Withstand voltage prediction and control method for post insulator with anti-icing umbrella skirts inserted between converter stations

Abstract

The invention discloses a voltage withstanding prediction and control method for a post insulator with an anti-icing umbrella skirt inserted between converter stations, and relates to the technical field of converter stations. The method comprises the following steps: carrying out flashover tests in different icing states according to the structure of a post insulator with a large umbrella skirt inserted in a station, obtaining the ice flashover voltage of the typical post insulator with the umbrella skirt inserted in the station under different ice ridge bridging gap distances and pollution grades, and establishing an ice flashover voltage fitting model; in consideration of monitoring conditions in the converter station, operation and maintenance personnel or a monitoring device shoot insulator pictures at multiple angles every other time period when icing is serious, and the icing development condition is comprehensively predicted in combination with meteorological conditions and environmental factors; and in combination with the icing development condition and the pollution condition, predicting an ice flashover voltage value by adopting an ice flashover voltage fitting model, determining the safe operation voltage of the next time period, and determining whether voltage reduction operation is needed or not and the voltage reduction amplitude.

Description

Technical Field

[0001] The invention relates to the technical field of converter stations, and in particular to a method for predicting and controlling the withstand voltage of a post insulator with anti-icing sheds inserted between converter stations. Background Art

[0002] In recent years, extreme weather has occurred frequently. Rainy and snowy weather in winter can easily cause ice on insulators in transmission lines, AC converter stations or DC converter stations. In the iced state, the ice flash voltage of the insulator may be lower than the normal operating voltage, causing tripping accidents, seriously affecting the safe and stable transmission of electric energy. When ice is severe and the risk of tripping is high, the AC line needs to be shut down. DC transmission is different from AC transmission lines. It can achieve voltage reduction operation through voltage regulation control, but it is necessary to combine the development of the ice state and predict the ice flash voltage of the insulator in order to make corresponding operation control in advance.

[0003] The main factors affecting insulator ice flashover are: 1. After the insulator string is covered with ice, the ice layer becomes thicker and may bridge the insulator shed, turning the discharge path along the edge of the insulator into an ice layer bridging path; 2. During the icing process, the ice water contains contaminants, and the conductivity of the ice layer surface will increase when it is about to melt, which will affect the flashover voltage.

[0004] Currently, most studies have conducted ice flashover tests on insulators suspended on transmission lines, including different types of glass insulators, porcelain insulators, and composite insulators, to analyze the impact of their structures on icing conditions and ice flashover voltage. Many existing patents also carry out risk assessment or ice flash prediction on ice flashover of transmission line insulators. For example, a method for assessing the degree of ice coverage based on the extension of the insulator image umbrella (patent number CN201610698549.1) analyzes the ice coverage status, takes photos of the insulator before and after ice coverage by a fixed camera, and assesses the degree of ice coverage of the insulator by the changes in the extension of the insulator umbrella before and after ice coverage; a patent for a method and device for evaluating the insulation performance of iced insulators (patent number CN202211338519.1) directly defines the degree of ice coverage as five levels: slight, mild ice coverage, moderate ice coverage, severe ice coverage, and serious ice coverage. The corresponding judgment criteria are no obvious ice layer, the length of the ice ridge reaches 1 / 3 of the gap distance, the length of the ice ridge reaches 1 / 2 of the gap distance, the length of the ice ridge reaches 2 / 3 of the gap distance, and except for the gap between the ice ridges at the high-voltage end, the rest of the ice ridges are basically bridged. The performance of insulators is mainly judged based on the degree of electric field unevenness, ice coverage, ice surface status, etc. The risk value is directly defined, and then the risk is calculated according to the formula. Patent: A method for predicting ice flash risk of insulators based on the deduction rule (patent number CN202011015869.5). Feature quantities such as ice coverage also use image processing methods. The feature parameters such as ice coverage type, ice coverage, bridging status and dirty surface are used to score, and then make a judgment. A method for assessing the risk of ice flashover of insulators in ultra-high voltage AC transmission lines (patent number CN201610217229.X) is mainly aimed at ice coverage of insulators in transmission lines. Due to the long line and the difficulty in arranging monitoring devices along the entire line, the ice coverage is mainly estimated based on meteorological data for the ice coverage, and then the line insulator flashover voltage correction formula is used to calculate the flashover voltage value under the ice coverage state.

[0005] The above-mentioned prior art is mainly aimed at line insulators, and ice coating comprehensively considers ice thickness, ice weight, bridging distance, etc. There are certain differences from the pillar insulators with interleaved ice umbrellas. The interleaved insulation umbrellas have larger umbrella skirts, and the pillar insulators are vertical. The small umbrellas in the middle of the interleaved ice umbrella skirts are less ice-covered. The main factor affecting the flashover voltage is the ice bridging of the large umbrella group, and the above-mentioned patents do not consider the prediction of ice development in combination with meteorological conditions. Moreover, the operation and maintenance personnel in the station can patrol and take pictures at regular intervals, or they can take pictures in combination with monitoring devices. It is more convenient and effective than monitoring of transmission lines, and the prediction of ice development can be dynamically corrected in combination with multiple monitoring results. Moreover, most of the existing patents are risk assessments, and the classification of influencing parameters is relatively rough. Risk assessments mostly use a scoring mechanism, which lacks accuracy and effectiveness. The correction formulas for ice flash of line insulators in some patents are also difficult to apply to the pillar insulators with interleaved ice umbrellas in converter stations. Summary of the invention

[0006] In order to solve the above technical problems, the purpose of the present invention is to provide a method for predicting and controlling the withstand voltage of the post insulators with anti-icing umbrella skirts inserted in the converter station. The method takes into account the ice flash characteristics of the post insulators with interposed ice umbrellas in the converter station, selects appropriate characteristic parameters in combination with experiments, and performs fitting correction on the ice flash voltage under different working conditions. At the same time, a prediction method for the development of icing is proposed to determine the corresponding safe operating voltage, decide whether voltage reduction operation is required and the voltage reduction amplitude, and ensure the safe and stable operation of DC transmission. Specifically, according to the structure of the post insulators with large sheds inserted in the station, flashover tests under different icing conditions are carried out to obtain the ice flashover voltage of typical interposed shed post insulators at different ice ridge bridging gap distances and pollution levels, and establish an ice flashover voltage fitting model; considering the monitoring conditions in the converter station, when the icing is serious, the operation and maintenance personnel or the monitoring device take pictures of the insulators from multiple angles at intervals, and the development of icing is comprehensively predicted in combination with meteorological conditions and environmental factors; combined with the development of icing and pollution conditions, the ice flashover voltage fitting model is used to predict the ice flashover voltage value, determine the safe operating voltage for the next time period, and decide whether voltage reduction operation is required and the voltage reduction amplitude.

[0007] The present invention provides a method for predicting and controlling the withstand voltage of a post insulator with anti-icing sheds inserted between converter stations, comprising the following steps:

[0008] Step 1: Use the test samples of the same structure as the post insulators with interleaved anti-icing sheds installed in the DC converter station, and perform contamination treatment on the post insulator test samples according to different pollution levels, and then simulate icing in the environmental climate laboratory to obtain the post insulator test samples under typical ice bridging conditions;

[0009] Step 2: Use the melting ice flashover test method to carry out ice flashover tests on post insulator specimens under typical pollution conditions and ice bridging conditions. Perform M tests for each condition and record and calculate the average flashover voltage.

