Power transmission line forest fire fault dynamic prediction analysis method and system

Abstract

The invention discloses a power transmission line forest fire fault dynamic prediction analysis method and system, and relates to the field of power transmission line forest fire prediction.The power transmission line forest fire fault dynamic prediction analysis method comprises the steps that area division is conducted according to a power transmission line, multiple sub-observation areas are obtained, tower information identification is conducted on each sub-observation area, and multiple key patrol areas are obtained; grading the plurality of key patrol areas to obtain a risk grade factor group of each key patrol area, and determining a risk coefficient based on the risk grade factor group; performing forest fire prediction on the line in each key patrol area by using the preliminary prediction model to obtain a corresponding forest fire prediction grade; the risk coefficient of each key patrol area is endowed with the forest fire prediction grade, and a corrected forest fire risk grade is obtained; the power transmission line area is graded, risk coefficients of different grades are determined, a model prediction result is given to the coefficients, and the model prediction result is further optimized, so that the accuracy of the final prediction result is improved.

Description

Technical Field

[0001] The present invention relates to the field of wildfire prediction for transmission lines, and more particularly, to a method and system for dynamic prediction and analysis of wildfire faults in transmission lines. Background Art

[0002] In the past, the prediction of wildfire faults in transmission lines mainly relied on relatively simple methods and limited data sources. In terms of early meteorological data acquisition, it mainly relied on sparsely distributed meteorological stations, and the spatial resolution of the data was low, making it difficult to accurately reflect the complex and changeable meteorological conditions along the transmission lines. For example, in mountainous areas, the distance between meteorological stations is relatively far, and it is impossible to effectively capture local microclimate differences, which have a key impact on the occurrence of wildfires. In terrain complex areas such as valleys or slopes, small changes in temperature and humidity may determine whether the vegetation is prone to catching fire.

[0003] For the monitoring of vegetation around transmission lines, the traditional method is mainly manual on-site inspection, which is not only inefficient but also difficult to achieve large-area and high-frequency monitoring. The subjective judgment of the dryness of vegetation by humans is strong and lacks accuracy. Moreover, it is difficult to comprehensively grasp information such as the types of vegetation, coverage, and spatial relationship with the transmission lines, which are extremely crucial for assessing wildfire risks.

[0004] With the development of technology, currently, the prediction of wildfire faults in transmission lines is through a big data system, which tracks and records meteorological data and plant data in real time, uses machine learning algorithms to collect a large amount of sample data related to wildfires in transmission lines, constructs a prediction model, and judges the risk level by importing the real-time tracked data into the prediction model. On the one hand, this method depends on a large amount of data to train the prediction model, and on the other hand, it also depends on the input learning method of the prediction model. If the learning method is incorrect, it will also affect the accuracy of the subsequent output results. In addition, the current predictions are all made separately using models, without the link of using historical data for actual prediction and then combining the actual prediction with the model prediction for data optimization, resulting in a relatively low accuracy of the risk prediction obtained by the prediction model.

[0005] In view of this, the present application is specifically proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide a method and system for dynamic prediction and analysis of wildfire faults in transmission lines, which divides the transmission line area into different grades to determine the risk coefficients of different grades, assigns the coefficients to the model prediction results, and further optimizes the model prediction results, thereby improving the accuracy of the final prediction results.

[0007] The embodiments of the present invention are implemented as follows: A method for dynamic prediction and analysis of wildfire faults in transmission lines includes the following steps: S100: Divide the area according to the transmission line to obtain multiple sub-observation areas, identify the tower information for each sub-observation area, and obtain multiple key inspection areas; wherein, the tower information includes: the number of towers information, the distribution information of towers, and the line status information between towers. S200: Classify the multiple key inspection areas to obtain a risk level factor group for each key inspection area, and determine the risk coefficient based on this risk level factor group. S300: Use the preliminary prediction model to predict wildfires for the lines in each key inspection area, and obtain the corresponding wildfire prediction level. S400: Assign the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level, and issue a warning for the wildfire fault information of the lines in this key inspection area based on the wildfire risk level. Wherein, when classifying the key inspection areas, the main level and the secondary level are respectively determined, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between towers, and the secondary level is determined by the number of tower information and the tower distribution information.

[0008] Further, the main level is determined by the historical short-circuit data between towers, and the secondary level is determined by the number of tower information and the tower distribution information, specifically including the following steps: Determine the main level according to the short-circuit point and the short-circuit frequency, determine the first additional number of the secondary level according to the number of tower information associated with the short circuit, and then determine the second additional number of the secondary level according to the tower distribution information in the key inspection area where the short circuit occurs. Assign the first additional number and the second additional number to the main level to obtain the preliminary level.

[0009] Further, obtaining the risk level factor group for each key inspection area includes the following steps: Determine the risk level factors in the key inspection area, screen the risk level factors based on the preliminary level, and use the screened risk level factors to construct the risk level factor group of this key inspection area. Among them, when screening the risk level factors, the following steps are followed: Trace the main risk factors according to the short-circuit point and the short-circuit frequency, and conduct a preliminary screening of the risk level factors based on the main risk factors to obtain a preliminary screening factor group. Determine the first secondary risk factor according to the number of tower information associated with the short circuit, and screen the preliminary screening factor group based on the first secondary risk factor to obtain the first screening factor group. Determine the second secondary risk factor according to the tower distribution information in the key inspection area where the short circuit occurs, and screen the first screening factor group based on the second secondary risk factor to obtain the second screening factor group. Construct a risk level factor group based on the risk level factors in the second screening factor group.

[0010] Further, determine the first secondary risk element based on the information on the number of poles and towers associated with the short circuit, including: marking the poles and towers where the short circuit occurs as risk poles and towers, obtaining the number of all risk poles and towers in the key inspection area, calculating the proportion of risk poles and towers in the key inspection area, and obtaining the first secondary risk element. Determine the second secondary risk element based on the pole and tower distribution information in the key inspection area where the short circuit occurs, including: analyzing all the risk pole and tower distribution information, obtaining the distribution shape of the risk poles and towers in the key inspection area, and obtaining the second secondary risk element according to the distribution shape of the risk poles and towers.

[0011] Further, the types of the main risk elements include meteorological elements, vegetation elements, line elements, and external factors. The types of the first secondary risk elements include A: 0 - 20%, B: 20% - 30%, C: 30% - 70%, D: 70% - 90%, E: 90% - 100%. The types of the second secondary risk elements include: linear style, circular style, arc style, fan style, or scattered point style.

[0012] Further, it also includes the step of judging whether the main risk element, the first secondary risk element, and the second secondary risk element are in a special combination situation: When the meteorological element in the main risk element of the key inspection area is strong wind, the first secondary risk element is A: 0 - 20%, and the second secondary risk element is linear style, the following analysis steps are also required: Obtain the directions of the lines on both sides of the short - circuited poles and towers, as well as the wind direction, and calculate the angle between the wind direction and the line direction. If the angle < 30°, obtain the vegetation height and the line height, and judge the difference between the line height and the vegetation height. If the difference is greater than 0 and the short - circuit frequency increases, it is determined that the short circuit is caused by the factor of bird flocks.

[0013] Further, the specific steps of assigning the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level are as follows: Obtain the wildfire prediction level of each key inspection area, then obtain the risk coefficient of this key inspection area, determine the combined weight, which includes the wildfire prediction level weight and the risk coefficient weight, and calculate the wildfire risk level by combining the result of assigning the wildfire prediction level weight to the wildfire prediction level and the result of assigning the risk coefficient weight to the risk coefficient.

