Insulator insulation performance dynamic evaluation and early warning method based on plant bridging identification

Abstract

The invention discloses an insulator insulation performance dynamic evaluation and early warning method based on plant bridging identification, and the method comprises the steps: carrying out the image semantic segmentation and vegetation identification through collecting insulator surface image data, and calculating a biological adhesion coverage rate and a composite bridging index; obtaining the moisture duration time, the leakage current effective value and the leakage current pulse count of the insulator, and inputting the composite bridging index, the biological adhesion coverage rate, the moisture duration time, the leakage current effective value and the leakage current pulse count into an insulator risk index model; and outputting the insulation performance risk index under plant bridging, converting the insulation performance risk index into an insulation performance risk score, comparing the insulation performance risk score with a multi-level insulation risk threshold, triggering a corresponding insulation early warning, and outputting an early warning level and plant bridging position information. By means of the method, special insulation faults caused by grass growing on the insulator can be found accurately in an early stage, conversion from passive response to active early warning is achieved, and the intelligent level and safety of operation and maintenance of a power grid are remarkably improved.

Description

Technical Field

[0001] This invention relates to the technical field of power transmission and distribution systems, and in particular to a method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification. Background Technology

[0002] Insulators are key equipment for ensuring the electrical insulation and mechanical fixation of transmission lines, and their operating status is directly related to the safety and stability of the power grid. In hot and humid regions, such as southern provinces of my country, dirt easily accumulates on the surface of insulators, and algae, moss, and even higher plants can grow.

[0003] A typical case involves an incident one summer when inspectors from the Southern Power Grid, conducting a detailed inspection of a 110kV line, discovered an anomaly in the operation of an insulator string on a remote mountain tower. Close-up drone footage revealed that lush grasses were growing between the sheds of the composite insulator. The roots of these plants were deeply embedded in the dust at the junction of the steel cap and the core rod, and their long stems and leaves had crossed the air gaps between adjacent insulator sheds, forming a stable "plant bridge" connecting multiple insulator plates. Under clear weather conditions, the line operated normally. However, after a period of dense fog, the line suddenly tripped. Fault recording equipment indicated a single-phase ground fault, and subsequent line inspection confirmed that the fault was indeed located on this grass-covered insulator string. The accident analysis concluded that in the humid environment, the bridging plant stems and leaves absorbed water, creating a high-conductivity path that significantly shortened the effective insulation creepage distance, ultimately leading to a flashover accident.

[0004] This case reveals the limitations of traditional operation and maintenance models. Currently, the monitoring and assessment of insulator condition mainly relies on regular manual inspections, measurements of equivalent salt density (ESDD) / ash density (NSDD), and online monitoring of leakage current. However, these traditional methods have significant shortcomings: manual inspections are inefficient and time-consuming, making it difficult to comprehensively cover such subtle but dangerous biological growth; the equivalent salt density method cannot effectively reflect the unique impact of biological adhesion on insulation performance; while leakage current monitoring can reflect the overall trend of insulation deterioration, it cannot distinguish the cause of the fault, and alarms are usually only triggered when insulation performance has severely deteriorated (e.g., plants have become sufficiently damp and conductive), resulting in a significant delay in early warning and an inability to specifically identify and provide early warning for the special high-risk condition of "plant bridging."

[0005] In the prior art, for example, patent application number CN202311585283.6, entitled "A Two-Stage Dispatch Method for Mobile Energy Storage to Enhance the Resilience of Distribution Networks", focuses on post-disaster resource dispatch and does not consider the special state of the equipment itself under complex weather conditions; patent application number CN202311635487.6, entitled "A Distribution Network Resilience Assessment Method Considering Distributed Power Supply Support", focuses on load recovery at the system level, but does not involve the identification and assessment of specific fault mechanisms at the component level. Summary of the Invention

[0006] To address the problem in existing technologies that fail to accurately identify and assess insulation performance degradation caused by plant bridging, this application provides an insulator condition monitoring and assessment method, particularly a dynamic assessment and early warning method for insulation performance degradation caused by plant growth bridging in insulators in humid and hot regions. This method can identify and assess insulation performance degradation caused by plant bridging early and accurately, and can issue specific early warnings in a timely manner before such "insulator overgrowth" develops into a serious fault.