[0010] Step 3: Combined with the pollution level and ice bridging situation, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained, and the ice flash voltage fitting model of the post insulator with interleaved anti-icing sheds is established;

[0011] Step 4: Measure the equivalent salt density of the spare insulators in the same environment next to the operating support insulators in the DC converter station to determine the pollution level;

[0012] Step 5: From the beginning of ice covering, T ps Through maintenance personnel or monitoring devices, take all-round images of ice coating on the pillar insulators to obtain the ice bridging gap situation;

[0013] Step 6: Combined with the monitoring data of meteorological factors in the converter station, an ice growth prediction model based on SVM is established. ps The model is trained and optimized based on the ice-covered insulator images until the prediction accuracy of icicle growth length is ≥η%, which is used for the dynamic prediction of subsequent icicle growth.

[0014] Step 7: When an icing disaster occurs at the converter station, the growth length of ice is predicted according to step 6. Based on the measured degree of contamination and the predicted ice bridging distance, combined with the post insulator ice flash voltage fitting model in step 3, the ice flash voltage of the post insulator is predicted. If the ice flash voltage value is higher than the rated voltage, normal operation continues; if it is lower than the corresponding voltage, reduced voltage operation is performed; a dynamic prediction is performed at regular intervals to determine the operating voltage value.

[0015] Preferably, in step 1, the pollution levels are divided into N kinds according to the preset pollution level, and the pollution is simulated according to the corresponding equivalent salt density and equivalent ash density, wherein each pollution level takes the middle value; the soluble substance is simulated by pure NaCl, and the insoluble substance is simulated by diatomaceous earth, in a ratio of 1:X h Mix the simulated contaminant into a solid layer and evenly apply the contaminant to the cleaned surface of the insulator manually. Place it in a dry and ventilated place. gz Time to stand by.

[0016] Preferably, the icing method adopts live icing. Tests show that when the rated operating phase voltage is energized for icing, flashover will occur during the icing process. Therefore, the rated voltage k is used. N % Live icing: Before spray icing, pre-treat the insulators and spray water on the surface of the insulators to form a layer of ice film to prevent the loss of dirt on the surface of the insulators. Then spray cooling water from the nozzles on both sides to the support insulators. Under the low temperature conditions of the environmental climate laboratory, an ice layer is gradually formed. When the ice layer bridges N different gap distances, the icing is stopped, that is, N test samples are obtained for each degree of dirt.

[0017] Preferably, in step 2, considering that there is water film during actual ice melting, flashover is more likely to occur, and the ice flash test considers more stringent conditions, that is, the ice melting flashover test method is adopted. After the ice coverage reaches the corresponding icicle length, the temperature in the climate environment room is controlled to rise to a certain temperature T1 close to 0 degrees Celsius, and then the temperature is increased at a temperature rise rate of St℃ / h. When there is a water film on the ice surface and water droplets are left on the icicle, the starting conditions of the ice flash test are met and pressurization is ready to begin.

[0018] Preferably, the following pressurization method is used: when the ice flash voltage approximate range has not yet been obtained during the first pressurization, the voltage range of the ice flash voltage is calculated according to U t kV / s speed until flashover occurs, and record the flashover voltage U at this time tc, the subsequent test first quickly increases the voltage to U tc K tc %, and then increase U each time tc K jy %, and maintain voltage tolerance time T ns , in T ns Then the voltage is boosted again until breakdown occurs. The ice flash voltage obtained in this way is closer to the actual situation.

[0019] Preferably, in step 3, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained in combination with the pollution level and the ice bridging situation, and an ice flash voltage fitting model of the post insulator with interleaved anti-icing sheds is established;

[0020] To ensure the validity of the test and the reliability of the data, the insulator pollution levels are taken as ESDD1, ESDD2, ESDD3, ESDD4, and ESDD5 respectively; the length of the insulator ice is expressed by the bridging ratio between the large sheds, which are p1, p2, p3, p4, and p5 respectively. Each insulator pollution level and ice bridging ratio are combined in pairs, and three repeated breakdown tests are carried out. After excluding abnormal conditions, the average breakdown voltage is taken to fit the model;

[0021]

[0022] Where p is the ice bridging ratio, l is the ice length of the insulator, and d is the distance between the insulator sheds;

[0023] The pollution flashover voltage of the insulator decreases with the increase of pollution degree, and the ice flashover voltage of the insulator decreases with the increase of ice length. The function is iterated multiple times in the way of least squares approximation to build a surface fitting mathematical model, and then the breakdown voltage corresponding to the pollution degree and bridging ratio in addition to the characteristic parameters can be calculated and predicted. The fitting degree of the surface is R 2 , the surface expression is shown in formula (2)

[0024]

[0025] Where U is the breakdown voltage under different salt density and ice bridging ratio, x is the ice bridging ratio, and y is the insulator salt density.

[0026] Preferably, in step 4, the normally operating post insulator cannot be removed for testing. Several post insulators specially used for testing are placed next to the normally operating insulator in the same natural environment and under the same pollution conditions. jc The contaminants in the area are collected and dissolved to a volume of V jcMeasure the conductivity of the solution in distilled water, then put pure NaCl into the same volume of water. When the conductivity of the solution is the same as the conductivity of the pollutant solution, record the milligrams of pure NaCl at this time and divide it by the area to get the equivalent salt density in mg / cm 2 , and determine the level of contamination.

[0027] Preferably, in step 5, the specific processing method is as follows:

[0028] S5.1: The images of iced insulators are taken by staff at regular intervals to obtain the omnidirectional icing conditions of the insulators at different monitoring stages. Before taking pictures, the structural characteristics and parameters of the pillar insulators are obtained to calibrate the taken images. The relationship between the image pixels and the real physical size of the object is established through the length, area, and feature point position of the object to avoid errors caused by viewing angles, lens distortion, etc., thereby improving the accuracy of image measurement and analysis. When taking icing images of insulators from multiple angles, avoid blocking and shaking during shooting, adjust the appropriate focal length and aperture, and take fill light and other measures when necessary to obtain omnidirectional and clear images of icing on insulators. In addition, the previous shooting orientation should be repeated at different time periods to avoid image processing errors.

[0029] S5.2: When processing the image of ice-covered insulators, the image features such as ice length and bridging ratio are extracted, and grayscale and median filtering algorithms are used to pre-process the image to enhance the target features and reduce the background noise, so as to facilitate the subsequent calibration and feature extraction; according to the color features of ice-covered insulators and the sensitivity of human eye perception, the weighted average method is used for grayscale processing, as shown in formula (3):

[0030] M=0.3R+0.59G+0.11B(3)

[0031] In the formula, M is the weighted gray value, R, G, and B are the components of red, green, and blue respectively;

[0032] In order to separate the insulator image from its background image, the pixels of the two are divided according to the gray value, and then the image threshold segmentation is realized according to the difference of gray level.

[0033]

[0034] Where f(x, y) is the original image, g(x, y) is the segmented image, t1 and t2 are the grayscale ranges of the segmentation;

[0035] The Canny edge detection algorithm is used to extract the contour and characteristic edge of the insulator. The calculation process includes image size and gradient direction calculation, non-maximum suppression, double threshold screening, and edge connection.

