Converter station insulation bushing defect identification method and system

Abstract

The application discloses a converter station insulating sleeve defect identification method and system, through alignment processing of obtained temperature data and field intensity data of the insulating sleeve to be identified, first temperature data and first field intensity data are obtained, if the first temperature data and the first field intensity data meet the judgment condition, the radial gradient, the axial gradient and the temperature anomaly index are calculated based on the first temperature data, the electric field anomaly index is calculated based on the first field intensity data, and the electric field temperature ratio is calculated according to the first temperature data and the first field intensity data; according to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, the scoring model is used for calculation, the score of each defect type is obtained, the defect type with the maximum score is selected as the defect identification result of the insulating sleeve to be identified, and the above method effectively improves the defect detection precision of the converter station insulating sleeve.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of high-voltage direct current converter station temperature monitoring, and in particular to a converter station insulation bushing defect identification method and system. BACKGROUND

[0002] High-voltage direct current (HVDC) converter stations are widely used for long-distance and large-capacity power transmission, and their safe and stable operation depends on the insulation performance of the bushings. Oil-immersed bushings are usually composed of a porcelain sleeve, an insulation core and an external metal cover, and are filled with insulation oil inside to achieve uniform electric field distribution and heat conduction. With the increase of operating time and the influence of environmental factors, the following typical defects may occur inside the bushing: 1) moisture, the insulation oil or paper insulation inside the bushing is affected by environmental humidity or manufacturing process, resulting in water content, forming local water accumulation, changing the dielectric constant and causing electric field distortion. 2) surface defects of the conducting rod, scratches, metal particle adhesion or corrosion on the surface of the conducting rod, resulting in partial discharge or abnormal electric heat concentration. 3) metal deposit simulation creeping on the inner wall of the porcelain sleeve: metal impurities or "metal deposit" marks are formed on the inner wall of the porcelain sleeve, simulating the creeping path, enhancing the current channel along the surface of the porcelain sleeve and producing local overheating points. 4) ground screen loss, i.e. virtual welding and ground loss of the thimble structure, loose connection of the ground screen or thimble, virtual welding of the welding point, resulting in loss of the equipotential ring, causing local field strength mutation and temperature anomaly. Therefore, it is necessary to monitor the inside of the bushing to improve the early warning capability.

[0003] However, in the existing scheme, only a small number of FBG (Fiber Bragg Grating) is usually arranged at the top of the bushing in a single section, and the calibration curve does not consider the complex structure inside the bushing, resulting in errors in identifying defects of the oil-immersed bushing and low accuracy. SUMMARY

[0004] To solve the above technical problems, the present application provides a converter station insulation bushing defect identification method and system to solve the technical problem of low temperature monitoring accuracy of the oil-immersed bushing in the prior art.

[0005] The first aspect of the present application provides a converter station insulation bushing defect identification method, which comprises:

[0006] obtaining the first temperature data and the first field strength data of the aligned insulation bushing to be identified;

[0007] comparing the first temperature data with the preset temperature threshold to obtain the first comparison result, and comparing the first field strength data with the preset electric field threshold to obtain the second comparison result;

[0008] When any one of the first comparison result and the second comparison result satisfies a preset condition, a defect recognition step is entered, wherein the defect recognition step comprises:

[0009] Based on the first temperature data, a radial gradient, an axial gradient and a temperature anomaly index are calculated, based on the first field strength data, an electric field anomaly index is calculated, and based on the first temperature data and the first field strength data, an electric field temperature ratio is calculated;

[0010] According to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed;

[0011] The feature vector is input into a scoring model corresponding to each defect type for calculation to obtain a score corresponding to each defect type, and a defect type with the largest score is selected as a defect recognition result of the insulating sleeve to be recognized.

[0012] In a possible implementation manner of the first aspect, the first temperature data is collected by a group of temperature detection points arranged inside the insulating sleeve to be recognized, and the group of temperature detection points are evenly arranged on any cross section of the insulating sleeve to be recognized along a radial direction of the insulating sleeve to be recognized.

[0013] The first field strength data is collected by electric field sampling electrodes installed in a plurality of key annular grooves inside a sleeve shell of the insulating sleeve to be recognized.

[0014] In a possible implementation manner of the first aspect, based on the first temperature data, the radial gradient and the axial gradient are calculated, comprising:

[0015] The first temperature data is extracted to obtain outer layer temperature data and inner layer temperature data of each layer;

[0016] Based on the outer layer temperature data and the inner layer temperature data of each layer, a radial gradient of each layer is obtained;

[0017] According to the maximum temperature value and the minimum temperature value of each layer in the first temperature data, an axial gradient of each layer is obtained.

