Automatic gas taking detection device and detection method for main transformer gas of transformer substation

Abstract

The invention discloses an automatic gas taking detection device and method for main transformer gas of a transformer substation. The automatic gas taking detection device comprises a control unit, an alarm signal synchronization unit, an image recognition module, an electromagnetic valve, a miniature vacuum pump, a gas sensor module, a display module, a communication module, a man-machine interaction module, a power supply module, a gas washing bottle and a gas storage bottle. According to the transformer substation main transformer gas automatic gas taking detection method using the device, based on an image recognition strategy of dual-wavelength optical characteristics, the image recognition module is used for obtaining image characteristics of insulating oil and gas distribution in the gas relay through the observation window, and therefore the internal gas state is recognized; and when the light gas action of the main transformer of the transformer substation alarms, gas taking detection is carried out, and whether the action is malfunction is judged. The method is high in operability, time-saving, labor-saving and high in working efficiency; the potential value is huge.

Description

Technical Field

[0001] The present invention relates to the field of structural design and application technology of a substation main transformer gas automatic gas extraction and gas composition detection and alarm device, and in particular to a substation main transformer gas automatic gas extraction detection device and a substation main transformer gas automatic gas extraction detection method using the device. Background Art

[0002] In the prior art, when the main transformer gas relay of the substation triggers a light gas operation, an alarm signal is issued, and the monitoring personnel are notified to go to the site to check whether there is gas inside the gas relay. If there is no gas, it is a light gas malfunction. The main reasons for the malfunction are:

[0003] (1) The transformer's oil filling, oil filtering, oil changing, and cooling systems are not tight, or air enters the transformer tank during the silicone replacement process;

[0004] (2) The oil level is lower than the light gas float of the gas relay due to temperature drop or oil leakage;

[0005] (3) A minor fault in the transformer produces a small amount of gas;

[0006] (4) When a through-fault short circuit occurs in the transformer, the oil flow speed in the oil gap is accelerated under the action of the through-fault current. When the pressure difference between the oil gap and the outside of the winding changes greatly, the gas relay may malfunction;

[0007] (5) Faults in the gas relay itself or the secondary circuit can also cause malfunctions and require further inspection. When there is gas inside the on-site gas relay, the operator must collect gas on-site for flammability testing, preliminarily determine the gas composition, and send the gas to a professional offline gas chromatograph for analysis. After the analysis results are available, a decision on whether to shut down the power supply is made.

[0008] The gas extraction process requires manual operation using a glass syringe, requiring two people to work together. This is complex and poses risks such as glass tube breakage and contamination from syringe oscillation during testing. Furthermore, the entire process from gas extraction to testing and analysis can take several hours. Transformer failures can easily progress from minor faults to major ones, missing the optimal time for troubleshooting and causing harm to equipment, personnel, and the power grid. A 1000 kV substation experienced a minor gas triggering event in the main transformer. During on-site inspection and gas extraction, operators experienced a transformer explosion, resulting in casualties. Automating gas extraction and analysis during this minor gas triggering event would effectively address these issues.

[0009] Chinese patent application publication number CN109459280A, published on March 12, 2019, entitled "A Transformer Gas On-site Sampling Device and Method," discloses a transformer gas on-site sampling device and method. The device has good sealing performance and can effectively prevent gas leakage or external air from entering and causing sample gas contamination. However, the device does not have an active gas sampling detection function, and has not truly solved the current problem of automatically performing gas sampling analysis for light gas action in substations to improve the quality and efficiency of main transformer abnormality processing.

[0010] In view of the above problems, in order to improve the gas extraction efficiency of the transformer gas relay and realize the rapid judgment of the gas detection composition, it is urgently necessary to develop a device that can automatically extract gas and quickly detect and alarm when light gas is actuated. The dispatcher can grasp the operation status of the main transformer on site at the first time, make a quick judgment, decide whether the main transformer needs to be shut down immediately, and evaluate whether the operating personnel can be arranged to go to the site for inspection, so as to effectively avoid the risk of personal injury, equipment and power grid failure. Summary of the Invention

[0011] The present invention provides a device and method for automatically detecting gas from a substation main transformer using the device. This device effectively addresses the time-consuming and labor-intensive manual gas extraction process, along with the long lead time between test results and results. This can lead to minor faults developing into serious ones, missing the optimal timeframe for troubleshooting and causing harm to equipment, personnel, and the power grid. The corresponding detection method offers low cost, strong operability, and excellent technical performance.

[0012] When the main transformer is in normal operation, the observation window of the gas relay is filled with transformer oil. When an internal fault occurs in the main transformer, the transformer oil decomposes to produce gas, and the gas relay is filled with gas. An obvious dividing line between the gas and the transformer oil will appear on the glass mirror of the gas relay observation window. By identifying the characteristics of the glass mirror of the gas relay observation window, it is possible to accurately determine whether the main transformer has a minor gas malfunction, and effectively avoid malfunction of the automatic detection device due to minor gas malfunction.

[0013] The technical solution of the present invention is:

[0014] The integrated device for automatic gas extraction and detection of the main transformer of the substation is used in conjunction with a gas relay 15 and a gas collecting box 19. The gas collecting box 19 is connected to the gas relay 15 through an air duct 17, and the gas relay 15 is connected to the main transformer body 14 of the substation. The key technology is that the integrated device for automatic gas extraction and detection of the main transformer of the substation includes: a control unit 1, an alarm signal synchronization unit 2, an image recognition module 3, a solenoid valve 4, a micro vacuum pump 5, a gas sensor module 6, a display module 7, a communication module 8, a human-computer interaction module 9, a power supply module 10, a gas washing bottle 11, and a gas storage bottle 12. Among them:

[0015] The alarm signal synchronization unit 2, image recognition module 3, solenoid valve 4, micro vacuum pump 5, gas sensor module 6, display module 7, communication module 8, human-computer interaction module 9, and power module 10 are all connected to the control unit 1. The gas collecting box 19 is connected to the gas washing bottle 11 via a pipeline, and the gas washing bottle 11 is connected to the gas storage bottle 12 via a pipeline. In turn, the gas storage bottle 12 is connected to the micro vacuum pump 5 via a pipeline. The gas sensor module 6 is arranged in the pipeline connecting the gas storage bottle 12 and the micro vacuum pump 5. The power module 10 is also connected to the image recognition module 3, solenoid valve 4, micro vacuum pump 5, gas sensor module 6, display module 7, and communication module 8. When the gas relay 15 connected to the main transformer of the substation triggers a light gas trigger and detects a small amount of gas, the alarm signal issued is synchronously transmitted to the alarm signal synchronization unit 2. The subsequent normal handling procedure requires: notifying the operating personnel to check whether there is gas inside the gas relay 15 on site. If there is no gas, it is a light gas malfunction; if there is gas, corresponding measures are taken to eliminate the safety hazard.

[0016] A gas collecting box liquid level sensor 19.1 is installed inside the gas collecting box 19; a manual oil drain valve 19.2 and a solenoid valve A4.1 for draining oil are provided below the gas collecting box 19; a manual air drain valve 19.3 and a solenoid valve B4.2 are provided in sequence on the ventilation pipe connecting the gas collecting box 19 to the gas washing bottle 11; the solenoid valve C4.3 has two channels A and B that can be independently controlled to be open or closed. Among them, channel A of the solenoid valve C4.3 is connected to the gas sensor module 6, and channel B of the solenoid valve C4.3 is connected to the gas storage bottle 12. The solenoid valve C4.3 is also connected to the gas washing bottle 11; a solenoid valve D4.4 is provided on the pipeline between the gas sensor module 6 and the micro vacuum pump 5.

