065] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device comprises determining that only one region exists within the determined pixel in the second infrared image.

[0066] In an example, determining that a hotspot is present in the device and / or that a hotspot is forming in the device comprises generating at least one binary image wherein determined pixels in the first and second infrared images are given different binary values ​​than remaining pixels in the at least one infrared image.

[0067] According to an example, determining that a hotspot exists in a device and / or that a hotspot is forming in a device includes utilizing a machine learning algorithm.

[0068] In an example, the machine learning algorithm is a trained neural network.

[0069] According to an example, a plurality of infrared images are acquired at different times. The processing unit is configured to: determine a pixel in each of the plurality of infrared images that is associated with a highest temperature in each of the plurality of infrared images. The processing unit is configured to: determine a threshold number of pixels in each of the plurality of infrared images that is associated with a temperature within a threshold temperature of the highest temperature in each of the plurality of infrared images. Determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes: determining a rate of change of the number of threshold pixels over time.

[0070] In an example, determining that a hotspot exists in the device and / or that a hotspot is developing in the device includes comparing a rate of change of a threshold number of pixels over time to a baseline rate of change of a threshold number of pixels over time.

[0071] In an example, the apparatus comprises at least a part of a medium voltage switchgear or comprises an electric machine.

[0072] In an example, the processing unit is configured to predict a temperature at a location of the device associated with a pixel having a highest temperature in the second infrared image, including utilizing a temperature magnitude and a rate of change of the temperature and a correlation. The correlation is a correlation of a plurality of temperature magnitudes and a plurality of temperature rates of change with a plurality of hot spot temperatures. Determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes utilizing the predicted temperature.

[0073] In an example, determining that a hot spot exists in the device and / or that a hot spot is developing in the device includes determining that a predicted temperature exceeds a threshold temperature.

[0074] In an example, the processing unit is configured to select a correlation from a plurality of correlations for different operating scenarios of the device. Each correlation in the plurality of correlations is a correlation of a plurality of temperature amplitudes and a plurality of temperature change rates at a sensor location with a plurality of hot spot temperatures at the location.

[0075] In an example, each of the plurality of correlations is determined experimentally or by simulation.

[0076] In an example, a method for monitoring a device includes:

[0077] Acquiring a plurality of infrared images of the device by an infrared camera, the plurality of infrared images comprising a first infrared image and a second infrared image, the second infrared image being acquired in a time period after the first infrared image;

[0078] determining, by the processing unit, a pixel having a highest temperature in the first infrared image and a pixel having a highest temperature in the second infrared image;

[0079] determining, by the processing unit, a first number of pixels in the first infrared image whose temperature is within a threshold temperature of a maximum temperature of the first infrared image, and determining a second number of pixels in the second infrared image whose temperature is within a threshold temperature of a maximum temperature of the second infrared image;

[0080] determining, by the processing unit, the temperature amplitude, including utilizing a maximum temperature in the second infrared image and / or utilizing a maximum temperature in a different infrared image of the plurality of infrared images acquired at a different time than the second infrared image;

[0081] determining, by the processing unit, a rate of change of temperature, including utilizing a maximum temperature in the second infrared image and utilizing a maximum temperature in a different infrared image, and utilizing a duration between acquisition of the second infrared image and acquisition of the different infrared image;

[0082] Determining, by the processing unit, that a hotspot exists in the device and / or that a hotspot is forming in the device includes utilizing:

[0083] a comparison between a first number of pixels and a second number of pixels; and

[0084] The temperature amplitude and the rate of change of temperature; and

[0085] An indication of a fault in the device is output by the output unit based on determining that a hot spot exists and / or is forming.

[0086] In an example, the temperature amplitude is the highest temperature in the second infrared image.

[0087] In the example, the temperature amplitude is the highest temperature in the different infrared images.

[0088] In an example, the temperature amplitude is an average of a maximum temperature in the second infrared image and a maximum temperature in a different infrared image.

[0089] In an example, the different infrared image is a first infrared image.

[0090] In an example, the different infrared image is not the first image, and the different infrared image is acquired at a time between the first infrared image and the second infrared image, or the second infrared image is acquired at a time between the first infrared image and the different infrared image.