[0018] In a possible implementation manner of the first aspect, based on the first temperature data, the temperature anomaly index is calculated, comprising:

[0019] Based on the first temperature data and the number of temperature detection points, a temperature average value is obtained;

[0020] According to the temperature average value and the first temperature data corresponding to each temperature detection point, a temperature anomaly index set is obtained;

[0021] The largest index in the temperature anomaly index set is selected as the temperature anomaly index.

[0022] In a possible implementation manner of the first aspect, the scoring model is obtained by using the sample feature vectors, and the obtaining includes:

[0023] obtaining a sample feature vector of the insulation sleeve to be identified;

[0024] dividing the sample feature vectors into corresponding defect types to obtain a plurality of feature sample libraries;

[0025] calculating an intra-class variation coefficient of each sample feature vector in each feature sample library;

[0026] based on the sample feature vectors, calculating an inter-class variation coefficient of any one feature sample library and the remaining feature sample libraries, and obtaining a discrimination index according to the intra-class variation coefficient and the inter-class variation coefficient;

[0027] normalizing the discrimination index to obtain a scoring model corresponding to each defect type.

[0028] In a possible implementation manner of the first aspect, after obtaining the defect identification result, the method further includes:

[0029] based on the first temperature data, obtaining temperature information of each temperature detection point, and obtaining a temperature deviation of each temperature detection point according to the temperature information of each temperature detection point and the temperature average value;

[0030] determining a weight of each temperature detection point according to the temperature deviation, and calculating a weighted barycenter based on the weight of each temperature detection point and the position coordinates to obtain the defect position.

[0031] To solve the same technical problem, a second aspect of the embodiment of the application provides an insulation sleeve defect identification system of a converter station, which includes:

[0032] an acquisition module, configured to acquire first temperature data and first field intensity data of an insulation sleeve to be identified after alignment;

[0033] a comparison module, configured to compare the first temperature data with a preset temperature threshold to obtain a first comparison result, and compare the first field intensity data with a preset electric field threshold to obtain a second comparison result;

[0034] a defect identification module, configured to enter a defect identification step when any one of the first comparison result and the second comparison result meets a preset condition, wherein the defect identification step includes:

[0035] based on the first temperature data, calculating a radial gradient, an axial gradient and a temperature anomaly index, based on the first field intensity data, calculating an electric field anomaly index, and based on the first temperature data and the first field intensity data, calculating an electric field temperature ratio;

[0036] According to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed;

[0037] The feature vector is input into a scoring model corresponding to each defect type for calculation to obtain a score corresponding to each defect type, and a defect type with the maximum score is selected as a defect recognition result of the insulating sleeve to be identified.

[0038] In a possible implementation manner of the second aspect, the first temperature data is acquired by a set of temperature detection points arranged inside the insulating sleeve to be identified, and the set of temperature detection points are evenly arranged on any cross section of the insulating sleeve to be identified along a radial direction of the insulating sleeve to be identified.

[0039] The first field strength data is acquired by an electric field sampling electrode installed inside a key annular groove of a sleeve shell of the insulating sleeve to be identified.

[0040] The third aspect of the embodiment of the present application provides a computer device, comprising:

[0041] A memory for storing a computer program;

[0042] A processor for executing the computer program to realize the steps of the converter station insulating sleeve defect identification method according to the first aspect.

[0043] The fourth aspect of the embodiment of the present application provides a storage medium, and the storage medium stores a computer program, and the computer program is executed by a processor to realize the steps of the converter station insulating sleeve defect identification method according to the first aspect.

[0044] The technical scheme of the present application has the following advantages:

[0045] The converter station insulating sleeve defect identification method provided by the embodiment of the present application, by using the temperature detection points and the electric field detection points arranged on the insulating sleeve to be identified to obtain temperature data and field strength data, after aligning processing of the temperature data and the field strength data, first temperature data and first field strength data are obtained, when any one of the first comparison result and the second comparison result meets the preset condition, defect identification is entered, based on the first temperature data, radial gradient, axial gradient and temperature anomaly index are calculated, based on the first field strength data, electric field anomaly index is calculated, and according to the first temperature data and the first field strength data, electric field temperature ratio is calculated; according to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed; the feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect identification result of the insulating sleeve to be identified, and the above method can improve the defect detection precision of the converter station insulating sleeve by multi-point monitoring and analysis on the inside of the converter station insulating sleeve. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the specific embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0047] Figure 1 The identification flowchart of the converter station insulating sleeve defect identification method in the embodiment of the present application;

[0048] Figure 2 The monitoring device arrangement structure schematic diagram of the converter station insulating sleeve defect identification method in the embodiment of the present application;

[0049] Figure 3 The defect identification flowchart of the converter station insulating sleeve defect identification method in the embodiment of the present application;

[0050] Figure 4 The structure block diagram of the converter station insulating sleeve defect identification system in the embodiment of the present application;

[0051] The drawings show that: 400, the converter station insulating sleeve defect identification system; 401, the acquisition module; 402, the comparison module; 403, the defect identification module. DETAILED DESCRIPTION

[0052] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0053] The method for identifying defects of an insulating bushing of a converter station provided by the embodiments of the present application is shown in Figure 1 , which is a flowchart of the method for identifying defects of an insulating bushing of a converter station, comprising steps S101 to S103, and each step is specifically as follows. Figure 1

[0054] S101, acquiring first temperature data and first field strength data of the aligned insulating bushing to be identified.

