A wind farm line fault monitoring and early warning method and system

Abstract

The present application discloses a wind farm line fault monitoring and early warning method and system, which relates to the field of line fault monitoring. The method includes monitoring the current and voltage values ​​of the line to obtain a numerical fluctuation curve, and making a first judgment on the line fault according to the numerical fluctuation curve; calculating the power loss of the power generation of the generator set and the transmission amount of the collector line, and making a second judgment on the line fault according to the power loss; acquiring images of the laying area of ​​the direct buried cable according to the judgment result to obtain regional image data; dividing the regional image data according to the pre-buried path to obtain multiple image blocks; analyzing the image blocks to obtain first fault image data with suspected fault points; comparing and analyzing the multiple image blocks to obtain second fault image data with suspected fault points; analyzing the fault image blocks according to the numerical fluctuation curve and the power loss to determine the fault point. The present application has the effect of improving the efficiency of line fault detection.

Description

Technical Field

[0001] The present application relates to the field of line fault monitoring, and in particular to a wind farm line fault monitoring and early warning method and system. Background Art

[0002] The collector line is an important part of the wind farm and is the carrier for the electric energy generated by the wind turbine. The direct buried cable is an important form of the collector line. By laying the cable underground, it reduces the influence of bad weather, such as lightning, ice and snow, and makes the power supply more reliable. However, since the direct buried cable is buried underground, it is difficult to confirm the fault point when the direct buried cable fails.

[0003] In the related art, when fault detection is performed on direct buried cables, the ground potential detection method is often used, that is, by using metal rods with two insulating handles, connected to the AC low voltage position of the multimeter, and gradually detecting along the direction of the direct buried cable, the change in voltage value is judged, thereby judging whether there is leakage at the corresponding position.

[0004] Regarding the above-mentioned related technologies, when conducting ground potential detection, workers are required to conduct step-by-step exploration along the direction of the directly buried cable, which is time-consuming and labor-intensive, resulting in low detection efficiency, and there is room for improvement. Summary of the invention

[0005] In order to improve the detection efficiency of line faults, the present application provides a wind farm line fault monitoring and early warning method and system.

[0006] In the first aspect, the present application provides a wind farm line fault monitoring and early warning method, which adopts the following technical solution:

[0007] A wind farm line fault monitoring and early warning method, comprising:

[0008] Monitor the current and voltage values ​​of the wind farm lines to obtain the value fluctuation curve, and make the first judgment on the line fault according to the value fluctuation curve;

[0009] If the result of the first judgment is that there is a line fault, the power loss of the generator set and the transmission amount of the collector line are calculated, and the line fault is judged a second time based on the power loss;

[0010] If the result of the second judgment is that there is a fault in the line, the laying area image is collected according to the pre-buried path of the direct buried cable to obtain the area image data;

[0011] Pre-embedded path marking is performed on the regional image data, and the marked regional image data is randomly divided into regions according to the direction of the pre-embedded path to obtain a plurality of image blocks;

[0012] Analyze the single image block to screen out image blocks with suspected fault points and record them as first fault image data;

[0013] Compare and analyze the multiple image blocks to obtain image blocks with suspected fault points, and record them as second fault image data;

[0014] The first fault image data and the second fault image data are integrated to obtain a fault image block, and the fault image block is analyzed according to the value fluctuation curve and the electric loss to determine the fault point.

[0015] Preferably, color information is extracted from the image block, and the color information of the image block is judged to be single, so as to determine whether the colors in the image block are of the same color system;

[0016] If the color information in the image block is of the same color system, it is determined that the image block does not have a suspected fault point;

[0017] If the color information in the image block is not of the same color system, the color information in the image block is matched with the built-in fault color;

[0018] If there is a faulty color in the color information of the image block, the gradient direction of the faulty color is determined according to the color brightness of the faulty color in different areas of the image block;

[0019] The positional relationship between the gradient direction of the fault color and the marked pre-embedded path is judged. If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path is correlated, it is determined that there is a suspected fault point in the image block;

[0020] If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path does not show correlation, it is determined that there is no suspected fault point in the image block.

[0021] Preferably, the color information of each image block is counted to obtain the main color tone of the color in the regional image data;

[0022] Comparing the color information in the image block with a single color system with the main color tone of the color in the regional image data to determine whether the color information in the image block with a single color system is the same as the main color tone in the regional image data;

[0023] If the color information in the image block of a single color system is the same as the main color tone in the regional image data, it is determined that there is no suspected fault point in the image block;

[0024] If the color information in the image block of a single color system is different from the main color tone in the regional image data, the adjacent image blocks are read for the image block until the color information in the adjacent image blocks is the same as the main color tone in the regional image data;

[0025] Integrate the read adjacent image blocks to obtain an integrated image block, which is recorded as a set image block;

[0026] Marking the fault color in the set image block to obtain the fault color area, and analyzing the fault color area to determine whether there is a suspected fault point in the set image block;

[0027] If it is determined that there is a suspected fault point in the set image block, the center point of the area covered by the fault color in the set image block is confirmed to obtain the center point position data;

[0028] The image block where the center point position data is located is marked as a suspected fault point.

