Multi-pole tower fault probability analysis method and system based on multi-information nodes

Abstract

This invention provides a method and system for multi-tower fault probability analysis based on multiple information nodes, belonging to the field of multi-tower fault analysis technology. The method includes: obtaining the fault range of power poles; obtaining the fault probability of each pole within the fault range; and obtaining the component fault probability of each pole within the fault range. The method and system provided by this invention first obtains the fault range of the power poles, then combines the fault probability of each pole within the fault range with the component fault probability of each pole to obtain a multi-tower fault probability analysis table within the fault range. Based on this table, potential faults within the fault range can be predicted, thereby increasing the inspection frequency and narrowing the inspection range, thus improving the inspection effect of power poles.

Description

Technical Field

[0001] This invention relates to the field of multi-tower fault analysis technology, and specifically to a multi-tower fault probability analysis method and system based on multiple information nodes. Background Technology

[0002] With the continuous construction and development of power grid systems, the stability and reliability of power system operation are facing increasing challenges. Power poles are an important component of power transmission circuits, and the hardware, conductors, and ground wires on power poles are all crucial components for ensuring reliable power transmission. Therefore, regular inspections of power poles are necessary.

[0003] Currently, manual inspection is generally used, but with the development of drone technology, drones are increasingly replacing manual inspection. However, due to the increasing number of power lines, drone inspection still requires a lot of manpower and resources. When a power pole fails, it can only identify the current fault and cannot promptly identify other faulty components, resulting in poor inspection effectiveness.

[0004] In the process of realizing this invention, the inventors of this application discovered that the above-mentioned solutions in the prior art have the defect of poor inspection effect. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for multi-tower fault probability analysis based on multiple information nodes, which has the function of good inspection effect.

[0006] To achieve the above objectives, embodiments of the present invention provide a multi-tower fault probability analysis method based on multiple information nodes, including:

[0007] Determine the fault range of the power pole;

[0008] Obtain the failure probability of each tower within the fault range;

[0009] Obtain the component failure probability of each tower within the fault range;

[0010] Based on the failure probability of each tower and the failure probability of each component within the failure range, output a multi-tower failure probability analysis table within the failure range.

[0011] Optionally, obtaining the fault probability of each tower within the fault range includes:

[0012] Obtain the number and type of towers within the fault range;

[0013] Obtain the fault normal distribution parameters of the tower according to the type of tower;

[0014] A multivariate normal distribution model is constructed based on the fault normal distribution parameters of the towers and the number of towers.

[0015] The failure probability of each tower is obtained based on the multivariate normal distribution model.

[0016] Optionally, the fault normal distribution parameters include the mean and standard deviation.

[0017] Optionally, obtaining the component failure probability of each tower within the fault range includes:

[0018] Obtain historical component failure data for power poles;

[0019] A knowledge graph of power pole component failures is constructed based on the historical component failure data.

[0020] Obtain information nodes for faulty components;

[0021] The component failure probability of each faulty component in the tower is obtained based on the information node.

[0022] Optionally, constructing a knowledge graph of power pole component failures based on the historical component failure data includes:

[0023] Key fault information is extracted from the historical component fault data;

[0024] The TextRank algorithm is used to extract keywords from the key fault information and generate a text summary.

[0025] Obtain the original dictionary of keywords;

[0026] Based on the original keyword dictionary, the keywords are merged into synonyms;

[0027] A knowledge graph of power pole component failures is constructed based on the keywords and the text summary.

[0028] Optionally, the information nodes include fault type / weather / terrain.

[0029] Optionally, the multi-tower fault probability analysis table within the fault range is output based on the fault probability of each tower and the component fault probability, including:

[0030] The failure probability of each component on each tower within the fault range is calculated according to formula (1).

[0031] P ij =P i ×P j (1)

[0032] Among them, P ijLet P be the failure probability of the j-th component on the i-th tower within the failure range. i Let P be the failure probability of the i-th tower within the failure range. j Let be the component failure probability of the j-th component on the tower.