[0017] The substation main transformer gas automatic gas extraction detection device is equipped with multiple solenoid valves 4, a gas collecting box liquid level sensor 19.1 is installed in the gas collecting box 19, and a solenoid valve B4.2 is installed on the pipeline connecting the manual air release valve 19.3 to the gas washing bottle 11; a solenoid valve A4.1 is installed on the pipeline with the manual oil drain valve 19.2 below the gas collecting box 19, which is used to realize electric control of oil draining and gas extraction; when draining oil, the gas collecting box 19 oil level position is unchanged according to the gas collecting box liquid level sensor 19.1. All the gas inside the gas relay 15 has been collected in the gas collecting box 19, and the solenoid valve A4.1 corresponding to the manual oil drain valve 19.2 can be closed; the manual air release valve 19.3 and the solenoid valve B4.2 are arranged on the top of the gas collecting box 19, and the oil level in the gas collecting box 19 can be judged by the gas collecting box liquid level sensor 19.1. Specifically, whether the gas in the gas collecting box 19 is full of oil can be judged to determine whether all the gas in the gas collecting box 19 has been released. The solenoid valve B4.2 on the same section of the pipeline as the manual air release valve 19.3 can be closed.

[0018] The gas washing bottle 11 is installed downstream of the outlet of the solenoid valve B4.2 and is used to filter the gas-oil-gas mixture and store transformer oil.

[0019] The gas storage bottle 12 is installed behind the gas washing bottle 11 and is connected with the gas washing bottle 11 through a pipeline. It is used to store a part of the gas and send it to the oil chemical professional for detection.

[0020] The integrated device for automatic gas extraction and detection of main transformer gas in a substation according to the present invention preferably claims the following technical contents:

[0021] The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements:

[0022] First, in the integrated device for automatic gas extraction and detection of main transformer gas in a substation, the image recognition module 3 is specifically a dual-wavelength optical camera and / or an infrared image camera;

[0023] Secondly, the gas sensor module 6 integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6, so as to detect the gas components and gas concentration.

[0024] The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements:

[0025] First, the control unit 1 collects light gas action circuit information and information from the image recognition module 3 through the alarm signal synchronization unit 2 to determine whether light gas action has actually occurred. It then completes gas extraction and detection by controlling the start and stop of the micro vacuum pump 5 and each solenoid valve 4 according to the logic of traditional manual gas extraction.

[0026] Secondly, the micro vacuum pump 5 is used to keep the gas extraction pipelines in a vacuum state;

[0027] Thirdly, the display module 7 is used to display information such as the composition, concentration and fault type of the gas;

[0028] Fourthly, the communication module 8 is used for communication between the control unit 1 and the substation monitoring background, so that the dispatcher and the operator can immediately grasp the specific situation of the main transformer light gas operation;

[0029] Fifth, each solenoid valve, the image recognition module 3 , the gas sensor module 6 , the micro vacuum pump 5 , the display module 7 , the communication module 8 , and the power supply module 10 are all connected to the control unit 1 .

[0030] The present invention also claims protection for a method for automatically detecting gas from a main transformer in a substation using the aforementioned integrated device for automatically detecting gas from a main transformer in a substation. The key technical aspects of the method are as follows: The method employs a method for identifying the presence of gas in the observation window of a transformer gas relay 15: The method employs an image recognition strategy based on dual-wavelength optical features, employing an image recognition module 3 to acquire image features of the insulating oil and gas distribution within the gas relay 15 through the gas relay observation window 15.1, thereby identifying the internal gas state;

[0031] When the substation main transformer light gas alarm is triggered, the device determines whether there is gas inside the gas relay 15 through the image recognition module 3. If there is gas inside the gas relay 15, the automatic gas extraction detection control process is started to perform gas extraction detection; if there is no gas inside the gas relay 15, it is judged as a false operation signal, which may be caused by external reasons such as a short circuit in the light gas signal circuit, and professional personnel are sent to the site for investigation.

[0032] The method for automatically extracting gas from a main transformer of a substation preferably requires protection of the following technical contents:

[0033] The image recognition module 3 is specifically a dual-wavelength optical camera or / and an infrared image camera; the image recognition module 3 is arranged on the front face of the gas relay observation window 15.1 on the gas relay 15, so as to observe whether there is gas in the gas relay 15 through the gas relay observation window 15.1; other modules are integrated in Figure 2 The dotted box on the right corresponds to the main unit, which is installed near the gas collecting box 19 outside the main transformer body;

[0034] The controller 1 is connected to a dual-wavelength optical camera and / or an infrared imaging camera. The camera's field of view includes a Buchholz relay observation window 15.1 corresponding to the Buchholz relay 15. The dual-wavelength optical camera is capable of recognizing both visible and infrared wavelengths. The former can perceive color distribution characteristics, while the latter can capture temperature distribution characteristics. The image characteristics of the insulating oil and gas distribution within the Buchholz relay 15 are compared, identified, and verified in the two wavelength optical images. This significantly improves the recognition success rate and reduces the false alarm rate.

[0035] When the main transformer is operating normally, the gas relay observation window 15.1 of the gas relay 15 is filled with transformer oil. When a fault occurs within the main transformer body 14, the transformer oil decomposes to produce gas, which fills the gas relay 15. The dual-wavelength optical camera detects a clear boundary between the gas and transformer oil on the glass surface of the gas relay observation window 15.1 of the gas relay 15. By identifying the presence of the boundary between the gas and transformer oil on the glass surface of the gas relay observation window 15.1, it can be determined whether the main transformer has a minor gas misoperation, effectively preventing malfunction of the automatic detection device due to a minor gas misoperation.

[0036] The identification algorithm for determining whether there is gas in the observation window of the transformer gas relay used in the automatic gas extraction detection method for the main transformer of the substation meets the following steps and content requirements:

[0037] Step 1: The control unit 1 obtains the dual-wavelength optical camera and infrared image camera data in the image recognition module 3;

[0038] Step 2: Normalize the brightness of the visible light image: Convert the image from RGB color space to HSV color space, calculate the average global image brightness, and adjust the bias value to make the overall brightness of the image tend to a fixed value M = 40% to 60% (generally set to 50%), thereby eliminating the influence of ambient light on the visible light image;

[0039] Step 3: Infrared image contrast correction: The infrared image color data acquired by the infrared camera is directly converted from temperature data. The upper and lower temperature limits within the global image are calculated and used as the boundaries of the color mapping display to ensure higher contrast between different temperatures within the image.

[0040] Step 4: Based on the circular feature, the position of the observation window in the optical high-definition image obtained by the dual-wavelength optical camera and the infrared image obtained by the infrared image camera is found, and the image within the circle is intercepted based on the circular boundary;

[0041] Step 5: Search for the optimal horizontal dividing line using the traversal method: First, filter out invalid RGB colors, then set the horizontal dividing line from top to bottom, calculate and record the color average difference between the upper and lower parts of the horizontal dividing line; after traversal, select the horizontal dividing line with the largest color average difference as the traversal method recognition results R1 and R2;

[0042] Step 6: Calculate the optimal horizontal segmentation line using the centroid method: First, filter out invalid RGB colors, then extract the upper and lower images using the contours, calculate the color centroid within the contours, and use the centroid to reversely calculate the position of the horizontal segmentation line, which are used as the centroid method recognition results R3 and R4.

[0043] Step 7: Comparison and judgment of recognition results: The horizontal segmentation line position recognition results R1, R2, R3 and R4 in Steps 5 and 6 are combined. If the difference between the results is within 5%, they are considered similar; the number of similar results in R1, R2, R3 and R4 is counted using the "result factor" N;

[0044] If the "result factor" N ≥ 2, the result is valid, and the average value of these N similar result sets is used as the horizontal dividing line recognition result; if two recognition results appear and the result factor N is 2, the recognition is judged to have failed and the next recognition is restarted.

[0045] The gas sensor module 6 integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6, so as to detect the gas components and gas concentration.