[0007] The above-mentioned inventive objective of this application is achieved through the following technical solutions: A method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification, the method comprising: Collect insulator surface image data including the overall view of the insulator string and multiple perspectives, perform image semantic segmentation and vegetation recognition on the insulator surface image data, and simultaneously calculate biological attachment coverage and composite bridging index. The operating environment data and leakage current data of the insulator are collected and preprocessed to obtain the wet duration, effective value of leakage current and leakage current pulse count of the insulator. The composite bridging index, bio-attachment coverage, wet duration, effective value of leakage current, and leakage current pulse count are input into the insulator risk index model under plant bridging, and the insulation performance risk index under plant bridging is output. After normalizing the insulation performance risk index, an insulation performance risk score is obtained. This score is then compared with a preset multi-level insulation risk threshold to trigger a corresponding insulation warning and output the warning level and plant bridging location information.

[0008] In a preferred embodiment, this application can be further configured as follows: the acquisition of insulator surface image data including the overall view of the insulator string and multiple perspectives, the image surface image data undergoing semantic segmentation and vegetation recognition, and the simultaneous calculation of biological attachment coverage and composite bridging index, including: The image semantic segmentation model U-Net is used to perform three-dimensional analysis on the insulator surface image data, and the image pixels are divided into background, insulator porcelain bottle and plants. Based on the results of image segmentation and 3D analysis, target features of several vegetation units that may form bridging are identified and extracted, and corresponding bridging unit division and parameter calculation strategies are selected according to each vegetation combination morphology. The bio-attachment coverage rate and composite bridging index are calculated simultaneously according to the selected bridging unit division and parameter calculation strategy.

[0009] In a preferred embodiment, this application can be further configured as follows: the vegetation identification process, which involves acquiring insulator surface image data including the overall view of the insulator string and multiple perspectives, performing image semantic segmentation and vegetation identification on the insulator surface image data, and simultaneously calculating biological attachment coverage and composite bridging index, includes: Simultaneous images of several insulator sides are acquired, and a three-dimensional point cloud of the insulator surface is calculated and generated using a stereo matching algorithm. Based on the image semantic segmentation results, the distance between the point cloud of the region where each vegetation unit is located and the reference surface of the insulator porcelain bottle is calculated to obtain the thickness of the vegetation unit. Based on the image semantic segmentation results, the length of the vegetation unit is obtained by calculating the projection size of the three-dimensional point cloud corresponding to each segmented vegetation unit onto the insulator string axis. Each image patch of vegetation region obtained from image semantic segmentation is input into a pre-trained vegetation classification sub-model, which outputs the probability distribution of vegetation type in the vegetation region and assigns corresponding basic electrical conductivity weights.

[0010] In a preferred embodiment, this application can be further configured as follows: The step of inputting image patches of each vegetation region obtained from image semantic segmentation into a pre-trained vegetation classification sub-model, outputting the probability distribution of vegetation types in the vegetation regions, and assigning corresponding basic electrical conductivity weights, further includes: Using a dataset of labeled images containing common vegetation types of insulators as training samples, a vegetation classification sub-model was constructed for vegetation type probability distribution analysis through training a convolutional neural network.

[0011] In a preferred embodiment, this application can be further configured as follows: based on image segmentation and 3D analysis results, identifying and extracting target features of several vegetation units that may form bridging units, and selecting corresponding bridging unit division and parameter calculation strategies according to each vegetation combination morphology, including: For uniformly covered moss vegetation, the continuously covered moss area is regarded as a whole bridging unit. The equivalent thickness of the bridging unit is the average estimated thickness of the continuously covered moss area. The length of the bridging unit is defined as the vertical distance between the maximum insulating skirts that the moss-covered area can cover along the insulator string axis. For single-stem discrete vegetation, each individual stem and leaf is regarded as an independent bridging unit. The thickness of each bridging unit is the average diameter of the individual stem and leaf, and the length is the natural extension length of the individual stem and leaf. For vegetation combinations with multiple overlapping vegetation, they are considered as a whole composite bridging unit. The thickness of the composite bridging unit is taken as the equivalent total thickness of the overlapping vegetation clumps perpendicular to the bridging direction, and the length is taken as the total length of the overlapping vegetation clumps. For other vegetation regenerating on moss, it is regarded as a layered composite bridging unit with moss and regenerated vegetation as two independent bridging units, and the bridging contribution value is accumulated when calculating the composite bridging index.