[0033] Optionally, the method of outputting a multi-tower fault probability analysis table within the fault range based on the fault probability of each tower and the component fault probability within the fault range further includes:

[0034] The failure probability of each component on each tower within the failure range is summarized to form a multi-tower failure probability set;

[0035] Normalize each element in the multi-tower failure probability set;

[0036] Sort each normalized element and output a multi-tower failure probability analysis table.

[0037] On the other hand, the present invention also provides a multi-tower fault probability analysis system based on multiple information nodes, comprising:

[0038] The fault location module is used to locate the fault range of power poles in the power system;

[0039] The controller is communicatively connected to the fault location module and is used to execute the multi-tower fault probability analysis method described above.

[0040] In another aspect, the present invention also provides a computer-readable storage medium storing instructions for being read by a machine to cause the machine to perform the multi-tower fault probability analysis method as described in any one of the above.

[0041] Through the above technical solution, the multi-tower fault probability analysis method and system based on multiple information nodes provided by the present invention first obtains the fault range of the power poles, and then combines the fault probability of each pole within the fault range with the component fault probability of each pole to obtain a multi-tower fault probability analysis table within the fault range. Based on the multi-tower fault probability analysis table, the faults that may occur within the fault range can be predicted, thereby increasing the inspection frequency and narrowing the inspection range, and thus improving the inspection effect of power poles.

[0042] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0043] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings:

[0044] Figure 1 This is a flowchart of a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention;

[0045] Figure 2 This is a flowchart illustrating the process of obtaining the failure probability of each tower in a multi-tower failure probability analysis method based on multiple information nodes according to an embodiment of the present invention.

[0046] Figure 3 This is a flowchart of obtaining component failure probability in a multi-tower failure probability analysis method based on multiple information nodes according to an embodiment of the present invention;

[0047] Figure 4 This is a flowchart of constructing a fault knowledge graph for power poles in a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention;

[0048] Figure 5 This is a flowchart of obtaining a multi-tower failure probability analysis table in a multi-tower failure probability analysis method based on multiple information nodes according to an embodiment of the present invention.

[0049] Figure 6 This is a schematic diagram of a keyword dictionary in a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention;

[0050] Figure 7 This is a schematic diagram of a keyword dictionary in a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention;

[0051] Figure 8 This is a structural block diagram of a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention.

[0052] Figure 9 This is a schematic diagram of launching the Neo4j database interface from the console in a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention.

[0053] Figure 10 This is a schematic diagram of the knowledge graph visualization interface in a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention. Detailed Implementation

[0054] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0055] Figure 1 This is a flowchart of a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention; Figure 8 This is a structural block diagram of a multi-tower fault probability analysis method based on multiple information nodes according to an embodiment of the present invention. Figure 1 and Figure 8 In this multi-tower failure probability analysis method, the following can be included:

[0056] In step S10, the fault range of the power poles is obtained. In the power grid system, if a component on a power pole fails, it will be detected in a timely manner by the system, and the corresponding fault range will be selected. Specifically, the fault range may include multiple power poles.

[0057] In step S11, the fault probability of each tower within the fault range is obtained. Specifically, after obtaining multiple power towers within the fault range, the fault probability of each tower is obtained separately, i.e., the fault probability of a single tower.

[0058] In step S12, the component failure probability of each tower within the fault range is obtained. The component failure probability of each tower within the fault range can be obtained by converting the historical failure probabilities of the components.

[0059] In step S13, a multi-tower fault probability analysis table is output based on the fault probability of each tower and the component fault probability within the fault range. Specifically, by combining the fault probability of each tower and the component fault probability on each tower, the component fault probability of each faulty component on each tower within the fault range can be obtained. Summarizing these probabilities yields the multi-tower fault probability analysis table.