[0014] Further, obtain historical actual wildfire information of transmission lines, compare the wildfire prediction level in this key inspection area with the risk coefficient of this key inspection area when wildfires occurred historically, respectively determine the accuracy of the wildfire prediction level and the risk coefficient, and determine the combined weight according to this accuracy.

[0015] Further, obtain multiple pieces of historical actual wildfire information of transmission lines, compare the wildfire prediction level in this key inspection area when wildfires occurred historically, and determine the accuracy of the wildfire prediction level; and obtain the line state information between the poles and towers in this key inspection area; Obtain the line state information between the poles and towers in the current key inspection area, compare the line state information between the poles and towers in the key area when a fire occurred historically, obtain a similarity factor, assign this similarity factor to the wildfire prediction level weight, and obtain the corrected wildfire prediction level weight.

[0016] A dynamic prediction and analysis system for wildfire faults in transmission lines includes: a regional division unit, a level division unit, a prediction unit, and a correction unit. The regional division unit is used to divide regions according to transmission lines, obtain multiple sub-observation regions, and identify pole and tower information for each sub-observation region to obtain multiple key inspection areas; wherein, the pole and tower information includes: the number of pole and tower information, the distribution information of pole and towers, and the line state information between the poles and towers; The level division unit is used to divide the levels of multiple key inspection areas, obtain a risk level factor group for each key inspection area, and determine the risk coefficient based on this risk level factor group; wherein, when dividing the levels of the key inspection areas, the main level and the secondary level are determined respectively, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between the poles and towers, and the secondary level is determined by the number of pole and tower information and the distribution information of pole and towers; The prediction unit uses a preliminary prediction model to predict wildfires on the lines in each key inspection area and obtain the corresponding wildfire prediction level; The correction unit is used to assign the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level, and issue a warning about the wildfire fault information of the lines in this key inspection area based on the wildfire risk level.

[0017] The beneficial effects of the embodiments of the present invention are: The dynamic prediction and analysis method and system for wildfire faults in transmission lines provided by the embodiments of the present invention use steps such as refined regional division, grading of key inspection areas, extraction of risk factor groups, determination of risk coefficients, preliminary prediction of wildfire risk levels, and correction and optimization of wildfire risk levels to achieve high-precision prediction of wildfire fault information in transmission lines. In particular, by determining key inspection areas based on tower information, grading the key inspection areas, further improving the screening conditions for prediction, and considering the impact of risk factors on wildfire faults in transmission lines, by obtaining the risk level factor group of each key inspection area, finally determining the risk coefficient, and assigning the risk coefficient to the wildfire prediction level preliminarily predicted by the preliminary prediction model, the corrected wildfire risk level is obtained. The key lies in taking into account the risk factors causing short circuits in transmission lines, and through grading and determination of risk coefficients, further correcting and optimizing the wildfire prediction level, significantly improving the reliability and accuracy of wildfire fault prediction in transmission lines, especially in areas with different tower information, thereby providing more accurate prediction services for the safe operation of transmission lines.

[0018] Generally speaking, the dynamic prediction and analysis method and system for wildfire faults in transmission lines provided by the embodiments of the present invention divide the transmission line area into grades, determine risk coefficients for different grades, assign the coefficients to the model prediction results, and further optimize the model prediction results, thereby improving the accuracy of the final prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can also be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a flowchart of the main steps of the analysis method provided by the embodiments of the present invention; Figure 2 For Figure 1 It is a flowchart of one of the steps S200 of the analysis method shown; Figure 3 For Figure 2 It is a flowchart of one of the steps S270 of the step 200 shown; Figure 4 For Figure 2 It is a flowchart of one of the steps S280 of the step 200 shown; Figure 5 For Figure 1 It is a flowchart of one of the steps S400 of the analysis method shown; Figure 6 For Figure 5 the flowchart of one of the steps S420 in the shown step 400; Figure 7 For Figure 6 the flowchart of one of the steps S423 in the shown step 420; Figure 8 is the modular schematic diagram of the analysis system provided by the embodiment of the present invention.

[0021] Icons: 500 - analysis system, 510 - area division unit, 520 - level division unit, 530 - prediction unit, 540 - correction unit. Detailed implementation manners

[0022] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some but not all of the embodiments of the present invention. Generally, the components of the embodiments of the present invention described and illustrated in the accompanying drawings here can be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed present invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0024] It should be understood that the "system", "device" and / or "module" used in the present invention is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the said words can be replaced by other expressions.

[0025] As shown in the present invention and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include plural. Generally, the terms "comprising" and "including" only indicate the inclusion of the clearly identified steps and elements, and these steps and elements do not constitute an exclusive list, and the method or device may also include other steps or elements.

[0026] Flowcharts are used in the present invention to illustrate the operations performed by the systems according to the embodiments of the present application. It should be understood that the operations before or after do not necessarily need to be executed precisely in sequence. On the contrary, they can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0027] Embodiment: Currently, the main factor causing wildfires due to transmission lines is short circuits, including short circuits caused by extreme weather, short circuits caused by vegetation contacting the lines, short circuits caused by human factors, short circuits caused by overload, etc. In short, short circuit factors are the key factors leading to wildfires on transmission lines; At present, the prediction of wildfire faults on transmission lines is carried out through a big data system, which tracks and records meteorological data and plant data in real time. Using machine learning algorithms, a large amount of sample data related to wildfires on transmission lines is collected to build a prediction model. By importing the real-time tracked data into the prediction model, the risk level is judged. On the one hand, this method depends on a large amount of data to train the prediction model. When there is no large amount of existing data, the prediction accuracy will be greatly affected. On the other hand, it also depends on the input learning method of the prediction model. If the learning method is incorrect, it will also affect the accuracy of the subsequent output results; In addition, in actual exploration, it is found that the occurrence of wildfires is affected by the complex interaction of multiple factors such as meteorology, vegetation, and terrain. Although the model can consider multiple variables, for some extremely complex non-linear interaction relationships, it may not be able to fully and accurately capture and model them, resulting in prediction deviations. Relying solely on meteorological data and plant data for prediction has a too narrow scope, ignoring the situations where other factors cause wildfire faults on transmission lines, thus further reducing the prediction accuracy.

[0028] In addition, prediction models are usually constructed based on certain assumptions and simplifications, while the actual wildfire occurrence process is very complex, with many uncertainties and random factors. For example, the model may assume that meteorological conditions are uniformly distributed in a certain area, but in reality, the local meteorological changes in mountainous areas may be very drastic, which does not conform to the model assumptions, and this will also affect the accuracy of the prediction model.

[0029] In addition to the method of using a prediction model for prediction, in the actual work process, on-site surveys are also used. By statistically analyzing the data collected through on-site surveys, the risk level of wildfires on transmission lines is obtained. This method has a higher accuracy than model prediction, but its limitations are relatively large, especially for real-time prediction situations, which require a large amount of manpower and material resources. Through comparative analysis, we found that the method of combining a prediction model with actual surveys can not only improve the prediction accuracy but also does not require a large amount of manpower and material resources.

[0030] Based on the above reasons and ideas, please refer to Figure 1 , this embodiment provides a dynamic prediction analysis method for wildfire faults on transmission lines, including the following steps: S100: Divide the area according to the transmission line to obtain multiple sub-observation areas, and identify the pole information for each sub-observation area to obtain multiple key inspection areas; among them, the pole information includes: the number of poles information, the pole distribution information, and the line status information between poles. For example, one kilometer to the left and right of a predetermined transmission line is included in the scope of wildfire detection, so as to obtain multiple sub-observation areas along the transmission line.

