Power equipment heat rapid positioning method based on semantic segmentation and topology reconstruction

Abstract

The application relates to a power equipment overheat fault rapid positioning method based on semantic segmentation and topology reconstruction. After power equipment is labeled according to categories, the same phase recognition of the power equipment is realized by using connected region analysis, whether the areas and abscissas of three-phase power equipment of all categories are equal difference series is calculated and judged, if the areas and abscissas of three-phase power equipment of all categories are equal difference series, the same phase power equipment of all categories is inferred according to the largest area phase and the smallest or largest abscissa phase in the power equipment of all categories, a local electrical topology diagram is formed, the same phase connection relationship of the power equipment of all categories in the local electrical topology diagram is verified whether correct by comparing the local electrical topology diagram with a transformer substation main wiring diagram; thus, the connection relationship topology reconstruction of the power equipment is realized, and the fault positioning is realized based on the local electrical topology diagram and the transformer substation main wiring diagram. The application can rapidly realize the fault correlation and calculation of the same phase equipment, provides a pre-diagnosis basis for fault diagnosis, and is convenient for rapid fault positioning.

Description

Technical Field

[0001] This invention relates to a method for rapid localization of heat generation in power equipment based on semantic segmentation and topology reconstruction, belonging to the field of image processing for power equipment in substations. Background Technology

[0002] In recent years, many monitoring technologies have been actively promoted and applied. Among these, infrared thermal imaging technology is favored due to its advantages such as no power outages, no sampling, and no disassembly. However, infrared image detection has limited ability to analyze and judge defects, thus restricting the level of intelligent detection equipment status. To improve the status of intelligent detection equipment, more intelligent image processing methods should be used to segment infrared images and extract the necessary power equipment areas for subsequent fault location and early warning. Since the color information in infrared images corresponds to the temperature information of the equipment, it can serve as an important basis for early warning of power equipment faults.

[0003] Current power equipment fault analysis and processing mainly relies on comparing fault images with normal images, using shape recognition to achieve fault diagnosis. However, it fails to consider the interrelationships between individual devices. In reality, equipment clusters are interconnected and have interrelationships. Exploring these interrelationships can facilitate rapid and accurate fault analysis. Summary of the Invention

[0004] In view of the above-mentioned technical problems, the present invention provides a method for rapid localization of heat generation in power equipment based on semantic segmentation and topology reconstruction.

[0005] The technical solution adopted in this invention is as follows: A method for rapid localization of heat generation in power equipment based on semantic segmentation and topology reconstruction, comprising the following steps:

[0006] Step 1: Use semantic segmentation to segment the power equipment in the infrared image, and label different categories of power equipment in the infrared image with different colors;

[0007] Step 2: Binarize the infrared image after semantic segmentation, and use connected component analysis to label the power equipment with different phase sequences according to the 1st, 2nd and 3rd phases respectively, so as to realize the same phase identification of the power equipment;

[0008] Step 3: Calculate the area of ​​similar power equipment in each phase, and solve for the arithmetic sequence of the areas of adjacent phases of similar power equipment;

[0009] Step 4: Find the abscissa of each phase of the same type of power equipment, and solve for the arithmetic sequence of the abscissas of adjacent phases of the same type of power equipment;

[0010] Step 5: If the area and x-coordinate of all three-phase electrical equipment form an arithmetic sequence, then based on the phase with the largest area and the smallest or largest x-coordinate among all types of electrical equipment, deduce the electrical equipment of the same phase and form a local electrical topology diagram. Compare the local electrical topology diagram with the substation main wiring diagram to verify whether the same-phase connection relationship of all types of electrical equipment in the local electrical topology diagram is correct. If correct, combine the local electrical topology diagram and the infrared image to form a power equipment topology association diagram. If incorrect, return to step 1 to perform the next set of infrared image analysis.

[0011] Step 6: Based on the fault diagnosis results of the infrared image, mark the faulty phase power equipment in the verified local electrical topology diagram, and mark the faulty phase power equipment in the substation main wiring diagram according to the matching relationship between the local electrical topology diagram and the substation main wiring diagram, thereby realizing the rapid location of the faulty phase power equipment.

[0012] Further preferred, the connected component analysis uses a two-pass algorithm. The first pass assigns a label to all pixels in the infrared image and records which connected component each label belongs to. The second pass marks each pixel as belonging to its connected component.

