Power distribution network fault differentiation operation and maintenance first-aid repair decision recommendation method, system and equipment based on knowledge graph, and medium

Abstract

The invention discloses a power distribution network fault differentiation operation and maintenance first-aid repair decision recommendation method, system and device based on a knowledge graph and a medium, and belongs to the technical field of operation and maintenance recommendation, and the method comprises the steps: obtaining the real-time monitoring data of a power distribution network, and constructing a fault feature data set; based on the fault feature data set, combining with a power distribution network knowledge graph to identify a historical fault entity; extracting a corresponding operation and maintenance first-aid repair strategy from the historical fault entity, presetting an evaluation index, and calculating a strategy effect score; evaluating a fault processing priority; and sorting the candidate first-aid repair strategies according to the strategy effect score and the fault processing priority, and screening and outputting a target operation and maintenance strategy suitable for the current fault and matters needing attention of the target operation and maintenance strategy. According to the method, the geographic information system is used for positioning the fault, the distance between the fault and the first-aid repair personnel is evaluated, the processing priority is determined in combination with the emergency degree of the fault, resource allocation is effectively optimized, and response efficiency is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of operation and maintenance recommendation, in particular to a power distribution network fault differentiated operation and maintenance repair decision recommendation method, system, device and medium based on a knowledge graph. BACKGROUND

[0002] The technical field of operation and maintenance recommendation includes maintenance and fault handling strategies for power systems, especially in power distribution networks. The core content of this field is to improve the reliability and efficiency of power distribution networks by implementing system-level fault prediction, diagnosis and rapid response mechanisms. The main research directions of the technical field include fault detection, optimization of operation and maintenance resources, and automated scheduling of repair work, aiming to reduce power outage time, optimize maintenance resources, and improve the operational stability and economic efficiency of power grids.

[0003] Among them, the power distribution network fault differentiated operation and maintenance repair decision recommendation method refers to a technical method that provides customized decision support for fault response and repair work in power distribution network operation and maintenance. This method covers the accurate identification of fault types, the rapid assessment of fault impact range, and the intelligent recommendation of repair priority and resource allocation. By analyzing historical data and real-time monitoring data, applying fault analysis and resource scheduling algorithms, targeted operation and maintenance strategies and repair action recommendations are formed.

[0004] Traditional recommendation methods still have room for improvement in terms of fault data processing efficiency, rapid and accurate identification of fault types, and assessment of fault impact range. For example, traditional fault handling strategies lack specificity and do not fully utilize historical fault data and real-time field data for in-depth analysis, resulting in suboptimal resource allocation and sometimes even resource waste. This situation is particularly evident during emergency failures, as the lack of effective strategy support and intelligent recommendations can result in inadequate response to important equipment failures, affecting the stable operation and economic benefits of the entire power grid and affecting the overall performance and user satisfaction of the power grid. SUMMARY

[0005] In view of the above problems, the present application is proposed.

[0006] Therefore, the technical problem solved by the present application is: how to realize dynamic power coordination of thermal power units and distributed energy storage, improve frequency regulation accuracy and system redundancy fault tolerance, and at the same time suppress high-voltage DC bus voltage fluctuations through segmented high-voltage current architecture and adaptive droop control strategy.

[0007] To solve the above technical problems, the present application provides the following technical solutions: a power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph, comprising,

[0008] Real-time monitoring data of a power distribution network is acquired, information related to a fault is processed, and a fault feature data set containing fault types, device characteristics, and operating states is constructed;

[0009] Based on the fault feature data set, in combination with a power distribution network knowledge graph, a historical fault entity that meets a preset similarity threshold in fault type, device attribute, or occurrence condition with a current fault is identified;

[0010] A corresponding operation and maintenance repair strategy is extracted from the historical fault entity, a preset evaluation index is set, and a strategy effect score is calculated;

[0011] The influence range, power emergency degree, and spatial position of the repair resource of the current fault are evaluated to determine the fault handling priority;

[0012] The strategy effect score and the fault handling priority are combined to sort the candidate repair strategies, and a target operation and maintenance strategy suitable for the current fault and its precautions are selected and output.

[0013] As a preferred scheme of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph, the information related to the fault is processed, including extracting the type, occurrence time, duration, and affected devices related to the fault of the fault device;

[0014] The information is data cleaned to remove non-fault time records and format abnormal entries;

[0015] The time format, device number, and type identification of the fault data are unified, and a structured standard fault feature data set is generated.

[0016] As a preferred scheme of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph, the historical fault entity that meets the preset similarity threshold includes establishing a feature comparison framework based on the type, occurrence time, and duration of the fault device, and retrieving the nodes related to the device and their edge attributes in the power distribution network knowledge graph.

[0017] The feature information of the current fault and the historical fault is compared according to the set similarity measurement standard, and the similarity value is calculated;

[0018] The historical fault entity with a similarity value higher than the preset threshold is selected as the similar fault recognition result.

[0019] As a preferred scheme of the power distribution network fault differential operation and repair decision recommendation method based on a knowledge graph, wherein: the strategy effect score includes extracting historical operation and repair records related to the similar fault from the power distribution network knowledge graph, obtaining the statistical results of the execution cost, execution time and success of each operation strategy in multiple historical cases;

[0020] And based on the average value of each index as the basis for calculating the strategy effect score.

[0021] As a preferred scheme of the power distribution network fault differential operation and repair decision recommendation method based on a knowledge graph, wherein: the calculation of the strategy effect score includes, for each operation strategy, calculating the difference between its average execution cost and the maximum execution cost among all candidate strategies, and multiplying the difference by the weight coefficient of the cost index;

[0022] Calculate the difference between the average execution time of the strategy and the maximum execution time among all strategies, and multiply the difference by the weight coefficient of the time index;

[0023] At the same time, multiply the average success rate of the strategy by the weight coefficient of the success rate index;

[0024] Add the above three calculation results, and divide the sum by the sum of the three index weight coefficients to get the final effect score value of the strategy.

[0025] As a preferred scheme of the power distribution network fault differential operation and repair decision recommendation method based on a knowledge graph, wherein: the evaluation of the fault handling priority includes, based on the geographic information system of the power distribution network, obtaining the geographic position coordinates of each fault point, and combining the current position of the repair personnel, calculating the expected distance and time required to reach each fault point;

[0026] At the same time, according to the number of users in each fault influence range and the emergency degree of the power load, set multiple priority influence factors;

[0027] By weighting calculation of the above factors according to the preset weight coefficient, the processing priority value of each fault is generated.