[0091] In an example, a time period between acquisition of the first infrared image and the second infrared image is greater than a time period between acquisition of the second infrared image and the different infrared image.

[0092] In an example, the threshold temperature is a fixed temperature offset.

[0093] In an example, the comparison between the first number of pixels and the second number of pixels includes determining a pixel count difference equal to subtracting the second number of pixels from the first number of pixels.

[0094] In an example, determining that a hotspot exists in the device and / or that a hotspot is developing in the device includes comparing a pixel count difference to a threshold pixel count number.

[0095] In an example, the threshold pixel count number is fixed.

[0096] In an example, the threshold pixel count number is a function of a maximum temperature in the first infrared image and / or the second infrared image.

[0097] In an example, determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes analysis of one or more of: a shape of at least one area within determined pixels in the first infrared image and the second infrared image; a size of at least one area within determined pixels in the first infrared image and the second infrared image; a location of at least one area within determined pixels in the first infrared image and the second infrared image.

[0098] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes determining that an area in the second infrared image is smaller than a corresponding area in the first infrared image.

[0099] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes determining that a number of regions in the second infrared image is less than a number of regions in the first infrared image.

[0100] In an example, determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes determining that a number of regions in the second image is one.

[0101] In an example, a device includes two or more components of the same type, which are loaded with substantially the same current; wherein determining that a hot spot exists in the device and / or a hot spot is being formed in the device includes one or more of the following: a shape of a first area within a determined pixel in a second infrared image is different from a shape of a second area within a determined pixel in the second infrared image; a size of the first area within a determined pixel in the second infrared image is different from a size of the second area within a determined pixel in the second infrared image; a position of the area within the determined pixel in the second infrared image is independent of the position of the two or more components; and the number of areas within the determined pixel in the second infrared image is less than the number of the two or more components.

[0102] In an example, determining that a hot spot exists in the device and / or that a hot spot is forming in the device comprises determining that only one region exists within the determined pixel in the second infrared image.

[0103] In an example, determining that a hotspot is present in the device and / or that a hotspot is forming in the device comprises generating at least one binary image wherein determined pixels in the first and second infrared images are given different binary values ​​than remaining pixels in the at least one infrared image.

[0104] In an example, determining that a hotspot exists in a device and / or that a hotspot is forming in a device includes utilizing a machine learning algorithm.

[0105] In an example, the machine learning algorithm is a trained neural network.

[0106] In an example, multiple infrared images are acquired at different times, and the method includes determining, by a processing unit, pixels in each of the multiple infrared images that are associated with a highest temperature in each of the multiple infrared images, and determining, by the processing unit, a threshold number of pixels in each of the multiple infrared images that are associated with a temperature within a threshold temperature of the highest temperature in each of the multiple infrared images, and determining that a hotspot exists in the device and / or that a hotspot is forming in the device includes determining a rate of change of the number of threshold pixels over time.

[0107] In an example, determining that a hotspot exists in the device and / or that a hotspot is developing in the device includes comparing a rate of change of a threshold number of pixels over time to a baseline rate of change of a threshold number of pixels over time.

[0108] In an example, the apparatus comprises at least a part of a medium voltage switchgear or comprises an electric machine.

[0109] In an example, a method includes predicting, by a processing unit, a temperature at a location of the device associated with a pixel having a highest temperature in a second infrared image, including utilizing temperature amplitudes and rates of change of temperature and correlations, wherein the correlations are correlations of multiple temperature amplitudes and multiple rates of temperature change with multiple hot spot temperatures, and wherein determining that a hot spot exists in the device and / or that a hot spot is forming in the device includes utilizing the predicted temperature.

[0110] In an example, determining that a hot spot exists in the device and / or that a hot spot is developing in the device includes determining that a predicted temperature exceeds a threshold temperature.

[0111] In an example, the method includes selecting, by the processing unit, a correlation from a plurality of correlations for different operating scenarios of the device, wherein each correlation of the plurality of correlations is a correlation of a plurality of temperature amplitudes and a plurality of temperature change rates with a plurality of hot spot temperatures.

[0112] In an example, each of the plurality of correlations is determined experimentally or by simulation.