[0029] Preferably, if the color information in the image block of a single color system is the same as the main color tone in the regional image data, the image block is identified based on the built-in vegetation category to determine the vegetation category present in the image block;

[0030] Match the identified vegetation category with the built-in vegetation information to determine the growth status data of the vegetation in the current image block;

[0031] The image block is divided based on the pre-embedded path and the image block boundary to obtain an edge image and an intermediate image;

[0032] Compare the growth state data of the vegetation in the edge image with the built-in state standard value to determine whether the growth state data of the vegetation in the edge image is greater than the state standard value;

[0033] If the growth state data of the vegetation in the edge image is greater than or equal to the state standard value, the growth state data of the vegetation in the edge image is compared with the growth state data of the vegetation in the middle image;

[0034] If the growth status data of the vegetation in the middle image is greater than or equal to the growth status data of the vegetation in the edge image, it is determined that there is no suspected fault point in the image block;

[0035] If the growth status data of the vegetation in the middle image is smaller than the growth status data of the vegetation in the edge image, it is determined that there is a suspected fault point in the image block;

[0036] If the growth state data of the vegetation in the edge image is less than the state standard value, it is determined that there is a suspected fault point in the image block.

[0037] Preferably, environmental judgment is performed on the regional image data to obtain overall environmental information;

[0038] Performing environmental judgment on the fault color area containing only the fault color in the set image block to obtain sub-environmental information;

[0039] comparing the total environment information with the sub-environment information to determine whether the sub-environment information is the same as the total environment information;

[0040] If the sub-environment information is the same as the total environment information, it indicates that the environment information of the set image block has not changed, and it is determined that there is a suspected fault point in the set image block;

[0041] If the sub-environment information is different from the total environment information, it indicates that the environment information of the set image block has changed, and the regularity of the fault color area in the set image block is judged;

[0042] If the fault color area in the set image block is regular, then there is no suspected fault point in the set image block;

[0043] If the fault color area in the set image block is irregular, it is determined that there is a suspected fault point in the set image block.

[0044] Preferably, when it is determined that the environmental information of the set image block has changed, a boundary line judgment is performed on the faulty color area in the set image block to determine whether the faulty color area in the set image block has an obvious boundary line;

[0045] Performing regular pattern judgment on the fault color area in the set image block to determine whether the pattern corresponding to the image area of ​​the fault color in the set image block is a regular pattern;

[0046] If it is determined that the faulty color region in the set image block has an obvious boundary line and the pattern corresponding to the faulty color region in the set image block is a regular pattern, then it is determined that the image region of the faulty color in the set image block has regularity;

[0047] If it is determined that the faulty color region in the set image block does not have an obvious boundary line and / or the pattern corresponding to the faulty color region in the set image block is not a regular pattern, then it is determined that the image region of the faulty color in the set image block has no regularity.

[0048] Preferably, the first fault image data and the second fault image data are statistically summarized to obtain the number of suspected fault points;

[0049] Obtain the number of historical fault points under historical fault states and the historical numerical fluctuation curve and historical power loss value corresponding to the number of historical fault points;

[0050] The historical fault points and the historical numerical fluctuation curve and the historical power loss value under the corresponding historical fault points are respectively subjected to feature value extraction to obtain the data relationship between the historical fault points and the historical numerical fluctuation curve and the historical power loss value, and obtain the data relationship formula;

[0051] Based on the data relationship, the numerical fluctuation curve and the power loss in the current state are evaluated to obtain the corresponding number of first fault points and the number of second fault points;

[0052] Calculate the mean of the first fault point number and the second fault point number to obtain an approximate integer value of the fault point, and record it as the expected fault point number;

[0053] The number of suspected fault points is compared with the expected number of fault points. If the number of suspected fault points is less than or equal to the expected number of fault points, all image blocks corresponding to the suspected fault points are marked as having fault points.

[0054] If the number of suspected fault points is greater than the expected number of fault points, the suspected fault points are sequentially marked according to the direction of the embedded path, and the detection sequence is obtained and output.

[0055] In the second aspect, the present application provides a wind farm line fault monitoring and early warning system, which adopts the following technical solution:

[0056] A wind farm line fault monitoring and early warning system comprises: a fault judgment module, an image acquisition module and a fault point positioning module;

[0057] The fault judgment module monitors the current and voltage values ​​of the wind farm line to obtain a numerical fluctuation curve, and makes a first judgment on the line fault according to the numerical fluctuation curve; if the first judgment result is that there is a line fault, the power generation of the generator set and the transmission amount of the collector line are calculated for power loss, and a second judgment on the line fault is made according to the power loss;

[0058] The image acquisition module, if the second judgment result is that there is a fault in the line, then performs laying area image acquisition according to the pre-buried path of the direct buried cable to obtain regional image data; the pre-buried path is marked on the regional image data, and the marked regional image data is randomly divided into regions according to the direction of the pre-buried path to obtain multiple image blocks;

[0059] The fault point location module analyzes a single image block to screen out image blocks with suspected fault points, and records them as first fault image data; compares and analyzes multiple image blocks to screen out image blocks with suspected fault points, and records them as second fault image data; integrates the first fault image data and the second fault image data to obtain a fault image block, and analyzes the fault image block according to the numerical fluctuation curve and the electrical loss to determine the fault point.

[0060] In summary, the present application includes at least one of the following beneficial technical effects:

[0061] By using the numerical fluctuation of current and voltage values ​​to make the first judgment on line faults, and by using the power loss of the line to make the second judgment on line faults, the accuracy of line fault judgment is improved. When it is determined that there is a fault in the line, the image of the laying area of ​​the direct buried cable is collected to determine the status information of the laying area of ​​the direct buried cable. By randomly dividing the collected regional image data into regions, the amount of data when the system analyzes the image data is reduced, and the analysis efficiency of the system is improved. At the same time, by randomly dividing the regional image data, the divided image blocks are irregular, thereby improving the accuracy of the system analysis results. First, by analyzing a single image block, and then by analyzing the joint effect between multiple image blocks, the suspected fault points are judged more comprehensively, and then combined with the numerical fluctuation curve of the current and voltage values ​​for comprehensive analysis, the suspected fault points are further screened and judged, so that the final judgment result is more accurate, thereby reducing the number of manual measurements and improving the detection efficiency of line faults.