[0013] Further optimization involves step 3, where the area of ​​the power equipment in phases 1, 2, and 3 is obtained using the Breadth-First Search (BFS) algorithm. Then, based on the arithmetic sequence determination formula, it is calculated whether the area of ​​the corresponding power equipment forms an arithmetic sequence, i.e., whether it simultaneously satisfies the following formula:

[0014]

[0015]

[0016] Where n represents the category of power equipment, S n1 S n2 S n3 These represent the areas of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let σ be the average area of ​​the nth type of three-phase electrical equipment. n Let be the area standard deviation of the nth type of electrical equipment.

[0017] Further optimization involves step 4, where the abscissas of the first, second, and third phase power equipment are obtained using the gray-scale centroid method. The abscissas of the power equipment are then used to determine whether they form an arithmetic sequence, i.e., whether they simultaneously satisfy the following formula:

[0018]

[0019]

[0020] Where, x n1 xn2 x n3 These represent the x-coordinates of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let γ be the average abscissa of the nth type of three-phase electrical equipment. n Let be the standard deviation of the x-axis of the nth type of power equipment.

[0021] This invention categorizes electrical equipment and uses connected component analysis to identify the same-phase power equipment. It calculates and determines whether the area and x-coordinate of each of the three-phase power equipment types form an arithmetic progression. If they do, it infers the same-phase power equipment by identifying the phase with the largest area and the smallest or largest x-coordinate among all types of power equipment, forming a local electrical topology diagram. This local electrical topology diagram is compared with the substation main wiring diagram to verify the correctness of the same-phase connections of the power equipment. This achieves topology reconstruction of the power equipment connections and enables fault location based on the local electrical topology diagram and the substation main wiring diagram. This invention creatively introduces topology reconstruction into power equipment early warning, making it easier to locate faulty phases within power equipment. Attached Figure Description

[0022] To gain a more complete understanding of the invention and its advantages, reference will now be made to the following description taken in conjunction with the accompanying drawings, wherein:

[0023] Figure 1 This is a flowchart of the method of the present invention;

[0024] Figure 2 This is a schematic diagram of the semantic segmentation results.

[0025] Figure 3 This is the main wiring diagram of the substation;

[0026] Figure 4 This is the power equipment topology diagram of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0028] like Figure 1 As shown, a method for rapid localization of heat generation in power equipment based on semantic segmentation and topology reconstruction includes the following steps:

[0029] Step 1: Use semantic segmentation to segment the power equipment in the infrared image, and label different categories of power equipment in the infrared image with different colors.

[0030] Step 2: Binarize the semantically segmented infrared image and use connected component analysis to label power equipment with different phase sequences according to phases 1, 2, and 3, respectively, to achieve same-phase identification of power equipment. The connected component analysis uses a two-pass algorithm. The first pass assigns a label to all pixels in the infrared image and records which connected component each label belongs to. The second pass labels each pixel with its corresponding connected component. For example, if a semantically segmented infrared image contains current transformers and circuit breakers, the current transformers for phases 1, 2, and 3 are labeled CT1, CT2, and CT3, respectively; and the circuit breakers for phases 1, 2, and 3 are labeled QF1, QF2, and QF3, respectively. Figure 2 As shown.

[0031] Step 3: Calculate the area of ​​similar power equipment in each phase, and solve for the arithmetic sequence of the areas of adjacent phases of similar power equipment;

[0032] The area of ​​the power equipment in phases 1, 2, and 3 is obtained using the breadth-first search (BFS) algorithm. Based on the arithmetic sequence determination formula, it is calculated whether the area of ​​the corresponding power equipment forms an arithmetic sequence. Considering the engineering error in image recognition, the arithmetic sequence determination formula allows an error not exceeding ±10% of the sample standard deviation, i.e., it must simultaneously satisfy the following formula:

[0033]

[0034]

[0035] Where n represents the category of power equipment, S n1 S n2 S n3 These represent the areas of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let σ be the average area of ​​the nth type of three-phase electrical equipment. n Let be the area standard deviation of the nth type of electrical equipment.

[0036] In this embodiment, the power equipment used is a current transformer and a circuit breaker. It is determined whether the area of ​​the three-phase current transformer and the area of ​​the three-phase circuit breaker form an arithmetic sequence within the error range.