[0055] In the embodiment, the internal temperature distribution is obtained by using the FBG sensor arranged in the radial-axial combination to obtain the first temperature data. The micro electric field sampling electrodes (E1-E4) are installed in the key annular grooves on the inside of the bushing shell, one electrode in each layer, and one reference electrode at the bottom, to monitor the local field strength in real time and obtain the first field strength data.

[0056] All sensing wires are led out to an integrated data acquisition and processing unit through the bottom multi-channel sealing joint, and the industrial bus is connected with the control system of the converter station. The temperature and electric field signals are synchronously collected according to the spatial coordinates and time stamps, the multi-dimensional representation of the electric-thermal coupling dynamic process is realized, the characteristic electric-thermal response under different defect types such as moisture and surface defects of the conductive rod is identified through data fusion and model analysis.

[0057] In an embodiment, the first temperature data is collected by a group of temperature detection points arranged in the internal part of the insulating bushing to be identified, and the group of temperature detection points are evenly arranged on any cross section of the insulating bushing to be identified along the radial direction of the insulating bushing to be identified.

[0058] The first field strength data is collected by the electric field sampling electrodes installed in the key annular grooves on the inside of the bushing shell of the insulating bushing to be identified.

[0059] In the embodiment, as shown in Figure 2 , the method for identifying defects of an insulating bushing of a converter station comprises the following steps. Figure 2 ​The oil-immersed transformer bushing built-in multi-point electric-thermal combined monitoring device of the application is characterized in that four FBG optical fibers are evenly arranged in the radial direction on each cross section, are arranged at equal intervals, and three groups of FBG sensors, i.e. temperature detection points, are arranged at 20%, 50% and 80% of the length of the bushing, with four in each layer, and each FBG sensor is a temperature detection point. One path of each optical fiber passes through the core and grinds the grating at the specified axial cross section to realize "point" measurement. The miniature electric field sampling electrodes, i.e. sampling electrode E1, sampling electrode E2, sampling electrode E3 and sampling electrode E4, are installed in the key annular groove on the inside of the bushing shell, with one electrode in each layer and one reference electrode at the bottom, to monitor the local field strength in real time, and the field strength data can also be electric field strength data.

[0060] As shown in Figure 3 , Figure 3 The defect identification flowchart of the defect identification method for the converter station insulation bushing. After reading the 12-way temperature and 4-way field strength data, time / space alignment denoising processing is required, and then each data is subjected to over-limit judgment. If it is judged to be over-limit, defect identification and positioning are carried out. Specifically, to ensure the consistency of the data collected by the two modules in time, GPS clock is used for synchronization. When the system starts, the FBG sensor and the electric field sampling module are connected to the GPS receiver to obtain accurate global positioning system time signals, which are used as the reference of the respective sampling time stamps. At the same time, to prevent time synchronization failure caused by loss of GPS signal, the system is also equipped with a local RTC (real-time clock) chip as a backup time source. When the GPS signal is normal, the local RTC time is updated using the GPS clock; when the GPS signal is interrupted, the system automatically switches to the local RTC to provide the time stamp, ensuring the continuity of time synchronization. Due to the different sampling frequencies of the two modules, the data time points obtained are not completely consistent. To fill in the temperature and electric field data to the common time grid, the linear interpolation method or the nearest neighbor matching method is used to fill in the first temperature data and the first electric field data to the common time grid. For example, assuming that the FBG sensor collects temperature data T1, T2, T3…… at t1, t2, t3……, and the electric field sampling module collects electric field data C1, C2, C3…… at s1, s2, s3……. First, a common time grid is determined, and the time interval of the grid is set according to the actual requirements, such as 50ms. Then, for each time point t in the common time grid, two adjacent time points and are found in the data time sequence of the FBG sensor , and the temperature value T corresponding to the time is calculated by the linear interpolation formula. For electric field data, the same method is used for linear interpolation processing, so that the temperature data and the electric field data are filled in to the common time grid, realizing time alignment.

[0061] In the spatial alignment, the coordinates are calibrated in advance for the sensing positions of the axial three layers and the radial four temperature detection points. In the collection process, the position information of all points is attached to the data record to ensure that the subsequent feature calculation is based on the same spatial reference, so that the first temperature data and the second electric field data after alignment are aligned.

[0062] S102, compare the first temperature data with the preset temperature threshold to obtain a first comparison result, and compare the first field strength data with the preset electric field threshold to obtain a second comparison result.