[0012] In a preferred embodiment, this application can be further configured such that: the simultaneous calculation of bioattachment coverage and composite bridging index according to the selected bridging unit division and parameter calculation strategy includes: Based on the image segmentation results, the biofilm coverage rate is obtained by calculating the proportion of plant pixels to the surface area of ​​the insulator porcelain insulator. The expression for the biofilm coverage rate is as follows: (1) in, Indicates biofouling coverage. This represents the total number of pixels in the image that are identified as plants. This represents the total number of pixels identified as insulator porcelain bottles; The formula for calculating the composite bridging index is as follows: (2) in, Indicates the composite bridging index. This represents the total number of gaps between adjacent insulators on an insulator string that are bridged by vegetation. This represents the sequence number of the gap currently being calculated. In the first The total number of all independent or composite bridging units identified on each gap, For the currently being calculated, located at the th The serial number of the bridging unit on each gap, For the first The length of the net air gap for each gap, For the first The length of each bridging unit For the first The equivalent thickness of each bridging unit This serves as the baseline value for thickness. For the first The vegetation type conductivity weight of each bridging unit is determined by a predefined vegetation type. and To adjust the index, This represents the total number of all adjacent insulator pairs in a string of insulators.

[0013] In a preferred embodiment, this application can be further configured as follows: the acquisition of the insulator's operating environment data and leakage current data, and the preprocessing of these data to obtain the insulator's wet duration, effective value of leakage current, and leakage current pulse count, includes: Monitor whether the relative humidity of the environment exceeds the preset humidity threshold within the preset collection period, and take the total number of unit times within the collection period that exceed the preset humidity threshold as the humidity duration. The number of times the instantaneous value of leakage current exceeds a preset current threshold within a unit of time is obtained to obtain the leakage current pulse count.

[0014] In a preferred embodiment, this application can be further configured such that the process of constructing the risk index model for insulators under plant bridging includes: Using image features, environmental features, leakage current features, and relevant data characterizing insulation performance status of insulators under various states as training samples, a gradient boosting decision tree model (XGBoost) is used for data training to construct a risk index model for insulators under plant bridging, with the final output as the insulation performance risk index.

[0015] In a preferred embodiment, this application can be further configured such that: the composite bridging index, bio-attachment coverage, wet duration, effective value of leakage current, and leakage current pulse count are input into the insulator risk index model under plant bridging, and the insulator outputs the insulation performance risk index under plant bridging, including: The composite bridging index and the bio-attachment coverage are weighted and the sum of the weighted composite bridging index and the bio-attachment coverage is used as the visual risk factor of the insulator. The calculation expression of the visual risk factor is as follows: (3) in, Indicates visual risk factors, , They represent the composite bridging index, respectively. and the bioattachment coverage Weighting coefficients ; Duration of dampness and attenuation coefficient The product of and is used as a natural constant The negative index measures the current relative humidity. Then, the environmental risk factors are calculated. The calculation expression is as follows: (4) in, Indicates environmental risk factors, Indicates the attenuation coefficient. Indicates the duration of dampness. , This indicates the current relative humidity and the baseline humidity value. The humidity effect index is an empirical constant that reflects the nonlinearity of the effect of humidity on insulation performance.

[0016] The ratio of the effective value of the leakage current to the set reference value of the leakage current and the ratio of the leakage current pulse count to the set reference value of the pulse count are calculated respectively, and the electrical risk factor is obtained by weighting them respectively. The calculation expression of the electrical risk factor is as follows: (5) in, Indicates electrical risk factors, , These are the effective values ​​of leakage current. Compared with the set leakage current reference value ratio Leakage current pulse counting Compared with the set pulse counting reference value ratio The weighting coefficients, ; Weights are assigned to visual risk factors, environmental risk factors, and electrical risk factors respectively, and the insulation performance risk index is calculated. The calculation expression for the insulation performance risk index is as follows: (6) in, Indicates the risk index of insulation performance. , , These represent the weighting coefficients for visual risk factors, environmental risk factors, and electrical risk factors, respectively. .

[0017] In a preferred embodiment, this application can be further configured as follows: after normalizing the insulation performance risk index to obtain an insulation performance risk score, it is compared with a preset multi-level insulation risk threshold, triggering a corresponding insulation warning and outputting the warning level and plant bridging location information, including: The insulation risk index is mapped to a preset, fixed numerical range and normalized to obtain an insulation performance risk score. The insulation performance risk score is compared with a preset multi-level insulation risk threshold, and the corresponding risk level is determined. The risk registration includes normal state, low risk, medium risk, and high risk. Based on the determined warning level, the system automatically generates and outputs operation and maintenance guidance suggestions that include the corresponding level information and the bridging location.