[0028] As a preferred scheme of the power distribution network fault differential operation and repair decision recommendation method based on a knowledge graph, wherein: the screening and output of the target operation strategy suitable for the current fault and its precautions include, based on the similarity between the current fault and the historical fault entity, and the effect score of the corresponding operation strategy, calculating the recommendation score of each candidate strategy;

[0029] Sort the candidate strategies according to the recommendation score, and according to the set recommendation quantity threshold, screen a number of strategies with high scores;

[0030] And in combination with the available state of repair resources and policy applicable conditions, output as a recommended result of the target strategy.

[0031] Another object of the present application is to provide a power distribution network fault differentiated operation and maintenance repair decision recommendation system based on a knowledge graph.

[0032] To solve the above technical problems, the present application provides the following technical scheme: a power distribution network fault differentiated operation and maintenance repair decision recommendation system based on a knowledge graph, comprising: a data acquisition module for acquiring real-time monitoring data of a power distribution network, processing information related to faults, and constructing a fault feature data set containing fault types, device characteristics and operating states;

[0033] A graph recognition module is used to identify historical fault entities that meet a preset similarity threshold in fault type, device attribute or occurrence condition in combination with a power distribution network knowledge graph based on the fault feature data set.

[0034] A strategy evaluation module is used to extract corresponding operation and maintenance repair strategies from the historical fault entities, preset evaluation indicators, and calculate strategy effect scores.

[0035] A priority analysis module is used to evaluate fault handling priority according to the influence range of the current fault, the emergency degree of electricity use and the spatial position of repair resources.

[0036] A decision recommendation module is used to sort, filter and output target operation strategies and their precautions suitable for the current fault by combining strategy effect scores and fault handling priority.

[0037] The present application provides a computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph when executing the computer program.

[0038] The present application provides a computer readable storage medium having a computer program stored thereon, characterized in that the computer program is executed by a processor to implement the steps of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph.

[0039] The application has the beneficial effects that: by positioning the fault through the geographic information system, evaluating the distance from the repair personnel, and determining the processing priority in combination with the emergency degree of the fault, the resource allocation is effectively optimized, the response efficiency is improved, through the similar fault analysis combined with the knowledge graph, the decision basis of the historical data support is provided for the operation and maintenance team, the accuracy of the fault processing and the pertinence of the strategy can be significantly improved, through the evaluation of the execution cost, time and success probability of various operation and maintenance repair strategies, a systematic recommendation mechanism is formed, the economic benefit of the repair work is improved, the stability and reliability of the distribution network are enhanced, the operation and maintenance quality of the distribution network is comprehensively improved, the frequency and duration of power outage events are reduced, and the satisfaction of consumers is enhanced. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0041] Figure 1 The overall flowchart of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0042] Figure 2 The acquisition flowchart of the power distribution network fault feature data set of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0043] Figure 3 The acquisition flowchart of the fault repair priority information of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0044] Figure 4 The acquisition flowchart of the similar fault identification result of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0045] Figure 5 The acquisition flowchart of the repair strategy effect evaluation information of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0046] Figure 6 The acquisition flowchart of the screened operation and maintenance repair strategy of the power distribution network fault differentiated operation and maintenance repair decision recommendation method based on the knowledge graph provided by an embodiment of the application.

[0047] Figure 7The method for recommending power distribution network fault differentiated operation and repair decision provided by an embodiment of the application comprises the following steps of: DETAILED DESCRIPTION

[0048] In order to make the above objectives, characteristics and advantages of the application more apparent, obvious and easy to understand, the specific embodiments of the application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor should fall within the protection scope of the application.

[0049] Embodiment 1, reference Figure 1 For an embodiment of the application, the embodiment provides a method for recommending power distribution network fault differentiated operation and repair decision based on a knowledge graph, which comprises the following steps of:

[0050] Step 1: acquiring real-time monitoring data of a power distribution network, processing information related to a fault, and constructing a fault feature data set containing fault types, device characteristics and operating states;

[0051] Step 2: based on the fault feature data set, combining a power distribution network knowledge graph, and identifying historical fault entities that meet a preset similarity threshold in fault types, device attributes or occurrence conditions with a current fault;

[0052] Step 3: extracting corresponding operation and repair strategies from the historical fault entities, presetting evaluation indexes, and calculating strategy effect scores;

[0053] Step 4: evaluating fault handling priorities according to an influence range of the current fault, an electricity emergency degree and a spatial position of repair resources;

[0054] Step 5: sorting candidate repair strategies, screening and outputting target operation strategies suitable for the current fault and matters needing attention by comprehensively considering the strategy effect scores and the fault handling priorities.

[0055] In step 1, the processing of the information related to the fault comprises extracting the types, occurrence times, durations and affected devices related to the fault of the fault device;

[0056] The information is subjected to data cleaning to remove non-fault time records and abnormal format entries;

[0057] and the time format, device number and type identification of the fault data are unified to generate a structured standard fault feature data set.

[0058] In an optional embodiment of the present application, the process of constructing the fault feature dataset can further introduce an edge computing-based on-site preliminary screening processing module. The module is deployed at the regional monitoring terminal of each power distribution network, and through the real-time collection of current, voltage, frequency and other key electrical parameters locally, the data is preliminarily screened by using a lightweight fault identification algorithm, and only the data with suspected fault features are uploaded to the central system, so as to realize the effective compression of the uplink data stream, reduce the communication burden, and improve the overall data processing efficiency of the system.

[0059] In a preferred embodiment of the present application, the processing of the fault-related information includes the following steps: extracting the type of the fault equipment, the fault occurrence time, the fault duration and the affected equipment information related to the fault; performing data cleaning on the above data to eliminate non-fault time period records, logical conflict data and format error entries; converting the time information into UTC timestamp format, converting the equipment number into a standard coding system, and uniformly labeling the fault type according to the pre-defined classification. The finally constructed fault feature dataset includes: fault equipment type, occurrence time, duration, affected equipment number and its operating state parameters, with unified structure and semantic specification.

[0060] The beneficial effects of the preferred technical solution are: through the systematic data preprocessing process, the accuracy, consistency and completeness of the fault data are effectively improved, and the interference of invalid and abnormal data on the subsequent analysis process is significantly reduced; the unified fault feature data format provides structured and standardized input support for subsequent knowledge graph matching, similarity evaluation and other key steps; at the same time, in the multi-source data fusion scene, the scheme also has good expansibility and adaptability, which is conducive to enhancing the generalization ability and intelligent decision basis of the system.