[0046] The method for automatically detecting gas extraction from the main transformer of a substation meets one or a combination of the following requirements:

[0047] First, the control unit 1 collects light gas action circuit information and information from the image recognition module 3 through the alarm signal synchronization unit 2 to determine whether light gas action has actually occurred. It then completes gas extraction and detection by controlling the start and stop of the micro vacuum pump 5 and each solenoid valve 4 according to the logic of traditional manual gas extraction.

[0048] Secondly, the micro vacuum pump 5 is used to keep the gas extraction pipelines in a vacuum state;

[0049] Thirdly, the display module 7 is used to display information such as the composition, concentration and fault type of the gas;

[0050] Fourthly, the communication module 8 is used for communication between the control unit 1 and the substation monitoring background, so that the dispatcher and the operator can immediately grasp the specific situation of the main transformer light gas operation;

[0051] Fifth, each solenoid valve, the image recognition module 3 , the gas sensor module 6 , the micro vacuum pump 5 , the display module 7 , the communication module 8 , and the power supply module 10 are all connected to the control unit 1 .

[0052] In the automatic gas extraction detection method for the main transformer of the substation, the automatic gas extraction detection control process meets the following steps and content requirements:

[0053] First, control solenoid valve A4.1 to drain the oil from the gas collecting box, directing the gas inside gas relay 15 downward to the upper portion of gas collecting box 19. The specific duration of oil draining is determined by the oil level sensor 19.1 inside gas collecting box 19 detecting that the oil level no longer changes, which serves as a criterion for closing manual oil drain valve 19.2. Simultaneously, control micro vacuum pump 5 to evacuate gas washing bottle 11, gas storage bottle 12, and pipelines to prevent gas contamination. Then, control solenoid valve B4.2 in the gas collecting box to open, allowing gas to flow into the pipeline of the integrated automatic gas extraction and detection device for the main transformer of the substation, until the liquid level sensor inside gas collecting box 19 detects that the position has returned to the top of the gas collecting box, at which point solenoid valve B4.2 is closed.

[0054] The gas-oil mixture first passes through the gas washer bottle 11 to separate the gas-oil mixture. At the same time, channel B of the solenoid valve C4.3 in the gas storage bottle 12 is controlled to be open while channel A is closed. The gas continues to pass through the pipeline with the filter 18 to first inflate the gas storage bottle 12. After a certain period of time, channel B of the solenoid valve C4.3 is controlled to be closed. Then, channel A of the solenoid valve C4.3 is controlled to be open. The solenoid valve C4.3 transmits the remaining gas to the gas sensor module 6 to facilitate the detection of the gas composition and concentration.

[0055] The automatic gas sampling detection method for the main transformer of the substation meets the following requirements: According to the "DLT722-2014 Guidelines for Analysis and Judgment of Dissolved Gases in Transformer Oil", the characteristic value and three-ratio method are used for comprehensive judgment:

[0056] First, the characteristic value method should be used to check whether H2, C2H2 and total hydrocarbons have reached the caution value. If not, it is judged that there is no abnormality, but the gas concentration value needs to be saved. The next time there is gas inside the gas relay, it will be used to calculate the gas production rate and increase the gas production rate caution value analysis;

[0057] If H2, C2H2 and total hydrocarbons reach the attention value (except for H2 which reaches the attention value), the three-ratio method is further used to make judgments and characterize the fault.

[0058] The image recognition module 3 is arranged on the front face of the gas relay observation window 15.1 on the gas relay 15, so as to observe whether there is gas in the gas relay 15 through the gas relay observation window 15.1; other modules are integrated in Figure 2The dotted box shown on the right corresponds to the main unit, which is installed near the air collecting box 19 outside the main transformer body.

[0059] The integrated device for automatic gas extraction and detection of main transformer gas in substation and the method for automatic gas extraction and detection of main transformer gas in substation using the device can effectively solve a series of problems such as low degree of automation, low efficiency, time-consuming and labor-intensive manual gas extraction and detection of light gas action of main transformer, usually requiring cooperation of two people and complicated operation, easy breakage of glass tube in gas extraction process, easy contamination of oscillating gas in syringe during inspection, etc.; it realizes rapid online detection of gas; it fundamentally improves the major technical difficulty of using traditional methods from gas extraction, inspection and testing to obtaining results, which is time-consuming and labor-intensive, and even transformer failure is likely to develop from minor failure to serious failure in a relatively lengthy handling process, missing the best decision-making fault handling time, causing harm to equipment, people and power grid. At the same time, the present invention is also suitable for auxiliary work support for daily inspections of operation and maintenance personnel; when the light gas protection is not activated, there is gas inside the gas relay that needs to be taken out for testing. The human-computer interaction module of the integrated device for automatic gas extraction and detection of the main transformer gas of the substation described in the present invention can be used to achieve manual start-up and automatic gas extraction. It has high adaptability, strong operability, flexible use, time-saving and labor-saving, and high work efficiency; it has potential and extremely huge economic value and social value. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a schematic diagram showing the core components of the integrated automatic gas extraction and detection device for the main transformer of a substation;

[0061] Figure 2 This is a schematic diagram of the internal structure of the integrated automatic gas extraction and detection device for the main transformer of the substation;

[0062] Figure 3 This is a flow chart of an algorithm for identifying gas in the observation window of the transformer gas relay 15;

[0063] Figure 4 This is a control flow chart of the automatic gas extraction detection method for the main transformer of the substation;

[0064] Figure 5 This is a schematic diagram of the control panel of the integrated automatic gas extraction and detection device for the main transformer of the substation;

[0065] Figure 6 This is a schematic diagram showing the connection and installation principles of some structures of the main transformer body 14 and the gas relay 15. DETAILED DESCRIPTION

[0066] The present invention will be further described below with reference to the embodiments and the accompanying drawings, but is not limited thereto.

[0067] The meanings of the accompanying symbols are as follows: control unit 1, alarm signal synchronization unit 2, image recognition module 3, solenoid valve 4, solenoid valve A4.1, solenoid valve B4.2, solenoid valve C4.3, solenoid valve D4.4; miniature vacuum pump 5, gas sensor module 6, display module 7, communication module 8, human-computer interaction module 9, power supply module 10, gas washing bottle 11, gas storage bottle 12; main transformer oil pillow 13, main transformer body 14, gas relay 15, gas relay observation window 15.1, gas collecting box observation window 16, air bleed pipe 17, filter 18; gas collecting box 19, gas collecting box liquid level sensor 19.1, manual oil drain valve 19.2, manual air drain valve 19.3.

[0068] Example 1

[0069] The integrated device for automatic gas extraction and detection of main transformer gas in substation, such as Figure 1 、 Figure 2 、 Figure 5 、 Figure 6 As shown; it is used in conjunction with a gas relay 15 and a gas collecting box 19. The gas collecting box 19 is connected to the gas relay 15 through an air duct 17, and the gas relay 15 is connected to the main transformer body 14 of the substation. The key technology is that the substation main transformer gas automatic gas extraction and detection integrated device includes: a control unit 1, an alarm signal synchronization unit 2, an image recognition module 3, a solenoid valve 4, a micro vacuum pump 5, a gas sensor module 6, a display module 7, a communication module 8, a human-computer interaction module 9, a power module 10, a gas washing bottle 11, and a gas storage bottle 12; wherein:

[0070] The alarm signal synchronization unit 2, image recognition module 3, solenoid valve 4, micro vacuum pump 5, gas sensor module 6, display module 7, communication module 8, human-computer interaction module 9, and power module 10 are all connected to the control unit 1. The gas collecting box 19 is connected to the gas washing bottle 11 via a pipeline, and the gas washing bottle 11 is connected to the gas storage bottle 12 via a pipeline. In turn, the gas storage bottle 12 is connected to the micro vacuum pump 5 via a pipeline. The gas sensor module 6 is arranged in the pipeline connecting the gas storage bottle 12 and the micro vacuum pump 5. The power module 10 is also connected to the image recognition module 3, solenoid valve 4, micro vacuum pump 5, gas sensor module 6, display module 7, and communication module 8. When the gas relay 15 connected to the main transformer of the substation triggers a light gas trigger and detects a small amount of gas, the alarm signal issued is synchronously transmitted to the alarm signal synchronization unit 2. The subsequent normal handling procedure requires: notifying the operating personnel to check whether there is gas inside the gas relay 15 on site. If there is no gas, it is a light gas malfunction; if there is gas, corresponding measures are taken to eliminate the safety hazard.