[0062] By comparing the color information in the same image block, it is determined whether the image block has color anomalies. If they are all in the same color system, it indicates that there is no color anomaly in the image block. When the color information in the image block is not in the same color system, it is necessary to determine the reason for not being in the same color system. By matching the color information in the image block with the built-in fault color, it is determined whether the color in the image block is normal. If the color is a fault color, the positional relationship between the color brightness of the fault color and the embedded path is determined, thereby further strengthening the judgment of the fault point of multiple image blocks and improving the judgment accuracy of the suspected fault point.

[0063] By comparing the environmental information of the set image block with the regional image data, it is determined whether the environmental information in the set image block has changed, and then corresponding analysis is performed according to different situations to improve the accuracy of the judgment on the set image block. When it is determined that the environment in which the set image block is located has not changed, that is, a fault color appears in the set image block, it indicates that the area where the set image block is located is affected by the fault of the direct buried cable. When it is determined that the environment in which the set image block is located has changed, the fault color in the set image block is manually judged to further determine whether the change in the environment in which the set image block is located is caused by human factors. If it is determined to be caused by human factors, the set image block is marked as having no suspected fault point, otherwise, the set image block is marked as a suspected fault point, thereby improving the accuracy of the judgment on whether there is a suspected fault point in the set image block. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a flow chart of the steps of the wind farm line fault monitoring and early warning method of this embodiment;

[0065] Figure 2 This is a module block diagram of the wind farm line fault monitoring and early warning system of this embodiment.

[0066] Figure numerals: 1. Fault judgment module; 2. Image acquisition module; 3. Fault point positioning module. DETAILED DESCRIPTION

[0067] The following is combined with Figure 1-Figure 2 This application is described in further detail.

[0068] The embodiments of the present application disclose a wind farm line fault monitoring and early warning method and system.

[0069] Example: Figure 1 As shown, the present invention provides a wind farm line fault monitoring and early warning method, comprising:

[0070] S100, monitoring the current and voltage values ​​of the wind farm line to obtain a numerical fluctuation curve, and making a first judgment on the line fault according to the numerical fluctuation curve; wherein, the line fault judgment is made by the numerical fluctuation curve of the current and voltage values, specifically: when the insulation layer of the line is broken, since the current in the line remains unchanged, the resistance increases, and then the voltage decreases, so the voltage fluctuation value can be effectively used to judge whether the line has a line fault.

[0071] S200, if the result of the first judgment is that there is a line fault, the power loss of the power generation unit and the transmission amount of the collector line are calculated, and the line fault is judged for the second time based on the power loss; wherein, the line fault is judged by the power loss, specifically: when the line is intact, the power loss of the line is in a stable state, and when the line fails, due to the increase in resistance, the power loss generated when the voltage and current flow through the line also increases, and then by comparing the current power loss with the past power loss, it can be effectively judged whether there is a line fault in the current state of the line.

[0072] S300, if the second judgment result is that there is a line fault, then collect images of the laying area according to the pre-buried path of the direct buried cable to obtain regional image data;

[0073] S400, marking the regional image data with a pre-embedded path, and randomly dividing the marked regional image data into regions according to the direction of the pre-embedded path to obtain a plurality of image blocks;

[0074] S500, analyzing a single image block to screen out an image block having a suspected fault point and recording it as first fault image data;

[0075] S600, comparing and analyzing the multiple image blocks to screen out image blocks with suspected fault points, and recording them as second fault image data;

[0076] S700, integrating the first fault image data and the second fault image data to obtain a fault image block, and analyzing the fault image block according to the value fluctuation curve and the electric loss to determine the fault point.

[0077] In this embodiment, when the insulation layer of the direct-buried cable is broken, since the direct-buried cable is buried underground, the broken part is in an external resistance state, so the current and voltage values ​​transmitted by the wind farm will fluctuate irregularly, and the line fault can be judged according to the fluctuation of the current and voltage values ​​of the wind farm. Similarly, when the insulation layer of the direct-buried cable is broken, the broken part is in an external resistance state, so the line will increase the power loss due to the increase in resistance when transmitting electric energy, and the line fault can be judged according to the change of line power loss of the wind farm. When it is determined that the insulation layer of the direct-buried cable in the wind farm line is broken, the direct-buried cable discharges to the soil, which will cause the moisture in the soil to be electrolyzed, and the increase in resistance will cause the temperature of the soil to rise, which will cause the soil composition and microorganisms in the soil to change, thereby affecting the growth state of the surface vegetation. Therefore, by collecting images of the laying area of ​​the direct-buried cable, and then analyzing the collected regional image data, the area where the vegetation changes caused by the change in soil composition due to the rupture of the insulation layer of the direct-buried cable is determined, and the corresponding fault point is determined.

[0078] The line fault is first judged by utilizing the numerical fluctuation of the current and voltage values, and the line fault is secondly judged by the power loss of the line, thereby improving the accuracy of the line fault judgment. When it is determined that there is a fault in the line, the image of the laying area of ​​the direct buried cable is collected to determine the status information of the laying area of ​​the direct buried cable. By randomly dividing the collected regional image data into regions, the amount of data when the system analyzes the image data is reduced, and the analysis efficiency of the system is improved. At the same time, by randomly dividing the regional image data, the divided image blocks are irregular, thereby improving the accuracy of the system analysis results. First, a single image block is analyzed, and then the joint effect between multiple image blocks is analyzed, so that the suspected fault points are judged more comprehensively, and then a comprehensive analysis is performed in combination with the numerical fluctuation curve of the current and voltage values, and the suspected fault points are further screened and judged, so that the final judgment result is more accurate.