[0061] In step 2, the historical fault entities satisfying the preset similarity threshold include: establishing a feature comparison framework based on the type of the fault equipment, the fault occurrence time and the duration, and retrieving the nodes and edge attributes related to the equipment in the power distribution network knowledge graph;

[0062] According to the set similarity measurement standard, the feature information of the current fault and the historical fault is compared, and the similarity value is calculated;

[0063] And the historical fault entities with a similarity value higher than the preset threshold are screened as the similar fault recognition results.

[0064] In an optional embodiment of the present application, when identifying the historical fault entities that meet the preset similarity threshold with the current fault, a representation learning method based on vector embedding can be used to encode the fault feature data and the device attributes in the knowledge graph through a graph neural network to generate a unified dimensional vector representation. In the high-dimensional vector space, the cosine similarity or Euclidean distance is used to evaluate the similarity between the current fault and the historical fault, so as to filter out the historical fault records with a similarity higher than the threshold. This method has stronger robustness and generalization capability when processing multi-source heterogeneous fault features.

[0065] In a preferred embodiment of the present application, identifying the historical fault entities that meet the preset similarity threshold with the current fault in terms of fault type, device attribute or occurrence condition comprises the following steps: constructing a feature comparison template based on the device type, fault occurrence time and duration in the current fault feature, and retrieving the nodes related to the target device and their edge attributes in the knowledge graph; comparing the current fault feature with the historical node features item by item according to the set similarity calculation standard, and calculating the overall similarity value; and finally filtering out the historical fault entities with a similarity value higher than the set threshold as the identification result.

[0066] The preferred technical solution has the following beneficial effects: through structured comparison and feature calculation, the fault instances with high similarity to the current fault in key attributes in history can be accurately identified, and the interpretability and accuracy of similarity calculation are improved; meanwhile, the entity is screened based on the associated attributes of the knowledge graph, so that the identification of the fault mode is more in line with the actual operation logic of the power distribution network, and the adaptability and decision support capability of the system recommendation strategy are enhanced.

[0067] In step 3, the strategy effect score comprises extracting the historical operation and maintenance repair records related to the similar fault from the power distribution network knowledge graph, obtaining the statistical results of the execution cost, execution time and success of each operation and maintenance strategy in multiple historical cases;

[0068] and taking the average value of each index as the basis for calculating the strategy effect score.

[0069] The calculation of the strategy effect score comprises, for each operation and maintenance strategy, calculating the difference between its average execution cost and the maximum execution cost among all candidate strategies, and multiplying the difference by the weight coefficient of the cost index;

[0070] calculating the difference between the average execution time of the strategy and the maximum execution time among all strategies, and multiplying the difference by the weight coefficient of the time index;

[0071] Meanwhile, the average success rate of the strategy is multiplied by the weight coefficient of the success rate index;

[0072] The three calculation results are added, and the sum is divided by the sum of the three index weight coefficients to obtain the final effect score value of the strategy.

[0073] In an optional embodiment of the present application, for the extraction and effect evaluation of historical operation and repair strategies, a case-based reasoning (CBR) method framework can be introduced. The system first constructs a case library containing historical fault features, treatment strategies and actual results, then matches the similarity of the current fault features with the faults in the case library, extracts the corresponding strategy and its execution effect, and uses the normalized multi-dimensional indicators (such as cost, time, success rate) for comprehensive sorting and scoring. This method can dynamically reason on the basis of existing knowledge, improving the system's autonomous learning and adaptability.

[0074] In a preferred embodiment of the present application, the process of extracting corresponding operation and repair strategies from historical fault entities and calculating effect scores includes: counting the execution cost, execution time and success or failure of each strategy in historical cases, and calculating the average cost, average time and average success rate respectively; then according to the set weight coefficient, the three indicators are weighted and synthesized to calculate the effect score of each strategy. The scoring model uses a normalized linear weighting formula to ensure that different dimension indicators have comparability under a unified scoring system.

[0075] The beneficial effects of the preferred technical solution are: through statistical analysis of the performance of strategies in real historical cases, objective quantitative evaluation of strategy effect is realized; the scoring results can effectively distinguish the actual application effect of different strategies under various faults, providing high-quality evaluation basis for subsequent decision recommendation; in addition, the clear structure of the weighted calculation model makes the evaluation process have good interpretability, facilitating system parameter adjustment and expert review, further improving the credibility and practicality of the system.

[0076] In step 4, the evaluation of fault handling priority includes obtaining the geographic coordinates of each fault point based on the geographic information system of the distribution network, and combining the current location of the repair personnel to calculate the expected distance and time required to reach each fault point;

[0077] At the same time, according to the number of users and the emergency degree of power load within the influence range of each fault, a plurality of priority influence factors are set;

[0078] By weighting the above factors according to the preset weight coefficient, the handling priority value of each fault is generated.

[0079] In an optional embodiment of the present application, when evaluating the fault handling priority, an artificial intelligence path optimization algorithm and a multi-factor weighted scoring model can be combined to dynamically predict the task load and reachable time of each repair personnel. By introducing an algorithm such as genetic algorithm or ant colony algorithm to optimize the inspection path of fault points and the resource allocation efficiency, the system can reasonably schedule repair resources in the scenario of concurrent multi-point faults, and evaluate the impact of different handling schemes on the overall repair efficiency, so as to dynamically adjust the fault handling priority and improve the emergency response capability.

[0080] In a preferred embodiment of the present application, the process of evaluating the fault handling priority according to the influence range of the current fault, the electricity emergency degree and the spatial position of the repair resource includes: obtaining the position information and the influence area range of each fault point by using the power distribution network geographic information system; combining the real-time position of the repair personnel and the traffic condition to calculate the expected time for reaching each fault point; further combining the number of users involved in the fault area, the user type (such as whether it is a hospital, an important enterprise, etc.) and the electricity urgency to set the weights of the time factor, the distance factor and the urgency factor, and to calculate the priority value by using the weighted scoring method.

[0081] The beneficial effects of the preferred technical solution are: through multi-source data fusion to evaluate the comprehensive handling priority of the fault, not only the response efficiency in terms of geography and time is considered, but also the urgency of user demand is introduced as an important decision basis to ensure the fairness and stability of power supply service; the scoring mechanism has dynamic adjustable ability, which can flexibly adjust the parameters according to the actual scene to improve the scheduling agility and the rationality of fault handling when the system responds to emergencies.