[0036]

[0037] In the formula, g x (x, y), g y (x, y) are the partial derivative functions of the image in the x and y directions respectively, g 45 (x, y), g 135 (x, y) are the partial derivative functions of the image at 45° and 135° respectively, e x 、e y is the first-order partial derivative of the image in the x and y directions, F(x, y) is the magnitude of the gradient, and g d (x, y) is the direction of the gradient;

[0038] S5.3: Through the above-mentioned image processing and edge detection method, the edge features of the image of the un-iced insulator in the same orientation are obtained, and the image edges of the iced insulator and the un-iced insulator are compared. The edge difference is the ice thickness of the iced insulator. Each pixel point of the edge of the umbrella skirt in the image of the un-iced insulator is covered on the image of the iced insulator, and this is used as the starting point of the growth of ice. The pixel point is the seed pixel point. If there is a pixel point that is the same as the seed pixel point below or around the seed pixel point in the image of the iced insulator, it is considered that the ice grows towards this point, and this pixel point is used as the new seed pixel point; repeat the above-mentioned growth process, continuously search for the pixel points of the iced insulator image, and combine with the image calibration in step S5.1 to obtain the ice length of the iced insulator, and then calculate the proportion of the ice bridging the insulator gap;

[0039] Repeat the above steps S5.1 and S5.2 to calculate the ice length and bridging condition of the ice-covered insulator in all directions, and take the maximum ice length to calculate the gap bridging ratio of the entire insulator.

[0040] Preferably, in step 6, an ice growth prediction model based on SVM is established in combination with the monitoring data of meteorological factors in the converter station, and the model is trained and optimized until the prediction accuracy of the ice growth length is less than a certain threshold, so as to be used for the dynamic prediction of subsequent ice growth;

[0041] The steps for building the ice growth prediction model are as follows:

[0042] S6.1: The ice growth process on the insulator surface is strongly correlated with temperature, humidity, rainfall, wind speed, and wind direction, and the ice growth rate is nonlinear. Lower temperature is a necessary condition for the phase change and freezing of liquid water or water mist. In the range of -25 to 0°C, the lower the temperature, the higher the probability of the insulator capturing and adsorbing liquid water, and the faster the ice rate. Humidity represents the content of liquid water in the atmosphere. The higher the humidity, the more water droplets are captured on the insulator surface. In the case of rainfall, the larger the diameter of the raindrops, the greater the probability of colliding with the insulator surface, and the greater the degree of ice coverage. The wind speed The wind direction directly affects the position and shape of ice. At the same time, the freezing rate of liquid water will be accelerated under the action of convective heat transfer. Meteorological data is obtained from the meteorological monitoring device of the converter station and the local meteorological platform. The meteorological monitoring device includes a host, temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, rainfall sensor, and air pressure sensor. It dynamically monitors the environmental conditions of the converter station and sends meteorological data to the management background in real time through the wireless network. It obtains the temperature, humidity, rainfall and other forecast data of the meteorological platform in real time to predict the growth of ice on the insulator.

[0043] S6.2: The SVM model has a good effect in dealing with the multi-dimensional factors that affect insulator icing in complex environments. Considering the factors of ambient temperature, humidity, wind speed, rainfall, and raindrop diameter, an insulator icing growth prediction model based on the SVM model is established.

[0044] The extracted characteristic parameters that affect ice growth have a direct impact on the prediction results of the model. Too many non-correlated or weakly correlated characteristic vectors will reduce the SVM prediction accuracy. Therefore, in order to improve the accuracy of the model, the Pearson correlation coefficient method is used to select the characteristic parameters. The calculation formula is as follows:

[0045]

[0046] In the formula, X i and Y i They represent the i-th value of a characteristic parameter X and the icicle length Y in unit time respectively;

[0047] Through feature screening, the model is trained, wherein the temperature change per unit time, the humidity change per unit time, the wind speed change per unit time, the rainfall change per unit time, and the raindrop diameter change per unit time are shown in equations (9)-(13);

[0048]

[0049] In the formula, Δt represents the unit time interval, which is 10 minutes, T i+1 , T i Respectively represent the temperature at the i-th moment and after Δt time, R i+1 , R iRespectively represent the humidity at the i-th moment and after Δt time, V i+1 、V i Respectively represent the wind speed at the i-th moment and after Δt time, J i+1 , J i represent the rainfall at the i-th moment and after Δt time, Y i+1 , Y i Respectively represent the raindrop diameter at the i-th moment and after Δt time;

[0050] In order to eliminate the influence of dimension, each characteristic parameter is normalized;

[0051]

[0052] In the formula, x i is the characteristic parameter of an input model, x min and x max are the maximum and minimum values ​​of the parameter, respectively, i ' is the normalized value;

[0053] S6.3: Use the SVM model to predict the length of ice on the insulator after a unit time. The acquired ice growth data of the insulator includes the structural characteristics of the insulator, the electrical characteristics of the insulator, the environmental characteristics, and the corresponding length of ice growth in a unit time. A large amount of sample data is divided into a training set and a test set. The training set is input into the initial SVM model to obtain the corresponding penalty coefficient and kernel function. Then, the test set data is input into the model to compare the ice length in the test set data with the length predicted by the SVM. When the prediction error is greater than k%, the improved GS algorithm is used to optimize the parameters of the support vector machine model until the error meets the expected requirements, and the final optimized SVM model is obtained.

[0054] The penalty coefficient of the final SVM model is C, the kernel function is γ, and the mean absolute percentage error of the test set is k1%;

[0055] Therefore, based on the current temperature, humidity, and wind speed meteorological data, combined with the meteorological data predicted by the meteorological platform, the meteorological changes after a unit of time can be obtained. Then, the meteorological change data is substituted into the SVM ice prediction model to obtain the ice growth length at the next moment, and then the ice bridging ratio can be calculated based on the ice length and the insulator gap distance.

[0056] S6.4: Before using the SVM model trained with historical data for actual prediction, the current environment and parameter characteristics can be obtained to perform targeted model corrections again. For example, when using the model for prediction, the initial meteorological data is first obtained and substituted into the model. The model is optimized again using the GS algorithm to obtain a new SVM prediction model. After the model is optimized in a targeted manner, its prediction results are closer to the actual situation. Then, based on the current temperature, humidity, and wind speed meteorological data, combined with the meteorological data predicted by the meteorological platform, the meteorological changes after a unit of time can be obtained, and then predictions can be made to obtain the ice growth length and gap bridging ratio at the next moment.