[0018] In summary, this application includes at least one of the following beneficial technical effects: 1. This invention uses image recognition technology to identify the thickness, length and type of vegetation, calculate the biological attachment coverage rate and composite bridging index, intuitively quantify the physical severity of plant bridging, and simply and accurately locate high-risk insulators that are prone to flashover. This is of practical significance for formulating precise operation and maintenance strategies and achieving proactive defense.

[0019] 2. Based on the multi-feature fusion assessment model, the relationship between various factors such as plant bridging, environmental humidity and leakage current and insulation performance is comprehensively analyzed, and the degree of influence of these factors on the performance of insulators is quantitatively assessed, which facilitates the accurate assessment of the risk of insulator flashover.

[0020] 3. This invention adopts a multi-level early warning mechanism based on risk scoring. By outputting specific early warning levels and bridging location information, it connects the dual objectives of early identification and precise operation and maintenance. It also considers the characteristics of plant bridging and the probability of insulation performance degradation, which can effectively reflect the risk level of insulators and assess the safe operation level of transmission lines in humid and hot environments. Attached Figure Description

[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0022] Figure 1 This is a flowchart illustrating the implementation of the dynamic evaluation and early warning method for insulator insulation performance based on plant bridging identification in this embodiment. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0025] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0027] In one embodiment, such as Figure 1 As shown, this application discloses a method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification, which specifically includes the following steps: Step S1: Collect insulator surface image data containing the overall view of the insulator string and multiple perspectives, perform image semantic segmentation and vegetation identification on the insulator surface image data, and simultaneously calculate biological attachment coverage and composite bridging index.

[0028] Specifically, image data of the insulator surface is collected by at least two high-definition cameras installed on the transmission tower. At least two high-definition cameras are deployed on both sides of the insulator string to ensure that the full view and multiple perspectives of the insulator string can be captured, and the collection frequency is at least once a day.

[0029] Specifically, step S1 includes: S1.1: The U-Net image semantic segmentation model is used to perform three-dimensional analysis on the insulator surface image data, and the image pixels are divided into background, insulator porcelain bottle and plants.

[0030] Specifically, a U-Net image semantic segmentation model is constructed based on deep learning. The U-Net model is used to identify the semantics of the image and to segment the insulator surface image into multiple image regions according to the recognition results. By performing three-dimensional analysis on each image, the image pixels are identified as the background, the insulator porcelain bottle and the plants.

[0031] S1.2: Based on the image segmentation and 3D analysis results, identify and extract target features of several vegetation units that may form bridging units, and select corresponding bridging unit division and parameter calculation strategies according to each vegetation combination morphology.

[0032] Specifically, target features of corresponding vegetation units are extracted based on different vegetation combination morphologies, and corresponding bridging unit partitioning and parameter calculation strategies are selected, including: S1.2.1: For uniformly covered moss vegetation, the continuously covered moss area is regarded as a whole bridging unit. The equivalent thickness of the bridging unit is the average estimated thickness of the continuously covered moss area. The length of the bridging unit is defined as the vertical distance between the maximum insulating skirts that the moss-covered area can cover along the insulator string axis.

[0033] S1.2.2: For single-stem discrete vegetation, each independent stem and leaf is regarded as an independent bridging unit. The thickness of each bridging unit is the average diameter of the independent stem and leaf, and the length is the natural extension length of the independent stem and leaf.

[0034] S1.2.3: For vegetation combinations with multiple overlapping vegetation, they are regarded as a whole composite bridging unit. The thickness of the composite bridging unit is taken as the equivalent total thickness of the overlapping vegetation clump in the direction perpendicular to the bridging direction, and the length is taken as the total length of the overlapping vegetation clump.

[0035] S1.2.4: For other vegetation regenerating on moss, it is regarded as a layered composite bridging unit with moss and regenerated vegetation as two independent bridging units, and the bridging contribution value is accumulated when calculating the composite bridging index.

[0036] S1.3: Simultaneously calculate the bio-attachment coverage rate and composite bridging index according to the selected bridging unit division and parameter calculation strategy.