[0060] In steps S10 to S13, the fault range of the power poles is first determined, and then the fault probability of each pole within the fault range is obtained. Combining this with the historical component fault probabilities of the poles, the component fault probability of each pole within the fault range is obtained. Finally, the above probabilities are combined and summarized to obtain a multi-pole fault probability analysis table within the fault range. Based on this multi-pole fault probability analysis table, the component fault probabilities within the fault range can be obtained, allowing for the prediction of potential faults within the fault range. Then, based on these component fault probabilities, key inspection planning or corresponding operation and maintenance strategies can be formulated.

[0061] Traditionally, power pole inspections are conducted manually. However, with the development of drone technology, drones are increasingly replacing manual inspections. However, due to the increasing number of power lines, drone inspections still require significant manpower and resources. When a power pole fails, only the current fault can be identified, failing to promptly diagnose other faulty components, resulting in poor inspection effectiveness. In this embodiment of the invention, the failure probability of multiple poles within the current fault range can be analyzed, enabling timely and early prediction of potential faults, i.e., potential fault risks. Appropriately allocating inspection resources and developing maintenance strategies can prevent further escalation of the fault and reduce fault risk. This method effectively narrows the inspection scope, reduces manpower and resource consumption, and significantly improves the inspection effectiveness and efficiency of power poles, ensuring the reliability of the power system operation.

[0062] In this embodiment of the invention, after obtaining the fault range, it is necessary to obtain the fault probability of each tower within the fault range. The specific steps for obtaining this probability are as follows: Figure 2 As shown. Specifically, in Figure 2 In addition, the multi-tower failure probability analysis method may also include:

[0063] In step S20, the number and type of towers within the fault range are obtained. The towers within the fault range can be numbered, for example, tower 1, tower 2, tower 3, etc.

[0064] In step S21, the fault normal distribution parameters of the tower are obtained according to the tower type. For the fault normal distribution parameters, expert experience in assessing faults in towers of different types, structures, and materials can be collected through questionnaires, interviews, and literature. The normal distribution parameters for each type of tower fault are then determined using statistical methods, i.e., the corresponding fault normal distribution parameters are obtained based on the tower type. Specifically, these normal distribution parameters may include the mean and standard deviation. Specifically, the probability density function of the normal distribution can be as shown in formula (2).

[0065]

[0066] Where f(x) is the probability, σ is the standard deviation, and μ is the mean.

[0067] Specifically, in the actual process of drone-based pole and tower fault inspection, we can treat the probability of pole and tower failure as a random variable, with a value ranging from 0 to 1. Assuming we have collected a certain amount of pole and tower fault data, we can calculate the mean and standard deviation of these data to estimate the probability distribution of pole and tower failures.

[0068] Based on the properties of the normal distribution, we can draw the following conclusions:

[0069] When the probability of failure is close to 0, the probability of failure is relatively small; when the probability of failure is close to 1, the probability of failure is relatively large.

[0070] The distribution of failure probability is symmetrical, meaning that the probability density functions on the left and right sides are equal.

[0071] The distribution of failure probability has the shape of a bell curve, that is, the probability density is highest near the mean, and the probability density gradually decreases on both sides of the mean.

[0072] We first collect a certain amount of tower failure data and calculate the mean and standard deviation of these data. Then, we use the probability density function of the normal distribution to calculate the tower failure probability. Finally, we compare the calculated tower failure probability with the actual failure probability to evaluate the applicability of the normal distribution.

[0073] Since tower fault data may contain outliers or missing values, we will preprocess the data in practical applications, such as removing outliers and filling in missing values, to improve the applicability of the normal distribution law in this project.

[0074] In step S22, a multivariate normal distribution model is constructed based on the fault normal distribution parameters of the towers and the number of towers. The fault probabilities of multiple towers generally exhibit a normal distribution, and a multivariate normal distribution model is constructed based on the aforementioned fault normal distribution parameters. Specifically, the operator extracts a corresponding subset from the multivariate normal distribution model based on the input fault range, obtaining a sub-multivariate normal distribution model. For instance, if the operator inputs three towers (tower 1, tower 2, and tower 3), a ternary normal distribution model can be extracted from the multivariate normal distribution model.