[0113] Continuing with the drawings, systems and methods for monitoring equipment such as switchgear or motors are described in more detail with respect to specific embodiments, with reference to monitoring switchgear.

[0114] Figure 1A detailed workflow related to an example of processing performed to determine a fault in a switchgear is shown. As shown, at "A", an IR image of the switchgear is acquired by an infrared camera. A preprocessing step may be used to improve the image quality. This may involve noise suppression, compensation for optical effects, edge removal, etc. At "B", the highest temperature point within the image is acquired or identified, for example 35°C. There may be several pixels with exactly the same maximum temperature, and one or two of them may be selected. Then, at "C", all pixels in the image within a threshold temperature range of the highest temperature are selected. For example, if the threshold temperature range is 7°C, all pixels between 28°C and 35°C are selected. Thus, a threshold is used to calculate a range of values ​​that are considered overheated in this particular case. The threshold is the maximum temperature minus a specific value. The value may be a fixed number of degrees Celsius or a percentage of the current average temperature of the image, or a percentage of the maximum temperature of the image, or a value that changes dynamically in degrees Celsius according to the average temperature or maximum temperature of the image. Thus, the threshold may take into account the effects of the ambient air temperature (inside the compartment and / or outside the switchgear), as well as any effects of the compartment type. The results are shown in the following images. Optionally, at "D", the connected regions are identified. At "E", the pixel count, number of regions, shape and / or size of each region are analyzed to determine if this is a fault. If it is determined to be a fault, at "F", the pixel map is overlaid on a visible light image of the switchgear (e.g., switchgear compartment) to identify the location of the fault and / or define which regions belong to which phase of a three-phase system.

[0115] Thus, a threshold is used to identify all camera pixels that are hotter than a threshold. An image with a very strong hot spot will show a rather small area. Without the hot spot, the area would be wider or broader, and in an ideal case would even show different parts of interest as different unconnected areas. The original IR image can be converted into a new binary image consisting of only two colors (e.g., black and white), where one color (e.g., black) is the pixels above the threshold (hot pixels) and the other color (e.g., white) is the pixels below the threshold (cold pixels) - this is just an example, of course the pixels can be reversed, i.e., white is "hot" and black is "cold". Typically, this conversion can produce a mask for further processing of the IR image, for example by a machine learning algorithm such as a trained neural network.

[0116] Therefore, the algorithm applies a dynamic threshold to the image and only returns pixels in the area of ​​the highest temperature. If the image shows a healthy device, the algorithm will separate the background (low temperature) from the foreground (high temperature). Therefore, because the number of pixels within the threshold of the maximum temperature begins to decrease, it can be determined whether a hot spot is forming. Therefore, a fault in the isolation device that causes the hot spot can be identified. In addition, for example, for a three-phase switchgear, the parts of each phase typically exhibit equal current and are heated to equal temperatures due to Joule heating. Therefore, the temperature difference between the same elements of the phase will be less than the threshold temperature range, and therefore all phases will be displayed and the number of pixels remains relatively stable. It is then further determined that a fault exists because the faulty part remains at the highest temperature, but other parts of the phase and other parts of other phases begin to disappear from the temperature threshold range.

[0117] Therefore, the method is to count the number of hot pixels, which is Figure 3 is shown in . As shown, a situation with a strong fault will show a sharp drop in the number of hot pixels as the temperature increases over time. The number of pixels used to determine a hot spot situation can be derived for each scene by counting the pixels belonging to a component of interest. If the number of hot pixels drops below this value, the image shows a hot spot. The number of pixels used to determine a hot spot situation can also be determined by experiment or simulation, or via a machine learning algorithm. Therefore, the number of pixels in the hot area compared to a baseline for the component can be used to indicate whether a fault exists, and the time change in the number of pixels in the threshold area when compared to such a baseline can also be used to indicate that the component is abnormally hot and a fault exists.

[0118] To catch more subtle faults, additional properties can be examined. For example, by taking the number and relative size of different regions, even less obvious hot spots can be detected. Figure 2 , where in the event of a fault, the shape of the pixel area within the threshold changes shape.