[0057] Preferably, in step 7, when an icing disaster occurs in the converter station, the growth length of ice is predicted according to step 6, and the ice flash voltage of the post insulator is predicted based on the measured degree of contamination and the predicted ice bridging distance, combined with the post insulator ice flash voltage fitting model in step 3. If the ice flash voltage value is higher than the rated voltage, normal operation continues; if it is lower than the corresponding voltage, reduced voltage operation is performed, and a dynamic prediction is performed at regular intervals to determine the operating voltage value.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1. According to the test, when the anti-icing sheds are inserted between the post insulators, the main ice and ice development is mainly on the large sheds, and its breakdown is mainly related to the pollution conditions and the ice bridging distance. The present invention is different from the ice flash risk prediction method used in the existing patents. It is no longer based on the multi-factor scoring mechanism. According to the anti-icing sheds inserted between the post insulators, ice flash tests are carried out. According to the test results, the influencing factors are reasonably simplified. Considering that only salt is easily brought in during the ice covering process after the outer side of the insulator is covered with ice, and affects the discharge during ice melting, only two parameters, equivalent salt density and ice bridging, are considered in combination with the test results. The ice flash voltage curve is obtained by fitting. Compared with the multi-factor scoring mechanism, it is more accurate to judge whether the post insulator has ice flash under the operating voltage;

[0060] 2. Existing patents are mainly aimed at identifying existing icing conditions, while DC control needs to predict the length of ice within a certain period of time and whether flashover will occur during normal operation. It is necessary to formulate a control strategy in combination with the corresponding flashover voltage value to achieve a special DC step-down operation mode. The step-down process takes a certain amount of time and safety needs to be ensured during this time. Therefore, based on existing image recognition, the present invention combines the characteristics of inspections in converter stations, considers that icing takes a long time to develop, and establishes an icing growth model by regularly monitoring icing growth at the beginning of an ice disaster. The model is dynamically modified in combination with actual conditions, thereby more accurately predicting icing growth and further predicting the ice flash characteristics of support insulators. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is the surface fitting diagram of the insulator breakdown voltage.

[0062] Figure 2 This is the flow chart of iced insulator image processing.

[0063] Figure 3 Flowchart for ice length prediction.

[0064] Figure 4 Flowchart for ice flash prediction and operation control of post insulators. DETAILED DESCRIPTION

[0065] The following is a clear and complete description of the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0066] like Figure 4 As shown, the present invention provides a method for predicting and controlling the withstand voltage of a post insulator with anti-icing sheds inserted between converter stations, comprising the following steps:

[0067] Step 1: Use the test samples of the same structure as the post insulators with interleaved anti-icing sheds installed in the DC converter station, and perform contamination treatment on the post insulator test samples according to different pollution levels, and then simulate icing in the environmental climate laboratory to obtain the post insulator test samples under typical ice bridging conditions;

[0068] Step 2: Use the melting ice flashover test method to carry out ice flashover tests on post insulator specimens under typical pollution conditions and ice bridging conditions. Perform three tests for each condition and record and calculate the average flashover voltage.

[0069] Step 3: Combined with the pollution level and ice bridging situation, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained, and the ice flash voltage fitting model of the post insulator with interleaved anti-icing sheds is established;

[0070] Step 4: Measure the equivalent salt density of the spare insulators in the same environment next to the operating support insulators in the DC converter station to determine the pollution level;

[0071] Step 5: From the beginning of ice covering, T ps Through operation and maintenance inspection personnel or monitoring devices, all-round images of ice coating on the pillar insulators are taken to obtain the ice bridging gap situation.

[0072] Step 6: Combined with the monitoring data of meteorological factors in the converter station, an ice growth prediction model based on SVM is established.ps The model is trained and optimized based on the ice-covered insulator images until the prediction accuracy of icicle growth length is ≥η%, which can be used for the dynamic prediction of subsequent icing and icicle growth.

[0073] Step 7: When an icing disaster occurs in the converter station, the ice growth length is predicted according to step 6. Based on the measured pollution degree and the predicted ice bridging distance, combined with the post insulator ice flash voltage fitting model in step 3, the post insulator ice flash voltage is predicted. If the ice flash voltage value is higher than the rated voltage, normal operation continues. If it is lower than the corresponding voltage, the voltage is reduced. A dynamic prediction is performed at regular intervals to determine the operating voltage value.

[0074] In step 1, the pollution level of the State Grid is divided into five levels of pollution, and the pollution is simulated according to the corresponding equivalent salt density and equivalent ash density, where each pollution level takes the middle value, i.e. 0.015 mg / cm 2 , 0.045mg / cm 2 , 0.08mg / cm 2 , 0.175mg / cm 2 , 0.3mg / cm 2 Soluble substances were simulated by pure NaCl, and insoluble substances were simulated by diatomaceous earth, in a ratio of 1:X h Mix it into simulated contamination material, apply the contamination evenly to the cleaned surface of the insulator manually by solid layer method, and place it in a dry and ventilated place for 7 days for use.

[0075] Table 1 Classification of pollution levels

[0076]

[0077] The icing method uses live icing. Tests show that live icing at rated operating phase voltage will cause flashover during the icing process, so live icing at 80% of the rated voltage is used. Before spraying icing, the insulator is pretreated, and a water film is sprayed on the surface of the insulator to form an ice film to prevent the dirt on the surface of the insulator from running away. Then, cooling water is sprayed on the post insulator from the nozzles on both sides, and an ice layer is gradually formed under the low temperature of the environmental climate laboratory. When the ice layer bridges 5 different gap distances, the icing is stopped, that is, 5 test samples are obtained for each degree of dirt.

[0078] In step 2, considering that there is a water film during actual ice melting, it is more likely to cause flashover. The ice flash test considers more stringent conditions, that is, the ice melting flashover test method is adopted. After the ice coverage reaches the corresponding icicle length, the temperature in the climate environment room is controlled to rise to -2°C, and then the temperature is increased at a temperature rise rate of 5°C / h. When there is a water film on the ice surface and water droplets are left on the icicle, the starting conditions of the ice flash test are met and pressurization is ready to begin.

[0079] Considering the characteristics of ice melting, ice flash may occur during the dripping process of water film. The following pressurization method is used: the approximate range of ice flash voltage has not been obtained during the first pressurization. At this time, the voltage is increased at a rate of 3kV / s until flashover occurs. The flashover voltage U at this time is recorded. tc The subsequent test first quickly increases the voltage to U tc 70% and then each time U tc 2% and maintain the voltage tolerance time for 5s. After 5s, the voltage is increased again until breakdown occurs. The ice flash voltage obtained in this way is closer to the actual situation.

[0080] In step 3, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained in combination with the pollution level and the ice bridging situation, and an ice flash voltage fitting model for the post insulator with interleaved anti-icing sheds is established.

[0081] In order to ensure the validity of the test and the reliability of the data, the insulator pollution level is set to 0mg / cm 2 , 0.045mg / cm 2 , 0.08mg / cm 2 , 0.175mg / cm 2 , 0.3mg / cm 2 The length of ice on the insulator is represented by the bridging ratio between the large sheds, which are 0%, 25%, 50%, 75%, and 100% respectively. The pollution degree and ice bridging ratio of each insulator are combined in pairs, and three repeated breakdown tests are carried out to exclude abnormal conditions, and the average breakdown voltage of the insulator is taken to fit the model, as shown in Table 2.

[0082]

[0083] Where p is the ice bridging ratio, l is the ice length of the insulator, and d is the spacing between the insulator sheds.

[0084] Table 2 Insulator breakdown voltage under different working conditions

[0085] Among them, when the insulators are not contaminated, the breakdown voltages of the insulators at different ice bridging ratios are 722.2kV, 675.5kV, 618.3kV, 550.8kV and 471.1kV respectively; when the insulators are not covered with ice, the breakdown voltages at different contamination degrees are 722.2kV, 689.3kV, 642.6kV, 580.2kV and 493.6kV respectively.