[0075] In step S23, the failure probability of each tower is obtained based on a multivariate normal distribution model. Specifically, workers calculate the probability of one or more towers failing within a certain time (or distance) range using this multivariate normal distribution model. This probability can be calculated using the cumulative distribution function (CDF) of the multivariate normal distribution. While the CDF of the multivariate normal distribution does not have a closed-form expression, it can be approximated using methods such as numerical integration or Monte Carlo simulation. Specifically, the `Constructing_Probability_Models()` function is used to calculate the probability distribution parameters for each level of tower. The inputs are the mean and standard deviation, and the output is a list of probability distribution parameters for the corresponding number of towers.

[0076] In steps S20 to S23, the number and type of towers within the fault range are first obtained, and the corresponding fault normal distribution parameters are obtained based on the tower type. Then, a corresponding multivariate normal distribution model is constructed based on the fault normal distribution parameters and the number of towers, and the fault probability of each tower is obtained based on the multivariate normal distribution model. This method can obtain the fault probability of each tower within the fault range relatively accurately.

[0077] In this embodiment of the invention, when obtaining the component failure probability of a multi-tower structure, it is also necessary to calculate the component failure probability of each faulty component. The specific calculation steps can be as follows: Figure 3 As shown. Specifically, in Figure 3 In addition, the multi-tower failure probability analysis method may also include:

[0078] In step S30, historical component failure data of the power poles is acquired. This historical component failure data can include textual data and empirical data. Specifically, the textual data can be obtained from multiple sources, such as the State Grid website, network failure reports, and network failure analysis reports. After obtaining the failure reports, the textual data is organized according to certain formats and standards, such as using standardized titles, numbers, and dates. The empirical data can include the collection and summarization of expert experience and opinions.

[0079] In step S31, a knowledge graph of power pole component failures is constructed based on historical component failure data.

[0080] In step S32, information nodes of the faulty component are obtained. These information nodes represent the search criteria for the faulty component. Specifically, the search criteria may include the fault type and / or weather and / or terrain.

[0081] In step S33, the component failure probability of each faulty component in the tower is obtained based on the information nodes. Specifically, a search is performed in the power tower component failure knowledge graph according to search constraints to obtain relevant component failure information. This information is analyzed and summarized to determine the occurrence count of each faulty component in the tower. The component failure probability of each faulty component is obtained by dividing the occurrence count of each faulty component by the total occurrence count of all faulty components. Specifically, the faulty component is the one that has been found to have failed. The `Constructing_Multi_Probabilit(P, result)` function calculates the failure probability of each level of tower based on the input faulty component probability dictionary `result` and the multi-tower probability model `P`. The `transform_list(B, count_row)` function maps the actual required tower faulty component probabilities to the list `actual_node_dicts`. Based on the `count_row` value, the required probabilities of different levels of tower faulty components are obtained and stored in `actual_node_dicts`. In addition, the generate_list_C(A, n) function is used to map actual_node_dicts to sorted_node_dicts according to the pole level order of the CSV file.

[0082] In steps S30 to S33, historical component failure data of the power pole is first collected, and then a knowledge graph of power pole component failures is constructed based on this historical component failure data. Pre-defined information nodes for faulty components, i.e., search constraints, are used to search the power pole component failure knowledge graph to obtain relevant component failure information. This information is then summarized and analyzed to obtain the component failure probability for each faulty component.