[0063] In this embodiment, after alignment and denoising, it is determined whether the specified limit is exceeded. Specifically, all temperature data of the first temperature data is compared with the preset temperature threshold to obtain a first comparison result, and all field strength data in the first field strength data is compared with the preset electric field threshold to obtain a second comparison result. The first comparison result of all temperature detection points and the second comparison result of all electric field measurement points are comprehensively analyzed. If the first comparison result of any one temperature detection point is out of the specified limit, or the second comparison result of any one electric field measurement point is out of the specified limit, the system determines that the specified limit is exceeded at the current time. For example, the temperature threshold is set to 50 degrees Celsius, and the electric field threshold is set to 10 kV / m. For example , If any temperature detection point or electric field measurement point , it is determined that the specified limit is exceeded, and the defect recognition and positioning link is entered, otherwise the data collection continues to be recycled.

[0064] S103, when any one of the first comparison result and the second comparison result meets the preset condition, the defect recognition step is entered, wherein the defect recognition step includes:

[0065] Based on the first temperature data, the radial gradient, the axial gradient and the temperature anomaly index are calculated, based on the first field strength data, the electric field anomaly index is calculated, and based on the first temperature data and the first field strength data, the electric field temperature ratio is calculated;

[0066] According to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed;

[0067] The feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect recognition result of the insulating sleeve to be identified.

[0068] ​In this embodiment, when the limit is determined, the defect recognition and positioning link is entered immediately, otherwise the data collection continues to circulate. Under normal operating conditions, there is a significant temperature gradient and electric field in the sleeve, that is, the electric field strength increases along the radial direction, and the field strength is the largest at the outermost shield. Using 3 layers of axial x 4 point radial total of 12 temperature detection points and 4 electric field measuring points, the temperature / electric field distribution anomalies caused by different defects are analyzed according to the characteristic parameters under each typical defect. By setting 12 FBG points covering the radial and axial multi-layer sections, a three-dimensional temperature cloud map is formed, the internal hot spot position can be accurately determined, and the risk of missed detection is reduced. In addition, the field strength and temperature data are obtained, which reveals the electric-thermal interaction mechanism caused by defects, and can distinguish between simple thermal anomalies and composite defects caused by electric field distortion.

[0069] Table 1 Temperature and electric field characteristics of various defects

[0070]

[0071] From the above table, different types of defects show their own unique characteristics in temperature distribution and electric field response, and there are obvious differences between them. Specifically:

[0072] 1) Moisture defect. When moisture slowly penetrates into the sleeve, it first stays near the outer layer of the capacitor core and the inner wall of the porcelain sleeve. Therefore, when it is wet, a hot spot often appears on the outer layer or the lower part, the corresponding radial temperature gradient increases significantly, and the temperature of the outer layer is significantly higher than that of the inner layer; the temperature rise ratio, i.e. the temperature difference between the outer layer and the inner layer, is significantly larger. Since humidity mainly increases dielectric loss, the field strength at the electric field sensor location may not change abruptly. Typical quantitative indicators can be manifested as an increase in the temperature difference between the outer layer and the inner layer, and a small change in the axial gradient.

[0073] 2) Defect on the surface of the conductive rod. When there is a defect on the surface of the conductive rod, local overheating will occur at the contact. Both simulation and field research show that a local hot spot will form at the defect, usually at the upper part of the sleeve where the conductive rod is joined, causing the temperature on one side of the axial layer to rise sharply, the radial and axial temperature gradients to change abruptly, and the temperature rise ratio, i.e. the temperature difference between the hot spot and the surrounding temperature, to increase significantly. Since the defect changes the local conductive path, the electric field distribution may also change abruptly, for example, the field strength near the defect may be abnormally high or asymmetric.

[0074] 3) Metal deposition on the inner wall of the porcelain sleeve. The attachment of metal particles or deposits on the inner wall of the porcelain sleeve is equivalent to forming a local conductive area on the insulating inner wall, which will cause local charge accumulation and local discharge. At this time, the hot spot often appears near the deposition position, and the temperature sensor reading is abnormally high, causing the radial temperature gradient at that position to increase abnormally. The electric field sensor near the deposition position may measure an abnormally high electric field strength, i.e. an electric field mutation, so that the electric field / temperature ratio at that point is significantly abnormal. The typical characteristics of this defect are the simultaneous occurrence of local temperature peak and electric field peak.

[0075] 4) Grounded shield loss, i.e. ground screen open circuit. Normally, the outer sleeve ground screen is grounded, and the electric field distribution gradually decreases along the radial direction. If the ground screen loses grounding, the outer screen will float, and the electric field distribution will be severely distorted. At this time, the electric field strength of the outer layer of the sleeve may abnormally increase, causing the overall temperature of the outer layer to rise, and even the entire outer layer to be at a relatively high temperature. The radial distribution shows uniform temperature rise, and the temperature of each radial point is close. The axial gradient may also increase. This defect can be characterized by the overall outer layer temperature rise amplitude and abnormal electric field distribution.

