CN-121982795-A - Substation inspection method, device, equipment and storage medium

CN121982795ACN 121982795 ACN121982795 ACN 121982795ACN-121982795-A

Abstract

The application discloses a substation inspection method, a device, equipment and a storage medium, and relates to the technical field of operation and maintenance of power systems. The method comprises the steps of acquiring an image of target equipment according to a task to obtain a first image, wherein the image comprises a knife switch image and an oil level gauge image, calculating the static geometrical gap of a contact surface of a moving contact tip and a static contact tip and the stability of a moving track of the moving contact tip based on the knife switch image, judging a knife switch state according to the static geometrical gap and the stability of the moving track, calculating a first oil level value based on the oil level gauge image, identifying texture change of a peripheral sealing area to generate oil seepage early warning, and generating a grading early warning signal according to the knife switch state, the first oil level value and the oil seepage early warning. The method improves the reading stability of the oil level gauge and the accuracy and timeliness of oil seepage early warning.

Inventors

YANG XU
GAO YUHUA
LIU NAN
SONG HAO
ZHAO YAJING
JIA TAO
ZHANG HUI
LIU YUHAN
Gou lei
WEN SHUAIZHAO
LI XUELIANG
JIANG WENXIAO
Cui Menglu
JIN MINGSHU

Assignees

国网冀北电力有限公司廊坊供电公司
国家电网有限公司

Dates

Publication Date: 20260505
Application Date: 20260127

Claims (10)

1. A power transformation inspection method, the method comprising: acquiring a video stream of the power transformation equipment, positioning a target equipment according to the video stream of the power transformation equipment, and generating a staring acquisition task; Acquiring an image of target equipment according to the staring acquisition task to obtain a first image, wherein the first image comprises a knife switch image containing a moving contact and a fixed contact area and an oil level gauge image containing a dial plate and a peripheral sealing area of an oil level gauge; Calculating a static geometric gap of a contact surface of a moving contact tip and a fixed contact and a motion trail stability of the moving contact tip based on the knife switch image containing the contact areas of the moving contact and the fixed contact, and judging a knife switch state according to the static geometric gap and the motion trail stability; Calculating a first oil level value based on the dial plate of the oil level gauge and an oil level gauge image of a peripheral sealing area, and recognizing texture change of the peripheral sealing area to generate an oil seepage early warning; and generating a grading early warning signal according to the state of the disconnecting link, the first oil level value and the oil seepage early warning.
2. The method of claim 1, wherein the image capturing of the target device according to the gaze capturing task results in a first image, comprising: Image acquisition is carried out on the target equipment according to the staring acquisition task and preset parameters, so that an initial image is obtained; and preprocessing the initial image to obtain a first image.
3. The method according to claim 1, wherein calculating a static geometrical gap of a moving contact tip and a static contact surface and a motion trajectory stability of the moving contact tip based on the knife gate image including the moving contact and the static contact area comprises: Sub-pixel edge detection is carried out on the disconnecting link image, and a moving contact tip contour line and a fixed contact surface contour line are extracted; Calculating the Euclidean distance between the characteristic point of the moving contact tip and the vertical projection point of the contour line of the contact surface of the static contact to obtain a static geometric gap; calculating standard deviation of the characteristic point positions based on the characteristic point positions of the moving contact tips in the continuous multi-frame disconnecting link images, performing spectrum analysis on the deviation of the characteristic point sequences in the horizontal and vertical directions, and calculating the proportion of the vibration energy in a preset frequency band to the total energy; And obtaining the stability of the motion track based on the standard deviation and the proportion of the vibration energy in the preset frequency band to the total energy.
4. The method of claim 1, wherein said determining a knife switch state from said static geometry gap and said motion profile stability comprises: Comparing the static geometric gap with a preset closing gap threshold value and a preset opening gap threshold value; Comparing the stability of the motion trail with a preset stability threshold; When the static geometric gap is smaller than the closing gap threshold and the stability of the motion trail is smaller than the stability threshold, determining that the state of the disconnecting link is stable closing; when the static geometric gap is smaller than the closing gap threshold but the stability of the motion trail is larger than or equal to the stability threshold, judging that the state of the disconnecting link is half-closing or poor in contact; And when the static geometric gap is larger than the opening gap threshold, judging that the disconnecting link state is opening.
5. The method of claim 1, wherein the calculating a first oil level value based on the oil level gauge image of the dial plate of the oil level gauge and the peripheral seal area and identifying a peripheral seal area texture change to generate an oil bleed warning comprises: Performing center positioning and polar coordinate transformation on a dial plate area in the oil level gauge image, extracting sub-pixel coordinates of a pointer tip, and calculating an included angle between the pointer and the center of the dial plate; obtaining an instantaneous oil level reading based on a preset first mapping relation between the included angle between the pointer and the center of the dial plate; Calculating a first oil level value by adopting a multi-frame fusion algorithm based on an instantaneous oil level reading sequence obtained by calculating continuous multi-frame oil level gauge images; Registering and aligning a peripheral sealing area in the oil level gauge image with a pre-stored reference image, calculating a multi-dimensional differential feature map, and fusing the multi-dimensional differential feature map to generate a comprehensive differential map; Performing self-adaptive threshold segmentation on the comprehensive difference map to obtain a suspected oil seepage region, and analyzing the area change, gravity center displacement and shape diffusion characteristics of the suspected oil seepage region in a plurality of inspection cycles; And calculating the oil seepage risk degree based on the area change, the gravity center displacement and the shape diffusion characteristics, and generating the oil seepage early warning when the oil seepage risk degree exceeds a preset oil seepage risk threshold.
6. The method of claim 1, wherein generating a classification warning signal based on the knife gate status, the first oil level value, and the oil bleed warning comprises: Matching the disconnecting link state, the first oil level value and the risk level of the oil seepage early warning according to a preset early warning rule base to generate a grading early warning signal, wherein the grading early warning signal comprises an emergency warning signal, an attention early warning signal and a normal state signal; When the disconnecting link state is half-closed or poor in contact, or the first oil level value exceeds a normal working range, or the oil seepage early warning is high risk, an emergency warning signal is generated; When the oil seepage early warning is low in risk, an attention early warning signal is generated; when all states are normal, a normal state signal is generated and data is archived.
7. The method of claim 2, wherein preprocessing the initial image to obtain a first image comprises: Based on the type of target equipment, a corresponding pre-training lightweight model is called to extract an interested area from the initial image, and an initial disconnecting link image containing a moving contact and a fixed contact area and an initial oil level gauge image containing a dial plate and a peripheral sealing area of the oil level gauge are obtained; And carrying out graying, histogram equalization and denoising filtering treatment on the extracted initial disconnecting link image and the initial oil level gauge image to obtain a first image.
8. A power transformation inspection device, the device comprising: the acquisition module is used for acquiring video streams of the power transformation equipment, positioning target equipment according to the video streams of the power transformation equipment and generating a staring acquisition task; The acquisition module is used for acquiring images of target equipment according to the staring acquisition task to obtain a first image, wherein the first image comprises a knife switch image containing a moving contact and a fixed contact area and an oil level gauge image containing a dial plate and a peripheral sealing area of the oil level gauge; The calculation module is used for calculating the static geometrical gap of the contact surface of the moving contact tip and the fixed contact and the motion track stability of the moving contact tip based on the knife switch image containing the contact area of the moving contact and the fixed contact, and judging the state of the knife switch according to the static geometrical gap and the motion track stability; And the early warning module is used for generating a grading early warning signal according to the state of the disconnecting link, the first oil level value and the oil seepage early warning.
9. A computing device comprising a memory and a processor; Wherein one or more computer programs are stored in the memory, the one or more computer programs comprising instructions, which when executed by the processor, cause the computing device to perform the method of any of claims 1-7.
10. A computer-readable storage medium, characterized in that the computer-readable storage medium is for storing a computer program, the computer program for performing the method of any one of claims 1 to 7.