[0083] In this embodiment of the invention, when constructing a knowledge graph of power pole component faults, it is necessary to process historical component fault data. Specific processing steps can be as follows: Figure 4 As shown. Specifically, in Figure 4 In addition, the multi-tower failure probability analysis method may also include:

[0084] In step S40, key fault information is extracted from historical component fault data. Specifically, when constructing a knowledge graph of power pole component faults, key information needs to be extracted from both textual and empirical data. Specifically, textual data is first manually analyzed and then extracted using artificial intelligence to obtain key fault information, which is then saved in a CSV file. Empirical data may include basic information about overhead transmission lines and towers, such as line length, direction, type, voltage level, and service life; it may also include questionnaire analyses by experts regarding the fault rate, fault cause distribution, fault component distribution, and fault influencing factor distribution for different lines. Specifically, the analytic hierarchy process (AHP) can be used to extract key information from expert questionnaire analyses. This AHP includes four steps: establishing a hierarchical structure model, constructing a judgment matrix, calculating weights, and verifying consistency.

[0085] In step S41, the TextRank algorithm is used to extract keywords from key fault information and generate text summaries. TextRank is a graph-based ranking algorithm for text, derived from Google's PageRank algorithm. It segments text into units (words, sentences) and builds a graph model, using a voting mechanism to rank important components. It can achieve keyword extraction and summarization using only information from a single document. Specifically, the TextRank algorithm includes two tasks: keyword extraction and text summarization. Keyword extraction refers to identifying terms that describe the meaning of a document. In this task, words are nodes in the graph, and edges between words are determined using co-occurrence relationships. Co-occurrence means words appearing together; words within a given sliding window are considered to co-occur, and edges exist between these words. Text summarization is the process of extracting short text summarizing the main idea and content of one or more documents. In this task, nodes are no longer words but sentences. The connections between sentences are determined using similarity. Similarity can be calculated by dividing the number of words in two sentences by the logarithmic length of the two sentences.

[0086] Specifically, the original text is split into sentences. Stop words are filtered out (optional) within each sentence, and only words with specified parts of speech (optional) are retained. This yields a set of sentences and a set of words. Each word is treated as a node in PageRank. A window size of k is set. Assume a sentence is composed of words "w1, w2, w3, w4, w5, ..., wn", where "w1, w2, ..., wk", "w2, w3, ..., wk+1", etc., are windows. An undirected, unweighted edge exists between any two word nodes within a window. Based on this graph, the importance of each word node can be calculated, and the most important words are selected as keywords.

[0087] After extracting several keywords, if several keywords are adjacent in the original text, these keywords can form a keyword group. Each sentence is treated as a node in a graph; if two sentences are similar, there is an undirected weighted edge between the corresponding nodes, with the weight representing the similarity. The sentences with the highest importance calculated using the PageRank algorithm can be used as a summary.

[0088] Specifically, taking the report "Lightning Strike Fault on 500kV Pingfei 5302 Line on August 24" as an example, the extracted data can be obtained as shown in Tables 1, 2, and 3.

[0089] Table 1 Summary of Report Keywords

[0090] Keywords: Importance score: Fault 0.036683575 line 0.018706546 Wire 0.017735498 Lightning 0.016789774 Grounding 0.015640442 tower 0.012843846 tower 0.01208235 insulator 0.011965438 Check line 0.011276398 Condition 0.011153903

[0091] Table 2 Summary of Key Phrases in the Report

[0092]

[0093]

[0094] Table 3 Summary of Key Information from the Report

[0095]

[0096]

[0097] Analyzing Tables 1, 2, and 3 can help obtain information such as fault type (lightning fault), weather conditions (thunderstorm), time of occurrence (August 24, 2020), location of occurrence (Wushan Town, Changfeng County, Hefei City), and tower type (tower #162 of the 500kV Pingfei 5302 line lightning fault), simplifying the report analysis process and effectively reducing the workload.

[0098] In step S42, a keyword dictionary is obtained. This dictionary can be understood as a list of commonly used synonyms for components of power poles, created based on practical experience. It is used to query keywords and output the results. Specifically, the keyword dictionary may include, for example... Figure 6 as well as Figure 7 As shown.