[0079] For example, if a wind power station is built on a grassland, the corresponding direct buried cable laying area is also on the grassland. When the direct buried cable leaks electricity, the high voltage current will damage the soil environment around the cable, causing the grass living in the area to grow poorly or even die. Therefore, by collecting images of the laying area of ​​the direct buried cable and comparing the collected images with each other, it is possible to determine whether there is an area in the image where the growth state of the grass is different from that of the surrounding area or there is an area where the grass suddenly dies. Based on this phenomenon, the area is marked and analyzed again to determine whether the area is caused by leakage, and then the final fault point is determined.

[0080] In step S500, a single image block is analyzed to screen out image blocks with suspected fault points and record them as first fault image data, including the following steps:

[0081] S510, extracting color information from the image block, and determining the uniformity of the color information of the image block to determine whether the colors in the image block are of the same color system; wherein the same color system is the same color category, for example, the green system may include dark green, light green, fluorescent green, etc.

[0082] S520, if the color information in the image block is of the same color system, determining that the image block does not have a suspected fault point;

[0083] S530, if the color information in the image block is not of the same color system, the color information in the image block is matched with a built-in fault color; wherein the fault color is a common environmental color after the environment changes due to leakage faults in direct buried cables in different environments.

[0084] S540, if there is a faulty color in the color information of the image block, determining a gradient direction of the faulty color according to the color brightness of the faulty color in different areas of the image block;

[0085] S550, determine the positional relationship between the gradient direction of the fault color and the marked pre-embedded path. If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path is correlated, it is determined that there is a suspected fault point in the image block; wherein the correlation is: the farther the area where the fault color is located is from the pre-embedded path, the brighter the color brightness of the corresponding fault color or the darker the color brightness, and it is determined that there is a correlation; conversely, if there is no correlation between the color brightness of the fault color and the distance between the area where the fault color is located and the pre-embedded path, it is determined that there is no correlation.

[0086] S560: If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path is not correlated, it is determined that there is no suspected fault point in the image block.

[0087] In this embodiment, when the insulation layer of the cable breaks, the result of the rupture increases from small to large until it causes a wind farm failure. During the gradual change, the current released from the rupture also gradually increases. Therefore, the impact on the surface vegetation gradually changes during the continuous process. Therefore, when analyzing a single image block, it is possible to determine to a certain extent whether the image block has a corresponding vegetation growth state phenomenon.

[0088] By performing color extraction on the image block and comparing the color information in the same image block, it is determined whether the image block has color anomalies. If they are all of the same color system, it indicates that there is no color anomaly in the image block. When the color information in the image block is not of the same color system, it is necessary to determine the reason for not being of the same color system. By matching the color information in the image block with the built-in fault color, it is determined whether the color in the image block is normal. If the color is a fault color, the positional relationship between the color brightness of the fault color and the embedded path is judged, thereby further strengthening the judgment of the fault points of multiple image blocks and improving the accuracy of judgment of suspected fault points.

[0089] For example, if a direct buried cable has leakage, the leakage will damage the soil environment around the leakage area, thereby preventing the grass in the soil environment from growing normally, thus causing the grass that originally grew normally in the soil environment to change from green to yellow. After dividing the soil environment, the colors on the divided image blocks may be all yellow, all green, and partial yellow + partial green. Among them, all yellow and all green cannot be judged as faults by a single image block, while the partial yellow + partial green situation indicates that one of the yellow or green colors is not the environmental color. Therefore, the color in this situation is compared with the fault color. For example, yellow is the fault color, and the partial yellow + partial green situation has the fault color: yellow, indicating that the image block may have a fault. Then, the color brightness of the area composed of the fault color is judged. If the starting point or end point of the gradient direction of the fault color coincides with the location of the cable, the appearance of yellow is likely to be caused by cable leakage, so the image block is marked as a suspected fault point.

[0090] In step S600, multiple image blocks are compared and analyzed with each other to screen out image blocks with suspected fault points and record them as second fault image data, including the following steps:

[0091] S610, performing statistics on the color information of each image block to obtain the main color tone of the color in the regional image data;

[0092] S620, comparing the color information in the image block with a single color system with the main color tone of the color in the regional image data to determine whether the color information in the image block with a single color system is the same as the main color tone in the regional image data;

[0093] S630, if the color information in the image block of the single color system is the same as the main color tone in the regional image data, it is determined that the image block does not have a suspected fault point;

[0094] S640, if the color information in the image block of the single color system is different from the main color tone in the regional image data, read the adjacent image blocks for the image block until the color information in the adjacent image blocks is the same as the main color tone in the regional image data;

[0095] S650, integrating the read adjacent image blocks to obtain an integrated image block, and recording it as a set image block;

[0096] S660, marking the fault color in the set image block to obtain a fault color area, and analyzing the fault color area to determine whether there is a suspected fault point in the set image block;

[0097] S670, if it is determined that there is a suspected fault point in the set image block, confirm the center point of the area covered by the fault color in the set image block to obtain center point position data;

[0098] S680: Mark the image block where the center point position data is located as a suspected fault point.

[0099] In this embodiment, after analyzing and judging a single image block, a first screening is performed to screen out image blocks with suspected fault points, and then a second screening is performed on the image blocks based on the relationship between multiple image blocks to obtain image blocks with suspected fault points. Through multiple screenings, the screened image blocks are made more accurate and comprehensive, thereby improving the accuracy of the judgment results of the fault area.

[0100] By statistically analyzing the color information of each image block, the main color tone of the environment in which the direct buried cable is located is clarified. By comparing the color of the image blocks of a single color system with the main color tone of the environment, the image blocks with the possibility of fault are determined, thereby reducing the possibility of misjudgment. By jointly analyzing the image blocks with the possibility of fault and the adjacent image blocks, the image blocks with the possibility of fault are further judged, thereby further improving the accuracy of the judgment results.