[0086] The pollution flashover voltage of the insulator decreases with the increase of pollution degree, and the ice flashover voltage of the insulator decreases with the increase of ice length. By iterating the function multiple times in the way of least squares approximation, a surface fitting mathematical model is constructed, and then the breakdown voltage corresponding to the pollution degree and bridge ratio in addition to the characteristic parameters can be calculated and predicted. The fitted surface is as follows Figure 1 As shown, the fitting degree of the surface is 0.997. The expression of the surface is shown in formula (2).

[0087]

[0088] Where U is the breakdown voltage under different salt density and ice bridging ratio, x is the ice bridging ratio, and y is the insulator salt density.

[0089] In step 4, the normally operating post insulator cannot be removed for testing. Several post insulators specially used for testing will be placed next to the normally operating insulator in the same natural environment and in the same contamination conditions. 2 Collect the pollutants in the area, dissolve them in 500mL of distilled water, and measure the conductivity of the solution. Then put pure NaCl into the same volume of water. When the conductivity of the solution is the same as the conductivity of the pollutants, record the milligrams of pure NaCl at this time and divide it by the area to get the equivalent salt density, the unit is mg / cm 2 , and determine the pollution level according to Table 1.

[0090] In step 5, the specific processing method is as follows:

[0091] S5.1: The image monitoring method can obtain a large amount of intuitive image information, and the technology is mature, the data is reliable, and the economy is high. At the same time, the processing process can effectively avoid interference such as corona and corona. It is widely used in monitoring scenarios such as converter stations and transmission lines. The images of iced insulators are taken by staff at regular intervals to obtain the full range of ice coverage of insulators at different monitoring stages. The data source is simple and reliable. The image processing process of iced insulators is as follows: Figure 3 shown.

[0092] Before shooting, the structural characteristics and parameters of the post insulator are obtained as shown in Table 2, so as to calibrate the captured image. The relationship between the image pixels and the real physical size of the object can be established through the object length, area, feature point position, etc., to avoid errors caused by viewing angle, lens distortion, etc., and thus improve the accuracy of image measurement and analysis.

[0093] Table 3 Insulator structural parameters

[0094] Insulator height Insulation distance Large shed diameter Large umbrella spacing Minimum arc distance 1315mm 1060mm 300mm 120mm 5525mm

[0095] When shooting ice-covered insulator images from multiple angles, avoid blocking and shaking, adjust the appropriate focal length and aperture, and take measures such as fill light when necessary to obtain a full range of clear ice-covered insulator images. In addition, the previous shooting position should be repeated at different time periods to avoid image processing errors.

[0096] S5.2: In a low temperature and high humidity environment, rain or water mist gathers in liquid form and freezes on the surface of the insulator. The irregular ice-covered water film hinders the uniform distribution of water. The water droplets are frozen in the process of moving downward under the influence of wind power and gravity, which promotes the formation and growth of icicles. As the proportion of insulators bridged by icicles increases, their leakage distance gradually shortens and their insulation performance gradually decreases, which can easily cause the electric field distortion of the insulator string, reduce the flashover voltage, and eventually cause ice flash accidents. Therefore, when processing the image of ice-covered insulators, it is necessary to focus on extracting image features such as icicle length and bridging ratio.

[0097] The grayscale and median filtering algorithms are used to preprocess the image, enhance the target features, and reduce the background noise, so as to facilitate the calibration and feature extraction in the later stage. Compared with the RGB channel of the color image, the grayscale image only contains one grayscale value channel, and the image noise is also converted into the fluctuation of the grayscale value. While greatly reducing the amount of data, it retains the key features such as the edge, texture, and shape of the insulator, which can effectively improve the calculation rate.

[0098] According to the color characteristics of iced insulators and the sensitivity of human eye perception, the weighted average method is used for grayscale processing, as shown in formula (3).

[0099] M=0.3R+0.59G+0.11B(3)

[0100] Where M is the weighted gray value, and R, G, and B are the components of red, green, and blue, respectively.

[0101] In order to separate the insulator image from its background image, the pixels of the two are divided according to the gray value, and then the gray value is divided according to the gray value.

[0102]

[0103] Where f(x, y) is the original image, g(x, y) is the segmented image, and t1 and t2 are the grayscale ranges of the segmentation.

[0104] The Canny edge detection algorithm is used to extract the contour and feature edges of the insulator. The calculation process includes image size and gradient direction calculation, non-maximum suppression, double threshold screening, and edge connection.

[0105]

[0106] In the formula, gx (x, y), g y (x, y) are the partial derivative functions of the image in the x and y directions respectively, g 45 (x, y), g 135 (x, y) are the partial derivative functions of the image at 45° and 135° respectively, e x 、e y is the first-order partial derivative of the image in the x and y directions, F(x, y) is the magnitude of the gradient, and g d (x, y) is the direction of the gradient.

[0107] S5.3: The edge features of the image of the un-iced insulator in the same orientation are obtained through the above-mentioned image processing and edge detection methods. The image edges of the iced insulator and the un-iced insulator are compared, and the edge difference is the ice thickness of the iced insulator. Each pixel point on the edge of the umbrella skirt in the image of the un-iced insulator is covered on the image of the iced insulator, and this is used as the starting point for the growth of ice (the pixel point is the seed pixel). If there is a pixel point that is the same as the seed pixel below or around the seed pixel point in the image of the iced insulator, it is considered that the ice is growing towards this point, and the pixel point is used as the new seed pixel point. Repeat the above growth process, continuously search for the pixel points of the iced insulator image, and combine with the image calibration in step S5.1 to obtain the ice length of the iced insulator. Furthermore, the proportion of ice bridging the insulator gap can be calculated.

[0108] Repeat the above steps S5.1 and S5.2 to calculate the icing length and bridging of the ice-covered insulator in all directions, and take the maximum icing length to calculate the gap bridging ratio of the entire insulator. Taking a certain monitoring result as an example, the monitoring is performed every 10 minutes, and the icing lengths are 17.3mm, 25.5mm, 36.1mm, and 51.2mm, respectively. Therefore, the icing bridging ratios are 14.4%, 21.3%, 30.1%, and 42.7%, respectively.

[0109] In step 6, an ice growth prediction model based on SVM is established in combination with the monitoring data of meteorological factors in the converter station, and the model is trained and optimized until the prediction accuracy of ice growth length is less than 5%, so as to be used for the dynamic prediction of subsequent ice growth.

[0110] The steps for building the ice growth prediction model are as follows:

[0111] S6.1: The ice growth process on the insulator surface is strongly correlated with factors such as temperature, humidity, rainfall, wind speed, and wind direction, and the ice growth rate is nonlinear. Lower temperature is a necessary condition for the phase change freezing of liquid water or water mist. In the range of -25 to 0°C, the lower the temperature, the higher the probability of the insulator capturing and adsorbing liquid water, and the faster the ice rate. Humidity characterizes the content of liquid water in the atmosphere. The higher the humidity, the more water droplets are captured on the insulator surface. In the case of rainfall, the larger the diameter of the raindrops, the greater the probability of colliding with the insulator surface, and the greater the degree of ice coverage. The wind speed and direction directly affect the orientation and shape of the ice. At the same time, under the action of convective heat transfer, the freezing rate of liquid water will be accelerated.