[0076] After time and space alignment, the first temperature data and the first field strength data are obtained, wherein the first temperature data includes temperature values of four radial temperature detection points and three axial temperature detection points of the insulating sleeve at different times, and the first temperature data can also include temperature values of four radial detection points and three axial detection points of the insulating sleeve at the same time; the first field strength data records electric field strength values of multiple electric field detection points around the insulating sleeve at the same time.

[0077] After obtaining the aligned first temperature data and first field strength data, the radial gradient, axial gradient, temperature anomaly index and electric field temperature ratio in the to-be-identified insulating sleeve are calculated, and a feature vector is constructed based on the above results. The feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect identification result of the to-be-identified insulating sleeve. The reason for the radial gradient and the axial gradient, i.e. the radial temperature gradient and the axial temperature gradient, is that the heat distribution of different defects has specificity, such as radial gradient for moisture and radial and axial gradient mutation for conductive rod defects. Gradient calculation can effectively distinguish the spatial mode of heat diffusion.

[0078] The field-temperature ratio, i.e. the electric field temperature ratio, relates the coupling relationship between electric field and temperature. Defects such as metal deposition can cause electric field distortion and temperature rise at the same time. The ratio can amplify the characteristics of such defects, such as a sudden increase in the electric field temperature ratio. The temperature anomaly index can quantify the degree of deviation from the normal state, avoiding the limitations of single threshold judgment.

[0079] In an embodiment, based on the first temperature data, the radial gradient and the axial gradient are calculated, including:

[0080] The first temperature data is extracted to obtain outer layer temperature data and inner layer temperature data of each layer;

[0081] Based on the outer layer temperature data and the inner layer temperature data of each layer, the radial gradient of each layer is obtained;

[0082] According to the maximum temperature value and the minimum temperature value of each layer in the first temperature data, the axial gradient of each layer is obtained.

[0083] In this embodiment, the heat distribution of different defects is specific. For example, moisture defects are dominated by radial gradients, while conductive rod defects exhibit abrupt changes in both radial and axial gradients. Gradient calculation can effectively distinguish the spatial patterns of heat diffusion. Specifically, the radial gradient can be calculated using the temperature difference between adjacent measuring points, and the specific calculation formula is as follows:

[0084]

[0085] In the formula, It is the first Radial temperature gradient of the layer, The outer layer temperature, This refers to the inner layer temperature.

[0086] When calculating the axial gradient, the temperature values ​​of all temperature detection points are determined from the first temperature data. The maximum and minimum temperature values ​​among these values ​​are found, and the difference between the maximum and minimum values ​​is calculated to obtain the axial gradient. The specific calculation formula is as follows:

[0087]

[0088] In the formula, This is the maximum temperature value. This is the minimum temperature value.

[0089] At the same time, for the first For each temperature detection point, calculate the ratio of its electric field to its temperature. The specific calculation formula is as follows:

[0090]

[0091] In the formula, For the first Temperature data from road temperature sensors. For the first Electric field intensity data at the road electric field sampling points.

[0092] In one embodiment, a temperature anomaly index is calculated based on first temperature data, including:

[0093] Based on the initial temperature data and the number of temperature detection points, the average temperature is obtained;

[0094] Based on the average temperature and the first temperature data corresponding to each temperature detection point, a set of temperature anomaly indices is obtained.

[0095] The index with the largest value in the set of temperature anomaly indices is selected as the temperature anomaly index.

[0096] In this embodiment, based on the first temperature data and the number of temperature detection points, an average temperature is obtained. Then, the average temperature is used in conjunction with the first... The temperature anomaly index is obtained from the temperature data of each temperature monitoring point. The formula for calculating the temperature anomaly index is:

[0097]

[0098] In the formula, For the first Temperature data from road temperature sensors. This represents the average temperature.

[0099] Similarly, the electric field anomaly index at each electric field sampling electrode is calculated based on the first field strength data. The formula for calculating the electric field anomaly index is as follows:

[0100]

[0101] In the formula, For the first Field strength data at road temperature sensing points This represents the average field strength.

[0102] Using the obtained radial gradient, axial gradient, temperature anomaly index, and electric field anomaly index, an eigenvector is constructed. .

[0103] In one embodiment, the scoring model is constructed using sample feature vectors, including:

[0104] Obtain the sample feature vector of the insulating sleeve to be identified;

[0105] The sample feature vectors are divided into corresponding defect types to obtain multiple feature sample libraries;

[0106] Calculate the intraclass coefficient of variation of the feature vectors of each sample in each feature sample library;

[0107] Based on the sample feature vectors, calculate the inter-class coefficient of variation between any feature sample library and the remaining feature sample libraries. Based on the intra-class coefficient of variation and the inter-class coefficient of variation, obtain the discrimination index.