[0037] Specifically, step S1.3 includes: Based on the image segmentation results, the biofilm coverage rate is obtained by calculating the proportion of plant pixels to the total number of pixels on the surface of the insulator porcelain insulator. The expression for the biofilm coverage rate is as follows: (1) in, Indicates biofouling coverage. This represents the total number of pixels in the image that are identified as plants. This represents the total number of pixels identified as insulator porcelain bottles.

[0038] The formula for calculating the composite bridging index is as follows: (2) in, This represents the total number of gaps between adjacent insulators on an insulator string that are bridged by vegetation. This is the sequence number of the gap currently being calculated; In the first The total number of all independent or composite bridging units identified on each gap; For the currently being calculated, located at the th The serial number of the bridging unit on each gap; For the first The length of the net air gap for each gap; For the first The length of each bridging unit; For the first The equivalent thickness of each bridging unit; The base value for thickness is 1.0 mm. For the first The vegetation type conductivity weight of each bridging unit is determined by the predefined vegetation type: 0.8 for mosses and algae, 0.5 for herbaceous plants, and 1.0 for vines. and To adjust the index, among which The value range is from 1.5 to 2.5. The value range is from 0.5 to 1.0; This represents the total number of all adjacent insulator pairs in a string of insulators.

[0039] In this embodiment, the vegetation identification process in step S1 includes: S1.4: Use at least two high-definition cameras to collect synchronous images of several sides of the insulator from both sides, and use a stereo matching algorithm to calculate and generate a three-dimensional point cloud of the insulator surface.

[0040] S1.5: Based on the image semantic segmentation results, calculate the distance between the point cloud of the region where each segmented vegetation unit is located and the reference surface of the insulator porcelain bottle to obtain the thickness of the vegetation unit.

[0041] S1.6: Based on the image semantic segmentation results, calculate the projection size of the three-dimensional point cloud corresponding to each segmented vegetation unit onto the insulator string axis to obtain the length of the vegetation unit.

[0042] S1.7: Input the image patch of each vegetation region obtained by image semantic segmentation into the pre-trained vegetation classification sub-model, output the probability distribution of vegetation type in the vegetation region, and assign the corresponding basic electrical conductivity weights.

[0043] Specifically, step S1.7 also includes: Using an annotated image dataset containing various common insulator vegetation such as mosses, algae, herbaceous plants, and vines as training samples, a vegetation classification sub-model was constructed for vegetation type probability distribution analysis through data training via a convolutional neural network.

[0044] Step S2: Collect the operating environment data and leakage current data of the insulator and preprocess them respectively to obtain the wet duration, effective value of leakage current and leakage current pulse count of the insulator.

[0045] Specifically, step S2 includes: S2.1: Monitor whether the relative humidity of the environment exceeds the preset humidity threshold within the preset collection period, and take the total number of unit times that exceed the preset humidity threshold within the collection period as the humidity duration.

[0046] Specifically, the input characteristic of the operating environment data is the duration of humidity. The system monitors the number of hours in the past 24 hours when the relative humidity of the environment exceeds a first preset humidity threshold, where the first preset humidity threshold is 80% of the relative humidity of the environment.

[0047] S2.2: Obtain the number of times the instantaneous value of leakage current exceeds the preset current threshold per unit time, and obtain the leakage current pulse count.

[0048] Specifically, the input characteristics of leakage current data include: the effective value of leakage current. and leakage current pulse count ,in The number of times the instantaneous value of the leakage current exceeds the second preset current threshold within a unit of time, wherein the second preset current threshold is 10mA.

[0049] Step S3: Input the composite bridging index, bio-attachment coverage, wet duration, effective value of leakage current, and leakage current pulse count into the insulator risk index model under plant bridging, and output the insulation performance risk index under plant bridging.

[0050] Specifically, insulation performance risk index The calculation, in its physical essence, comprehensively considers visual risk factors. Environmental risk factors Electrical risk factors The impact of these three categories of risk factors. Step S3 includes: Step S3.1: Assign weights to the composite bridging index and bio-attachment coverage rate. The sum of the weighted composite bridging index and bio-attachment coverage rate is used as the visual risk factor of the insulator. The calculation expression for the visual risk factor is as follows: (3) in, Indicates visual risk factors, , They represent the composite bridging index, respectively. and bioattachment coverage Weighting coefficients In this embodiment , The weighting ratio is 7:3, and the sum of the weighting coefficients is 1, that is... =0.7, =0.3.