[0082] In step 5, the screening and output of the target operation and maintenance strategy suitable for the current fault and its precautions include: based on the similarity between the current fault and the historical fault entity, and the effect score of the corresponding operation and maintenance strategy, calculating the recommendation score of each candidate strategy;

[0083] The candidate strategies are sorted according to the recommendation score, and a certain number of strategies with high scores are screened according to the set threshold of the recommended number;

[0084] And combined with the available state of the repair resource and the strategy application condition, the recommended result of the target strategy is output.

[0085] In an optional embodiment of the present application, when screening and recommending the operation and maintenance strategy suitable for the current fault, a reinforcement learning model can be introduced to optimize the recommended strategy online. The system takes the current fault state as the environment input, the candidate strategy as the selectable action, and the strategy score as the immediate reward value, and through the continuous interaction and feedback mechanism, trains the selection model of the recommended strategy, so as to continuously improve the consistency between the recommended result and the actual execution effect, and realize the adaptive update and enhancement of the strategy recommendation.

[0086] In a preferred embodiment of the present application, the process of ranking candidate repair strategies and outputting recommended results by using the comprehensive strategy effect score and the fault handling priority includes: first, jointly calculating each candidate strategy according to the similarity between the historical fault associated with the candidate strategy and the current fault, the strategy effect score, and the handling priority of the current fault to form a comprehensive recommendation score; then, according to a preset strategy quantity threshold, screening strategies with high scores from the score results, and under the premise of meeting the repair resource allocation feasibility and the strategy applicability conditions, outputting the target strategy and the corresponding operation precautions.

[0087] The preferred technical solution has the beneficial effects that: in the strategy screening process, the historical data evaluation, real-time fault characteristics, and resource allocation state are fully integrated to ensure the consistency of the recommended results in terms of theoretical effectiveness and execution feasibility; the comprehensive scoring mechanism enhances the scientificity of the recommendation ranking, enabling the system to make more targeted recommendations under the condition of multiple strategy competition, further improving the repair efficiency and the applicability accuracy of the repair strategy.

[0088] Embodiment 2, refer to Figures 1-7 For an embodiment of the present application, a power distribution network fault differentiated operation and maintenance repair decision recommendation method based on a knowledge graph is provided based on the previous embodiment, including:

[0089] S1: Based on the real-time monitoring data of the power distribution network, the power distribution network fault data is sorted and classified, and the key information including the fault equipment type, the fault occurrence time, the fault duration, and the fault affected equipment is extracted, the fault equipment type includes circuit breaker fault, line overload, and insulation damage, and the invalid and error data is removed, the data is standardized and formatted, and the power distribution network fault feature data set is obtained;

[0090] S2: Based on the power distribution network fault feature data set, the location of the power grid fault is located through the geographic information system, the distance between the fault power grid equipment and the nearest repair personnel is analyzed, the handling priority of each power grid fault is evaluated in combination with the fault influence range and the emergency degree of power consumption, and the fault repair priority information is obtained;

[0091] S3: Based on the power distribution network fault feature data set, the device nodes and edge attributes of the power distribution network knowledge graph are analyzed in combination with the power distribution network knowledge graph, the matching degree between the current fault characteristics and the known device nodes and edge attributes of the power distribution network knowledge graph is analyzed, the similarity is calculated in combination with the association between the fault type and the power distribution network equipment type, the similarity between the current fault and the known fault entity is evaluated, the fault entities with similarity exceeding a preset threshold are screened, and the similar fault identification result is obtained;

[0092] S4: Based on the similar fault identification result, extract the operation and maintenance repair strategy record of the similar fault, including extracting the execution cost, time and success probability of the operation and maintenance repair strategy, evaluating the effect score of each operation and maintenance repair strategy, and obtaining the repair strategy effect evaluation information;

[0093] S5: Based on the repair strategy effect evaluation information and the similar fault identification result, according to the similarity of each fault entity and the effect score of the operation and maintenance repair strategy, evaluate the recommendation score of each operation and maintenance repair strategy, sort the operation and maintenance repair strategies, select a target number of operation and maintenance repair strategies, and obtain the selected operation and maintenance repair strategies;

[0094] S6: Based on the selected operation and maintenance repair strategy and the fault repair priority information, collect the matters needing attention of the operation and maintenance repair strategy, including power grid operation safety matters, repair equipment needed and corresponding power distribution network equipment operation matters, and according to the fault repair priority, implement the power distribution network repair strategy recommendation to the repair personnel, and obtain the repair strategy recommendation information.

[0095] The power distribution network fault feature data set includes fault occurrence time, duration, fault device type, associated device and associated device running state data, fault repair priority information includes fault geographical location, influence range, emergency degree sorting and distance between repair personnel and fault location, similar fault identification result includes fault type, fault mode matching degree, fault occurrence environment condition and fault case comparison information, repair strategy effect evaluation information includes average cost information, expected time consumption information, success rate statistical information and expected benefit of repair strategy, selected operation and maintenance repair strategy includes selected strategy measures, applicable conditions, expected solution efficiency and strategy priority, repair strategy recommendation information includes safety operation procedures, detailed list of required repair equipment, function information of each equipment and repair operation step information.

[0096] Please refer to Figure 2 , based on the real-time monitoring data of the power distribution network, the power distribution network fault data is sorted and classified, and the key information including the fault device type, the fault occurrence time, the fault duration and the fault affected device is extracted, the fault device type includes the circuit breaker fault, the line overload and the insulation damage, and the invalid and error data is removed, the data is standardized and the format is unified, and the steps of obtaining the power distribution network fault feature data set are as follows:

[0097] S101: Based on the real-time monitoring data of the power distribution network, the data records associated with the fault are selected, including the type of the fault device, the time of the fault occurrence, the duration of the fault and the list of the devices affected by the fault, and the basic fault information is obtained;

[0098] Based on the real-time monitoring data of the power distribution network, the data records associated with the fault are screened, and the type of the faulty equipment can be screened through the identification number of the equipment. The screening method includes matching the equipment number with the known fault equipment type, the time of the fault occurrence is determined by the timestamp in the event log, usually involving the calibration of the time format of the log file and the system time, the duration of the fault is calculated by comparing the start and end timestamps of the fault, in addition, the list of devices affected by the fault is established by analyzing the network connection and dependency graph between devices, for example, if a main transformer fails, all downstream devices that depend on this transformer will be marked as affected, the process involves analysis of the network topology and logical judgment of device dependencies, and the data records are sorted into basic fault information.