[0031] This step means that the transmission line within the prediction range is recorded as the line to be monitored, and the area width is delimited according to one kilometer to the left and right of the transmission line to obtain multiple sub-observation areas, so as to determine the range of the transmission line that needs to be detected currently. In addition, according to the pole information of the transmission line, the sub-observation areas are further divided, that is, through the number of poles information, the pole distribution information, and the line status information between poles, key inspection areas are divided within multiple sub-observation areas, so as to narrow the key detection range, facilitate in-depth monitoring and analysis of wildfire faults caused by transmission lines, and provide a basic guarantee for improving the prediction accuracy; among them, the number of poles information mainly represents the specific number of poles in this key inspection area, and the pole distribution information mainly represents the distribution pattern formed by all poles in this key inspection area, such as: linear distribution, concentrated distribution, and scattered distribution; the line status information between poles mainly represents the transmitted current of the line, the line type, the wiring method, and the line direction, etc. In other embodiments, the pole model, the pole wiring method, etc. can also be considered.

[0032] S200: Classify multiple key inspection areas to obtain a risk level factor group for each key inspection area, and determine the risk coefficient based on this risk level factor group; this step means that first, a risk level representation is made for the already divided key inspection areas, that is, classification is carried out, and each corresponding level divided covers different risk situations, for example, it is reflected by the risk level factors it covers, so that for the key inspection areas divided into corresponding levels, multiple risk level factors it covers can be obtained to form a risk level factor group, which is used as the possible risk points for causing transmission line faults in this key inspection area.

[0033] Considering that each key inspection area can be divided into different levels (different risk situations), it is necessary to clarify the different risk situations or degrees covered by different levels during its division. Therefore, different situation data representations are required for different levels of division. On the one hand, it is convenient to grasp the different degrees of risk situations, and on the other hand, it is convenient for subsequent traceability analysis. Specifically, when classifying the key inspection areas, the main level and the secondary level are determined respectively, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between poles, and the secondary level is determined by the number of poles information and the pole distribution information.

[0034] This step indicates that different (risk) levels are represented by a combination of main and sub-levels, which means that the main level determines the main risk level representation, and the sub-level determines the auxiliary representation of the key risk level. For example, for the main level, the towers in the key inspection area are marked with levels using existing information such as historical short-circuit data and towers, so as to obtain the frequency values of repeated short-circuit failures that cause transmission line fires in each key inspection area, and determine different main levels through the frequency values of transmission line short-circuits; for example, for the sub-level, the factors that indirectly affect transmission line fires can be obtained through the proportion of short-circuited towers and the distribution of short-circuited towers, thereby determining different sub-levels.

[0035] The risk factors that cause short circuits are obtained through different levels. The risk factors corresponding to the levels of different key inspection areas are determined by screening the risk factors, and finally the risk coefficient of the key inspection area is obtained. The accuracy of the prediction is improved from another perspective through historical data.

[0036] S300: using the preliminary prediction model to predict wildfires on the lines in each key inspection area, and obtaining corresponding wildfire prediction levels; The preliminary prediction model refers to the existing transmission line wildfire prediction model, which predicts transmission line wildfires through real-time analysis of meteorology and vegetation; this step means that the existing model is used to make a preliminary prediction of the current transmission line wildfire failure, so as to obtain the wildfire prediction level of each key inspection area, and use the wildfire prediction level as the basis for the final prediction data. In some embodiments, the preliminary prediction model can be implemented through the following steps: S1 Preliminary Data Collection: S11: Establish data connection with professional meteorological departments to obtain real-time and historical meteorological information such as temperature, relative humidity, wind speed, precipitation, air pressure and lightning activities along the transmission lines at a stable frequency of every hour.

[0037] S12: Use high-resolution satellite remote sensing technology to monitor vegetation conditions around transmission lines on a quarterly basis, calibrate data accuracy through field surveys, and identify vegetation types in detail.

[0038] S13: The load current, voltage, conductor temperature and other operating parameters are collected every 15 minutes through sensors installed on the line.

[0039] S2 data preprocessing: S21: Comprehensively check all types of collected data for missing values and outliers. For missing values, the mean filling method is used, that is, the gaps are filled according to the average value of the historical data of the feature to ensure data integrity; for outliers, the 3 times standard deviation method is used to identify and correct them, and eliminate erroneous or extreme data; S22: Normalize the meteorological data, convert data such as temperature and humidity to the range of 0 to 1 for easy model calculation and comparison; perform one-hot encoding on the vegetation types and convert them into binary vector forms to highlight the differences in the impacts of different vegetation types on wildfire risks. At the same time, calculate the vegetation dryness index based on the temperature and humidity data to intuitively reflect the flammability of the vegetation.

[0040] S3 Data calculation: After the model training and optimization are completed, continuously input the meteorological, vegetation, and line operation data collected in real time. The model quickly calculates the wildfire occurrence probabilities along the transmission lines for time periods such as the next 24 hours and 48 hours. Once the probability exceeds the 30% threshold, immediately trigger an alarm and send an alarm message to the operation and maintenance personnel, accompanied by a detailed risk analysis indicating the inducing factors.

[0041] In this embodiment, the wildfire prediction levels are displayed in the form of percentages, i.e., 0% - 100%. The higher the percentage, the higher the probability of the transmission line catching fire; classify the data according to the fire occurrence probabilities calculated by the preliminary prediction model and the wildfire prediction level classifications to obtain the wildfire prediction levels.

[0042] S400: Assign the risk coefficients of each key inspection area to the wildfire prediction levels to obtain the corrected wildfire risk levels, and issue early warnings for the wildfire fault information of the lines within the key inspection areas based on the wildfire risk levels.

[0043] This step means that by analyzing the risk factors, assign the determined risk coefficients to the wildfire prediction levels predicted by the preliminary model of the key inspection area, and then further correct and optimize the wildfire prediction levels to determine the final wildfire risk levels.

[0044] This embodiment combines the model prediction with the actual risk detection results. It not only utilizes the advantages of the prediction model, such as rapid response, wide range, real-time response, and resource conservation, but also combines the advantages of high accuracy in actual prediction. By effectively combining the two, it not only ensures the overall rapid response of the prediction but also improves the accuracy of the prediction.

[0045] Through the above technical solutions, through steps such as refined area division, classification of key inspection areas, extraction of risk factor groups, determination of risk coefficients, preliminary prediction of wildfire risk levels, and correction and optimization of wildfire risk levels, high-precision prediction of wildfire fault information for transmission lines is achieved. In particular, by determining key inspection areas based on tower information, classifying the key inspection areas, the screening conditions for prediction are further improved. Considering the impact of risk factors on wildfires in transmission lines, by obtaining the risk level factor group for each key inspection area, the risk coefficient is finally determined, and the risk coefficient is assigned to the wildfire prediction level preliminarily predicted by the preliminary prediction model to obtain the corrected wildfire risk level. The key lies in taking into account the risk factors causing short circuits in transmission lines. Through classification and determination of risk coefficients, the wildfire prediction level is further corrected and optimized, significantly improving the reliability and accuracy of wildfire fault prediction for transmission lines, especially in areas with different tower information, thereby providing more accurate prediction services for the safe operation of transmission lines.

[0046] Considering that the main inducing factor for wildfires in transmission lines is short circuits in the transmission lines, when determining the main level, the main level is determined by historical short circuit data between towers. In addition, when determining the secondary level, considering the auxiliary representation of the key risk degree, some other factors are also reasons for wildfires in transmission lines. For example, when the number of towers with short circuits in a unit area is relatively large, the probability of wildfires in transmission lines occurring in that area will increase accordingly; another example is that when the distance between towers with short circuits is too close and dense, the probability of wildfires in transmission lines occurring in that area will also increase accordingly.