[0101] Exemplary, for example, if the environment in which the direct buried cable is located is grassland, the main color of the corresponding environment is green. Since the main color of the environment is green, when there is no leakage in the direct buried cable, the colors of all the corresponding collected image blocks are all single green. On the contrary, when leakage occurs in the direct buried cable, there may be image blocks with yellow colors in the corresponding collected image blocks. Therefore, by comparing the image blocks with the main color of the environment, it is possible to effectively determine whether the color block is an abnormal color block. For example, if there is a single yellow image block, when it is compared with the main color of the environment, the comparison result is a different color, then the image block needs to be marked. At the same time, since the image block is all yellow, it is impossible to determine whether the yellow color in the image block is caused by cable leakage. Therefore, further judgment is required, by reading the image blocks adjacent to the image block, and then performing color gradient analysis on the collective image block that combines multiple image blocks, and then judging whether the yellow color of the area corresponding to the collective image block is caused by cable leakage, thereby improving the accuracy of judgment.

[0102] In step S630, if the color information in the image block of a single color system is the same as the main color tone in the regional image data, it is determined that there is no suspected fault point in the image block, including the following steps:

[0103] S631, if the color information in the image block of a single color system is the same as the main color tone in the regional image data, the image block is identified based on the built-in vegetation category to determine the vegetation category existing in the image block;

[0104] S632, matching the identified vegetation category with the built-in vegetation information to determine the growth status data of the vegetation in the current image block;

[0105] S633, dividing the image block based on the pre-embedded path and the image block boundary to obtain an edge image and an intermediate image;

[0106] S634, comparing the growth status data of the vegetation in the edge image with the built-in status standard value to determine whether the growth status data of the vegetation in the edge image is greater than the status standard value; wherein the status standard value may be a mean value calculated based on the vegetation growth status of the surrounding image blocks.

[0107] S635, if the growth status data of the vegetation in the edge image is greater than or equal to the status standard value, then comparing the growth status data of the vegetation in the edge image with the growth status data of the vegetation in the intermediate image;

[0108] S636, if the growth status data of the vegetation in the middle image is greater than or equal to the growth status data of the vegetation in the edge image, it is determined that there is no suspected fault point in the image block;

[0109] S637, if the growth status data of the vegetation in the middle image is smaller than the growth status data of the vegetation in the edge image, it is determined that there is a suspected fault point in the image block;

[0110] S638: If the growth status data of the vegetation in the edge image is less than the status standard value, it is determined that there is a suspected fault point in the image block.

[0111] In this embodiment, since there is a time difference from the beginning of the cable leakage to the effect on the vegetation, when it is determined that the color information in the image block of a single color system is the same as the main color tone in the regional image data, it is necessary to further judge the image block of the single color system. When it is determined that the color in the image block of a single color system is the same as the main color tone in the regional image data, the vegetation state in the image block is identified to determine whether the growth state of the vegetation in the region is normal, and the region is further judged by the growth state of the vegetation in the region. The image block is divided with the pre-buried path as the center line and the boundary of the image block as the edge curve to obtain an edge image and an intermediate image. By comparing the growth state data of the vegetation in the edge image with the state standard value, and comparing the growth state data of the vegetation in the edge image with the growth state data of the vegetation in the intermediate image, it is determined whether the image block has a suspected fault point, thereby improving the accuracy of the judgment.

[0112] For example, since cable leakage can damage the soil, and the damaged soil needs a certain time period to act on the surface vegetation, when there is a certain time difference between the detection period of the cable leakage phenomenon and the surface vegetation status, the growth state of the surface vegetation will change but not completely change, and the cable leakage phenomenon has been detected. For example, the complete manifestation of the surface vegetation to leakage is that the vegetation turns yellow and withers, but due to the time difference, the surface vegetation is shrunk but not dead. At this time, the color of the surface vegetation is the same as the main color of the area. Therefore, the growth state of the surface vegetation in the area is determined by identification, and then the area is divided. When the cable in the area leaks, the vegetation closer to the cable is more affected, and the farther away from the cable, the less affected the vegetation is. Therefore, the state standard value of the growth state data of the vegetation in the edge image is judged to determine whether the area is affected by the cable leakage. For example, 0~10 is used as the evaluation standard of vegetation status, where 10 indicates the best growth status and 0 indicates that the vegetation is dead. Therefore, when the growth status of the vegetation in the surrounding area is 10, the vegetation growth status data at the edge position of the area is 6, which indicates that the area is affected by cable leakage. When the growth status of the vegetation in the surrounding area is 6 and the vegetation growth status data at the edge position of the current area is 7, it indicates that the edge position is in a normal state. Therefore, further judgment is required. By comparing the vegetation growth status data of the edge image and the intermediate image, if the vegetation growth status data in the intermediate image is greater than or equal to the vegetation growth status data in the edge image, it indicates that the area closer to the cable is not affected by the cable, and then it can be judged that there is no leakage in the cable of the image block. If the vegetation growth status data in the intermediate image is less than the vegetation growth status data in the edge image, since the growth status of the intermediate image is affected first, it is judged that the image block is affected by the cable, and then it is judged that there is leakage in the cable of the image block.