[0071] A gas collecting box liquid level sensor 19.1 is installed inside the gas collecting box 19; a manual oil drain valve 19.2 and a solenoid valve A4.1 for draining oil are provided below the gas collecting box 19; a manual air drain valve 19.3 and a solenoid valve B4.2 are provided in sequence on the pipeline connecting the gas collecting box 19 to the gas washing bottle 11; the solenoid valve C4.3 has two channels, A and B, which can be independently controlled to be open or closed. Among them, channel A of the solenoid valve C4.3 is connected to the gas sensor module 6, and channel B of the solenoid valve C4.3 is connected to the gas storage bottle 12. The solenoid valve C4.3 is also connected to the gas washing bottle 11; a solenoid valve D4.4 is provided on the pipeline between the gas sensor module 6 and the micro vacuum pump 5.

[0072] The substation main transformer gas automatic gas extraction detection device is equipped with multiple solenoid valves 4, a gas collecting box liquid level sensor 19.1 is installed in the gas collecting box 19, and a solenoid valve B4.2 is installed on the pipeline connected to the manual air release valve 19.3 and the manual oil release valve 19.2 is arranged below the gas collecting box 19. A solenoid valve A4.1 is installed on the pipeline for realizing electric control of oil release and gas extraction; when draining oil, the gas collecting box 19 oil level position is judged to be unchanged according to the gas collecting box liquid level sensor 19.1. This indicates that all the gas inside the gas relay 15 has been collected in the gas box 19, and the solenoid valve corresponding to the manual oil drain valve 19.2 can be closed; the manual air release valve 19.3 and the solenoid valve B4.2 are arranged on the top of the gas box 19, and the oil level in the gas box 19 can be determined by the gas box liquid level sensor 19.1. Specifically, whether the gas in the gas box 19 is full of oil can be determined to determine whether all the gas in the gas box 19 has been released. The solenoid valve B4.2 on the same section of the pipeline as the manual air release valve 19.3 can be closed.

[0073] The gas washing bottle 11 is installed downstream of the outlet of the solenoid valve B4.2 and is used to filter the gas-oil-gas mixture and store transformer oil.

[0074] The gas storage bottle 12 is installed behind the gas washing bottle 11 and is connected with the gas washing bottle 11 through a pipeline. It is used to store a part of the gas and send it to the oil chemical professional for detection.

[0075] The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements:

[0076] First, in the integrated device for automatic gas extraction and detection of main transformer gas in a substation, the image recognition module 3 is specifically a dual-wavelength optical camera and / or an infrared image camera;

[0077] Secondly, the gas sensor module 6 integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6, so as to detect the gas components and gas concentration.

[0078] The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements:

[0079] First, the control unit 1 collects light gas action circuit information and information from the image recognition module 3 through the alarm signal synchronization unit 2 to determine whether light gas action has actually occurred. It then completes gas extraction and detection by controlling the start and stop of the micro vacuum pump 5 and each solenoid valve 4 according to the logic of traditional manual gas extraction.

[0080] Secondly, the micro vacuum pump 5 is used to keep the gas extraction pipelines in a vacuum state;

[0081] Thirdly, the display module 7 is used to display information such as the composition, concentration and fault type of the gas;

[0082] Fourthly, the communication module 8 is used for communication between the control unit 1 and the substation monitoring background, so that the dispatcher and the operator can immediately grasp the specific situation of the main transformer light gas operation;

[0083] Fifth, each solenoid valve, the image recognition module 3 , the gas sensor module 6 , the micro vacuum pump 5 , the display module 7 , the communication module 8 , and the power supply module 10 are all connected to the control unit 1 .

[0084] Example 2

[0085] The key technical aspects of the method for automatically detecting gas from a main transformer in a substation using the integrated device for automatically detecting gas from a main transformer in a substation described in Example 1 are as follows: The method employs a method for identifying the presence of gas in the observation window of a transformer gas relay 15 : This method employs an image recognition strategy based on dual-wavelength optical features, using an image recognition module 3 to acquire image features of the insulating oil and gas distribution within the gas relay 15 through the gas relay observation window 15.1, thereby identifying the internal gas state;

[0086] When the substation main transformer light gas alarm is triggered, the device determines whether there is gas inside the gas relay 15 through the image recognition module 3. If there is gas inside the gas relay 15, the automatic gas extraction detection control process is started to perform gas extraction detection; if there is no gas inside the gas relay 15, it is judged as a false operation signal, which may be caused by external reasons such as a short circuit in the light gas signal circuit, and professional personnel are sent to the site for investigation.

[0087] The method for automatically extracting gas from a main transformer of a substation preferably requires protection of the following technical contents:

[0088] The image recognition module 3 is specifically a dual-wavelength optical camera or / and an infrared image camera; the image recognition module 3 is arranged on the front face of the gas relay observation window 15.1 on the gas relay 15, so as to observe whether there is gas in the gas relay 15 through the gas relay observation window 15.1; other modules are integrated in Figure 2 The dotted box on the right corresponds to the main unit, which is installed near the gas collecting box 19 outside the main transformer body;

[0089] The controller 1 is connected to a dual-wavelength optical camera and / or an infrared imaging camera. The camera's field of view includes a Buchholz relay observation window 15.1 corresponding to the Buchholz relay 15. The dual-wavelength optical camera is capable of recognizing both visible and infrared wavelengths. The former can perceive color distribution characteristics, while the latter can capture temperature distribution characteristics. The image characteristics of the insulating oil and gas distribution within the Buchholz relay 15 are compared, identified, and verified in the two wavelength optical images. This significantly improves the recognition success rate and reduces the false alarm rate.

[0090] When the main transformer is operating normally, the gas relay observation window 15.1 of the gas relay 15 is filled with transformer oil. When a fault occurs within the main transformer body 14, the transformer oil decomposes to produce gas, which fills the gas relay 15. The dual-wavelength optical camera detects a clear boundary between the gas and transformer oil on the glass surface of the gas relay observation window 15.1 of the gas relay 15. By identifying the presence of the boundary between the gas and transformer oil on the glass surface of the gas relay observation window 15.1, it can be determined whether the main transformer has a minor gas misoperation, effectively preventing malfunction of the automatic detection device due to a minor gas misoperation.

[0091] The identification algorithm for determining whether there is gas in the observation window of the transformer gas relay used in the automatic gas extraction detection method for the main transformer of the substation meets the following steps and content requirements:

[0092] Step 1: The control unit 1 acquires and analyzes the dual-wavelength optical camera and infrared image camera data in the image recognition module 3;

[0093] Step 2: Normalize the visible light image brightness: Convert the image from RGB color space to HSV color space, calculate the global image brightness average, and adjust the bias value to make the overall image brightness tend to a fixed value M = 50%, thereby eliminating the influence of ambient light on the visible light image;

[0094] Step 3: Infrared image contrast correction: The infrared image color data acquired by the infrared camera is directly converted from temperature data. The upper and lower temperature limits within the global image are calculated and used as the boundaries of the color mapping display to ensure higher contrast between different temperatures within the image.