[0047] For details, please refer to Figure 2 , in this embodiment, the secondary level is determined by the tower number information and the tower distribution information, and specifically includes the following steps: S210: Determine the main level based on the short circuit point and the short circuit frequency; S220: Determine the first additional number of the secondary level based on the tower number information related to the short circuit; S230: Then determine the second additional number of the secondary level based on the tower distribution information in the key inspection area where the short circuit occurs; S240: Assign the first additional number and the second additional number to the main level to obtain the preliminary level.

[0048] Through the above steps, the main level can be determined based on the short - circuit point and the short - circuit frequency, the first additional number of the secondary level can be determined based on the number information of the poles and towers associated with the short - circuit, and then the second additional number of the secondary level can be determined based on the pole and tower distribution information in the key inspection area where the short - circuit occurs. Thus, the key factors and secondary factors can be effectively combined, with the key factors taking the main position and the secondary factors following, ensuring the accuracy of the area division. And when dividing the secondary level, the influence of multiple secondary factors is considered. Therefore, by setting additional numbers, multiple secondary factors are separately counted, ensuring the breadth and accuracy of the division data.

[0049] Specifically, in some embodiments, the calculation of the preliminary level can be achieved in the following way: Preliminary level = Main level * 60%+First additional number of secondary level * 30%+Second additional number of secondary level * 10%. When determining the main level, it is obtained according to the number of short - circuit points and the maximum value of the short - circuit frequency. A: Maximum value ≤ 3; B: 4 < Maximum value ≤ 9; C: 10 < Maximum value ≤ 15; D: Maximum value ≥ 16. When calculating the preliminary level, each classification in the main level accounts for 25%. For example, if the number of short - circuit points in a key inspection area is 8 and the short - circuit frequency is 12, the main level is C. When calculating the preliminary level: Preliminary level = 75% * 60%+First additional number of secondary level * 30%+Second additional number of secondary level * 10%, that is, Preliminary level = 45%+First additional number of secondary level * 30%+Second additional number of secondary level * 10%.

[0050] At this time, when determining the first additional number of the secondary level, a: Number of associated poles and towers ≤ 3; b: 4 < Number of associated poles and towers ≤ 10; c: 11 < Number of associated poles and towers ≤ 20; d: Number of associated poles and towers ≥ 21. Each classification accounts for 25% of the proportion of the first additional number. For example, if the number of poles and towers associated with the short - circuit in a key inspection area is 9, the first additional number level is b and its proportion is 50%. That is, Preliminary level = 45%+50% * 30%+Second additional number of secondary level * 10%.

[0051] Then, when determining the second additional number of the secondary level, it is determined according to the pole and tower distribution information in the key inspection area where the short - circuit occurs. This distribution information can be directly determined based on the relative positions of each pole and tower in the key inspection area. For example, it can be divided into linear style, circular style, arc style, fan - shaped style or scattered - point style. When dividing the weight of each style, it can be adjusted according to the actual situation. For example: Linear style: 20%; Circular style: 40%; Arc style: 60%; Fan - shaped style: 80%; Scattered - point style: 90%. When the distribution style of the poles and towers in a key inspection area is arc style, at this time Preliminary level = 45%+15%+60% * 10%; The final preliminary level = 66%.

[0052] It should be noted that in the actual operation process, other calculation methods can be adopted according to different scenarios, with the purpose of distinguishing different acquisition methods of risk levels. Of course, the weights of the main level, secondary level, first additional number of the secondary level, and second additional number of the secondary level can also be adjusted adaptively according to the actual situation to adapt to the changes in different regions and further improve the data accuracy.

[0053] Another thing to note is that in some embodiments, the overhead line wildfire fault can be directly predicted based on this preliminary level, using this preliminary level to represent the final prediction level. Similarly, in some other embodiments, the wildfire prediction level predicted by the preliminary prediction model can also be directly used to represent the final prediction level.

[0054] In some other embodiments, in order to further improve the accuracy and precision of the prediction data, the method of combining model prediction with the preliminary level can also be adopted. By assigning different weights to the model prediction and the preliminary level, and then combining the data of the two, the accuracy of the data can be improved. Taking the above example data as an example, assuming that a weight of 60% is assigned to the model prediction and a weight of 40% is assigned to the preliminary level. When the wildfire prediction level of the model prediction is 70%, at this time, the final wildfire level = 70% * 60% + 66% * 40%; that is, the final wildfire level = 42% + 26.4% = 68.4%. In some embodiments, this final wildfire level can also be used as the final prediction level. It should be noted that when using this method, the weights assigned to the prediction model and the preliminary level can be adjusted adaptively according to the actual situation. In addition, in the actual operation process, other calculation methods can be adopted according to different scenarios, with the purpose of distinguishing different acquisition methods of risk levels.

[0055] In this embodiment, in order to further improve the accuracy of the prediction data, please refer to Figure 2 , by obtaining the risk level factor group of each key inspection area, determining the risk coefficient based on this risk level factor group, assigning the risk coefficient of each key inspection area to the wildfire prediction level, obtaining the corrected wildfire risk level, and predicting the overhead line wildfire fault based on this wildfire risk level. In this process, first, the risk level factor group of each key inspection area needs to be obtained, which specifically includes the following steps: S250: Determine the risk level factors in the key inspection area, screen the risk level factors based on the preliminary level, and use the screened risk level factors to construct the risk level factor group of this key inspection area. Among them, when screening the risk level factors, the following steps are taken: S260: Trace the main risk factors based on the short - circuit point and short - circuit frequency, and conduct a preliminary screening of the risk level factors based on this main risk factor to obtain the preliminary screening factor group; S270: Determine the first secondary risk factor based on the information about the number of poles and towers associated with the short circuit. Screen the initial screening factor group based on this first secondary risk factor to obtain the first screening factor group; S280: Determine the second secondary risk factor based on the pole and tower distribution information within the key inspection area where the short circuit occurred. Screen the first screening factor group based on this second secondary risk factor to obtain the second screening factor group; S290: Construct a risk level factor group based on the risk level factors in the second screening factor group.

[0056] This step means that the risk level factors within the key inspection area are screened in a refined manner using different conditions. Through the screening of the main risk factor, the first secondary risk factor, and the second secondary risk factor, the risk level factor group is finally determined, thus realizing the preliminary screening of the data. The purpose of this design is to be able to quickly screen the data, and through hierarchical screening, the way of screening layer by layer can screen the data more deeply, thereby further ensuring the accuracy of the final data, and further providing refined support for improving the accuracy of the prediction level.

[0057] Among them, when tracing the main risk factor based on the short circuit point, it is mainly based on the specific coordinate position of the short circuit point. On the one hand, it depends on what kind of area its geographical location is. For example, near a city or a densely populated area: If the short circuit point is near the edge of the city or a densely populated area, once a wildfire occurs, it may quickly spread to the urban area, threatening the lives and property of residents, and the risk level should be judged as high. For example, the short circuit point is in the suburban area, with a large number of residential houses and commercial areas around. Once a wildfire is triggered by a short circuit, the fire is likely to quickly expand with the help of surrounding buildings and vegetation.

[0058] For example, if the short circuit point is located in a remote mountainous area with inconvenient transportation, it is difficult for fire fighting and rescue forces to arrive quickly, and the wildfire may spread wantonly, burning large areas of mountain forests, and the risk level is also relatively high. For example, the short circuit point is deep in the mountains and old forests, far from the nearest town, with rugged roads, and it is difficult for fire trucks to pass quickly, making it difficult to fight the wildfire in the initial stage.