[0113] In step S660, the fault color in the set image block is marked to obtain a fault color area, and the fault color area is analyzed to determine whether there is a suspected fault point in the set image block, including the following steps:

[0114] S661, performing environmental judgment on the regional image data to obtain overall environmental information;

[0115] S662, performing environmental judgment on the faulty color region containing only the faulty color in the set image block to obtain sub-environmental information;

[0116] S663, comparing the total environment information with the sub-environment information to determine whether the sub-environment information is the same as the total environment information;

[0117] S664, if the sub-environment information is the same as the total environment information, it indicates that the environment information of the set image block has not changed, and it is determined that there is a suspected fault point in the set image block;

[0118] S665, if the sub-environmental information is different from the total environmental information, it indicates that the environmental information of the collective image block has changed, and a regularity judgment is performed on the faulty color area in the collective image block; when the environmental information of the collective image block changes, for example, the original grassland turns into wasteland, it may be due to cable leakage causing a large number of grasslands to die, thus evolving into wasteland, or it may be due to human weeding or walking on the grassland causing the grassland in the area to die, thus evolving into wasteland, so it is necessary to perform a regularity judgment on the image area of ​​the faulty color in the collective image block.

[0119] S666, if the fault color region in the set image block has regularity, then the set image block does not have a suspected fault point;

[0120] S667: If the fault color region in the set image block is irregular, it is determined that there is a suspected fault point in the set image block.

[0121] In this embodiment, environmental information is compared between the collective image block and the regional image data to determine whether the environmental information in the collective image block has changed, and then corresponding analysis is performed according to different situations to improve the accuracy of judgment on the collective image block. When it is determined that the environment in which the collective image block is located has not changed, that is, a fault color appears in the collective image block, it indicates that the area where the collective image block is located is affected by the fault of the direct buried cable. When it is determined that the environment in which the collective image block is located has changed, the fault color in the collective image block is manually judged to further determine whether the change in the environment in which the collective image block is located is caused by human factors. If it is determined to be caused by human factors, the collective image block is marked as having no suspected fault point. Otherwise, the collective image block is marked as a suspected fault point.

[0122] Exemplarily, for example, if the image area of ​​the faulty color in the set image block is grassland, and the regional image data is also grassland, it indicates that the environment in which the set image block is located has not changed. On the contrary, when the image area of ​​the faulty color in the set image block is wasteland, it indicates that the environment in which the set image block is located has changed, so it is necessary to re-analyze the cause of the environmental change. Since when humans destroy the environment, it has a certain purpose, so the wasteland in the set image block tends to be more regular. Therefore, by judging the regularity of the wasteland, it is determined whether it is human destruction, and then an accurate judgment can be made on the wasteland in the set image block, thereby improving the accuracy of the judgment result.

[0123] In step S665, if the sub-environment information is different from the total environment information, it indicates that the environment information of the set image block has changed, and the regularity of the fault color area in the set image block is judged, including the following steps:

[0124] S6651, when it is determined that the environmental information of the set image block has changed, a boundary line judgment is performed on the faulty color region in the set image block to determine whether the faulty color region in the set image block has an obvious boundary line;

[0125] S6652, performing regular pattern judgment on the faulty color area in the set image block to determine whether the pattern corresponding to the image area of ​​the faulty color in the set image block is a regular pattern; wherein the regular pattern may be a circle, rectangle, diamond, trapezoid or other pattern that can find an obvious pattern.

[0126] S6653, if it is determined that the faulty color region in the set image block has an obvious boundary line and the pattern corresponding to the faulty color region in the set image block is a regular pattern, then it is determined that the image region of the faulty color in the set image block has regularity;

[0127] S6654: If it is determined that the faulty color area in the set image block does not have an obvious boundary line and / or the pattern corresponding to the faulty color area in the set image block is not a regular pattern, it is determined that the image area of ​​the faulty color in the set image block is irregular.

[0128] In this embodiment, edge boundary line judgment and graphic regularity judgment are performed on the image area of ​​the fault color in the collective image block to determine whether the formation of the fault color in the collective image block is human-caused, thereby improving the accuracy of judging whether the collective image block has suspected fault points.

[0129] In step S700, the first fault image data and the second fault image data are integrated to obtain a fault image block, and the fault image block is analyzed according to the value fluctuation curve and the power loss to determine the fault point, including the following steps:

[0130] S710, performing statistics aggregation on the first fault image data and the second fault image data to obtain the number of suspected fault points;

[0131] S720, obtaining the number of historical fault points under the historical fault state and the historical numerical fluctuation curve and the historical power loss value corresponding to the number of historical fault points;

[0132] S730, extracting characteristic values ​​of the historical fault points and the historical numerical fluctuation curve and the historical power loss value under the corresponding historical fault points, respectively, to obtain a data relationship between the historical fault points and the historical numerical fluctuation curve and the historical power loss value, and to obtain a data relationship formula;

[0133] S740, evaluating the value fluctuation curve and the power loss in the current state based on the data relationship formula to obtain the corresponding first fault point number and second fault point number;

[0134] S750, calculating the mean of the first number of fault points and the second number of fault points to obtain an approximate integer value of the fault points, and recording the approximate integer value as the estimated number of fault points;

[0135] S760, comparing the number of suspected fault points with the expected number of fault points, if the number of suspected fault points is less than or equal to the expected number of fault points, marking all image blocks corresponding to the suspected fault points as having fault points;

[0136] S770, if the number of suspected fault points is greater than the expected number of fault points, the suspected fault points are sequentially marked according to the direction of the embedded path, and a detection sequence is obtained and output. The detection personnel can perform targeted detection on the suspected fault points according to the detection sequence, thereby reducing the workload of the detection personnel and improving the efficiency of line detection.