[0095] Step 4: Based on the circular feature, the position of the observation window in the optical high-definition image obtained by the dual-wavelength optical camera and the infrared image obtained by the infrared image camera is found, and the image within the circle is intercepted based on the circular boundary;

[0096] Step 5: Search for the optimal horizontal dividing line using the traversal method: First, filter out invalid RGB colors, then set the horizontal dividing line from top to bottom, calculate and record the color average difference between the upper and lower parts of the horizontal dividing line; after traversal, select the horizontal dividing line with the largest color average difference as the traversal method recognition results R1 and R2;

[0097] Step 6: Calculate the optimal horizontal segmentation line using the centroid method: First, filter out invalid RGB colors, then extract the upper and lower images using the contours, calculate the color centroid within the contours, and use the centroid to reversely calculate the position of the horizontal segmentation line, which are used as the centroid method recognition results R3 and R4.

[0098] Step 7: Comparison and judgment of recognition results: The horizontal segmentation line position recognition results R1, R2, R3 and R4 in Steps 5 and 6 are combined. If the difference between the results is within 5%, they are considered similar; the number of similar results in R1, R2, R3 and R4 is counted using the "result factor" N;

[0099] If the "result factor" N ≥ 2, the result is valid, and the average value of these N similar result sets is used as the horizontal dividing line recognition result; if two recognition results appear and the result factor N is 2, the recognition is judged to have failed and the next recognition is restarted.

[0100] The gas sensor module 6 integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6, so as to detect the gas components and gas concentration.

[0101] The method for automatically detecting gas extraction from the main transformer of a substation meets one or a combination of the following requirements:

[0102] First, the control unit 1 collects light gas action circuit information and information from the image recognition module 3 through the alarm signal synchronization unit 2 to determine whether light gas action has actually occurred. It then completes gas extraction and detection by controlling the start and stop of the micro vacuum pump 5 and each solenoid valve 4 according to the logic of traditional manual gas extraction.

[0103] Secondly, the micro vacuum pump 5 is used to keep the gas extraction pipelines in a vacuum state;

[0104] Thirdly, the display module 7 is used to display information such as the composition, concentration and fault type of the gas;

[0105] Fourthly, the communication module 8 is used for communication between the control unit 1 and the substation monitoring background, so that the dispatcher and the operator can immediately grasp the specific situation of the main transformer light gas operation;

[0106] Fifth, each solenoid valve, the image recognition module 3 , the gas sensor module 6 , the micro vacuum pump 5 , the display module 7 , the communication module 8 , and the power supply module 10 are all connected to the control unit 1 .

[0107] In the automatic gas extraction detection method for the main transformer of the substation, the automatic gas extraction detection control process meets the following steps and content requirements:

[0108] First, control solenoid valve A4.1 to drain the oil from the gas collecting box, directing the gas inside gas relay 15 downward to the upper part of gas collecting box 19. The specific duration of oil draining is determined by the oil level sensor 19.1 in gas collecting box 19 no longer changing, which serves as the criterion for closing solenoid valve A4.1. Simultaneously, control the micro vacuum pump 5 to evacuate the gas washing bottle 11, gas storage bottle 12, and pipeline to prevent gas contamination. Then, control the solenoid valve B4.2 in the gas collecting box to open, allowing gas to flow into the pipeline of the integrated automatic gas extraction and detection device for the main transformer of the substation, until the liquid level sensor in the gas collecting box 19 detects that the position has returned to the top of the gas collecting box, at which point the solenoid valve B4.2 is closed.

[0109] The gas-oil mixture first passes through the gas washer bottle 11 to separate the gas-oil mixture. At the same time, channel B of the solenoid valve C4.3 in the gas storage bottle 12 is controlled to be open while channel A is closed. The gas continues to pass through the pipeline with the filter 18 to first inflate the gas storage bottle 12. After a fixed time, channel B of the solenoid valve C4.3 is controlled to be closed. Then, channel A of the solenoid valve C4.3 is connected. The solenoid valve C4.3 transmits the remaining gas to the gas sensor module 6 to facilitate the detection of the gas composition and concentration.

[0110] The automatic gas sampling detection method for the main transformer of the substation meets the following requirements: According to the "DLT722-2014 Guidelines for Analysis and Judgment of Dissolved Gases in Transformer Oil", the characteristic value and three-ratio method are used for comprehensive judgment:

[0111] First, the characteristic value method should be used to check whether H2, C2H2 and total hydrocarbons have reached the caution value. If not, it is judged that there is no abnormality, but the gas concentration value needs to be saved. The next time there is gas inside the gas relay, it will be used to calculate the gas production rate and increase the gas production rate caution value analysis;

[0112] If H2, C2H2 and total hydrocarbons reach the attention value (except for H2 which reaches the attention value), the three-ratio method is further used to make judgments and characterize the fault.

[0113] The image recognition module 3 is arranged on the front face of the gas relay observation window 15.1 on the gas relay 15, so as to observe whether there is gas in the gas relay 15 through the gas relay observation window 15.1; other modules are integrated in Figure 2 The dotted box shown on the right corresponds to the main unit, which is installed near the air collecting box 19 outside the main transformer body.

[0114] The integrated device for automatic gas extraction and detection of main transformer gas in substations used in Examples 1 and 2 and the method for automatic gas extraction and detection of main transformer gas in substations using the device can effectively solve a series of problems such as low degree of automation, low efficiency, time-consuming and labor-intensive manual gas extraction and detection of light gas action in main transformers, which usually require the cooperation of two people and are complicated to operate. At the same time, the glass tube is easily broken during the gas extraction process, and the gas is easily contaminated by the oscillating syringe during the inspection process. It realizes rapid online detection of gas and fundamentally improves the traditional method of taking gas, sending it for inspection and testing, and then obtaining results, which is time-consuming and labor-intensive. Even transformer failures are likely to develop from minor failures to serious failures during the relatively lengthy handling process, missing the best decision-making fault handling time, and causing major technical difficulties in harming equipment, personnel, and power grids. At the same time, Example 1 and Example 2 are also suitable for auxiliary work support for daily inspections of operation and maintenance personnel; when the light gas protection is not activated, there is gas inside the gas relay that needs to be taken out for testing. The human-computer interaction module of the integrated device for automatic gas extraction and detection of the main transformer gas of the substation described in Example 1 can be used to achieve manual start-up and automatic gas extraction. It has high adaptability, strong operability, flexible usage, time-saving and labor-saving, and high work efficiency; it has potential extremely huge economic value and social value.

[0115] Example 3

[0116] An integrated device for automatic gas extraction and testing of the main transformer of a substation effectively solves the time-consuming and labor-intensive manual extraction of gas during the main transformer's light gas operation, requiring two people to work together, resulting in complex operation, the easy breakage of the glass tube during the extraction process, and the easy contamination of the gas from the oscillating syringe during inspection. This device enables online gas detection, avoiding the traditional method of taking gas, sending it for inspection, and obtaining results, which takes a long time. Transformer failures are likely to develop from minor faults to serious faults, missing the optimal decision-making time for troubleshooting, and causing harm to equipment, personnel, and the power grid. At the same time, during routine inspections, when the light gas protection is not activated, operations and maintenance personnel may also find that there is gas inside the gas relay 15 that needs to be extracted and tested. Through the human-computer interaction module of this device, electric gas extraction can be achieved, saving time and effort and improving work efficiency.

[0117] Figure 1The middle alarm signal synchronization unit 2 originates from the secondary circuit of the light gas action signal, which is used as a sufficient condition for the start of the device and is directly connected to the control unit 1;

[0118] Figure 1 The image recognition module 3 uses an algorithm for identifying whether there is gas in the observation window 15.1 of the transformer gas relay, and reports the information of whether there is gas in the observation window 15.1 to the control unit in real time.