[0051] Step S3.2: Due to the duration of dampness The effect on insulation performance increases exponentially and tends to saturate, therefore its effect on insulation is not linear. The duration of moisture... and attenuation coefficient The product of and is used as a natural constant The negative index measures the current relative humidity. Then, the environmental risk factors are calculated. The calculation expression is as follows: (4) in, Indicates environmental risk factors, Indicates the attenuation coefficient. Indicates the duration of dampness. , This indicates the current relative humidity and the baseline humidity value. The humidity effect index is an empirical constant that reflects the nonlinearity of the effect of humidity on insulation performance.

[0052] Step S3.3: Calculate the ratio of the effective value of the leakage current to the set reference value of the leakage current, and the ratio of the leakage current pulse count to the set reference value of the pulse count, respectively, and perform weight allocation to obtain the electrical risk factor. The calculation expression of the electrical risk factor is as follows: (5) in, Indicates electrical risk factors, , These are the effective values ​​of leakage current. Compared with the set leakage current reference value ratio Leakage current pulse counting Compared with the set pulse counting reference value ratio The weighting coefficients, The weighting ratio of the two ratios is 4:6, and the sum of the weighting coefficients is 1. =0.4, =0.6. Step S3.4: Assign weights to the visual risk factor, environmental risk factor, and electrical risk factor respectively, and calculate the insulation performance risk index. The calculation expression for the insulation performance risk index is as follows: (6) in, Indicates the risk index of insulation performance. , , These represent the weighting coefficients for visual risk factors, environmental risk factors, and electrical risk factors, respectively. , , , The weighting ratio is 5:2:3, that is... =0.5, =0.2, =0.3.

[0053] The construction process of the insulator risk index model under plant bridging in this embodiment includes: using image features, environmental features, leakage current features, and relevant data used to characterize the insulation performance status of insulators under various states as training samples, training the data through the gradient boosting decision tree model XGBoost, and constructing an insulator risk index model under plant bridging with the final output as the insulation performance risk index.

[0054] Specifically, insulation performance risk index The insulation performance risk assessment model is derived using the XGBoost (Gradient Boosting Decision Tree) machine learning model. It is trained on a historical dataset, which includes a large amount of image features, environmental features, leakage current features, and other relevant data characterizing the insulation performance status of insulators under various conditions. The final output of this model is the insulation performance risk index. .

[0055] Step S4: After normalizing the insulation performance risk index, the insulation performance risk score is obtained. It is then compared with the preset multi-level insulation risk threshold, triggering the corresponding insulation warning and outputting the warning level and plant bridging location information.

[0056] Specifically, step S4 includes: Step S4.1: Insulation risk index The data is mapped to a preset, fixed numerical range and normalized to obtain an insulation performance risk score. ; Step S4.2: Assess the risk of insulation performance The risk level is compared with the preset multi-level insulation risk threshold and the corresponding risk level is determined. The risk registration includes normal state, low risk, medium risk, and high risk, as shown below: when When this occurs, it is considered a normal state; when At that time, it was determined to be a low-risk warning level; when At that time, it was determined to be at a medium-risk warning level; when At that time, it was determined to be at a high-risk warning level.

[0057] in, Indicates the risk score of insulation performance. This indicates the first-level insulation risk threshold. This indicates the second-level insulation risk threshold. This indicates the third-level insulation risk threshold.

[0058] Step S4.3: Based on the determined warning level, automatically generate and output operation and maintenance guidance suggestions containing the corresponding level information and the bridging location. The operation and maintenance guidance suggestions include: For low-risk warning levels, the operation and maintenance guidance information is an observation suggestion, reminding you to pay special attention during the next inspection. For medium-risk warning levels, the maintenance guidance information is a planned cleaning warning, which will be included in the recent maintenance plan; For high-risk warning levels, the operation and maintenance guidelines recommend handling the alarm as an emergency, immediately notifying operation and maintenance personnel to handle it on-site, and highlighting the specific bridging location in the system.

[0059] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0060] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application of the technical solution and the constraints involved. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0061] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical functional division. In actual implementation, there may be other division methods, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored.