[0037] Step 4: Calculate the abscissa of each phase of the same type of power equipment, and solve for the arithmetic sequence of the abscissas of adjacent phases of the same type of power equipment; use the gray-scale centroid method to obtain the abscissas of the 1st, 2nd, and 3rd phase power equipment. The gray-scale centroid positioning formula is:

[0038]

[0039] In the formula, i is the horizontal coordinate of the pixel, i = 1 to M, j is the vertical coordinate of the pixel, j = 1 to N, M × N is the pixel size of the photo, G(i, j) is the gray value of each pixel, and x is the horizontal coordinate of the power equipment to be determined.

[0040] The arithmetic sequence determination formula is used to calculate whether the horizontal coordinates of the power equipment form an arithmetic sequence. Considering the engineering error of image recognition, the arithmetic sequence determination formula allows for an error of no more than ±10% of the sample standard deviation, that is, it must simultaneously satisfy the following formula:

[0041]

[0042]

[0043] Where, x n1 x n2 x n3 These represent the x-coordinates of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let γ be the average abscissa of the nth type of three-phase electrical equipment. n Let be the standard deviation of the x-axis of the nth type of power equipment.

[0044] Step 5: If the area and x-coordinate of all three-phase electrical equipment are in an arithmetic sequence, then based on the phase with the largest area and the phase with the smallest x-coordinate among all types of electrical equipment, infer the electrical equipment of the same phase and form a local electrical topology diagram. Compare the local electrical topology diagram with the substation main wiring diagram to verify whether the connection relationship of the same phase of all types of electrical equipment in the local electrical topology diagram is correct. If it is correct, the local electrical topology diagram and the infrared image are combined to form a power equipment topology association diagram. If it is incorrect, return to step 1 to perform the next set of infrared image analysis.

[0045] The phase sequence of electrical equipment is determined based on the principle that closer electrical equipment occupies a larger area in the infrared image. Taking a current transformer as an example, the position when taking the infrared image is designated as position Y, which is also marked as position Y on the wiring diagram. The distances from position Y to the three-phase current transformer in the wiring diagram are measured as L. CT1 L CT2 L CT3 The distances measured in this embodiment are related as follows: L CT1 <L CT2 <L CT3 The closer the distance when capturing infrared images, the larger the area of ​​the current transformer. Given that the location is closest to phase A when the infrared image is captured, it can be deduced that the current transformer with the largest area in the infrared image (phase 1 current transformer CT1) is the phase A current transformer, i.e., CT1. A The current transformer of the second largest phase (the second-phase current transformer CT2) is a phase B current transformer, i.e., CT... BThe current transformer with the smallest area in the phase (the third phase current transformer CT3) is a C-phase current transformer, i.e., CT. C .

[0046] In this embodiment, the calculated areas and abscissa values ​​of the current transformers and circuit breakers both form an arithmetic sequence. Furthermore, the phase with the largest area in both the current transformers and circuit breakers, and the phase with the largest area in the circuit breaker, also has the smallest abscissa. We can infer that the current transformer (phase 1 current transformer CT1) and the circuit breaker (phase 1 circuit breaker QF1) in the infrared image, which have the largest area and the smallest abscissa, are in-phase devices. Based on the measured shooting distance, we infer that phase 1 current transformer CT1 is phase A current transformer. Therefore, we can conclude that phase 1 circuit breaker QF1 is phase A circuit breaker QF1. A Next, based on the second largest phase area and the second smallest phase on the horizontal axis in the current transformer and circuit breaker, it can be deduced that the second-phase current transformer CT2 and the second-phase circuit breaker QF2 are in-phase devices, that is, the second-phase circuit breaker QF2 is the B-phase circuit breaker QF. B The remaining third-phase circuit breaker QF3 is the C-phase circuit breaker QF. C .

[0047] Based on the obtained correspondence between current transformers and circuit breakers, a local electrical topology diagram is drawn, as follows: Figure 4 In the middle left section, straight lines are used to represent the correspondence between electrical equipment in the local electrical topology diagram and electrical equipment in the infrared image, where QS A QS B QS C QS A QS B QS C These are A, B, and C phase-operated knife switches, SA A SA B SA C SA A SA B SA C These are A, B, and C phase changeover switches, PT A PT B PT C These are voltage transformers for phases A, B, and C, respectively, FU A FU B FU C These are fuses for phases A, B, and C, respectively. Next, we will compare the partial electrical topology diagram with the substation main wiring diagram. Figure 4 By comparison, verify whether the same-phase connection relationship of various power devices in the local electrical topology diagram is correct. If correct, the local electrical topology diagram and the infrared image are combined to form the power device topology association diagram. If incorrect, return to step 1 to perform the next set of infrared image analysis.