[0119] A method for identifying the presence of gas within the observation window 15.1 of a gas relay located near the main transformer body 14 employs an image recognition strategy based on dual-wavelength optical signatures. This strategy uses an image recognition module 3 to capture image signatures of the insulating oil and gas distribution within the gas relay 15 through the observation window 15.1, thereby identifying the internal gas state. This identification method requires no modifications to the existing mechanical structure and electrical wiring within the gas relay 15, completely simulating the operator's daily inspection routine to complete the gas state identification process. To improve recognition accuracy and reduce the impact of ambient lighting and oil color on visible light image recognition accuracy, a dual-wavelength optical camera was selected, capable of sensing both visible and infrared wavelengths. The former detects color distribution signatures, while the latter captures temperature distribution signatures. The image signatures of the insulating oil and gas distribution within the gas relay 15 are compared and verified in the two wavelength optical images, significantly improving recognition success rates and reducing false alarm rates.

[0120] The image recognition algorithm based on dual-wavelength optical features is as follows Figure 3 As shown:

[0121] Step 1: The algorithm control unit acquires and analyzes the visible light high-definition camera and infrared image camera data;

[0122] Step 2: Normalize the brightness of the visible light image. Convert the image from RGB color space to HSV color space, calculate the average brightness of the global image, and adjust the bias value to make the overall brightness of the image tend to a fixed value M (usually set to 50%), so as to eliminate the influence of ambient light on the visible light image.

[0123] Step 3: Infrared image contrast correction. The infrared image color data is directly converted from the temperature data. The upper and lower temperature limits in the global image are calculated and used as the boundaries of the color mapping display to ensure a higher contrast between different temperatures in the image.

[0124] Step 4: Find the position of the Buchholz relay observation window 15.1 in the visible light high-definition image and the infrared image based on the circular feature, and intercept the image within the circle based on the circular boundary;

[0125] Step 5: Use the traversal method to search for the optimal horizontal dividing line: First, filter out invalid RGB colors, then set the horizontal dividing line from top to bottom, and calculate and record the color average difference between the upper and lower parts of the horizontal dividing line. After traversal, select the horizontal dividing line with the largest color average difference as the traversal method recognition results R1 and R2;

[0126] Step 6: Calculate the optimal horizontal segmentation line using the centroid method: First, filter out invalid RGB colors, then extract the upper and lower images using the contours, calculate the color centroid within the contours, and use the centroid to reversely calculate the position of the horizontal segmentation line, which are used as the centroid method recognition results R3 and R4.

[0127] Step 7: Compare and judge the recognition results: Combine the horizontal segmentation line position recognition results R1, R2, R3, and R4 from steps 5 and 6. If the difference between the results is within 5%, they are considered similar. The number of similar results in R1, R2, R3, and R4 is counted using the result factor. If the result factor N ≥ 2, the result is valid, and the average of these N similar result sets is used as the horizontal segmentation line recognition result. If two recognition results appear and the result factor N is 2, the recognition is considered a failure and the next recognition is restarted.

[0128] Figure 1 The middle solenoid valve 4 is a pneumatic electromagnetic control valve, which can control the opening and closing of the solenoid valve inlet and outlet. The 24V AirTac brand solenoid valve is selected here. It has good air tightness and can achieve a sensitive response of 0.05s. It is made of all-aluminum alloy, with good quality and long service life.

[0129] Figure 1 The micro vacuum pump 5 is used to keep the gas pipelines in a vacuum state. The micro vacuum pump selected here is the KLVP1-SB12 model of the Kammer brand. The product has stable performance, simple control, and sensitive response. It is widely used in gas analysis, air pressure massage, and chemical analysis.

[0130] The gas sensor 6 is a gas sensor integrated module, including characteristic gas sensors such as H2, CH4, C2H2, C2H4, C2H6, etc., which can detect gas composition and gas concentration in uL / L.

[0131] The display module 7 is a serial port LCD screen used to display gas composition, gas concentration, three ratio results and fault type analysis results, etc. There are many models available on the market, which are not listed in detail.

[0132] Communication module 8 uses the IEC61850 communication protocol to connect to the substation background monitoring system, so that the dispatcher can obtain the gas extraction analysis results of the substation site through the internal dispatching technical support system of the State Grid in the first time, so as to understand the on-site situation in advance and improve the disposal efficiency.

[0133] Figure 2 Gas wash bottle 11 and gas storage bottle 12 are made of glass with rubber stoppers, ensuring a good seal and preventing gas contamination. Gas wash bottle 11 has a capacity of 300ml, while gas storage bottle 12 has a capacity of 100ml. Gas wash bottle 11 is used to filter and separate the gas-oil mixture, storing transformer oil. Gas storage bottle 12 is installed behind the gas wash bottle, connected by a glass conduit, to store a portion of the gas for delivery to the oil chemical testing department.

[0134] Control unit 1 is a single-chip microcomputer (MCU) directly connected to the light gas action signal 2, image recognition module 3, solenoid valve 4, micro vacuum pump 5, gas sensor module 6, display module 7, communication module 8, human-computer interaction module 9, and gas box level sensor 19.1. The STM32H7 series MCU is used here. Its main frequency can reach 480MHz, and it supports a hardware double-precision floating-point unit and DSP instruction set. It is suitable for complex algorithms, has a rich interface, is highly scalable, and offers low-latency interrupt performance to meet real-time control requirements (such as the fast response requirements of the micro vacuum pump motor and solenoid valve).

[0135] The human-machine interaction module 9 can have three buttons: a manual button, a RES button, and an ON / OFF button. If, during routine inspections, operators discover gas inside a gas relay, but it may not have reached the light gas trigger level, they should release the gas from the relay for testing. The gas detection function can be activated using the manual button. The RES button initializes the device and clears historical records. It is typically used after a main transformer overhaul. The ON / OFF button is the device's power switch.

[0136] Figure 2 This is a schematic diagram of the internal structure of the integrated automatic gas extraction detection and alarm device for the main transformer of the substation.

[0137] Example 4

[0138] An integrated method for automatic gas extraction and detection of main transformer gas in a substation. Figure 4 This is the flow chart of automatic gas extraction detection control when light gas is actuated.

[0139] (1) When the substation main transformer light gas alarm is triggered, the device determines whether there is gas inside the gas relay 15 through the image recognition module 3. If there is no gas inside the gas relay 15, it is judged as a false operation signal, which may be caused by external reasons such as a short circuit in the light gas signal circuit. It is necessary to ask relay, maintenance and other professionals to check.

[0140] (2) When the substation main transformer light gas alarm is triggered, the alarm signal synchronization unit 2 transmits the information to the control unit 1. The device determines whether there is gas inside the gas relay 15 through the image recognition module 3. If there is gas, the automatic gas extraction detection control process is started to perform gas extraction detection.

[0141] (3) Automatic gas extraction detection control process: First, open the solenoid valve A4.1 (the manual oil drain valve is always open) to drain the oil and guide the gas inside the gas relay 15 to the upper part of the gas collecting box 19. When the gas collecting box liquid level sensor 19.1 detects that the oil level no longer changes, it means that the gas inside the gas relay 15 has all entered the gas collecting box 19. Close the solenoid valve A4.1 below the gas collecting box, and at the same time open the solenoid valve C4.3 and solenoid valve D4.4, start the micro vacuum pump 5, and evacuate the gas washing bottle 11, gas storage bottle 1, and pipeline to avoid gas contamination. The solenoid valve C4.3 has an A / B dual channel, and the two channels can independently control their conduction and closing. Then open the solenoid valve B4.2 until the gas collecting box liquid level sensor 19.1 in the gas collecting box 19 detects that the position is reset to the top of the gas collecting box 19. Gas flows into the device's conduit and first passes through the gas wash bottle 11, separating the oil-gas mixture. Channel B of solenoid valve C4.3 is opened, and channel A of solenoid valve C4.3 is closed. After a reasonable period of time, channel B of solenoid valve C4.3 is closed, and solenoid valve A4.3 is opened, passing the remaining gas to the gas sensor module 6 for detection of gas composition and concentration. The device uses characteristic values and the three-ratio method for comprehensive analysis in accordance with the "DLT722-2014 Guidelines for Analysis and Judgment of Dissolved Gases in Transformer Oil." The characteristic value method first checks whether H2, C2H2, and total hydrocarbons have reached cautionary values. If not, it can be determined that no abnormality has been detected, but the gas concentration values must be saved for use in calculating the gas production rate and increasing the gas production rate cautionary value the next time gas is present inside the gas relay 15. If H2, C2H2, and total hydrocarbons have reached cautionary values (except for H2), the three-ratio method can be further combined for diagnosis and fault characterization.