[0062] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0063] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0064] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification, characterized in that, The method includes: Collect insulator surface image data including the overall view of the insulator string and multiple perspectives, perform image semantic segmentation and vegetation recognition on the insulator surface image data, and simultaneously calculate biological attachment coverage and composite bridging index. The operating environment data and leakage current data of the insulator are collected and preprocessed to obtain the wet duration, effective value of leakage current and leakage current pulse count of the insulator. The composite bridging index, bio-attachment coverage, wet duration, effective value of leakage current, and leakage current pulse count are input into the insulator risk index model under plant bridging, and the insulation performance risk index under plant bridging is output. After normalizing the insulation performance risk index, an insulation performance risk score is obtained. This score is then compared with a preset multi-level insulation risk threshold to trigger a corresponding insulation warning and output the warning level and plant bridging location information.

2. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 1, characterized in that, The acquisition includes insulator surface image data encompassing the entire insulator string and multiple perspectives. Image semantic segmentation and vegetation identification are performed on the insulator surface image data, along with simultaneous calculations of biological attachment coverage and composite bridging index, including: The image semantic segmentation model U-Net is used to perform three-dimensional analysis on the insulator surface image data, and the image pixels are divided into background, insulator porcelain bottle and plants. Based on the results of image segmentation and 3D analysis, target features of several vegetation units that may form bridging are identified and extracted, and corresponding bridging unit division and parameter calculation strategies are selected according to each vegetation combination morphology. The bio-attachment coverage rate and composite bridging index are calculated simultaneously according to the selected bridging unit division and parameter calculation strategy.

3. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 2, characterized in that, The acquisition of insulator surface image data, including the overall view of the insulator string and multiple perspectives, and the processing of image semantic segmentation and vegetation identification on the insulator surface image data, as well as the simultaneous calculation of biological attachment coverage and composite bridging index, includes the following vegetation identification process: Simultaneous images of several insulator sides are acquired, and a three-dimensional point cloud of the insulator surface is calculated and generated using a stereo matching algorithm. Based on the image semantic segmentation results, the distance between the point cloud of the region where each segmented vegetation unit is located and the reference surface of the insulator porcelain bottle is calculated to obtain the thickness of the vegetation unit. Based on the image semantic segmentation results, the length of the vegetation unit is obtained by calculating the projection size of the three-dimensional point cloud corresponding to each segmented vegetation unit onto the insulator string axis. Each image patch of vegetation region obtained from image semantic segmentation is input into a pre-trained vegetation classification sub-model, which outputs the probability distribution of vegetation type in the vegetation region and assigns corresponding basic electrical conductivity weights.

4. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 3, characterized in that, The step of inputting image patches of each vegetation region obtained from image semantic segmentation into a pre-trained vegetation classification sub-model, outputting the probability distribution of vegetation types in the vegetation region, and assigning corresponding basic electrical conductivity weights, further includes: Using a dataset of labeled images containing common vegetation types of insulators as training samples, a vegetation classification sub-model was constructed for vegetation type probability distribution analysis through training a convolutional neural network.

5. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 2, characterized in that, Based on image segmentation and 3D analysis results, target features of several vegetation units that may form bridging units are identified and extracted. A corresponding bridging unit division and parameter calculation strategy is selected according to each vegetation combination morphology, including: For uniformly covered moss vegetation, the continuously covered moss area is regarded as a whole bridging unit. The equivalent thickness of the bridging unit is the average estimated thickness of the continuously covered moss area. The length of the bridging unit is defined as the vertical distance between the maximum insulating skirts that the moss-covered area can cover along the insulator string axis. For single-stem discrete vegetation, each individual stem and leaf is regarded as an independent bridging unit. The thickness of each bridging unit is the average diameter of the individual stem and leaf, and the length is the natural extension length of the individual stem and leaf. For vegetation combinations with multiple overlapping vegetation, they are considered as a whole composite bridging unit. The thickness of the composite bridging unit is taken as the equivalent total thickness of the overlapping vegetation clumps perpendicular to the bridging direction, and the length is taken as the total length of the overlapping vegetation clumps. For other vegetation regenerating on moss, it is regarded as a layered composite bridging unit with moss and regenerated vegetation as two independent bridging units, and the bridging contribution value is accumulated when calculating the composite bridging index.

6. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 2, characterized in that, The simultaneous calculation of bioattachment coverage and composite bridging index according to the selected bridging unit division and parameter calculation strategy includes: Based on the image segmentation results, the biofilm coverage rate is obtained by calculating the proportion of plant pixels to the surface area of ​​the insulator porcelain insulator. The expression for the biofilm coverage rate is as follows: （1） in, Indicates biofouling coverage. This represents the total number of pixels in the image that are identified as plants. This represents the total number of pixels identified as insulator porcelain bottles; The formula for calculating the composite bridging index is as follows: （2） in, Indicates the composite bridging index. This represents the total number of gaps between adjacent insulators on an insulator string that are bridged by vegetation. This represents the sequence number of the gap currently being calculated. In the first The total number of all independent or composite bridging units identified on each gap, For the currently being calculated, located at the th The serial number of the bridging unit on each gap, For the first The length of the net air gap for each gap, For the first The length of each bridging unit For the first The equivalent thickness of each bridging unit This serves as the baseline value for thickness. For the first The vegetation type conductivity weight of each bridging unit is determined by a predefined vegetation type. and To adjust the index, This represents the total number of all adjacent insulator pairs in a string of insulators.

7. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 1, characterized in that, The process involves collecting and preprocessing the insulator's operating environment data and leakage current data to obtain the insulator's wet duration, effective leakage current value, and leakage current pulse count, including: Monitor whether the relative humidity of the environment exceeds the preset humidity threshold within the preset collection period, and take the total number of unit times within the collection period that exceed the preset humidity threshold as the humidity duration. The number of times the instantaneous value of leakage current exceeds a preset current threshold within a unit of time is obtained to obtain the leakage current pulse count.

8. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 1, characterized in that, The process of constructing the risk index model for insulators under plant bridging includes: Using image features, environmental features, leakage current features, and relevant data characterizing insulation performance status of insulators under various states as training samples, a gradient boosting decision tree model (XGBoost) is used for data training to construct a risk index model for insulators under plant bridging, with the final output as the insulation performance risk index.

9. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 1, characterized in that, The composite bridging index, biofilm coverage, wet duration, effective value of leakage current, and leakage current pulse count are input into the insulator risk index model under plant bridging, and the output insulation performance risk index under plant bridging includes: The composite bridging index and the bio-attachment coverage are weighted and the sum of the weighted composite bridging index and the bio-attachment coverage is used as the visual risk factor of the insulator. The calculation expression of the visual risk factor is as follows: （3） in, Indicates visual risk factors, , They represent the composite bridging index, respectively. and the bioattachment coverage Weighting coefficients ; Duration of dampness and attenuation coefficient The product of and is used as a natural constant The negative index measures the current relative humidity. Then, the environmental risk factors are calculated. The calculation expression is as follows: （4） in, Indicates environmental risk factors, Indicates the attenuation coefficient. Indicates the duration of dampness. , This indicates the current relative humidity and the baseline humidity value. The humidity effect index represents the non-linearity of the effect of humidity on insulation performance. The ratio of the effective value of the leakage current to the set reference value of the leakage current and the ratio of the leakage current pulse count to the set reference value of the pulse count are calculated respectively, and the electrical risk factor is obtained by weighting them respectively. The calculation expression of the electrical risk factor is as follows: （5） in, Indicates electrical risk factors, , These are the effective values ​​of leakage current. Compared with the set leakage current reference value ratio Leakage current pulse counting Compared with the set pulse counting reference value ratio The weighting coefficients, ; Weights are assigned to visual risk factors, environmental risk factors, and electrical risk factors respectively, and the insulation performance risk index is calculated. The calculation expression for the insulation performance risk index is as follows: （6） in, Indicates the risk index of insulation performance. , , These represent the weighting coefficients for visual risk factors, environmental risk factors, and electrical risk factors, respectively. .

10. The method for dynamic evaluation and early warning of insulator insulation performance based on plant bridging identification according to claim 1, characterized in that, The process of normalizing the insulation performance risk index to obtain an insulation performance risk score, comparing it with a preset multi-level insulation risk threshold, triggering a corresponding insulation warning, and outputting the warning level and plant bridging location information includes: The insulation risk index is mapped to a preset, fixed numerical range and normalized to obtain an insulation performance risk score. The insulation performance risk score is compared with a preset multi-level insulation risk threshold, and the corresponding risk level is determined. The risk registration includes normal state, low risk, medium risk, and high risk. Based on the determined warning level, the system automatically generates and outputs operation and maintenance guidance suggestions that include the corresponding level information and the bridging location.

Images