[0108] The discrimination index is normalized to obtain the scoring model corresponding to each defect type.

[0109] In this embodiment, three-dimensional models of four typical defects—moisture absorption, surface defects on the conductive rod, metal deposition on the inner wall of the porcelain bushing, and grounding failure—are established. The electrostatic field equation (for calculating electric field distribution) and the thermal diffusion equation (for calculating temperature distribution) are coupled to simulate characteristic parameter data under different defect levels. The model includes key structures such as the porcelain bushing, insulating core, and insulating oil, and its geometric parameters are consistent with those of the actual bushing.

[0110] For each set of simulation results, such as a certain type of defect, a certain severity scene, extract the defined feature parameters: radial temperature gradient, axial temperature gradient, field-temperature ratio, temperature anomaly index, electric field anomaly index, summarize all the feature parameters of the simulation, form the feature sample library of 4 types of defects, at least 30 samples for each type of defect, covering different severity. Calculate the discriminability index of the feature parameters of the 4 types of defects in the sample library, i.e. the sample feature vector.

[0111] It should be noted that the discriminability index is a quantitative measure of the uniqueness of a certain feature parameter to a specific defect. That is, the difference between this feature and other defects in the target defect, specifically measured by "intra-class coefficient of variation / inter-class coefficient of variation". The intra-class coefficient of variation refers to the dispersion of a certain feature parameter within the same defect type, which can also be within the same feature sample library, reflecting the stability of the feature in the same defect. Its calculation formula is:

[0112]

[0113] In the formula, is the th defect type, is the feature parameter, is the standard deviation of the feature parameter of the th defect type, is the mean value, is the intra-class coefficient of variation of the th defect type.

[0114] The inter-class coefficient of variation refers to the difference between a certain feature parameter in the target feature sample library and other feature sample libraries, reflecting the uniqueness of the feature to the target defect. Its calculation formula is:

[0115]

[0116] In the formula, is the mean value of the feature parameter in other feature sample libraries, is the inter-class coefficient of variation of the th defect type.

[0117] It should be noted that each feature sample library corresponds to a type of defect, and each feature sample library includes multiple sample feature vectors, i.e. feature parameters.

[0118] The discriminability index refers to the combination of intra-class stability and inter-class uniqueness. The higher the value, the stronger the feature's ability to distinguish defects. Its calculation formula is:

[0119]

[0120] In the formula, is the discrimination index, is the coefficient of variation between classes for the th defect type, is the coefficient of variation within classes for the th defect type.

[0121] The discrimination index is normalized to the range, and the weight formula is as follows:

[0122]

[0123] wherein, is the dimension of the feature vector, is the weight value of the feature parameter in the th defect type.

[0124] The final scoring model is:

[0125]

[0126] wherein, is the weight value of the feature parameter in the th defect type, is the th feature parameter.

[0127] The is calculated for all classes of defects, respectively, and the largest scoring model value is selected, for example, when the of the "conductive rod defect" is higher than the other three, it is determined that the current most likely is a conductive rod surface defect.

[0128] It should be noted that the weight value of the feature parameter in each defect type is based on the electrical-thermal characteristics of the casing defect, and the discrimination ability of different features to a specific defect is quantitatively analyzed, such as the change of the radial temperature gradient when damp is more significant, which can truly reflect the correlation degree of the feature and the defect.

[0129] In an embodiment, after obtaining the defect recognition result, it further includes:

[0130] Based on the first temperature data, the temperature information of each temperature detection point is obtained, and based on the temperature information of each temperature detection point and the temperature average value, the temperature deviation of each temperature detection point is obtained;

[0131] According to the temperature deviation, the weight of each temperature detection point is determined, the weighted barycenter is calculated based on the weight and position coordinates of each temperature detection point, and the defect position is obtained.

[0132] In the embodiment, by using the heat diffusion principle, it is assumed that the additional heat released at the defect diffuses stably in the insulating medium, and the temperature of the measuring point closer to the defect increases more obviously, the temperature deviation of each temperature detection point is calculated, and the specific calculation formula is:

[0133]

[0134] In the formula, is the temperature data of the temperature sensing point of the first layer, is the temperature average value.

[0135] Then, in the sleeve own coordinate system, the three-dimensional coordinates of each temperature detection point are calibrated in advance as , wherein, corresponds to the radial position, corresponds to the axial layer, such as 0.2L, 0.5L, 0.8L.

[0136] The temperature deviation is taken as the weight, only the positive deviation is given weight , according to the obtained weight, the weighted center of gravity of all temperature detection points is calculated, that is, the approximate position of the defect most likely, the defect position is obtained, and the specific calculation formula is:

[0137]

[0138] In the formula, is the weight, is the three-dimensional coordinate of each temperature detection point, is the approximate position calculated by each coordinate.