[0142] The control unit 1 can Figure 4 When light gas is activated, the automatic gas extraction detection control process extracts gas from the gas relay 15 and analyzes the information collected by the gas sensor 6. The fault type is analyzed according to the gas analysis method, and the gas composition, concentration, and fault type are transmitted to the display module 7. The fault information is transmitted to the substation background monitoring system through the communication module 8, which can realize the understanding of the on-site situation in advance and improve the disposal efficiency.

[0143] Figure 5 The diagram below shows the external structure of the device. When the device is operating normally, the operation indicator light is on and the alarm light is off. When the substation main transformer light gas alarm is activated or the device is abnormal, the alarm light is on.

[0144] The device can be applied to the following situations but is not limited to the following situations:

[0145] In case (1), the main transformer light gas action alarm is triggered, but the device recognizes that there is no gas inside the gas relay 15, and the device does not start the automatic gas extraction detection control process. The operating personnel arrive at the scene to confirm. If there is indeed no gas inside the gas relay 15, contact relevant professionals to investigate other reasons. However, if there is gas inside the gas relay 15, it may be an image recognition error, and it is necessary to enable the automatic gas extraction detection function of the device through the manual button of the human-computer interaction module 9 of the device.

[0146] In case (2), the main transformer light gas alarm is triggered. At the same time, the device recognizes that there is gas inside the gas relay 15. The device performs gas extraction detection and analysis according to the automatic gas extraction detection control process, realizes automatic gas extraction and transmits the gas composition and fault type to the substation background monitoring system, thereby improving the gas extraction efficiency and fault analysis decision speed.

[0147] In case (3), the main transformer light gas alarm does not trigger, but the device recognizes that there is gas inside the gas relay 15 and transmits the above situation to the monitoring background. In this way, the operator arrives at the site for a special inspection and activates the automatic gas extraction detection function of the device through the manual button of the human-machine interaction module 9 of the device. The device realizes real-time monitoring and effectively prevents the situation where a minor fault inside the main transformer develops into a serious fault. Because the gas production rate of a minor fault is low and the gas volume is not enough to trigger the light gas alarm, but the substation has just been inspected, the next manual inspection will take about one week to two weeks. During this period, the minor internal fault will gradually develop into a major fault.

[0148] This device effectively addresses the time-consuming and labor-intensive manual gas extraction process for the main transformer's light gas protection, requiring two people to work together and resulting in complex operations. The device also addresses issues such as the risk of glass tube breakage during the extraction process and contamination of the oscillating gas in the syringe during testing. It enables online gas detection, avoiding the traditional method of time-consuming gas extraction, testing, and result generation. Transformer faults can easily progress from minor faults to major faults, missing the optimal time for decision-making and troubleshooting, and causing harm to equipment, personnel, and the power grid. Furthermore, during routine inspections, when the light gas protection is not activated, operations and maintenance personnel may also need to extract and test gas inside the gas relay 15. This device's human-computer interaction module allows for automated gas extraction, saving time and effort, improving gas extraction efficiency, and accelerating fault analysis and decision-making.

[0149] The above are merely examples of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the full scope of the present invention to be approved.

Claims

1. A device and method for automatically detecting gas from a main transformer of a substation, which is used in conjunction with a gas relay (15) and a gas collecting box (19). The gas collecting box (19) is connected to the gas relay (15) through an air duct (17). The gas relay (15) is connected to the main transformer body (14) of the substation. The device is characterized in that: The integrated device for automatic gas extraction detection of a main transformer in a substation comprises: a control unit (1), an alarm signal synchronization unit (2), an image recognition module (3), a solenoid valve (4), a micro vacuum pump (5), a gas sensor module (6), a display module (7), a communication module (8), a human-computer interaction module (9), a power supply module (10), a gas washing bottle (11), and a gas storage bottle (12); wherein: The alarm signal synchronization unit (2), the image recognition module (3), the electromagnetic valve (4), the micro vacuum pump (5), the gas sensor module (6), the display module (7), the communication module (8), the human-computer interaction module (9), and the power module (10) are all connected to the control unit (1); the gas collecting box (19) is connected to the gas washing bottle (11) through a pipeline, the gas washing bottle (11) is connected to the gas storage bottle (12) through a pipeline, and the gas storage bottle (12) is further connected to the micro vacuum pump (5) through a pipeline; the gas sensor module (6) is arranged in the pipeline connecting the gas storage bottle (12) to the micro vacuum pump (5); the power module (10) is also connected to the image recognition module (3), the electromagnetic valve (4), the micro vacuum pump (5), the gas sensor module (6), the display module (7), and the communication module (8) respectively; A gas collecting box liquid level sensor (19.1) is installed inside the gas collecting box (19); a manual oil drain valve (19.2) and a solenoid valve A (4.1) for draining oil are provided below the gas collecting box (19); a manual air drain valve (19.3) and a solenoid valve B (4.2) are provided in sequence on the ventilation pipe connecting the gas collecting box (19) to the gas washing bottle (11); the solenoid valve C (4.3) has two channels A and B, each of which can be independently controlled to be open or closed, wherein the channel A of the solenoid valve C (4.3) is connected to the gas sensor module (6), and the channel B of the solenoid valve C (4.3) is connected to the gas storage bottle (12). The solenoid valve C (4.3) is also connected to the gas washing bottle (11); and a solenoid valve D (4.4) is provided on the pipe between the gas sensor module (6) and the micro vacuum pump (5).

2. The integrated device for automatic gas extraction and detection of main transformer gas in a substation according to claim 1, characterized in that: The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements: First, in the integrated device for automatic gas extraction and detection of main transformer gas in a substation, the image recognition module (3) is specifically a dual-wavelength optical camera and / or an infrared image camera; Secondly, the gas sensor module (6) integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6.

3. The integrated device for automatic gas extraction and detection of main transformer gas in a substation according to claim 1 is characterized in that: The integrated device for automatic gas extraction and detection of main transformer gas in substations meets one or a combination of the following requirements: First, the control unit (1) collects light gas action circuit information and image recognition module (3) information through the alarm signal synchronization unit (2) to determine whether light gas action actually occurs, and completes gas extraction and detection by controlling the start and stop of the micro vacuum pump (5) and the solenoid valve D (4.4); Secondly, the micro vacuum pump (5) is used to maintain the vacuum state of each gas extraction pipeline; Thirdly, the display module (7) is used to display the composition, concentration and fault type information of the gas; Fourthly, the communication module (8) is used for communication between the control unit (1) and the substation monitoring background; Fifth, each solenoid valve as well as the image recognition module (3), the gas sensor module (6), the micro vacuum pump (5), the display module (7), the communication module (8), and the power supply module (10) are all connected to the control unit (1).