[0099] S102: Based on the basic fault information, invalid and error records are removed through data cleaning processing, including non-fault time records and format error data entries, to optimize the accuracy and availability of the data, and to obtain the cleaned fault information;

[0100] Based on the basic fault information, invalid and error records are removed through data cleaning processing, for example, the screening of non-fault time records depends on the comparison of the fault timestamp and the maintenance and repair time recorded in the log, if the fault timestamp overlaps with the maintenance time, the record is marked as invalid, format error data entries are usually identified by regular expression detection, the pre-defined standard of data format is matched with the actual recorded format, and the records that do not match are considered as format errors and excluded, the process usually needs to be executed with automated scripts to improve processing speed and accuracy, after cleaning the data, the accuracy and availability of the data can be significantly improved, and the cleaned fault information is obtained.

[0101] S103: Based on the cleaned fault information, unified formatting processing is performed, including unified time format, equipment number and fault type marking, to optimize the consistency of data format, and to obtain the power distribution network fault feature data set;

[0102] Based on the cleaned fault information, unified formatting processing is performed, for example, the unification of time format involves converting various time data from different sources into a unified standard format, the unification of equipment number requires checking and updating the information in the equipment database to ensure that the equipment numbers in all records are up-to-date and consistent, and the unification of fault type marking is achieved by establishing a standardized classification system containing all fault types, each fault type has a corresponding code, and the code is marked in a consistent manner in the record, through this formatting processing, the consistency of the data and the efficiency of the subsequent processing can be greatly improved.

[0103] Please refer to Figure 3, based on the power distribution network fault feature dataset, through the geographic information system, the location of the power grid fault is located, the distance between the fault power grid equipment and the nearest repair personnel is analyzed, combined with the fault influence range and the emergency degree of power consumption, the processing priority of each power grid fault is evaluated, and the fault repair priority information is obtained. The steps are as follows:

[0104] S201: Based on the power distribution network fault feature dataset, the geographic information system is used to input the location coordinates of the fault occurrence, record the geographic position of each fault point, and obtain the fault position dataset;

[0105] Based on the power distribution network fault feature dataset, the geographic information system is used to input the coordinate data of the fault occurrence position into the geographic information system, and the accurate position of these coordinates on the map is marked through the spatial analysis function of the geographic information system. Such positioning not only depends on the accurate input of fault data, but also needs the matching accuracy of the map data and coordinate system built in the geographic information system software. For example, the latitude and longitude data of the coordinates need to be converted into standard coordinates in the geographic information system through a conversion formula to ensure the geographic accuracy of the data. In addition, the geographic information system can also provide relative views and environmental information between geographic positions, such as important facilities and natural geographic environment nearby. The recorded geographic position of each fault point constitutes the fault position dataset.

[0106] S202: Based on the fault position dataset, according to the position of the repair personnel, the distance and time evaluation is carried out, the distance from the nearest repair personnel to the fault site and the estimated arrival time are analyzed, and the distance and time evaluation data are obtained;

[0107] Based on the fault position dataset, according to the position of the repair personnel, the shortest path and the estimated arrival time from the current position of the repair personnel to the fault site are calculated. The calculation process usually uses path planning algorithms such as Dijkstra or A* algorithm. The algorithm can consider various road condition information, including road type, traffic condition and distance, so as to provide an optimized route in terms of time and distance. In addition, the calculation of the estimated arrival time also needs to consider the average speed of the vehicle and possible delay factors such as traffic congestion during peak hours. When calculating the parameters, real-time traffic monitoring data can be used to adjust the estimated time to ensure its accuracy as much as possible. This can provide effective travel route and time arrangement for the repair team to ensure that they can reach the fault site in the shortest time.

[0108] S203: Based on the distance and time evaluation data, combined with the influence range of the power grid fault and the emergency degree of power consumption in the range, the processing priority of each power grid fault is evaluated, and the fault repair priority information is obtained.

[0109] The formula for evaluating the processing priority of each power grid fault is:

[0110]

[0111] where P is the handling priority of the power grid fault, d represents the actual distance from the nearest repair personnel to the fault location, v represents the average moving speed of the repair personnel, w d represents the weight coefficient of distance, T u represents the emergency time from the alarm to the current, T max represents the maximum allowed response time, w t represents the weight coefficient of time, I represents the emergency degree of electricity use within the influence range of the power grid fault, w i represents the weight coefficient of influence range, and P represents the fault handling priority value.

[0112] Formula:

[0113]

[0114] Formula parameter details and acquisition method:

[0115] d represents the actual distance from the nearest repair personnel to the fault location. The value can be calculated by a geographic information system by inputting the coordinates of the repair personnel and the fault location, and using the path analysis tool of the GIS software to measure the actual distance between the two points.

[0116] v represents the average moving speed of the repair team. The value is usually estimated according to the vehicle type and road conditions, for example, it can be set to 30 kilometers per hour in the city.

[0117] w d represents the weight coefficient of distance. It is an empirical value that can be determined according to past fault handling records and repair efficiency analysis, such as setting it to 0.5, indicating that distance plays a certain proportion in priority evaluation.

[0118] T u represents the emergency time from the alarm to the current. This value is the difference between the timestamp obtained from the fault alarm system and the current time, used to measure the emergency degree after the fault occurs.

[0119] T max represents the maximum allowed response time, which is usually set according to power grid operation standards and regulations, such as 2 hours.

[0120] w t represents the weight coefficient of time. Similar to w d , it is also an empirical value based on historical data analysis, set to 0.3, indicating the importance of time urgency.

[0121] I represents the urgency of power supply within the affected area of ​​a power grid failure. The value can be obtained from a failure impact analysis system and calculated based on the number of affected users and the extent to which critical infrastructure is impacted.

[0122] w i The weighting coefficient representing the scope of impact is set based on statistical data of power grid operation and expert opinions. For example, it can be set to 0.2 to reflect the impact of the fault's scope on priority.

[0123] Calculation example:

[0124] Settings: Distance from the repair personnel to the fault location d = 15 km; average speed of the repair team v = 30 km / h; time T from the alarm to the present. u = 45 minutes; Maximum allowable response time T max = 120 minutes; the urgency level of the fault's impact area I = 0.75 (assuming this is a scale from 0 to 1, where 1 represents the most urgent).