[0059] In addition, if the short circuit point is located in an open plain area with relatively few surrounding vegetation and is conducive to the development of fire fighting operations, the possibility and speed of wildfire spread are relatively low, and the risk level can be judged as medium or low.

[0060] By judging the position of the short circuit point, it is possible to further analyze the influence range of the wildfire, the development trend of the wildfire, and the difficulty of fighting the fire. In addition, based on the short circuit point, it is also possible to further analyze what causes different short circuit points and what the main risk factor is, so as to make it as the data basis for statistics and provide data basis guarantee for the subsequent analysis of the main risk factor of the short circuit.

[0061] In the main risk analysis based on the short - circuit frequency, it mainly depends on the number of short - circuits in a certain key inspection area; by tracing the number of short - circuits, the main risk factors causing the short - circuits are identified. For example, short - circuits caused by the condition of the transmission line itself (line aging, line defects), short - circuits caused by the surrounding environment (vegetation, meteorology), and short - circuits caused by electrical equipment failures (insulator failures, transformer failures). By combining the short - circuit points and the short - circuit frequency, the main risk factors causing the short - circuits can be effectively analyzed. By collecting historical short - circuit data of the transmission line per unit area, including information such as the time and location of the short - circuit, the type of short - circuit (such as phase - to - phase short - circuit, ground - to - ground short - circuit, etc.), the environmental conditions at that time (meteorology, vegetation, etc.), and the status of line equipment, a detailed database is established. By analyzing the data in the database, the pattern of short - circuit occurrence can be found. For example, by counting the short - circuit frequencies in different seasons and different weather conditions, the high - incidence periods and conditions of short - circuits can be identified, providing data support for tracing the main risk factors.

[0062] Specifically, the main risk factors are mainly divided into: meteorological factors, vegetation factors, line factors, and external factors.

[0063] By collecting data on short - circuit points and short - circuit frequencies, a main risk database is established. By analyzing the main risk database, the relationship between the corresponding short - circuit points, short - circuit frequencies, and short - circuit risk levels is obtained, thus forming a main risk data correspondence table. In the main risk data correspondence table, the short - circuit points, short - circuit frequencies, and main risk factors are corresponding. Different short - circuit points and / or short - circuit frequencies correspond to different main risk factors, and thus different main risk coefficients are also corresponding; furthermore, different main risk factors can correspond to different main risk coefficients.

[0064] It should be noted that the main risk coefficient represents the probability value of a short - circuit occurring. When the main risk coefficient is higher, the probability of a short - circuit occurring is higher; conversely, when the main risk coefficient is lower, the probability of a short - circuit occurring is lower.

[0065] Another thing to note is that in addition to the above - mentioned method of corresponding the relationship between short - circuit points, short - circuit frequencies, main risk factors, and main risk coefficients using a database and a main risk data correspondence table, other conventional technical means such as big - data models can also be used, which will not be elaborated here; the purpose is only to find the relationship between short - circuit points, short - circuit frequencies, main risk factors, and main risk coefficients through massive data, so as to correspond the data, and thus in the subsequent work process, the main risk coefficient can be inversely deduced only through the main risk factors.

[0066] Please refer to Figure 3 and Figure 4In this embodiment, when determining the first secondary risk factor based on the number of towers associated with the short circuit, it includes: S271: marking the tower where the short circuit occurs as a risk tower, S272: obtaining the number of all risk towers in the key inspection area, S273: calculating the proportion of risk towers in the key inspection area to obtain the first secondary risk factor; determining the second secondary risk factor based on the tower distribution information in the key inspection area where the short circuit occurs includes: S281: analyzing all risk tower distribution information, S282: obtaining the distribution shape of risk towers in the key inspection area, S283: obtaining the second secondary risk factor according to the distribution shape of risk towers.

[0067] For example, if the short circuit is only related to a few towers, it may be caused by problems in the local towers themselves. For example, damage to the insulators on the towers is a common cause. Insulators are exposed to the outdoors for a long time and are affected by environmental factors, which may cause cracks, dirt, etc., resulting in reduced insulation performance and causing short circuits. In addition, loose electrical connections on the towers may also cause short circuits. When current passes through loose connection points, arcs are generated, which in turn cause short circuits.

[0068] When a short circuit is associated with many towers, it may be due to an overall line problem or external factors. Line aging is an important factor. As the service life increases, the insulation layer of the conductor ages and wears, and the insulation performance of the line section between multiple towers may decrease, causing a short circuit. In addition, severe weather such as lightning strikes and strong winds may affect multiple towers. Lightning strikes may directly hit the line, causing insulation breakdown; strong winds may cause the conductor to swing too much, resulting in discharge short circuits between conductors or between conductors and towers.

[0069] In addition, short-circuit-related poles and towers may also be affected by environmental factors (vegetation factors, terrain factors); for example, if the vegetation around the related poles and towers is dense, especially if tall trees are close to the lines, short circuits may be caused by tree fall, branches touching the wires, etc. For short-circuit-related poles and towers, it is necessary to investigate the surrounding vegetation to see if there is a safe distance between the trees and the lines. In addition, the seasonal growth of vegetation may also affect the occurrence of short circuits. For example, during the peak season for tree growth, the distance between vegetation and the lines may be shortened, increasing the risk of short circuits.

[0070] In addition, the topography of the tower may also affect the short circuit. For example, the foundation of the tower in a low-lying and humid area is susceptible to corrosion, which affects the stability of the tower and may cause the tower to tilt, causing the distance between the conductor and the tower or other objects to change and cause a short circuit. If the tower is located in a mountainous area, it is also necessary to consider the impact of geological disasters such as falling rocks and landslides on the line, which may damage the line and cause a short circuit.

[0071] By collecting data related to short circuits, including information such as the time of short circuit occurrence, the number of associated poles and towers, the locations of poles and towers, the surrounding environment, and the type of short circuit, a first secondary risk database is established. Through long-term accumulation and analysis of these data, the above information is combined with the number of associated poles and towers, so as to correspond different risk levels of short circuits through the information of associated poles and towers. For example, by counting the changes in the number of poles and towers associated with short circuits in different seasons and different environments, the conditions and areas with high incidences of short circuits are found, and different short circuit risk levels corresponding to different numbers of poles and towers associated with short circuits are deduced.

[0072] Specifically, the first secondary risk factors include A: 0 - 20%, B: 20% - 30%, C: 30% - 70%, D: 70% - 90%, E: 90% - 100%; In the specific implementation process, based on the data in the first secondary risk database, a first secondary risk data correspondence table can be established to correspond the proportion of poles and towers associated with short circuits with the first secondary risk coefficients, so as to obtain different first secondary risk coefficients through different numbers of poles and towers associated with short circuits.

[0073] It should be noted that the first secondary risk coefficient represents the probability value of a short circuit occurring. When the first secondary risk coefficient is higher, the probability of a short circuit occurring is higher. Conversely, when the first secondary risk coefficient is lower, the probability of a short circuit occurring is lower.

[0074] Another thing to note is that in addition to the above method of corresponding the number of poles and towers associated with short circuits with the first secondary risk factors and the first secondary risk coefficients using the first secondary database and the first secondary risk data correspondence table, other conventional technical means such as big data models can also be used, which will not be elaborated here; the purpose is only to find the relationship between the number of poles and towers associated with short circuits, the first secondary risk factors, and the first secondary risk coefficients through massive data, so as to correspond the data, and thus only through the first secondary risk factors, the first secondary risk coefficients can be deduced reversely in the subsequent work process.