[0137] In this embodiment, the first fault image data determined by a single image block and the second fault image data jointly analyzed by multiple image blocks are statistically summarized to obtain the number of image blocks determined to have suspected fault points. The relationship between the historical number of fault points and the historical numerical fluctuation curve and the historical power loss value under the corresponding historical number of fault points is determined by performing feature analysis. The numerical fluctuation curve and the power loss in the current state are evaluated for the number of fault points based on the data relationship to determine the number of possible fault points in the current state. The expected number of fault points is compared with the number of suspected fault points. A corresponding judgment is made based on the relationship between the expected number of fault points and the number of suspected fault points to screen out the final fault point location, thereby improving the accuracy of fault point judgment.

[0138] Exemplarily, for example, the first fault image data includes image block a and image block b, and the second fault image data includes image block b and image block c. The statistical summary result is three suspected fault points of image block a, image block b, and image block c, that is, the number of suspected fault points is 3. When the number of historical fault points under the historical fault state and the historical numerical fluctuation curve and historical power loss value corresponding to the number of historical fault points are judged to be z=ax, where z is the number of fault points, a is the relationship coefficient of the numerical fluctuation curve, and x is the fluctuation value of the numerical fluctuation curve, the number of fault points and the power loss value are z=by, where b is the relationship coefficient of the power loss value, and y is the power loss value. By combining the current numerical fluctuation curve, the power loss Substitute the consumption value into the corresponding data relationship formula to obtain the number of the first fault points and the number of the second fault points in the current state. For example, the number of the first fault points is 3 and the number of the second fault points is 4, then the calculated mean is 3.5, and the integer value is 4, that is, the number of fault points in the current situation is expected to be 4, and the number of suspected fault points screened by the above steps is 3, 3<4, so all suspected fault points are fault points, and if the calculated number of fault points in the current situation is expected to be 2, then 3>2, then it is necessary to re-screen the suspected fault points, and use the existing ground potential detection method according to the direction of the embedded path to perform targeted analysis on the suspected fault points, thereby reducing the number of manual judgments and improving the efficiency of line fault detection.

[0139] Based on the description of the above wind farm line fault monitoring and early warning method embodiment, the embodiment of the present invention also discloses a wind farm line fault monitoring and early warning system:

[0140] like Figure 2 As shown, a wind farm line fault monitoring and early warning system, by applying the wind farm line fault monitoring and early warning method as described above, comprises: a fault judgment module 1, an image acquisition module 2 and a fault point positioning module 3;

[0141] Fault judgment module 1 monitors the current and voltage values ​​of the wind farm line to obtain a numerical fluctuation curve, and makes a first judgment on the line fault according to the numerical fluctuation curve; if the first judgment result is that there is a line fault, the power generation of the generator set and the transmission amount of the collector line are calculated for power loss, and a second judgment on the line fault is made according to the power loss;

[0142] Image acquisition module 2, if the second judgment result is that there is a fault in the line, then the laying area image is acquired according to the pre-buried path of the direct buried cable to obtain regional image data; the pre-buried path is marked on the regional image data, and the marked regional image data is randomly divided into regions according to the direction of the pre-buried path to obtain multiple image blocks;

[0143] The fault point location module 3 analyzes a single image block to screen out an image block with a suspected fault point, and records it as the first fault image data; compares and analyzes multiple image blocks to screen out an image block with a suspected fault point, and records it as the second fault image data; integrates the first fault image data and the second fault image data to obtain a fault image block, and analyzes the fault image block according to the numerical fluctuation curve and the electrical loss to determine the fault point.

[0144] Compared with the existing wind farm line fault monitoring and early warning method and system, the present invention improves the detection efficiency of line faults.

[0145] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A wind farm line fault monitoring and early warning method, characterized in that: include: Monitor the current and voltage values ​​of the wind farm lines to obtain the value fluctuation curve, and make the first judgment on the line fault according to the value fluctuation curve; If the result of the first judgment is that there is a line fault, the power loss of the generator set and the transmission amount of the collector line are calculated, and the line fault is judged a second time based on the power loss; If the result of the second judgment is that there is a fault in the line, the laying area image is collected according to the pre-buried path of the direct buried cable to obtain the area image data; Pre-embedded path marking is performed on the regional image data, and the marked regional image data is randomly divided into regions according to the direction of the pre-embedded path to obtain a plurality of image blocks; Analyze the single image block to screen out image blocks with suspected fault points and record them as first fault image data; Extracting color information from the image block and judging the uniformity of the color information of the image block to determine whether the colors in the image block are of the same color system; If the color information in the image block is of the same color system, it is determined that the image block does not have a suspected fault point; If the color information in the image block is not of the same color system, the color information in the image block is matched with the built-in fault color; If there is a faulty color in the color information of the image block, the gradient direction of the faulty color is determined according to the color brightness of the faulty color in different areas of the image block; The positional relationship between the gradient direction of the fault color and the marked pre-embedded path is judged. If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path is correlated, it is determined that there is a suspected fault point in the image block; If the positional relationship between the gradient direction of the fault color and the marked pre-embedded path does not show correlation, it is determined that there is no suspected fault point in the image block; Compare and analyze the multiple image blocks to obtain image blocks with suspected fault points, and record them as second fault image data; The color information of each image block is counted to obtain the main color tone of the color in the regional image data; Comparing the color information in the image block with a single color system with the main color tone of the color in the regional image data to determine whether the color information in the image block with a single color system is the same as the main color tone in the regional image data; If the color information in the image block of a single color system is the same as the main color tone in the regional image data, it is determined that there is no suspected fault point in the image block; If the color information in the image block of a single color system is the same as the main color tone in the regional image data, the image block is identified based on the built-in vegetation category to determine the vegetation category existing in the image block; Match the identified vegetation category with the built-in vegetation information to determine the growth status data of the vegetation in the current image block; The image block is divided based on the pre-embedded path and the image block boundary to obtain an edge image and an intermediate image; Compare the growth state data of the vegetation in the edge image with the built-in state standard value to determine whether the growth state data of the vegetation in the edge image is greater than the state standard value; If the growth state data of the vegetation in the edge image is greater than or equal to the state standard value, the growth state data of the vegetation in the edge image is compared with the growth state data of the vegetation in the middle image; If the growth status data of the vegetation in the middle image is greater than or equal to the growth status data of the vegetation in the edge image, it is determined that there is no suspected fault point in the image block; If the growth status data of the vegetation in the middle image is smaller than the growth status data of the vegetation in the edge image, it is determined that there is a suspected fault point in the image block; If the growth state data of the vegetation in the edge image is less than the state standard value, it is determined that there is a suspected fault point in the image block; If the color information in the image block of a single color system is different from the main color tone in the regional image data, the adjacent image blocks are read for the image block until the color information in the adjacent image blocks is the same as the main color tone in the regional image data; Integrate the read adjacent image blocks to obtain an integrated image block, which is recorded as a set image block; Marking the fault color in the set image block to obtain the fault color area, and analyzing the fault color area to determine whether there is a suspected fault point in the set image block; Perform environmental judgment on regional image data to obtain overall environmental information; Performing environmental judgment on the fault color area containing only the fault color in the set image block to obtain sub-environmental information; comparing the total environment information with the sub-environment information to determine whether the sub-environment information is the same as the total environment information; If the sub-environment information is the same as the total environment information, it indicates that the environment information of the set image block has not changed, and it is determined that there is a suspected fault point in the set image block; If the sub-environment information is different from the total environment information, it indicates that the environment information of the set image block has changed, and the regularity of the fault color area in the set image block is judged; If the fault color area in the set image block is regular, then there is no suspected fault point in the set image block; If the fault color area in the set image block is irregular, it is determined that there is a suspected fault point in the set image block; If it is determined that there is a suspected fault point in the set image block, the center point of the area covered by the fault color in the set image block is confirmed to obtain the center point position data; The image block where the center point position data is located is marked as a suspected fault point; The first fault image data and the second fault image data are integrated to obtain a fault image block, and the fault image block is analyzed according to the value fluctuation curve and the electric loss to determine the fault point.