4. A method for automatically detecting gas from a main transformer of a substation using the integrated device for automatically detecting gas from a main transformer of a substation according to claim 1, characterized in that: The automatic gas extraction detection method for the main transformer of the substation is based on an image recognition strategy of dual-wavelength optical characteristics, and uses an image recognition module (3) to obtain image characteristics of the insulating oil and gas distribution in the gas relay (15) through the gas relay observation window (15.1), thereby identifying the internal gas state; When the main transformer of the substation generates a light gas alarm, the device determines whether there is gas inside the gas relay (15) through the image recognition module (3). If there is gas inside the gas relay (15), the device starts the automatic gas extraction detection control process to perform gas extraction detection; if there is no gas inside the gas relay (15), it is determined to be a false operation signal.

5. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 4, characterized in that: The image recognition module (3) is specifically a dual-wavelength optical camera or / and an infrared image camera; the image recognition module (3) is arranged on the front face of the gas relay observation window (15.1) on the gas relay (15), so as to facilitate observation of whether there is gas in the gas relay (15) through the gas relay observation window (15.1); The controller (1) is connected to a dual-wavelength optical camera or / and an infrared image camera; wherein: the lens field of view of the dual-wavelength optical camera or / and the infrared image camera includes a gas relay observation window (15.1) corresponding to the gas relay (15), and the dual-wavelength optical camera has the ability to simultaneously recognize visible light wavelengths and infrared light wavelengths; the former can perceive color distribution characteristics, and the latter can capture temperature distribution characteristics; the image characteristics of the insulating oil and gas distribution in the gas relay (15) are mutually compared, identified, and verified under the two wavelength optical images; When the main transformer is in a normal operating state, the gas relay observation window (15.1) of the gas relay (15) is filled with transformer oil; when a fault occurs inside the main transformer body (14), the transformer oil decomposes to produce gas, and the gas relay (15) is filled with gas. The dual-wavelength optical camera identifies an obvious boundary line between the gas and the transformer oil on the glass mirror of the gas relay observation window (15.1) of the gas relay (15). By identifying the feature of the boundary line between the gas and the transformer oil on the glass mirror of the gas relay observation window (15.1), it is possible to determine whether the main transformer has a minor gas malfunction, thereby effectively avoiding malfunction of the automatic detection device caused by the minor gas malfunction.

6. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 5, characterized in that: The identification algorithm for determining whether there is gas in the observation window of the transformer gas relay used in the automatic gas extraction detection method for the main transformer of the substation meets the following steps and content requirements: Step 1: The control unit (1) acquires and analyzes the dual-wavelength optical camera and infrared image camera data in the image recognition module (3); Step 2: Normalize the brightness of the visible light image: Convert the image from RGB color space to HSV color space, calculate the average global image brightness, and adjust the bias value to keep the overall brightness of the image close to a fixed value M = 40% to 60%, thereby eliminating the influence of ambient light on the visible light image. Step 3: Infrared image contrast correction: The infrared image color data acquired by the infrared camera is directly converted from temperature data. The upper and lower temperature limits within the global image are calculated and used as the boundaries of the color mapping display to ensure higher contrast between different temperatures within the image. Step 4: Based on the circular feature, the position of the observation window in the optical high-definition image obtained by the dual-wavelength optical camera and the infrared image obtained by the infrared image camera is found, and the image within the circle is intercepted based on the circular boundary; Step 5: Search for the optimal horizontal dividing line using the traversal method: First, filter out invalid RGB colors, then set the horizontal dividing line from top to bottom, calculate and record the color average difference between the upper and lower parts of the horizontal dividing line; after traversal, select the horizontal dividing line with the largest color average difference as the traversal method recognition results R1 and R2; Step 6: Calculate the optimal horizontal segmentation line using the centroid method: First, filter out invalid RGB colors, then extract the upper and lower images using the contours, calculate the color centroid within the contours, and use the centroid to reversely calculate the position of the horizontal segmentation line, which are used as the centroid method recognition results R3 and R4. Step 7: Comparison and judgment of recognition results: The horizontal segmentation line position recognition results R1, R2, R3 and R4 in steps 5 and 6 are combined. If the difference between the results is within 5%, they are considered similar. The number of similar results in R1, R2, R3 and R4 is counted using the "result factor" N. If the "result factor" N ≥ 2, the result is valid, and the average value of these N similar result sets is used as the horizontal dividing line to identify the result.

7. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 3, characterized in that: The gas sensor module (6) integrates sensors for detecting the following characteristic gas components and concentrations: H2, CH4, C2H2, C2H4, C2H6.

8. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 6 or 7, characterized in that: The method for automatically detecting gas extraction from the main transformer of a substation meets one or a combination of the following requirements: First, the control unit (1) collects light gas action circuit information and image recognition module (3) information through the alarm signal synchronization unit (2) to determine whether light gas action actually occurs, and completes gas extraction and detection by controlling the start and stop of the micro vacuum pump (5) and the solenoid valve (4); Secondly, the micro vacuum pump (5) is used to maintain the vacuum state of each gas extraction pipeline; Thirdly, the display module (7) is used to display the composition, concentration and fault type information of the gas; Fourthly, the communication module (8) is used for communication between the control unit (1) and the substation monitoring background, so that the dispatcher and the operator can immediately grasp the specific situation of the main transformer light gas operation; Fifth, each solenoid valve, image recognition module (3), gas sensor module (6), micro vacuum pump (5), display module (7), communication module (8), and power module (10) are all connected to the control unit (1).

9. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 8, characterized in that: In the automatic gas extraction detection method for the main transformer of the substation, the automatic gas extraction detection control process meets the following steps and content requirements: First, the gas collecting box is drained of oil, and the gas inside the gas relay (15) is led down to the upper part of the gas collecting box (19). The specific time for draining the oil is to stop the solenoid valve A (4.1). The gas collecting box liquid level sensor (19.1) in the gas collecting box (19) detects that the oil level no longer changes, which is used as a criterion for closing the solenoid valve A (4.1); at the same time, the micro vacuum pump (5) is controlled to evacuate the gas washing bottle (11), the gas storage bottle (12) and the pipeline to avoid gas contamination; then the solenoid valve B (4.2) of the gas collecting box is controlled to open, and the gas flows into the pipeline of the automatic gas extraction and detection integrated device of the main transformer of the substation until the liquid level sensor in the gas collecting box (19) detects that the position is reset to the top of the gas collecting box, and the solenoid valve B (4.2) is closed; The gas-oil mixture first passes through the gas washing bottle (11) to separate the gas-oil mixture, and at the same time, the B channel of the solenoid valve C (4.3) of the gas storage bottle (12) is controlled to be open while the A channel is closed; the gas continues to pass through the pipeline with the filter (18) to first inflate the gas storage bottle (12), and after a fixed time, the B channel of the solenoid valve C (4.3) is controlled to be closed, and then the A channel of the solenoid valve C (4.3) is opened, and the solenoid valve C (4.3) transmits the remaining gas to the gas sensor module (6) to facilitate the detection of the gas composition and concentration.

10. The method for automatically extracting and detecting gas from a main transformer in a substation according to claim 9, characterized in that: The automatic gas sampling detection method for the main transformer of the substation meets the following requirements: According to the "DLT722-2014 Guidelines for Analysis and Judgment of Dissolved Gases in Transformer Oil", the characteristic value and three-ratio method are used for comprehensive judgment: First, the characteristic value method should be used to check whether H2, C2H2 and total hydrocarbons have reached the caution value. If not, it is judged that there is no abnormality, but the gas concentration value needs to be saved. The next time there is gas inside the gas relay, it will be used to calculate the gas production rate and increase the gas production rate caution value analysis; If H2, C2H2 and total hydrocarbons reach the attention value, the three-ratio method is further used to make judgments and characterize the fault. The image recognition module (3) is arranged on the front face of the gas relay observation window (15.1) on the gas relay (15), so as to facilitate observation of whether there is gas in the gas relay (15) through the gas relay observation window (15.1).

Images