[0075] In addition, when the second secondary risk factor is obtained through the distribution pattern of risk towers, the distribution pattern of risk towers can be divided into: linear distribution, concentrated distribution and dispersed distribution; for example, when the distribution pattern of risk towers is linear distribution, the factors that cause short circuits may be line faults or lightning strikes. For example, the wires along the line are damaged by external forces, such as being hit by flying stones, scratched by construction equipment, etc., resulting in damage to the insulation layer and causing short circuits. In addition, the wires may be overheated and aged when they run under high load for a long time, resulting in a decrease in insulation performance. In this case, short circuits may occur continuously within a certain length of line section, making the associated towers linearly distributed. In addition, lightning may propagate along the line and hit the linearly distributed towers and the wires connected to them. Especially in areas with many thunderstorms, if the lightning protection measures of the line are not perfect, such as improper installation of lightning conductors or failure of lightning arresters, lightning strikes may generate overvoltages on the lines between multiple towers, break through insulation, and cause short circuits.

[0076] For example, when the concentration pattern of risky poles is concentrated distribution, when the poles associated with short circuits are concentrated in a certain area, it may be caused by the special environment of the area. For example, the vegetation in the area is too dense, and the growing trees may come into contact with the wires, causing short circuits. In addition, if there are pollution sources such as factories and mines in the area, the insulation performance of the insulator surface may be reduced due to the accumulation of dirt, causing short circuits. In addition, concentrated poles may have common design, construction or maintenance problems. For example, during the construction of poles, improper foundation treatment causes some poles to tilt, causing the distance between the wires and the poles or other objects to change, causing short circuits. In addition, if the electrical equipment on the poles (such as insulators, hardware, etc.) is purchased from the same batch and has quality defects, it may also cause short circuit problems on multiple poles at the same time.

[0077] For example, when the concentration pattern of risky poles and towers is dispersed, the short circuit of dispersed poles and towers may be caused by some random factors. For example, individual poles and towers are affected by bird nesting, and bird nests may contain conductive materials such as metal wires. When the weather changes (such as strong winds), these conductive materials may overlap the wires and cause short circuits. In addition, some unforeseen external force damage, such as kite entanglement and balloon touch, may also cause individual poles and towers to short circuit. In addition, the short-circuit-related poles and towers seem to be dispersed, which may be due to errors in data collection or transmission, resulting in inaccurate tower location information. For example, sensor failure, communication line interference, etc. may cause the short-circuit location information to be wrong, resulting in misjudgment of the tower distribution. In addition, the malfunction of the relay protection device in the power system may also cause some normal poles and towers to be misjudged as short-circuit-related poles and towers.

[0078] In the specific implementation process, a second secondary risk database can be established based on the above data. According to the data in the second secondary risk database, a corresponding table for the second secondary risk data is established to correspond the distribution pattern of risk poles and towers with the second secondary risk coefficient, so as to obtain different second secondary risk coefficients through different distribution patterns of risk poles and towers.

[0079] Specifically, the second secondary risk factors are divided into: linear pattern, circular pattern, arc pattern, fan pattern or scattered point pattern.

[0080] In the specific implementation process, it should be noted that the second secondary risk coefficient represents the probability value of a short circuit. When the second secondary risk coefficient is higher, the probability of a short circuit is higher. Conversely, when the second secondary risk coefficient is lower, the probability of a short circuit is lower.

[0081] Another thing to note is that in addition to the above method of corresponding the distribution pattern of risk poles and towers with the second secondary risk factors and the second secondary risk coefficient by using the second secondary database and the corresponding table for the second secondary risk data, other conventional technical means such as big data models can also be used, which will not be elaborated here; the purpose is only to find the relationship between the distribution pattern of risk poles and towers, the second secondary risk factors and the second secondary risk coefficient through massive data, so as to correspond the data, so that in the subsequent work process, the second secondary risk coefficient can be inversely deduced only through the second secondary risk factors.

[0082] In addition to the above-described situations, there are also some situations that are not easy to analyze and judge. For example: when the meteorological element in the key inspection area is strong wind, the first secondary risk factor is A: 0-20%, and the second secondary risk factor is a linear pattern. In this case, since the first secondary risk factor is at a low level, the strong wind is usually used as the main judgment basis in the conventional judgment. In addition, since the second secondary risk factor is a linear pattern, the probability of the lines affecting each other due to strong wind is relatively low. Therefore, in the conventional judgment, this situation will only determine the final prediction level according to the level of the strong wind, that is, the stronger the wind force, the higher the predicted level; however, the relationship between the wind direction and the line direction, as well as the influence of special situations such as bird flocks, are ignored in this process. As is common knowledge, the included angle between the wind direction and the line direction is 0-90°. On the premise of the same wind force, when the included angle is larger, the influence of the wind on the line is greater, that is, the swing amplitude of the line caused by the strong wind is larger. When the included angle is smaller, the influence of the wind on the line will also decrease, that is, the swing amplitude of the line caused by the strong wind is smaller; therefore, in this special situation, the following steps are also required: Obtain the directions of the lines on both sides of the short-circuit pole and tower, as well as the wind direction, and calculate the included angle between the wind direction and the line direction. If the included angle < 30°, obtain the vegetation height and the line height, and judge the difference between the line height and the vegetation height. If the difference is greater than 0 and the short - circuit frequency increases, it is determined that the short - circuit is caused by the bird flock factor.

[0083] The purpose of this design is not to simply rely on the main risk factor, which is strong wind, to determine that the short - circuit must be caused by strong wind. Instead, a further overall judgment is made (i.e., the wind direction and the line orientation), so as to further determine the root cause of the short - circuit. Through the determination of this special situation, the prediction accuracy and the determination depth of the prediction in special situations can be further improved.

[0084] Specifically, when correcting the wildfire prediction level predicted by the preliminary prediction model, it is necessary to first determine the risk coefficient and assign the risk coefficient to the wildfire prediction level. When determining the risk coefficient, the risk coefficient mainly includes the main risk coefficient, the first secondary risk coefficient, and the second secondary risk coefficient. Among them, the main risk coefficient accounts for 60%, the first secondary risk coefficient accounts for 30%, and the second secondary risk coefficient accounts for 10%.

[0085] When calculating the risk coefficient, the risk coefficient = main risk coefficient * 60% + first secondary risk coefficient * 30% + second secondary risk coefficient * 10%.

[0086] It should be noted that the proportions of the main risk coefficient, the first secondary risk coefficient, and the second secondary risk coefficient in the risk coefficient are not fixed values, but can be adjusted adaptively according to the actual situation. The specific proportions are adjusted according to the actual situation.

[0087] Please refer to Figure 5 , in this embodiment, the specific steps of assigning the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level are as follows: S410: Obtain the wildfire prediction level of each key inspection area, and then obtain the risk coefficient of this key inspection area. S420: Determine the combined weight. S430: The combined weight includes the wildfire prediction level weight and the risk coefficient weight. The result of assigning the wildfire prediction level weight to the two values of the wildfire prediction level is combined with the result of assigning the risk coefficient weight to the risk coefficient, and then the wildfire prediction level is calculated by combining with the weights.