[0139] It should be noted that the defect positioning is based on the weighted center of gravity of the temperature deviation, and the logical basis is the physical law of heat diffusion: the defect as a heat source, the temperature of the measuring point closer to the defect increases more obviously (the temperature deviation is larger), therefore, the spatial center of gravity is calculated by taking the temperature deviation as the weight, which can approximate the actual position of the defect.

[0140] After the defect identification result is obtained by the above method, the defect position, type and cause can be determined, blind maintenance can be avoided, if the oil paper sleeve insulation is identified as being damp, oil quality detection, vacuum oil injection or replacement of insulating oil can be carried out during maintenance, instead of disassembling the entire sleeve. If the internal metal foreign matter is identified, the position of the foreign matter can be located by an endoscope, and the foreign matter can be taken out by using a special tool, so as to reduce the damage to the sleeve body. If the ground screen loses the ground, it will cause serious distortion of the electric field distribution in the sleeve, and the local field strength may be suddenly increased to instantaneously break down the insulation, so the corresponding equipment must be immediately shut down to avoid defect expansion.

[0141] The defect identification system of the insulating sleeve of the converter station provided in the embodiment, as shown in Figure 4 , comprises a temperature sensor, a temperature data acquisition module, a temperature deviation calculation module, a three-dimensional coordinate calibration module, a weight calculation module and a weighted center of gravity calculation module. Figure 4 ​A system block diagram of the converter station insulation bushing defect identification system 400 in the embodiment of the present application, comprising:

[0142] The acquisition module 401 is configured to acquire the first temperature data and the first field strength data of the aligned insulation bushing to be identified.

[0143] The comparison module 402 is configured to compare the first temperature data with a preset temperature threshold to obtain a first comparison result, and compare the first field strength data with a preset electric field threshold to obtain a second comparison result.

[0144] The defect identification module 403 is configured to enter a defect identification step when any one of the first comparison result and the second comparison result satisfies a preset condition, wherein the defect identification step comprises:

[0145] Based on the first temperature data, the radial gradient, the axial gradient and the temperature anomaly index are calculated, based on the first field strength data, the electric field anomaly index is calculated, and based on the first temperature data and the first field strength data, the electric field temperature ratio is calculated.

[0146] According to the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed.

[0147] The feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect identification result of the insulation bushing to be identified.

[0148] In an embodiment, the first temperature data is collected by a group of temperature detection points arranged inside the insulation bushing to be identified, and the group of temperature detection points are evenly arranged on any cross section of the insulation bushing to be identified along the radial direction of the insulation bushing to be identified.

[0149] The first field strength data is collected by an electric field sampling electrode installed in a key annular groove on the inner side of the bushing shell of the insulation bushing to be identified.

[0150] The specific implementation of the converter station insulation bushing defect identification system is basically the same as the specific embodiments of the converter station insulation bushing defect identification method described above, and will not be repeated here.

[0151] In an embodiment of the present application, a computer device is provided, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the above steps when executing the computer program; the computer device provided in the embodiment has similar implementation principles and technical effects to the above method embodiments, and will not be repeated here.

[0152] In an embodiment of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the above steps; the computer readable storage medium provided by the embodiment has similar implementation principles and technical effects to the above method embodiments, and will not be described here.

[0153] The technical features of the above embodiments can be combined in any manner. To make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present disclosure.

[0154] The above specific embodiments further illustrate the purpose, technical solutions and advantages of the present application. It should be understood that the above description is only for specific embodiments of the present application and is not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for identifying defects of an insulation bushing of a converter station, characterized in that, The method comprises the following steps: obtaining first temperature data and first field strength data of an aligned insulation sleeve to be identified; comparing the first temperature data with a preset temperature threshold to obtain a first comparison result, and comparing the first field strength data with a preset electric field threshold to obtain a second comparison result; when either the first comparison result or the second comparison result meets a preset condition, entering a defect identification step, wherein the defect identification step comprises: based on the first temperature data, calculating a radial gradient, an axial gradient and a temperature anomaly index, based on the first field strength data, calculating an electric field anomaly index, and based on the first temperature data and the first field strength data, calculating an electric field temperature ratio; based on the radial gradient, the axial gradient, the temperature anomaly index, the electric field anomaly index and the electric field temperature ratio, a feature vector is constructed; the feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect identification result of the insulation sleeve to be identified; based on the first temperature data, a radial gradient and an axial gradient are calculated, comprising: extracting the first temperature data to obtain outer layer temperature data and inner layer temperature data of each layer; based on the outer layer temperature data and the inner layer temperature data of each layer, a radial gradient of each layer is obtained, wherein the calculation formula of the radial gradient is: wherein is the first radial temperature gradient of the layer, outer layer temperature, inner layer temperature; based on the maximum temperature value and the minimum temperature value of each layer in the first temperature data, an axial gradient of each layer is obtained, wherein the calculation formula of the axial gradient is: wherein is the maximum temperature value, is the minimum temperature value; based on the first temperature data, a temperature anomaly index is calculated, comprising: based on the first temperature data and the number of temperature detection points, a temperature average value is obtained; based on the temperature average value and the first temperature data corresponding to each temperature detection point, a temperature anomaly index set is obtained; the maximum index in the temperature anomaly index set is selected as the temperature anomaly index, wherein the calculation formula of the temperature anomaly index is: In the formula, is the first temperature data of the temperature sensing point, is the temperature average value.