[0099] In step S43, synonyms are merged based on the keyword dictionary. In practice, staff may describe the same fault type, terrain, or weather in different ways. Directly searching or matching keywords in the knowledge graph would significantly reduce accuracy and success rate. Therefore, we choose to use a fuzzy matching algorithm or a dictionary of keyword sets for synonym merging. Specifically, the fuzzy matching algorithm calculates the edit distance between two strings and performs fuzzy matching. Edit distance refers to the minimum number of editing operations required to transform one string into another, including insertion, deletion, and replacement. Fuzzy matching finds the candidate string most similar to the input string based on the edit distance.

[0100] In step S44, a knowledge graph of power pole component faults is constructed based on keywords and text summaries. For the construction of the power pole fault knowledge graph, the Neo4j graph database can be used, which supports fast querying and analysis, making it suitable for large-scale datasets. Specifically, in a knowledge graph or graph database, entities are typically represented by nodes, while relationships are represented by edges or lines connecting these nodes. An entity group refers to a set of nodes related to a specific context, along with the set of relationships between them. The main purpose of this structure is to organize and present information, enabling us to more clearly understand the connections and patterns between different entities.

[0101] Specifically, the data sources for knowledge graph construction, taking semi-structured historical fault data of towers as an example, can be shown in Table 4.

[0102] Table 4. Examples of Semi-structured Data

[0103]

[0104]

[0105] For the semi-structured data in Table 4, the Python programming language and the Neo4j graph database can be used to parse, analyze, and establish relationships between entity groups in order to better understand the connections and patterns between the data.

[0106] Data input is typically in the form of a CSV file, with each column representing different fields, including line name, time, location, latitude and longitude coordinates, tower type, voltage level, etc. Each column header describes the content of the corresponding data, such as "Weather" or "Terrain".

[0107] When retrieving relational entity groups, the process first obtains all CSV files under a specified folder path, then uses Python's csv library to read the file contents into a two-dimensional array. The CSV data array is then iterated through to extract the master node data. Master node data typically represents fields of the entity group, such as "fault type" or "weather." Custom data preprocessing functions are used to appropriately transform the data based on field types, which helps ensure data consistency and queryability.

[0108] When graph data needs to be generated, the Neo4j graph database is used to construct the knowledge graph. It represents entities as nodes in the graph database, relationships between entities as relations in the graph database, and uses attributes to describe the characteristics of the relations. Each relation also includes an "occurrences" attribute to track the frequency of relation occurrences. The constructed knowledge graph is stored in the Neo4j graph database for subsequent querying and analysis.

[0109] Specifically, the relationships between entity groups can include:

[0110] 1. Relationship between fault types and other entities: The model creates relationships to represent the connections between fault types and other entities, such as the relationship between lightning strike faults and the load at the time of the fault.

[0111] 2. Relationship between terrain and weather: The model creates relationships to represent the connection between different weather conditions and terrain, such as the relationship between thunderstorms and mountainous terrain.

[0112] 3. Relationship between weather and faulty components: The model constructs relationships to represent the connection between different weather conditions and faulty components, such as the relationship between thunderstorms and lightning strikes.

[0113] 4. Relationship between terrain and faulty components: The model creates relationships to represent the connection between different terrains and faulty components, such as the relationship between mountainous terrain and faulty components.

[0114] Specifically, for the installation and configuration of Neo4j, taking Neo4j-4.4.4 as an example, you need to download JDK 11 to be compatible. You can download the community edition installer from the official website. After decompression, the main functional folders may include the following:

[0115] The bin directory is used to store the Neo4j executable program;

[0116] The conf directory contains configuration files used to control Neo4j startup.

[0117] The data directory is used to store the core database files.

[0118] The plugins directory is used to store Neo4j plugins.

[0119] Next, go to Environment Variables and configure them. Create a new system variable in the System variables section. Add the paths to the folders containing JDK and NEO4J to the variable value field.

[0120] Enter the command prompt (cmd) interface, then navigate to your NEO4J's bin directory. Type "cmd" to open the console, and use the command `neo4j.bat console` to open it. Figure 9 As shown. Then go to http: / / localhost:7474 / and enter the corresponding command to visualize the entire knowledge graph. The specific knowledge graph visualization interface can be seen as follows: Figure 10 As shown.