[0088] For example, when the meteorological data in a key inspection area is strong wind and thunderstorm weather, the proportion of risk towers is 75%, and the tower distribution pattern is a fan pattern. Assuming that the main risk factor is brought into the main risk correspondence table to obtain a main risk coefficient of 70%, the first secondary risk factor is brought into the first secondary risk correspondence table to obtain a first secondary risk coefficient of 80%, and the second secondary risk factor is brought into the second secondary risk correspondence table to obtain a second secondary risk coefficient of 70%. At this time, the risk coefficient = 70% * 60% + 80% * 30% + 70 * 10% = 73%. If the wildfire prediction level predicted by the preliminary prediction model is 80% at this time, and the risk coefficient is assigned to the wildfire prediction level, if the combined weights each account for 50%, that is, the wildfire risk level = 73% * 50% + 80% * 50% = 76.5%, thus the wildfire prediction level is further corrected.

[0089] It should be noted that the above example is only one way of this embodiment. There are various calculation methods for assigning the risk coefficient to the wildfire prediction level in the specific implementation process. Another example: If the combined weight of the wildfire prediction level accounts for 60% and the risk coefficient accounts for 40%, and the above example data is continued to be substituted. At this time, the wildfire risk level = 73% * 60% + 80% * 40% = 75.8%.

[0090] Another thing to note is that the combined weight is not a fixed weight, but can be adaptively adjusted according to the actual situation. Different weight distributions can be set manually under different usage conditions or different weights can be assigned according to past data experience.

[0091] In addition, it is found in actual work that when predicting wildfire risks through the preliminary prediction model, there is a certain error in its accuracy. For example, sometimes the wildfire risk prediction level predicts a wildfire risk of 80%, but a wildfire actually occurs in reality. And this error leads to an error in the prediction of the final wildfire risk level, resulting in low data accuracy.

[0092] Based on the above situation, please refer to Figure 6 , in this embodiment, in order to further improve the accuracy of wildfire prediction, the following steps are taken when determining the combined weight of the risk coefficient and the wildfire prediction level: S421: First, obtain the historical actual wildfire information of the transmission line. S422: Compare the wildfire prediction level in the key inspection area with the risk coefficient in the key inspection area when a wildfire occurred historically. S423: Determine the accuracy of the wildfire prediction level and the risk coefficient respectively, and determine the combined weight according to this accuracy. For example, in historical data, the prediction accuracy of the wildfire prediction level is lower than that of the risk coefficient. Then, when determining the combined weight, the weight proportion of the risk coefficient will necessarily be higher than that of the wildfire prediction level.

[0093] It should be noted that the above accuracy represents the average value of multiple data. For example, the prediction levels of multiple wildfires are compared with the actual wildfire occurrences to determine the accuracy of the wildfire prediction levels, and then the average value of multiple accuracies is calculated.

[0094] In addition, please refer to Figure 7 , to further determine the weight ratio of the wildfire prediction levels and improve the accuracy of the final wildfire risk level prediction data, the following steps are taken when determining the weight ratio of the wildfire prediction levels: S4231: First, obtain multiple pieces of historical actual transmission line wildfire information; S4232: Compare the wildfire prediction levels in the key inspection area when historical wildfires occurred to determine the accuracy of the wildfire prediction levels; S4233: And obtain the line state information between the poles and towers in the key inspection area; S4234: Obtain the line state information between the poles and towers in the current key inspection area; S4235: Compare the line state information between the poles and towers in the key area when historical fires occurred to obtain a similarity factor; S4236: Assign this similarity factor to the weight of the wildfire prediction level to obtain the corrected weight of the wildfire prediction level.

[0095] Specifically, a database of wildfire prediction accuracy and line state information between poles and towers can be established, and the historical wildfire prediction accuracy is corresponding to the line state information between poles and towers one by one. Before the next prediction and weight combination, first input the line state information between the poles and towers in the area to be predicted into the database to obtain a similarity factor, and match the corresponding wildfire prediction accuracy, so as to determine the combined weight according to this wildfire prediction accuracy and risk coefficient accuracy.

[0096] For example, if the similarity between the current line state information between poles and towers and the line state information between poles and towers that is closest to the historical one is 85%, then this similarity factor is 85%, and the accuracy of this wildfire prediction is 85% of the wildfire prediction accuracy corresponding to the line state information between poles and towers that is closest to the historical one. Thus, the combined weight of the wildfire prediction level is further optimized according to this similarity factor, so as to further improve the data accuracy.

[0097] Through the above design, the optimization of the combined weight of the wildfire prediction level and the risk coefficient can, on the one hand, enable the combination of the prediction model and the actual prediction to more accurately reflect the possibility and development trend of wildfires according to the changes of various factors in different periods, making the prediction results more in line with the actual situation; on the other hand, through weight optimization, the respective errors and uncertainties of the prediction model and the actual prediction can be balanced, making the error of the final prediction result smaller and the credibility higher. After weight optimization, the model has stronger adaptability under different regions and environmental conditions and can better handle various complex situations; in addition, through weight optimization, it can be clarified under which circumstances the model prediction is more reliable and under which circumstances more reliance on the actual prediction is needed. This helps to reasonably allocate monitoring resources and avoid unnecessary waste of manpower, material resources and financial resources.

[0098] In this embodiment, a dynamic prediction analysis system 500 for transmission line wildfire faults is also provided. Please refer to Figure 8 the modular schematic diagram of the dynamic prediction analysis system for transmission line wildfire faults in [reference], which is mainly used to divide the function modules of the dynamic prediction analysis system for transmission line wildfire faults according to the embodiments of the above method. For example, each function module can be divided, or two or more functions can be integrated into one processing module. The above integrated modules can be implemented in the form of hardware or in the form of software function modules. It should be noted that the division of modules in the present invention is schematic, only a logical function division, and there may be other division methods in actual implementation. For example, in the case of dividing each function module corresponding to each function, Figure 8 only a system / device schematic diagram is shown.

[0099] Specifically, the dynamic prediction analysis system for transmission line wildfire faults includes: a region division unit 510, a level division unit 520, a prediction unit 530, and a correction unit 540. Among them, the region division unit 510 is used to divide regions according to the transmission line, obtain multiple sub-observation regions, and identify tower information for each sub-observation region to obtain multiple key inspection regions; among them, the tower information includes: tower quantity information, tower distribution information, and line state information between towers; The grading unit 520 is used to grade multiple key inspection areas, obtain the risk level factor group of each key inspection area, and determine the risk coefficient based on this risk level factor group. Among them, when grading the key inspection areas, the main level and the secondary level are determined respectively, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between poles, and the secondary level is determined by the pole number information and the pole distribution information. In some embodiments, the main level is determined according to the short-circuit point and the short-circuit frequency, the first additional number of the secondary level is determined according to the pole number information associated with the short circuit, and then the second additional number of the secondary level is determined according to the pole distribution information within the key inspection area where the short circuit occurs. The first additional number and the second additional number are assigned to the main level to obtain the preliminary level. In addition, it is also used to determine the risk level factors within the key inspection area, screen the risk level factors based on the preliminary level, and use the screened risk level factors to construct the risk level factor group of this key inspection area. Among them, when screening the risk level factors, the main risk elements are traced according to the short-circuit point and the short-circuit frequency, and the risk level factors are preliminarily screened based on the main risk elements to obtain the initially screened factor group. The first secondary risk element is determined according to the pole number information associated with the short circuit, and the initially screened factor group is screened based on the first secondary risk element to obtain the first screened factor group. The second secondary risk element is determined according to the pole distribution information within the key inspection area where the short circuit occurs, and the first screened factor group is screened based on the second secondary risk element to obtain the second screened factor group. The risk level factor group is constructed based on the risk level factors in the second screened factor group. In some embodiments, the poles where short circuits occur are marked as risk poles, the number of all risk poles within this key inspection area is obtained, and the proportion of risk poles within this key inspection area is calculated to obtain the first secondary risk element. The distribution information of all risk poles is analyzed to obtain the distribution shape of risk poles within this key inspection area, and the second secondary risk element is obtained according to the distribution shape of risk poles.