[0048] Step 6: Based on the fault diagnosis results from the infrared images, mark the faulty phase power equipment in the verified local electrical topology diagram. (Local electrical topology diagram) Figure 4 The equipment selected in the middle left section is the faulty phase power equipment. Based on the matching relationship between the local electrical topology diagram and the substation main wiring diagram, the faulty phase power equipment is highlighted in the substation main wiring diagram. Figure 3 This allows for the rapid location of faulty power equipment.

[0049] The above description merely illustrates preferred embodiments of the present invention and is not intended to limit the invention in any other way. Any person skilled in the art may make modifications or alterations to the above-disclosed content to create equivalent embodiments. However, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A method for rapid localization of heat generation in power equipment based on semantic segmentation and topology reconstruction, characterized in that, Includes the following steps: Step 1: Use semantic segmentation to segment the power equipment in the infrared image, and label different categories of power equipment in the infrared image with different colors; Step 2: Binarize the infrared image after semantic segmentation, and use connected component analysis to label the power equipment with different phase sequences according to the 1st, 2nd and 3rd phases respectively, so as to realize the same phase identification of the power equipment; Step 3: Calculate the area of ​​similar power equipment in each phase, and solve for the arithmetic sequence of the areas of adjacent phases of similar power equipment; Step 4: Find the abscissa of each phase of the same type of power equipment, and solve for the arithmetic sequence of the abscissas of adjacent phases of the same type of power equipment; Step 5: If the area and x-coordinate of all three-phase electrical equipment are in an arithmetic sequence, then based on the phase with the largest area and the phase with the smallest x-coordinate among all types of electrical equipment, infer the electrical equipment of the same phase and form a local electrical topology diagram. Compare the local electrical topology diagram with the substation main wiring diagram to verify whether the connection relationship of the same phase of all types of electrical equipment in the local electrical topology diagram is correct. If it is correct, the local electrical topology diagram and the infrared image are combined to form a power equipment topology association diagram. If it is incorrect, return to step 1 to perform the next set of infrared image analysis. Step 6: Based on the fault diagnosis results of the infrared image, mark the faulty phase power equipment in the verified local electrical topology diagram, and mark the faulty phase power equipment in the substation main wiring diagram according to the matching relationship between the local electrical topology diagram and the substation main wiring diagram, thereby realizing the rapid location of the faulty phase power equipment.

2. The method for rapid location of heat generation in power equipment based on semantic segmentation and topology reconstruction according to claim 1, characterized in that, The connected component analysis uses a two-pass algorithm. The first pass assigns a label to all pixels in the infrared image and records which connected component each label belongs to. The second pass marks each pixel with its corresponding connected component.

3. The method for rapid location of heat generation in power equipment based on semantic segmentation and topology reconstruction according to claim 1, characterized in that, In step 3, the area of ​​the power equipment in phases 1, 2, and 3 is obtained using the breadth-first search (BFS) algorithm. Then, based on the arithmetic sequence determination formula, it is calculated whether the area of ​​the corresponding power equipment forms an arithmetic sequence, i.e., whether it simultaneously satisfies the following formula: |S n1 -S n2 |=|S n2 -S n3 | Where n represents the category of power equipment, S n1 S n2 S n3 These represent the areas of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let σ be the average area of ​​the nth type of three-phase electrical equipment. n Let be the area standard deviation of the nth type of electrical equipment.

4. The method for rapid location of heat generation in power equipment based on semantic segmentation and topology reconstruction according to claim 1, characterized in that, In step 4, the abscissas of the first, second, and third phase power equipment are obtained using the gray centroid method. Then, the abscissas of the power equipment are calculated according to the arithmetic sequence determination formula to determine whether they form an arithmetic sequence, i.e., whether they simultaneously satisfy the following formula: |x n1 -x n2 |=|x n2 -x n3 | Where, x n1 x n2 x n3 These represent the x-coordinates of the nth type of electrical equipment in phases 1, 2, and 3, respectively. Let γ be the average abscissa of the nth type of three-phase electrical equipment. n Let be the standard deviation of the x-axis of the nth type of power equipment.

Images