[0121] In knowledge graphs or graph databases, entities are typically represented by nodes, while relationships are represented by edges or lines connecting these nodes. An entity group refers to a set of nodes related to a specific context, along with the relationships between them. The primary purpose of this structure is to organize and present information, enabling us to more clearly understand the connections and patterns between different entities. Entity groups are a core concept in knowledge graphs, providing us with deep insights into domains, topics, or concepts by aggregating related entities and their connections. In the preceding series of preprocessing steps, semi-structured historical fault data of poles was extracted from expert reports using a large language model; this data served as the primary source for constructing the knowledge graph and obtaining relational entity groups.

[0122] In this embodiment of the invention, Python 3.x is used, and the py2neo library is used to interact with the Neo4j database. Other relevant Python libraries include csv for processing CSV files, and custom data preprocessing functions (such as type_match, weather_match, terrain_match, component_match, etc.) for standardizing data and transforming it into entities in a knowledge graph. Specifically, semi-structured data is saved as a CSV file, ensuring the data format matches the example data. The code first configures the Neo4j database connection information, including the database URL and authentication information. Then, it ensures the Neo4j database is started and running, runs the data processing model, and specifies the folder path containing the CSV file as an input parameter. The model constructs a knowledge graph, and the Neo4j database tools or programming interface are used to query and analyze the results.

[0123] In steps S40 to S44, key fault information is extracted from both textual data and empirical data. Keywords are then extracted from this key fault information, and a text summary is generated. During this process, synonyms for the keywords are merged according to a keyword dictionary to ensure that synonyms can be matched during searches, thus improving search accuracy. Finally, a knowledge graph of power pole component faults is constructed based on the keywords and the text summary.

[0124] In this embodiment of the invention, in order to obtain the failure probability of each component on each tower within the failure range, it is also necessary to combine the failure probability of a single tower within the failure range with the failure probability of the faulty component. Specifically, the combination method can be as follows: Figure 5 As shown. Specifically, in Figure 5 In addition, the multi-tower failure probability analysis method may also include:

[0125] In step S50, the failure probability of each component on each tower within the failure range is calculated according to formula (1).

[0126] P ij =P i ×P j (1)

[0127] Among them, P ij Let P be the failure probability of the j-th component on the i-th tower within the failure range. i Let P be the failure probability of the i-th tower within the failure range. j Let be the component failure probability of the j-th component on the tower.

[0128] In step S51, the failure probability of each component on each tower within the fault range is summarized to form a multi-tower failure probability set.

[0129] In step S52, each element in the multi-tower failure probability set is normalized. Specifically, the failure probability of each component on each tower within the calculated failure range needs normalization due to different dimensions. Common normalization methods include interval scaling and z-score standardization. Interval scaling transforms the feature values ​​of the samples to the same dimension, mapping the data to the interval [0, 1] or [-1, 1]. Z-score standardization processes the data according to the columns of the feature matrix, converting it to a standard normal distribution by calculating the z-score, which is related to the overall sample distribution. Using normalization allows for better analysis of the health status of UAV towers, providing a basis for equipment maintenance and inspection strategies, and enabling the rational allocation of maintenance cycles and resources to reduce equipment failure rates. Specifically, the `normalize_probabilities(dicts)` function normalizes each element in the input list so that the sum of all probability values ​​is 1.

[0130] In step S53, each normalized element is sorted and a multi-tower failure probability analysis table is output.

[0131] In steps S50 to S53, the failure probability of each component on each tower within the fault range is first calculated. Then, these failure probabilities are summarized and normalized to finally output a multi-tower failure probability analysis table. Based on this table, failure risks can be predicted, facilitating the coordinated allocation of inspection and maintenance resources and improving inspection efficiency.

[0132] On the other hand, the present invention also provides a multi-tower fault probability analysis system based on multiple information nodes. Specifically, the multi-tower fault probability analysis system may include a fault location module and a controller. Specifically, the fault location module is used to locate the fault range of power towers in the power system, and the controller is communicatively connected to the fault location module to execute the multi-tower fault probability analysis method described above.