Description

Substation inspection method, device, equipment and storage medium Technical Field The application relates to the technical field of operation and maintenance of power systems, in particular to a power transformation inspection method, a device, equipment and a storage medium. Background Traditional substation inspection mainly relies on manual carrying instruments to carry out on-site inspection and recording, and the problems of high labor intensity, low efficiency, obvious influence by subjective factors of personnel and the like exist. In recent years, with the development of computer vision technology, an automated inspection system based on a robot or a fixed camera has emerged. In the existing method, an image processing algorithm is mostly adopted to identify the state of equipment, for example, a straight line or a pointer is detected through Hough transformation to read a meter angle or judge the position of a knife switch. However, the existing method has poor robustness in coping with complex environments such as illumination change and background interference, the existing recognition algorithm has insufficient precision and reliability, cannot meet the high standard requirement of unmanned intelligent inspection, still depends on manual review, and has limited intelligent degree. Disclosure of Invention The application provides a substation inspection method, a device, equipment and a storage medium, which can solve the problems of low recognition precision, poor reliability and early warning lag of the readings of a disconnecting link and an oil level meter in the prior art, and realize accurate, real-time and automatic monitoring and early warning of the state of substation equipment. In order to achieve the above purpose, the application adopts the following technical scheme: in a first aspect, the present application provides a power transformation inspection method, the method comprising: acquiring a video stream of the power transformation equipment, positioning a target equipment according to the video stream of the power transformation equipment, and generating a staring acquisition task; Acquiring an image of target equipment according to the staring acquisition task to obtain a first image, wherein the first image comprises a knife switch image containing a moving contact and a fixed contact area and an oil level gauge image containing a dial plate and a peripheral sealing area of an oil level gauge; Calculating a static geometric gap of a contact surface of a moving contact tip and a fixed contact and a motion trail stability of the moving contact tip based on the knife switch image containing the contact areas of the moving contact and the fixed contact, and judging a knife switch state according to the static geometric gap and the motion trail stability; Calculating a first oil level value based on the dial plate of the oil level gauge and an oil level gauge image of a peripheral sealing area, and recognizing texture change of the peripheral sealing area to generate an oil seepage early warning; and generating a grading early warning signal according to the state of the disconnecting link, the first oil level value and the oil seepage early warning. In some possible implementations, the capturing an image of the target device according to the gaze capturing task, to obtain a first image, includes: And performing image acquisition on the target equipment according to the staring acquisition task and preset parameters to obtain an initial image, and preprocessing the initial image to obtain a first image. In some possible implementations, the calculating the static geometrical gap between the contact surface of the moving contact tip and the static contact and the motion track stability of the moving contact tip based on the knife switch image including the contact area of the moving contact and the contact area of the static contact includes: The method comprises the steps of carrying out sub-pixel edge detection on a knife switch image, extracting a moving contact tip outline and a fixed contact surface outline, calculating Euclidean distance between a moving contact tip characteristic point and a vertical projection point of the fixed contact surface outline to obtain a static geometric gap, calculating standard deviation of the characteristic point position based on the moving contact tip characteristic point position in the knife switch image, carrying out frequency spectrum analysis on deviation of a characteristic point sequence in horizontal and vertical directions, calculating the proportion of vibration energy in a preset frequency band to total energy, and obtaining the stability of a motion track based on the proportion of the standard deviation to the vibration energy in the preset frequency band to total energy. In some possible implementations, the determining the state of the knife switch according to the static geometrical gap and the motion trajectory stability inc