[0125] Use formula Perform the calculation:

[0126] Calculation of distance factor:

[0127] Calculation of time urgency factors:

[0128] Calculation of influencing factors: 0.75 × 0.2 = 0.15;

[0129] Adding these three parts together gives the final processing priority P:

[0130] P=0.25+0.1125+0.15=0.5125;

[0131] The result indicates a fault handling priority of 0.5125, with a higher value indicating a higher priority for handling the fault. This score helps the emergency response team decide which faults should be addressed first, especially when multiple faults occur simultaneously.

[0132] Please see Figure 4 Based on a distribution network fault feature dataset and combined with a distribution network knowledge graph, this paper analyzes the device nodes and edge attributes of the distribution network knowledge graph, analyzes the matching degree between the current fault features and the known device nodes and edge attributes of the distribution network knowledge graph, and calculates similarity by combining the association between fault type and distribution network equipment type. The paper evaluates the similarity between the current fault and known fault entities, and filters out fault entities with similarity exceeding a preset threshold to obtain similar fault identification results. The specific steps are as follows:

[0133] S301: Based on the power distribution network fault feature dataset, a comparison framework of fault equipment type, occurrence time and duration information is established, the records of associated equipment nodes and edge attributes in the knowledge graph are retrieved, and the fault feature and equipment node comparison data are obtained;

[0134] Based on the power distribution network fault feature dataset, a comparison framework of fault equipment type, occurrence time and duration information is established, the framework relies on the integration of fault logs and historical fault data, the fault equipment type and the corresponding fault occurrence time and duration are extracted from the dataset, the extraction process involves data query and screening operations, among which SQL query statements or similar technologies are used to locate and extract relevant information, then the data is used to construct a multidimensional data model which includes the association of fault type, time and duration, the records of associated equipment nodes and edge attributes in the knowledge graph are retrieved, which involves graph database traversal algorithms such as depth-first search or breadth-first search to find historical fault instances matching the current fault features, and the fault feature and equipment node comparison data are obtained, which is crucial for understanding the pattern of the fault and preventing similar events in the future.

[0135] S302: Based on the fault feature and equipment node comparison data, analyze the similarity of the fault feature and the device type and fault record in the existing knowledge graph, set the similarity measurement standard, calculate the similarity, and obtain the fault similarity analysis result;

[0136] Based on the fault feature and equipment node comparison data, analyze the similarity of the fault feature and the device type and fault record in the existing knowledge graph, set the similarity measurement standard, which usually involves selecting appropriate similarity calculation formulas such as cosine similarity, Jaccard similarity or Pearson correlation coefficient, the standard considers multiple dimensions of data such as fault type, occurrence time and duration, by calculating the similarity between different fault instances, the calculation process needs to handle a large number of data points, usually requiring efficient algorithms and possible optimization measures such as parallel computing or data index optimization, through similarity calculation, the fault similarity analysis result can be obtained, which helps to identify similar historical events to the current fault, providing basis for subsequent fault handling and prevention.

[0137] S303: Based on the fault similarity analysis result, filter the fault entities with similarity higher than the preset threshold, organize the data, and obtain the similar fault recognition result;

[0138] Based on the similarity analysis results, filter the fault entities with similarity higher than a preset threshold. First, a reasonable similarity threshold needs to be set during the filtering process. The threshold is set based on historical data and expert experience to ensure that only highly relevant fault instances are considered. The threshold is used to filter the similarity results, which usually involves writing a specific filtering algorithm or using database query functions to exclude records with a similarity lower than the threshold. The filtered data needs to be further organized for better understanding and use, which includes data cleaning, formatting, and possible reclassification. Through this organization process, similar fault identification results can be obtained.

[0139] Please refer to Figure 5 Based on the similar fault identification results, extract the operation and maintenance repair strategy records of similar faults, including the execution cost, time and success probability of the operation and maintenance repair strategy, evaluate the effect score of each operation and maintenance repair strategy, and obtain the steps of the repair strategy effect evaluation information:

[0140] S401: Based on the similar fault identification results, extract the operation and maintenance repair records related to similar faults from the knowledge graph, extract the execution cost and time of each repair, and obtain the repair record dataset;

[0141] Based on the similar fault identification results, extract the operation and maintenance repair records related to similar faults from the knowledge graph, and the extraction process uses the query and analysis functions of the knowledge graph to query all related nodes and edges of similar faults through graph query languages such as SPARQL or Cypher, and filter out the operation and maintenance repair record nodes directly related to the fault from them. The records include detailed data of historical repair events, such as repair time, execution cost, materials and personnel used, etc. The information is summarized and organized through programming scripts or automated tools to ensure the accuracy and availability of the data, especially the accurate recording of execution cost and time, which is crucial for subsequent analysis. Obtain the repair record dataset.

[0142] S402: Based on the repair record dataset, evaluate the success probability of each operation and maintenance repair strategy in the case records, calculate the average execution cost, success rate and time of each operation and maintenance repair strategy, and obtain the average data analysis results;

[0143] Based on the repair record dataset, the success probability of each operation and maintenance repair strategy in the case record is evaluated, the total number of cases and the number of successful cases of each repair strategy are counted, which is completed through database query and data processing script, and the success probability, average execution cost and average time of each strategy are calculated using data, which is usually calculated by simple mathematical formulas such as average value calculation and proportion calculation. The success probability is the number of successful cases divided by the total number of cases, and the average cost and time are the average values of all related case costs and times. Through calculation, the average data analysis result is obtained, which not only shows the economic and time efficiency of each repair strategy, but also provides the success rate in actual operation, providing data support for strategy optimization.

[0144] S403: Based on the average data analysis result, the repair strategy is evaluated according to the average cost, time and success probability of each strategy, the effect score of each operation and maintenance repair strategy is evaluated, and the repair strategy effect evaluation information is obtained.

[0145] The formula for evaluating the effect score of each operation and maintenance repair strategy is:

[0146]

[0147] Where S is the effect score of the operation and maintenance repair strategy, C represents the average execution cost of the strategy, C max represents the highest repair cost in all strategies, J represents the average execution time of the strategy, J max represents the longest repair time in all strategies, R represents the average success rate of the strategy, and w c , w J and w r are weight coefficients.

[0148] Formula:

[0149]

[0150] Parameter details and acquisition method: C represents the average execution cost of the strategy, which is calculated by averaging the cost data of all related repair activities in the repair record dataset. C max : indicates the highest repair cost in all strategies, which is the maximum cost value directly obtained from the repair record dataset, used for cost effect standardization. J represents the average execution time of the strategy, which is calculated by collecting the time data of all related repair activities in the repair record and then averaging. J max : represents the longest repair time in all strategies, which is the longest time value directly obtained from the repair record dataset, used for time effect standardization. R represents the average success rate of the strategy, which is calculated by counting the success and failure times in the repair record. w c , w J and wr : weight coefficients of cost, time and success rate respectively, usually set according to management experience and priority, for example, w c = 0.3, w J = 0.3 and w r = 0.4, reflecting the importance of each index in the total score.