[0112] Meteorological data is obtained from the meteorological monitoring device of the converter station and the local meteorological platform. The meteorological monitoring device is mainly composed of a host, temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, rainfall sensor, air pressure sensor and other units. It dynamically monitors the environmental conditions of the converter station and sends meteorological data to the management background in real time through the wireless network. It obtains temperature, humidity, rainfall and other forecast data from the meteorological platform in real time to predict the growth of ice on insulators.

[0113] S6.2: The SVM model has a good effect in dealing with the multi-dimensional factors that affect insulator icing in complex environments. Considering factors such as ambient temperature, humidity, wind speed, rainfall, and raindrop diameter, an insulator icing and icicle growth prediction model based on the SVM model is established.

[0114] The extracted characteristic parameters affecting ice growth are shown in Table 4.

[0115] Table 4 Characteristic parameters

[0116]

[0117]

[0118] The characteristic parameters have a direct impact on the prediction results of the model. Too many non-correlated or weakly correlated characteristic vectors will reduce the prediction accuracy of SVM. Therefore, in order to improve the accuracy of the model, the Pearson correlation coefficient method is used to select the characteristic parameters. The calculation formula is as follows:

[0119]

[0120] In the formula, X i and Y i They respectively represent the i-th value of a characteristic parameter X and the icicle length Y in unit time.

[0121] Table 5 Pearson correlation coefficients of characteristic parameters

[0122] parameter H D1 D2 H1 H2 H3 coefficient 0.214 0.397 0.147 0.386 0.208 0.168 parameter U I Tt ΔT R ΔR coefficient -0.584 -0.203 0.613 0.584 0.637 0.604 parameter Vt ΔV Jt ΔJ Y ΔY coefficient 0.588 0.542 0.638 0.513 0.522 0.487

[0123] Through feature screening, the feature parameters used to train the model are shown in Table 6.

[0124] Table 6 Feature parameters of SVM model input

[0125]

[0126] Among them, the temperature change per unit time, humidity change per unit time, wind speed change per unit time, rainfall change per unit time, and raindrop diameter change per unit time are shown in equations (9)-(13).

[0127]

[0128]

[0129] In the formula, Δt represents the unit time interval, which is 10 minutes, T i+1 , T i Respectively represent the temperature at the i-th moment and after Δt time, R i+1 , R i Respectively represent the humidity at the i-th moment and after Δt time, V i+1 、V i Respectively represent the wind speed at the i-th moment and after Δt time, J i+1 , J i represent the rainfall at the i-th moment and after Δt time, Y i+1 , Y i They represent the raindrop diameter at the i-th moment and after Δt time respectively.

[0130] In order to eliminate the influence of dimension, each characteristic parameter is normalized.

[0131]

[0132] In the formula, x i is the characteristic parameter of an input model, x min and x max are the maximum and minimum values ​​of the parameter, respectively, i ' is the normalized value.

[0133] S6.3: The process of using the SVM model to predict the length of ice on the insulator after a unit time is as follows Figure 2As shown, the acquired ice growth data of insulators include insulator structural characteristics, insulator electrical characteristics, environmental characteristics, and the corresponding length of ice growth per unit time. A large amount of sample data is divided into a training set and a test set, wherein the training set is input into the initial SVM model to obtain the corresponding penalty coefficient and kernel function, and then the test set data is input into the model to compare the ice length in the test set data with the length predicted by SVM. When the prediction error is greater than 5%, the improved GS algorithm is used to optimize the parameters of the support vector machine model until the error meets the expected requirements, and the final optimized SVM model can be obtained.

[0134] Taking the ice growth of the first shed at the top of the insulator as an example, at the initial monitoring moment, the temperature was -12℃, the humidity was 92%, the wind speed was 3.3m / s, the rainfall was 5mm, the raindrop diameter was about 2.2mm, the ice length was 17.3mm, the voltage on the insulator was 500kV, and the leakage current was negligible. Around the same moment, the meteorological factors of the insulator environment were monitored 5 times, and the average was shown in Table 5. The ice length at this time was recorded, and a series of data was input into the SVM model to train and optimize the model. Finally, the penalty coefficient C of the SVM model was 56.42, the kernel function γ was 0.0183, and the average absolute percentage error of the test set was 3.1%.

[0135] Table 7 Icicle growth process of insulators during a monitoring

[0136]

[0137] S6.4: Before using the SVM model trained with historical data for actual prediction, the current environment and parameter characteristics can be obtained to perform targeted model corrections again. For example, when the model is used for prediction, the initial condition is that the length of the icicle is 12.2 mm, the ambient temperature is -6°C, the humidity is 93%, the wind speed is 2.7 m / s, the rainfall is 7 mm, and the raindrop diameter is 2.4 mm. The five sets of data obtained at the same time are substituted into the model, and the model is optimized again using the GS algorithm. The new SVM model penalty coefficient C is 54.27, the kernel function γ is 0.0181, and the future data obtained from the meteorological platform are: ambient temperature -5°C, humidity 92%, wind speed 2.9 m / s, rainfall 6 mm, and raindrop diameter 2.2 mm. The model before optimization is used for prediction, and the length of the icicle is 15.4 mm; the model after targeted optimization is used for prediction, and the length of the icicle is 15.7 mm.

[0138] After 10 minutes, the ice coating on the insulator was photographed, and the length of the ice was calculated to be 16.1 mm through image processing. It can be seen that the prediction error of the model before optimization was 4.3%, which met the requirements; but the prediction error of the model after optimization was 2.5%, and the prediction result was closer to the actual situation.

[0139] Therefore, substituting the current meteorological data into the SVM ice model for targeted optimization can obtain a model with higher prediction accuracy. Then, based on the current meteorological data such as temperature, humidity, wind speed, etc., combined with the meteorological data predicted by the meteorological platform, the meteorological changes after unit time can be obtained, and then predictions can be made to obtain the ice growth length and gap bridging ratio at the next moment.

[0140] In step 7, when an icing disaster occurs in the 500kV DC converter station, the growth length of ice is predicted according to step 6. Based on the measured degree of contamination and the predicted ice bridging distance, combined with the ice flash voltage fitting model of the post insulator in step 3, the ice flash voltage of the post insulator is predicted. If the ice flash voltage value is higher than the rated voltage, normal operation continues. If it is lower than the corresponding voltage, reduced voltage operation is performed. A dynamic prediction is performed at regular intervals to determine the operating voltage value.

[0141] For example, the contamination level of the insulator is measured to be 0.08 mg / cm2. During the monitoring process, when the ice bridging ratio reaches 50%, the breakdown voltage is 542.1 kV, which is within the safety range. However, when predicting the ice length in the next stage, when the ice bridging ratio reaches 75%, the breakdown voltage obtained by the fitting model is 476.3 kV, which exceeds the safety range. Therefore, in order to ensure the safety and stability of the power grid, the converter station needs to reduce the voltage to 350 kV for operation.