[0133] In another aspect, the present invention also provides a computer-readable storage medium storing instructions for being read by a machine to cause the machine to execute the above-described multi-tower fault probability analysis method.

[0134] Through the above technical solution, the multi-tower fault probability analysis method and system based on multiple information nodes provided by the present invention first obtains the fault range of the power poles, and then combines the fault probability of each pole within the fault range with the component fault probability of each pole to obtain a multi-tower fault probability analysis table within the fault range. Based on the multi-tower fault probability analysis table, the faults that may occur within the fault range can be predicted, thereby increasing the inspection frequency and narrowing the inspection range, and thus improving the inspection effect of power poles.

[0135] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0137] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0138] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0139] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0140] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0141] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0142] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0143] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for analyzing the probability of multi-tower failures based on multiple information nodes, characterized in that, include: Determine the fault range of the power pole; Obtain the failure probability of each tower within the fault range; Obtain the component failure probability of each tower within the fault range; Based on the failure probability of each tower and the failure probability of components within the failure range, output a multi-tower failure probability analysis table within the failure range; Obtaining the fault probability of each tower within the fault range includes: Obtain the number and type of towers within the fault range; Obtain the fault normal distribution parameters of the tower according to the type of tower; A multivariate normal distribution model is constructed based on the fault normal distribution parameters of the towers and the number of towers. The failure probability of each tower is obtained based on the multivariate normal distribution model; Obtaining the component failure probability of each tower within the fault range includes: Obtain historical component failure data for power poles; A knowledge graph of power pole component failures is constructed based on the historical component failure data. Obtain information nodes for faulty components; The component failure probability of each faulty component in the tower is obtained based on the information node; The search is performed in the knowledge graph of power pole component failures according to the search criteria to obtain relevant component failure information. The component failure information is analyzed and summarized to obtain the occurrence number of each faulty component in the pole. The component failure probability of each faulty component is obtained by dividing the occurrence number of each faulty component by the total occurrence number of faulty components. Based on the failure probability of each tower and the failure probability of its components within the failure range, a multi-tower failure probability analysis table is output, including: The failure probability of each component on each tower within the failure range is calculated according to formula (1). ，（1） in, Within the range of the fault On the first of the aforementioned towers The failure probability of each component Within the range of the fault The failure probability of the aforementioned tower, For the first on the tower The probability of component failure for each part.

2. The multi-tower failure probability analysis method according to claim 1, characterized in that, The fault normal distribution parameters include the mean and standard deviation.

3. The multi-tower failure probability analysis method according to claim 1, characterized in that, Based on the historical component failure data, a knowledge graph of power pole component failures is constructed, including: Key fault information is extracted from the historical component fault data; The TextRank algorithm is used to extract keywords from the key fault information and generate a text summary. Obtain the original dictionary of keywords; Based on the original keyword dictionary, the keywords are merged into synonyms; A knowledge graph of power pole component failures is constructed based on the keywords and the text summary.

4. The multi-tower failure probability analysis method according to claim 1, characterized in that, The information nodes include fault type, weather, and terrain.

5. The multi-tower failure probability analysis method according to claim 1, characterized in that, The multi-tower fault probability analysis table, which outputs the fault probability of each tower and component within the fault range, also includes: The failure probability of each component on each tower within the failure range is summarized to form a multi-tower failure probability set; Normalize each element in the multi-tower failure probability set; Sort each normalized element and output a multi-tower failure probability analysis table.

6. A multi-tower fault probability analysis system based on multiple information nodes, characterized in that, include: The fault location module is used to locate the fault range of power poles in the power system; The controller is communicatively connected to the fault location module and is used to execute the multi-tower fault probability analysis method as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that are read by a machine to cause the machine to perform the multi-tower failure probability analysis method as described in any one of claims 1 to 5.

Images