[0151] Calculation example:

[0152] Set the following data: average execution cost C = 5000 yuan; highest repair cost C max = 8000 yuan; average execution time J = 120 minutes; longest repair time J max = 180 minutes; success rate R = 0.85 (i.e. 85% success rate).

[0153] Use the formula to calculate:

[0154] Cost score: points;

[0155] Time score: points;

[0156] Success rate score: points

[0157] The comprehensive score S is 900 + 18 + 0.34 = 918.34 points.

[0158] The score S = 918.34 shows the overall effect of the strategy, and a higher score indicates that the strategy performs well in terms of cost, time and success rate, and is suitable as a priority repair strategy.

[0159] Please refer to Figure 6 , based on the repair strategy effect evaluation information and the similar fault identification result, according to the similarity of each fault entity and the effect score of the operation and maintenance repair strategy, the recommended score of each operation and maintenance repair strategy is evaluated, the operation and maintenance repair strategies are sorted, a target number of operation and maintenance repair strategies are screened, and the steps of the screened operation and maintenance repair strategies are as follows:

[0160] S501: Based on the repair strategy effect evaluation information and the similar fault identification result, the similarity of each fault entity and the effect score of the corresponding repair strategy are analyzed, the recommended score of each repair strategy is evaluated, and the recommended score data is obtained;

[0161] Based on the repair strategy effect evaluation information and the similar fault recognition result, the similarity of each fault entity and the effect score of the corresponding repair strategy are analyzed. The analysis depends on the fusion processing of the two groups of data, that is, matching the fault entity in the similar fault recognition result with the repair strategy effect evaluation information. The matching algorithm finds the corresponding item of each fault entity in the repair record and obtains the effect score of the related repair strategy. The process involves data query and correlation analysis, and usually needs to use SQL query or data processing script to realize. After obtaining the matching result, the recommended score of each repair strategy is evaluated by weighted calculation, including calculating the average effect score corresponding to each strategy and the number of similar fault entities, so as to evaluate the applicability and effect of the repair strategy under similar fault conditions. The calculation provides decision support and helps to optimize the repair response. The recommended score data is obtained, which provides a basis for subsequent strategy selection and resource allocation.

[0162] S502: Based on the recommended score data, the operation and maintenance repair strategy is sorted, and the strategy is prioritized according to the score, to obtain the strategy priority sorting result;

[0163] Based on the recommended score data, the operation and maintenance repair strategy is sorted, and the strategy is prioritized according to the score, to obtain the strategy priority sorting result;

[0164] S503: Based on the strategy priority sorting result, according to the actual demand and resource situation of the repair personnel, a target number of repair strategies are selected, and the repair strategies with high priority are selected, to obtain the screened operation and maintenance repair strategy;

[0165] Based on the strategy priority sorting result, according to the actual demand and resource situation of the repair personnel, a target number of repair strategies are selected, and the repair strategies with high priority are selected, to obtain the screened operation and maintenance repair strategy;

[0166] Please refer to Figure 7, based on the screened operation and maintenance repair strategy and fault repair priority information, collect the notes of the operation and maintenance repair strategy, including power grid operation safety matters, required repair equipment and corresponding distribution network equipment operation matters, and implement distribution network repair strategy recommendation for repair personnel according to the fault repair priority, and the steps of obtaining the repair strategy recommendation information are as follows:

[0167] S601: Based on the screened operation and maintenance repair strategy and fault repair priority information, collect the operation matters included in each strategy, including power grid operation safety matters and required repair equipment list, and obtain the repair strategy safety and equipment list;

[0168] Based on the screened operation and maintenance repair strategy and fault repair priority information, collect the operation matters included in each strategy, including power grid operation safety matters and required repair equipment list, and collect the process involving cross-department cooperation, safety matters provided by the safety management department, to ensure that all repair operations comply with the latest safety specifications, usually including wearing appropriate personal protective equipment, ensuring that all tools and equipment are in good condition, and complying with specific safety protocols for high-voltage operations, etc. The required repair equipment list is prepared by the material supply department according to the fault type and specific requirements of the repair strategy, including necessary tools, replacement parts and special equipment. The equipment must be inspected and prepared before repair to avoid equipment shortage or failure in emergency situations. After integrating the information, the repair strategy safety and equipment list is formed, providing all the information and resources required for the repair team to perform the task, ensuring that they can safely and effectively complete the task.

[0169] S602: Based on the repair strategy safety and equipment list, combined with the repair priority of each fault, sort the repair operations of each fault in priority order to obtain the priority sorted repair operation list;

[0170] Based on the repair strategy safety and equipment list, combined with the repair priority of each fault, sort the repair operations of each fault in priority order to obtain the priority sorted repair operation list. The sorting process takes into account the impact and urgency of the fault, and determines which faults need to be handled first by analyzing the data provided by the system. It usually involves algorithm sorting, such as priority queue or heap sorting, to ensure the correctness and efficiency of the processing order. In addition, the availability of resources is also considered, such as the location and time arrangement of the repair team, and the preparation of the required equipment, to ensure the flow and standardization of the repair operation process.

[0171] S603: Based on the priority sorted repair operation list, provide repair strategy recommendation for repair personnel through data transmission display, including operation steps, required equipment and safety matters, and obtain repair strategy recommendation information;

[0172] Based on the priority ranking repair operation list, through data transmission display, the repair personnel are provided with repair strategy recommendations, including operation steps, required equipment and safety matters, the information transmission process is carried out through management software or mobile application, ensuring that all repair personnel can receive the latest operation information in real time, including detailed step-by-step instructions, equipment usage guide and necessary safety warnings, and the recommended information can also be dynamically updated according to the changes of the field conditions to adapt to the sudden situation or resource changes, ensuring that the repair personnel maintain the highest efficiency and safety throughout the operation process, and the repair strategy recommendation information provides strong support for field operation, ensuring that the fault is quickly and safely handled.

[0173] Embodiment 3 is an embodiment of the present application, which provides a power distribution network fault differentiated operation and repair decision recommendation system based on a knowledge graph, comprising:

[0174] A data acquisition module is configured to acquire real-time monitoring data of the power distribution network, process information related to faults, and construct a fault feature data set containing fault types, device characteristics and operating states.