2. A wind farm line fault monitoring and early warning method according to claim 1, characterized in that: If the sub-environment information is different from the total environment information, it indicates that the environment information of the set image block has changed, and the regularity of the fault color area in the set image block is judged, specifically: When it is determined that the environmental information of the set image block has changed, a boundary line judgment is performed on the faulty color area in the set image block to determine whether the faulty color area in the set image block has an obvious boundary line; Performing regular pattern judgment on the fault color area in the set image block to determine whether the pattern corresponding to the image area of ​​the fault color in the set image block is a regular pattern; If it is determined that the faulty color region in the set image block has an obvious boundary line and the pattern corresponding to the faulty color region in the set image block is a regular pattern, then it is determined that the image region of the faulty color in the set image block has regularity; If it is determined that the faulty color region in the set image block does not have an obvious boundary line and / or the pattern corresponding to the faulty color region in the set image block is not a regular pattern, then it is determined that the image region of the faulty color in the set image block has no regularity.

3. A wind farm line fault monitoring and early warning method according to claim 1, characterized in that: The first fault image data and the second fault image data are integrated to obtain a fault image block, and the fault image block is analyzed according to the value fluctuation curve and the power loss to determine the fault point, specifically: Performing statistics on the first fault image data and the second fault image data to obtain the number of suspected fault points; Obtain the number of historical fault points under historical fault states and the historical numerical fluctuation curve and historical power loss value corresponding to the number of historical fault points; The historical fault points and the historical numerical fluctuation curve and the historical power loss value under the corresponding historical fault points are respectively subjected to feature value extraction to obtain the data relationship between the historical fault points and the historical numerical fluctuation curve and the historical power loss value, and obtain the data relationship formula; Based on the data relationship, the numerical fluctuation curve and the power loss in the current state are evaluated to obtain the corresponding number of first fault points and the number of second fault points; Calculate the mean of the first fault point number and the second fault point number to obtain an approximate integer value of the fault point, and record it as the expected fault point number; The number of suspected fault points is compared with the expected number of fault points. If the number of suspected fault points is less than or equal to the expected number of fault points, all image blocks corresponding to the suspected fault points are marked as having fault points. If the number of suspected fault points is greater than the expected number of fault points, the suspected fault points are sequentially marked according to the direction of the embedded path, and the detection sequence is obtained and output.

4. A wind farm line fault monitoring and early warning system, characterized in that: The system is used to implement a wind farm line fault monitoring and early warning method as described in any one of claims 1 to 3: comprising: a fault judgment module, an image acquisition module and a fault point positioning module; The fault judgment module monitors the current and voltage values ​​of the wind farm line to obtain a numerical fluctuation curve, and makes a first judgment on the line fault according to the numerical fluctuation curve; if the first judgment result is that there is a line fault, the power generation of the generator set and the transmission amount of the collector line are calculated for power loss, and a second judgment on the line fault is made according to the power loss; The image acquisition module, if the second judgment result is that there is a fault in the line, then performs laying area image acquisition according to the pre-buried path of the direct buried cable to obtain regional image data; the pre-buried path is marked on the regional image data, and the marked regional image data is randomly divided into regions according to the direction of the pre-buried path to obtain multiple image blocks; The fault point location module analyzes a single image block to screen out image blocks with suspected fault points, and records them as first fault image data; compares and analyzes multiple image blocks to screen out image blocks with suspected fault points, and records them as second fault image data; integrates the first fault image data and the second fault image data to obtain a fault image block, and analyzes the fault image block according to the numerical fluctuation curve and the electrical loss to determine the fault point.

Images