[0119] The image (e.g., converted binary image) can be further processed by machine learning. The machine learning model can be trained using typical patterns that show the layout of hot pixels of an image without hot spots, the layout of hot pixels of an image with hot spots on a first specific component, the layout of hot pixels of an image with hot spots on a second component, and so on. The image can be fed to the model and the model classifies it as, for example, "no hot spots", "hot spots on the first component or the second component". Training data can be easily created for each new scene. Subject matter experts can identify components on IR or visible camera images and create typical patterns of black and white images that are expected for healthy images and images with faults in different components. Training data can be created in the office without the need for expensive and time-consuming experiments or simulations. The IR images discussed above can be composed of several separate images or segments of separate images, which are, for example, from different cameras in a panel or queue, or from different time steps.

[0120] However, the inventors have determined that the maximum temperature and the rate of change of its temperature provide additional information that improves the accuracy of fault determination based on the number of pixels within a threshold of the maximum temperature.

[0121] It has been determined that the determined values ​​of T and dT / dt provide values ​​related to the steady state temperature that would occur in the absence of any changes.

[0122] It has been determined that the maximum temperature derivative, when combined with the maximum temperature, produces two distinct clusters for healthy and faulty data.

[0123] This is Figure 4 is shown in Figure 4 Only the two features T and dT / dt are plotted for the hottest pixels. In a particular case, if a temperature of 45°C and a temperature rate of change of 1.1 are determined, it is determined that under steady state conditions, the temperature will not build up to a critical temperature if nothing else happens. However, if a temperature of 65°C and a temperature rate of change of 1.75 are measured, even though the temperature at that point is not critical, it has been determined that without changes, the temperature at that location will build up to a critical temperature under steady state conditions, and remedial action should be taken before that critical temperature builds up. Figure 2 Different lines are shown in , so that the different correlations for different scenarios can be more easily seen. Figure 2In the figure, for each correlation centered on a line, the temperature in the hottest pixel of the IR camera and its rate of rise are shown, where different lines represent several cases. For each case in the lower temperature range, the formation temperature of the hot spot T_HS will be insignificant. However, in the higher temperature range of each correlation, the hot spot temperature T_HS will form to a temperature above the threshold and overheating will occur, and mitigation actions will be required to prevent this from happening.

[0124] Therefore, when this information is combined with the number of pixels within the threshold of the maximum temperature, this actually creates three-dimensional information and provides more cluster discrimination.

[0125] Typically, the threshold between healthy and faulty areas can be defined by considering a constant value for the number of pixels within a threshold temperature. If the number of pixels is below this value, it can be determined that a fault has occurred. On the other hand, if this line is a function of the maximum temperature in the image, a more flexible boundary appears and problems can be identified at an early stage. The exact position of the line can be determined by a support vector machine (SVM) or similar algorithms. Alternatives to SVM are other machine learning classification or clustering algorithms.

[0126] Therefore, a support vector machine or similar ML algorithm can be trained to specify whether a device is healthy or faulty by using the maximum temperature and the number of pixels within a threshold range of the maximum temperature and the temperature derivative as features.

Claims

1. A system for monitoring equipment, the system comprising: Infrared camera; Processing unit; and output unit; wherein the infrared camera is configured to: acquire a plurality of infrared images of the device, wherein the plurality of infrared images include a first infrared image and a second infrared image, the second infrared image being acquired in a time period after the first infrared image; Wherein the processing unit is configured to: determine the pixel with the highest temperature in the first infrared image, and determine the pixel with the highest temperature in the second infrared image; wherein the processing unit is configured to: determine a first number of pixels in the first infrared image having a temperature within a threshold temperature of the maximum temperature of the first infrared image, and determine a second number of pixels in the second infrared image having a temperature within a threshold temperature of the maximum temperature of the second infrared image; wherein the processing unit is configured to: determine the temperature amplitude, including utilizing the maximum temperature in the second infrared image, and / or utilizing the maximum temperature in a different infrared image of the plurality of infrared images acquired at a different time than the second infrared image; wherein the processing unit is configured to: determine a rate of change of temperature, including utilization of the maximum temperature in the second infrared image and utilization of the maximum temperature in the different infrared image, and utilization of a duration between acquisition of the second infrared image and acquisition of the different infrared image; wherein the processing unit is configured to: determine that a hot spot exists in the device and / or that a hot spot is forming in the device, including utilizing a comparison between the first number of pixels and the second number of pixels, and utilizing the temperature amplitude and the rate of change of the temperature; as well as The output unit is configured to output an indication of a fault in the device based on determining that a hot spot exists and / or is forming. The system of claim 1 , wherein the different infrared image is the first infrared image. 3 . The system of claim 1 , wherein the different infrared image is acquired at a time between the first infrared image and the second infrared image, or the second infrared image is acquired at a time between the first infrared image and the different infrared image. 4 . The system of claim 3 , wherein a time period between acquisition of the first infrared image and acquisition of the second infrared image is greater than a time period between acquisition of the second infrared image and acquisition of the different infrared image.