[0100] The prediction unit 530 uses the preliminary prediction model to predict wildfires on the lines within each key inspection area and obtains the corresponding wildfire prediction level. The correction unit 540 is used to assign the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level, and issue a warning about the wildfire fault information of the lines within this key inspection area based on the wildfire risk level.

[0101] In the above embodiments, the more specific working processes of the functional units can refer to the corresponding content disclosed in the foregoing method embodiments. In addition, the functional units can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired or wireless manner. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid state disk (SSD)), etc.

[0102] The embodiments of the present application are described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device generate means for implementing the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0103] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the process Figure 1 in one process or multiple processes and / or boxes Figure 1 steps for the functions specified in one box or multiple boxes.

[0105] The foregoing are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A dynamic prediction and analysis method for wildfire faults in transmission lines, characterized in that, It includes the following steps: S100: Divide the area according to the transmission line to obtain multiple sub-observation areas, identify the tower information for each sub-observation area, and obtain multiple key inspection areas; wherein, the tower information includes: the number of towers information, the tower distribution information, and the line state information between towers; S200: Classify the multiple key inspection areas to obtain the risk level factor group for each key inspection area, and determine the risk coefficient based on this risk level factor group; S300: Use the preliminary prediction model to predict the wildfire for the lines in each key inspection area to obtain the corresponding wildfire prediction level; S400: Assign the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level, and issue a warning about the wildfire fault information of the lines in this key inspection area based on the wildfire risk level; Wherein, when classifying the key inspection areas, the main level and the secondary level are determined respectively, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between towers, and the secondary level is determined by the number of towers information and the tower distribution information.

2. The dynamic prediction and analysis method for mountain fire faults in transmission lines according to claim 1, wherein The main level is determined by the historical short-circuit data between towers, and the secondary level is determined by the number of towers information and the tower distribution information, which specifically includes the following steps: Determine the main level according to the short-circuit point and the short-circuit frequency, determine the first additional number of the secondary level according to the number of towers information associated with the short-circuit, and then determine the second additional number of the secondary level according to the tower distribution information in the key inspection area where the short-circuit occurs. Assign the first additional number and the second additional number to the main level to obtain the preliminary level.

3. The dynamic prediction and analysis method for mountain fire faults of transmission lines according to claim 2, characterized in that, The step of obtaining the risk level factor group for each key inspection area includes the following steps: Determine the risk level factors in the key inspection area, screen the risk level factors based on the preliminary level, and use the screened risk level factors to construct the risk level factor group for this key inspection area. Among them, when screening the risk level factors, follow the following steps: Trace the main risk factors according to the short-circuit point and the short-circuit frequency, and conduct a preliminary screening of the risk level factors based on this main risk factor to obtain a preliminary screening factor group; Determine the first secondary risk factor according to the number of towers information associated with the short-circuit, and screen the preliminary screening factor group based on this first secondary risk factor to obtain the first screening factor group; Determine the second secondary risk factor according to the tower distribution information in the key inspection area where the short-circuit occurs, and screen the first screening factor group based on this second secondary risk factor to obtain the second screening factor group; Construct the risk level factor group based on the risk level factors in the second screening factor group.

4. The dynamic prediction and analysis method for wildfire faults in transmission lines according to claim 3, characterized in that Determining the first secondary risk factor according to the number of towers information associated with the short-circuit includes: Marking the towers where the short-circuit occurs as risk towers, obtaining the number of all risk towers in this key inspection area, and calculating the proportion of risk towers in this key inspection area to obtain the first secondary risk factor; The second secondary risk factor is determined based on the tower distribution information in the key inspection area where the short circuit occurs, including: analyzing all the risk tower distribution information to obtain the distribution shape of the risk towers in the key inspection area, and obtaining the second secondary risk factor according to the distribution shape of the risk towers.

5. The transmission line wildfire fault dynamic prediction and analysis method according to claim 4, characterized in that The types of the main risk factors include meteorological factors, vegetation factors, line factors and external factors; The types of the first secondary risk factor include A: 0-20%, B: 20%-30%, C: 30%-70%, D: 70%-90%, E: 90%-100%; The types of the second secondary risk factor include: linear style, circular style, arc style, fan style or scattered point style.

6. The dynamic prediction and analysis method for mountain fire faults in transmission lines according to claim 5, characterized in that, It further includes the step of judging whether the main risk factor, the first secondary risk factor and the second secondary risk factor are in a special combination situation: When the meteorological factor in the main risk factors of the key inspection area is strong wind, the first secondary risk factor is A: 0-20%, and the second secondary risk factor is linear style, the following analysis steps are also required: Obtain the directions of the lines on both sides of the short-circuit tower and the wind direction, and calculate the included angle between the wind direction and the line direction; If the included angle < 30°, obtain the vegetation height and the line height, and judge the difference between the line height and the vegetation height; If the difference is greater than 0 and the short-circuit frequency increases, it is determined that the short circuit is caused by the bird flock factor.

7. The dynamic prediction and analysis method for wildfire faults in transmission lines according to claim 1, wherein The specific steps of assigning the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level are as follows: Obtain the wildfire prediction level of each key inspection area, then obtain the risk coefficient of this key inspection area, determine the combined weight, which includes the wildfire prediction level weight and the risk coefficient weight, and calculate the wildfire risk level by combining the result of assigning the wildfire prediction level weight to the wildfire prediction level and the result of assigning the risk coefficient weight to the risk coefficient.

8. The dynamic prediction and analysis method for mountain fire faults in transmission lines according to claim 7, characterized in that Obtain the historical actual transmission line wildfire information, compare the wildfire prediction level and the risk coefficient of this key inspection area when the wildfire occurred historically, respectively determine the accuracy of the wildfire prediction level and the risk coefficient, and determine the combined weight according to this accuracy.

9. The dynamic prediction and analysis method for mountain fire faults of transmission lines according to claim 8, wherein Obtain multiple historical actual transmission line wildfire information, compare the wildfire prediction level of this key inspection area when the wildfire occurred historically, and determine the accuracy of the wildfire prediction level; and obtain the line state information between the towers in this key inspection area; Obtain the line state information between the towers in the current key inspection area, compare the line state information between the towers in the key area when the fire occurred historically, obtain the similarity factor, and assign this similarity factor to the wildfire prediction level weight to obtain the corrected wildfire prediction level weight.

10. A transmission line wildfire fault dynamic prediction and analysis system, characterized in that Area division unit, which is used to divide areas according to transmission lines, obtain multiple sub-observation areas, identify tower information for each sub-observation area, and obtain multiple key inspection areas; wherein, the tower information includes: the number of towers information, the distribution information of towers, and the line status information between towers. Level division unit, which is used to divide the levels of multiple key inspection areas, obtain the risk level factor group of each key inspection area, and determine the risk coefficient based on this risk level factor group; wherein, when dividing the levels of the key inspection areas, the main level and the secondary level are determined respectively, and the main level and the secondary level are combined into a preliminary level. The main level is determined by the historical short-circuit data between towers, and the secondary level is determined by the number of towers information and the distribution information of towers. Prediction unit, which uses a preliminary prediction model to predict wildfires for the lines in each key inspection area and obtain the corresponding wildfire prediction level. Correction unit, which is used to assign the risk coefficient of each key inspection area to the wildfire prediction level to obtain the corrected wildfire risk level, and issue a warning about the wildfire fault information of the lines in this key inspection area based on the wildfire risk level.

Images