[0175] A graph recognition module is configured to recognize historical fault entities that meet a preset similarity threshold in fault type, device attribute or occurrence condition based on the fault feature data set and in combination with a power distribution network knowledge graph.

[0176] A strategy evaluation module is configured to extract corresponding operation and repair strategies from the historical fault entities, preset evaluation indicators, and calculate strategy effect scores.

[0177] A priority analysis module is configured to evaluate fault handling priority according to the influence range of the current fault, the emergency degree of electricity use and the spatial position of repair resources.

[0178] A decision recommendation module is configured to sort, filter and output target operation strategies and matters needing attention suitable for the current fault by comprehensively considering the strategy effect scores and the fault handling priority.

[0179] The embodiment also provides an electronic device suitable for the power distribution network fault differentiated operation and repair decision recommendation method based on a knowledge graph, comprising a memory and a processor; the memory is configured to store computer executable instructions, and the processor is configured to execute the computer executable instructions to implement the power distribution network fault differentiated operation and repair decision recommendation method based on a knowledge graph proposed in the above embodiment.

[0180] The embodiment also provides a storage medium having a computer program stored thereon, which is executed by a processor to implement the power distribution network fault differentiated operation and repair decision recommendation method based on a knowledge graph proposed in the above embodiment.

[0181] The storage medium proposed in the embodiment belongs to the same inventive concept as the method for realizing the differentiated operation and maintenance repair decision recommendation of the power distribution network fault based on the knowledge graph proposed in the above embodiment. The technical details not described in the embodiment can be seen from the above embodiment, and the embodiment has the same beneficial effects as the above embodiment.

[0182] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary general hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, and the computer software product can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk or an optical disk, etc., including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the method of each embodiment of the present application.

[0183] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A knowledge graph-based method for differentiated operation and maintenance emergency repair decisions in distribution networks, characterized by: include, Acquire real-time monitoring data of the power distribution network, process fault-related information, and construct a fault feature dataset that includes fault type, equipment characteristics, and operating status; Based on the fault feature dataset and combined with the distribution network knowledge graph, historical fault entities that meet the preset similarity threshold in terms of fault type, equipment attributes, or occurrence conditions with the current fault are identified. Extract the corresponding operation and maintenance emergency repair strategies from the historical fault entities, preset evaluation indicators, and calculate the strategy effectiveness score; Assess the priority of fault handling based on the current impact range of the fault, the urgency of power demand, and the spatial location of emergency repair resources; Based on the comprehensive strategy effectiveness score and fault handling priority, candidate emergency repair strategies are ranked, and the target operation and maintenance strategy applicable to the current fault and its precautions are selected and output.

2. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 1, characterized in that: The processing of fault-related information includes extracting the type of faulty device, the time of occurrence, the duration, and the affected devices related to the fault. The information is cleaned to remove non-faulty time records and entries with abnormal formats; It also standardizes the time format, equipment number, and type identifier of fault data to generate a structured standard fault feature dataset.

3. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 2, characterized in that: The historical fault entities that meet the preset similarity threshold include those that establish a feature comparison framework based on the type of faulty equipment, the time of fault occurrence, and the duration of fault, and retrieve nodes and their edge attributes related to the equipment in the power distribution network knowledge graph. The similarity value is calculated by comparing the feature information of the current fault with that of historical faults according to the set similarity measurement standard. Historical fault entities with similarity values ​​higher than a preset threshold are selected as similar fault identification results.

4. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 3, characterized in that: The strategy effectiveness evaluation includes extracting historical operation and maintenance emergency repair records related to the similar faults from the distribution network knowledge graph, and obtaining statistical results of the execution cost, execution time and success of each operation and maintenance strategy in multiple historical cases. The average value of each indicator is used as the basis for calculating the strategy effectiveness score.

5. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 4, characterized in that: The calculation of the strategy effectiveness score includes, for each operation and maintenance strategy, calculating the difference between its average execution cost and the maximum execution cost among all candidate strategies, and multiplying the difference by the weighting coefficient of the cost index. Calculate the difference between the average execution time of the strategy and the maximum execution time among all strategies, and multiply the difference by the weighting coefficient of the time metric; At the same time, the average success rate of this strategy is multiplied by the weighting coefficient of the success rate indicator; Add the three calculation results together and divide the sum by the sum of the weight coefficients of the three indicators to obtain the final performance score of the strategy.

6. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 5, characterized in that: The assessment of fault handling priorities includes obtaining the geographical coordinates of each fault point based on the geographic information system of the power distribution network, and calculating the estimated distance and time required for the repair personnel to reach each fault point in combination with their current location. At the same time, multiple priority impact factors are set according to the number of users within the affected area of ​​each fault and the urgency of the power load; By weighting the above factors according to preset weighting coefficients, a processing priority value for each fault is generated.

7. The knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in claim 6, characterized in that: The process of filtering and outputting target operation and maintenance strategies and their considerations applicable to the current fault includes calculating a recommendation score for each candidate strategy based on the similarity between the current fault and historical fault entities, as well as the effectiveness score of the corresponding operation and maintenance strategy. Candidate strategies are sorted according to their recommendation scores, and the top-scoring strategies are selected based on a set threshold for the number of recommendations. Based on the availability of emergency repair resources and the applicable conditions of the strategy, the system outputs a recommended result as the target strategy.

8. A knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation system, employing the knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in any one of claims 1 to 7, characterized in that, include: The data acquisition module is used to acquire real-time monitoring data of the power distribution network, process fault-related information, and construct a fault feature dataset containing fault type, equipment characteristics, and operating status. The graph recognition module is used to identify historical fault entities that meet a preset similarity threshold in terms of fault type, equipment attributes, or occurrence conditions, based on the fault feature dataset and in combination with the distribution network knowledge graph. The strategy evaluation module is used to extract the corresponding operation and maintenance emergency repair strategies from the historical fault entities, preset evaluation indicators, and calculate the strategy effectiveness score. The priority analysis module is used to assess the priority of fault handling based on the current impact range of the fault, the urgency of power use, and the spatial location of emergency repair resources; The decision recommendation module is used to combine the comprehensive strategy effectiveness score and fault handling priority to sort the candidate emergency repair strategies, filter and output the target operation and maintenance strategy applicable to the current fault and its precautions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the knowledge graph-based distribution network fault differentiated operation and maintenance emergency repair decision recommendation method as described in any one of claims 1 to 7.

Images