5. The system of any one of claims 1 to 4, wherein the threshold temperature is a fixed temperature offset.

6. The system of any one of claims 1 to 4, wherein the comparison between the first number of pixels and the second number of pixels comprises: Determining a pixel count difference is equal to subtracting the second number of pixels from the first number of pixels.

7. The system of claim 6, wherein determining that the hotspot exists in the device and / or that the hotspot is forming in the device comprises: A comparison of the pixel count difference to a threshold pixel count number.

8. The system of claim 7, wherein the threshold pixel count number is fixed.

9. The system of claim 7, wherein the threshold pixel count number is a function of the highest temperature in the first infrared image and / or the second infrared image.

10. A system according to any one of claims 1 to 4, wherein determining that the hotspot exists in the device and / or that the hotspot is being formed in the device includes analyzing one or more of the following: the shape of at least one area within the determined pixels in the first infrared image and the second infrared image; the size of the at least one area within the determined pixels in the first infrared image and the second infrared image; the position of the at least one area within the determined pixels in the first infrared image and the second infrared image.

11. The system of claim 10, wherein determining that the hotspot exists in the device and / or that the hotspot is forming in the device comprises: It is determined that an area in the second infrared image is smaller than a corresponding area in the first infrared image.

12. The system of claim 10, wherein determining that the hotspot exists in the device and / or that the hotspot is forming in the device comprises: It is determined that the number of regions in the second infrared image is smaller than the number of regions in the first infrared image.

13. The system according to any one of claims 1 to 4, wherein determining that the hotspot exists in the device and / or that the hotspot is forming in the device comprises: Utilization of machine learning algorithms.

14. The system of any one of claims 1 to 4, wherein the plurality of infrared images are acquired at different times, wherein the processing unit is configured to: determine a pixel in each of the plurality of infrared images that is associated with a highest temperature in each of the plurality of infrared images, wherein the processing unit is configured to: determine a threshold number of pixels in each of the plurality of infrared images that is associated with a temperature within a threshold temperature of the highest temperature in each of the plurality of infrared images, and wherein determining that the hotspot exists in the device and / or that the hotspot is forming in the device comprises: A rate of change of the number of threshold pixels over time is determined.

15. A method for monitoring a device, the method comprising: Acquire a plurality of infrared images of the device by an infrared camera, wherein the plurality of infrared images include a first infrared image and a second infrared image, wherein the second infrared image is acquired in a time period after the first infrared image; determining, by a processing unit, pixels having the highest temperature in the first infrared image and pixels having the highest temperature in the second infrared image; determining, by the processing unit, a first number of pixels in the first infrared image having a temperature within a threshold temperature of the maximum temperature of the first infrared image, and determining a second number of pixels in the second infrared image having a temperature within a threshold temperature of the maximum temperature of the second infrared image; determining, by a processing unit, a temperature amplitude, including utilizing the maximum temperature in the second infrared image and / or utilizing the maximum temperature in a different infrared image of the plurality of infrared images acquired at a different time than the second infrared image; determining, by the processing unit, a rate of change of temperature, including utilizing the maximum temperature in the second infrared image and utilizing the maximum temperature in the different infrared image, and utilizing a duration between acquisition of the second infrared image and acquisition of the different infrared image; Determining, by the processing unit, that a hotspot exists in the device and / or that a hotspot is forming in the device comprises utilizing: a comparison between the first number of pixels and the second number of pixels; as well as the temperature amplitude and the rate of change of the temperature; as well as An indication of a fault in the device is output by the output unit based on determining that a hot spot exists and / or is forming.

Images