[0142] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A method for predicting and controlling the withstand voltage of post insulators with anti-icing sheds inserted between converter stations, characterized in that: The following steps are involved: Step 1: Use the test samples of the same structure as the post insulators with interleaved anti-icing sheds installed in the DC converter station, and perform contamination treatment on the post insulator test samples according to different pollution levels, and then simulate icing in the environmental climate laboratory to obtain the post insulator test samples under typical ice bridging conditions; Step 2: Use the melting ice flashover test method to carry out ice flashover tests on post insulator specimens under typical pollution conditions and ice bridging conditions. Perform M tests for each condition and record and calculate the average flashover voltage. Step 3: Combined with the pollution level and ice bridging situation, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained, and the ice flash voltage fitting model of the post insulator with interleaved anti-icing sheds is established; Step 4: Measure the equivalent salt density of the spare insulators in the same environment next to the operating support insulators in the DC converter station to determine the pollution level; Step 5: From the beginning of ice covering, T ps Through maintenance personnel or monitoring devices, take all-round images of ice coating on the pillar insulators to obtain the ice bridging gap situation; Step 6: Combined with the monitoring data of meteorological factors in the converter station, an ice growth prediction model based on SVM is established. ps The model is trained and optimized based on the ice-covered insulator images until the prediction accuracy of icicle growth length is ≥η%, which is used for the dynamic prediction of subsequent icicle growth. Step 7: When an icing disaster occurs at the converter station, the growth length of ice is predicted according to step 6. Based on the measured degree of contamination and the predicted ice bridging distance, combined with the post insulator ice flash voltage fitting model in step 3, the ice flash voltage of the post insulator is predicted. If the ice flash voltage value is higher than the rated voltage, normal operation continues; if it is lower than the corresponding voltage, reduced voltage operation is performed; a dynamic prediction is performed at regular intervals to determine the operating voltage value.

2. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 1 is characterized in that: In step 1, the pollution level is divided into N kinds of pollution levels according to the preset pollution level, and the pollution is simulated according to the corresponding equivalent salt density and equivalent ash density, wherein each pollution level takes the middle value; the soluble substance is simulated by pure NaCl, and the insoluble substance is simulated by diatomaceous earth, in a ratio of 1:X h Mix the simulated pollutants and apply the pollutants evenly to the cleaned insulator surface by solid layer method and place them in a dry and ventilated place. gz Time to stand by.

3. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 2 is characterized in that: The icing method described above adopts live icing. Tests show that when the rated operating phase voltage is used for live icing, flashover will occur during the icing process. Therefore, the rated voltage k is used. N % Live icing: Before spray icing, pre-treat the insulators and spray water on the surface of the insulators to form a layer of ice film to prevent the loss of dirt on the surface of the insulators. Then spray cooling water from the nozzles on both sides to the support insulators. Under the low temperature conditions of the environmental climate laboratory, an ice layer is gradually formed. When the ice layer bridges N different gap distances, the icing is stopped, that is, N test samples are obtained for each degree of dirt.

4. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 3 is characterized in that: In step 2, considering that there is water film during actual ice melting, flashover is more likely to occur. The ice flash test considers more stringent conditions, that is, the ice melting flashover test method is adopted. After the ice coverage reaches the corresponding icicle length, the temperature in the climate environment room is controlled to rise to a certain temperature T1 close to 0 degrees Celsius, and then the temperature is increased at a temperature rise rate of St℃ / h. When there is a water film on the ice surface and water droplets are left on the icicle, the starting conditions of the ice flash test are met and pressurization is ready to begin.

5. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 4 is characterized in that: The following pressurization method is used: During the first pressurization, the approximate range of ice flash voltage has not yet been obtained. t kV / s speed until flashover occurs, and record the flashover voltage U at this time tc , the subsequent test first quickly increases the voltage to U tc K tc %, and then increase U each time tc K jy %, and maintain voltage tolerance time T ns , in T ns Then the voltage is boosted again until breakdown occurs. The ice flash voltage obtained in this way is closer to the actual situation.

6. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 5 is characterized in that: In the step 3, the mapping relationship between the two main influencing factors and the ice flash voltage is obtained by combining the pollution level and the ice bridging situation, and an ice flash voltage fitting model of the post insulator with interleaved anti-icing sheds is established; To ensure the validity of the test and the reliability of the data, the insulator pollution levels are taken as ESDD1, ESDD2, ESDD3, ESDD4, and ESDD5 respectively; the length of the insulator ice is expressed by the bridging ratio between the large sheds, which are p1, p2, p3, p4, and p5 respectively. Each insulator pollution level and ice bridging ratio are combined in pairs, and three repeated breakdown tests are carried out. After excluding abnormal conditions, the average breakdown voltage is taken to fit the model; Where, p is the ice bridging ratio, l is the ice length of the insulator, and d is the distance between the insulator sheds; The pollution flashover voltage of the insulator decreases with the increase of pollution degree, and the ice flashover voltage of the insulator decreases with the increase of ice length. The function is iterated multiple times in the way of least squares approximation to build a surface fitting mathematical model, and then the breakdown voltage corresponding to the pollution degree and bridging ratio in addition to the characteristic parameters can be calculated and predicted. The fitting degree of the surface is R 2 , the surface expression is shown in formula (2) Where U is the breakdown voltage under different salt density and ice bridging ratio, x is the ice bridging ratio, and y is the insulator salt density.

7. A method for predicting and controlling the withstand voltage of post insulators with anti-icing sheds inserted between converter stations according to claim 6, characterized in that: In step 4, the normally operating post insulator cannot be removed for testing. Several post insulators specially used for testing are placed next to the normally operating insulator in the same natural environment and under the same pollution conditions. jc The dirt in the area is collected and dissolved to a volume of V jc Measure the conductivity of the solution in distilled water, then put pure NaCl into the same volume of water. When the conductivity of the solution is the same as the conductivity of the pollutant solution, record the milligrams of pure NaCl at this time and divide it by the area to get the equivalent salt density in mg / cm 2 , and determine the level of contamination.

8. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 7 is characterized in that: In step 5, the specific processing method is as follows: S5.1: The images of iced insulators are taken by staff at regular intervals to obtain the omnidirectional icing conditions of the insulators at different monitoring stages. Before taking pictures, the structural characteristics and parameters of the pillar insulators are obtained to calibrate the taken images. The relationship between the image pixels and the real physical size of the object is established through the length, area, and feature point position of the object to avoid errors caused by viewing angles, lens distortion, etc., thereby improving the accuracy of image measurement and analysis. When taking icing images of insulators from multiple angles, avoid blocking and shaking during shooting, adjust the appropriate focal length and aperture, and take fill light and other measures when necessary to obtain omnidirectional and clear images of icing on insulators. In addition, the previous shooting orientation should be repeated at different time periods to avoid image processing errors. S5.2: When processing the image of ice-covered insulators, the image features such as ice length and bridging ratio are extracted, and grayscale and median filtering algorithms are used to pre-process the image to enhance the target features and reduce the background noise, so as to facilitate the subsequent calibration and feature extraction; according to the color features of ice-covered insulators and the sensitivity of human eye perception, the weighted average method is used for grayscale processing, as shown in formula (3): M=0.3R+0.59G+0.11B(3) In the formula, M is the weighted gray value, R, G, and B are the components of red, green, and blue respectively; In order to separate the insulator image from its background image, the pixels of the two are divided according to the gray value, and then the image threshold segmentation is realized according to the difference of gray level. Where f(x, y) is the original image, g(x, y) is the segmented image, t1 and t2 are the grayscale ranges of the segmentation; The Canny edge detection algorithm is used to extract the contour and characteristic edge of the insulator. The calculation process includes image size and gradient direction calculation, non-maximum suppression, double threshold screening, and edge connection. In the formula, g x (x, y), g y (x, y) are the partial derivative functions of the image in the x and y directions respectively, g 45 (x, y), g 135 (x, y) are the partial derivative functions of the image at 45° and 135° respectively, e x 、e y is the first-order partial derivative of the image in the x and y directions, F(x, y) is the magnitude of the gradient, and g d (x, y) is the direction of the gradient; S5.3: Through the above-mentioned image processing and edge detection method, the edge features of the image of the un-iced insulator in the same orientation are obtained, and the image edges of the iced insulator and the un-iced insulator are compared. The edge difference is the ice thickness of the iced insulator. Each pixel point of the edge of the umbrella skirt in the image of the un-iced insulator is covered on the image of the iced insulator, and this is used as the starting point of the growth of ice. The pixel point is the seed pixel point. If there is a pixel point that is the same as the seed pixel point below or around the seed pixel point in the image of the iced insulator, it is considered that the ice grows towards this point, and this pixel point is used as the new seed pixel point; repeat the above-mentioned growth process, continuously search for the pixel points of the iced insulator image, and combine with the image calibration in step S5.1 to obtain the ice length of the iced insulator, and then calculate the proportion of the ice bridging the insulator gap; Repeat the above steps S5.1 and S5.2 to calculate the ice length and bridging condition of the ice-covered insulator in all directions, and take the maximum ice length to calculate the gap bridging ratio of the entire insulator.

9. A method for predicting and controlling the withstand voltage of post insulators with anti-icing sheds inserted between converter stations according to claim 8, characterized in that: In step 6, an ice growth prediction model based on SVM is established in combination with the monitoring data of meteorological factors in the converter station, and the model is trained and optimized until the prediction accuracy of the ice growth length is less than a certain threshold, so as to be used for the dynamic prediction of subsequent ice growth; The steps for building the ice growth prediction model are as follows: S6.1: The ice growth process on the insulator surface is strongly correlated with temperature, humidity, rainfall, wind speed, and wind direction, and the ice growth rate is nonlinear. Lower temperature is a necessary condition for the phase change and freezing of liquid water or water mist. In the range of -25 to 0°C, the lower the temperature, the higher the probability of the insulator capturing and adsorbing liquid water, and the faster the ice rate. Humidity represents the content of liquid water in the atmosphere. The higher the humidity, the more water droplets are captured on the insulator surface. In the case of rainfall, the larger the diameter of the raindrops, the greater the probability of colliding with the insulator surface, and the greater the degree of ice coverage. The wind speed The wind direction directly affects the position and shape of ice. At the same time, the freezing rate of liquid water will be accelerated under the action of convective heat transfer. Meteorological data is obtained from the meteorological monitoring device of the converter station and the local meteorological platform. The meteorological monitoring device includes a host, temperature sensor, humidity sensor, wind speed sensor, wind direction sensor, rainfall sensor, and air pressure sensor. It dynamically monitors the environmental conditions of the converter station and sends meteorological data to the management background in real time through the wireless network. It obtains the temperature, humidity, rainfall and other forecast data of the meteorological platform in real time to predict the growth of ice on the insulator. S6.2: The SVM model has a good effect in dealing with the multi-dimensional factors that affect insulator icing in complex environments. Considering the factors of ambient temperature, humidity, wind speed, rainfall, and raindrop diameter, an insulator icing growth prediction model based on the SVM model is established. The extracted characteristic parameters that affect ice growth have a direct impact on the prediction results of the model. Too many non-correlated or weakly correlated characteristic vectors will reduce the SVM prediction accuracy. Therefore, in order to improve the accuracy of the model, the Pearson correlation coefficient method is used to select the characteristic parameters. The calculation formula is as follows: Where, X i and Y i They represent the i-th value of a characteristic parameter X and the icicle length Y in unit time respectively; Through feature screening, the model is trained, wherein the temperature change per unit time, the humidity change per unit time, the wind speed change per unit time, the rainfall change per unit time, and the raindrop diameter change per unit time are shown in equations (9)-(13); In the formula, Δt represents the unit time interval, which is 10 minutes, T i+1 , T i Respectively represent the temperature at the i-th moment and after Δt time, R i+1 , R i Respectively represent the humidity at the i-th moment and after Δt time, V i+1 、V i Respectively represent the wind speed at the i-th moment and after Δt time, J i+1 , J i represent the rainfall at the i-th moment and after Δt time, Y i+1 , Y i Respectively represent the raindrop diameter at the i-th moment and after Δt time; In order to eliminate the influence of dimension, each characteristic parameter is normalized; In the formula, x i is the characteristic parameter of an input model, x min and x max are the maximum and minimum values ​​of the parameter, respectively, i ' is the normalized value; S6.3: Use the SVM model to predict the length of ice on the insulator after a unit time. The acquired ice growth data of the insulator includes the structural characteristics of the insulator, the electrical characteristics of the insulator, the environmental characteristics, and the corresponding length of ice growth in a unit time. A large amount of sample data is divided into a training set and a test set. The training set is input into the initial SVM model to obtain the corresponding penalty coefficient and kernel function. Then, the test set data is input into the model to compare the ice length in the test set data with the length predicted by the SVM. When the prediction error is greater than k%, the improved GS algorithm is used to optimize the parameters of the support vector machine model until the error meets the expected requirements, and the final optimized SVM model is obtained. The penalty coefficient of the final SVM model is C, the kernel function is γ, and the mean absolute percentage error of the test set is k1%; Therefore, based on the current temperature, humidity, and wind speed meteorological data, combined with the meteorological data predicted by the meteorological platform, the meteorological changes after a unit of time can be obtained. Then, the meteorological change data is substituted into the SVM ice prediction model to obtain the ice growth length at the next moment, and then the ice bridging ratio can be calculated based on the ice length and the insulator gap distance. S6.4: Before using the SVM model trained with historical data for actual prediction, the current environment and parameter characteristics can be obtained to perform targeted model corrections again. For example, when using the model for prediction, the initial meteorological data is first obtained and substituted into the model. The model is optimized again using the GS algorithm to obtain a new SVM prediction model. After the model is optimized in a targeted manner, its prediction results are closer to the actual situation. Then, based on the current temperature, humidity, and wind speed meteorological data, combined with the meteorological data predicted by the meteorological platform, the meteorological changes after a unit of time can be obtained, and then predictions can be made to obtain the ice growth length and gap bridging ratio at the next moment.

10. The method for predicting and controlling the withstand voltage of the post insulators with anti-icing sheds inserted between converter stations according to claim 9, characterized in that: In step 7, when an icing disaster occurs in the converter station, the growth length of ice is predicted according to step 6, and the ice flash voltage of the post insulator is predicted based on the measured degree of contamination and the predicted ice bridging distance, combined with the post insulator ice flash voltage fitting model in step 3. If the ice flash voltage value is higher than the rated voltage, normal operation continues; if it is lower than the corresponding voltage, reduced voltage operation is performed, and a dynamic prediction is performed at regular intervals to determine the operating voltage value.

Images