2. The converter station insulation bushing defect identification method of claim 1, wherein, the first temperature data is collected by a group of temperature detection points arranged inside the insulation sleeve to be identified, and the group of temperature detection points are evenly arranged on any cross section of the insulation sleeve to be identified along the radial direction of the insulation sleeve to be identified; the first field strength data is collected by electric field sampling electrodes installed in a plurality of key annular grooves inside the shell of the insulation sleeve to be identified.

3. The converter station insulation bushing defect identification method of claim 1, wherein, The scoring model is constructed by using sample feature vectors, comprising: obtaining sample feature vectors of the insulation sleeve to be identified; dividing the sample feature vectors into corresponding defect types to obtain a plurality of feature sample libraries; calculating the within-class variation coefficient of each sample feature vector in each feature sample library, wherein the calculation formula of the within-class variation coefficient is: wherein is the th defect type, is the characteristic parameter, is the th defect type, is the standard deviation of the characteristic parameter is the mean, is the intra-class coefficient of variation for the th defect type; based on the sample feature vectors, calculating the between-class variation coefficient of any one of the feature sample libraries and the remaining feature sample libraries, and according to the within-class variation coefficient and the between-class variation coefficient, obtaining a discrimination index, wherein the calculation formula of the between-class variation coefficient is: wherein is the mean of the feature parameters of the other several classes of feature sample libraries, is the inter-class coefficient of variation of the defect type. The calculation formula of the discrimination index is: wherein is the index of discrimination, is the coefficient of interclass variation for the th defect type, is the coefficient of intraclass variation for the th defect type; The discrimination index is normalized to obtain a scoring model corresponding to each defect type.

4. The converter station insulation bushing defect identification method of claim 1, wherein, After obtaining the defect identification result, further comprising: Based on the first temperature data, obtain the temperature information of each temperature detection point, and according to the temperature information of each temperature detection point and the temperature average value, obtain the temperature deviation of each temperature detection point; According to the temperature deviation, determine the weight of each temperature detection point, and calculate the weighted center of gravity based on the weight and position coordinates of each temperature detection point to obtain the defect position.

5. A converter station insulation bushing defect identification system, characterized by, For realizing the converter station insulating sleeve defect identification method as claimed in any one of claims 1-4, comprising: An acquisition module is configured to acquire first temperature data and first field strength data of an insulating sleeve to be identified after alignment; A comparison module is configured to compare the first temperature data with a preset temperature threshold to obtain a first comparison result, and compare the first field strength data with a preset electric field threshold to obtain a second comparison result; A defect identification module is configured to enter a defect identification step when any one of the first comparison result and the second comparison result meets a preset condition, wherein the defect identification step comprises: Based on the first temperature data, calculate the radial gradient, axial gradient and temperature anomaly index, based on the first field strength data, calculate the electric field anomaly index, and based on the first temperature data and the first field strength data, calculate the electric field temperature ratio; according to the radial gradient, axial gradient, temperature anomaly index, electric field anomaly index and electric field temperature ratio, a feature vector is constructed; the feature vector is input into each defect type corresponding scoring model for calculation to obtain a score corresponding to each defect type, and the defect type with the maximum score is selected as the defect identification result of the insulating sleeve to be identified.

6. The converter station insulation bushing defect identification system of claim 5, wherein, The first temperature data is collected by a group of temperature detection points arranged inside the insulating sleeve to be identified, and the group of temperature detection points are evenly arranged on any cross section of the insulating sleeve to be identified along the radial direction of the insulating sleeve to be identified; The first field strength data is collected by an electric field sampling electrode installed in a key annular groove inside the sleeve shell of the insulating sleeve to be identified.

7. A computer device, characterized by Comprising: A memory for storing a computer program; A processor for executing the computer program to realize the converter station insulating sleeve defect identification method as claimed in any one of claims 1-4.

8. A storage medium, characterized by The storage medium has a computer program stored thereon, and the computer program is executed by the processor to realize the steps of the converter station insulating sleeve defect identification method as claimed in any one of claims 1-4.

Images