Fault repair method and device, electronic equipment, storage medium and product

Abstract

The invention provides a fault repair method and device, electronic equipment, a storage medium and a product, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining monitoring data of an electric power communication system; inputting the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; if it is determined that the power communication system has a fault based on the fault detection result, inputting the monitoring data to a fault positioning model, and obtaining a fault positioning result output by the fault positioning model; and based on the fault positioning result, performing fault repair on the electric power communication system. By means of the mode, automation of fault detection and fault positioning can be achieved through the model, manual intervention in the fault repairing process is avoided, and the labor cost of fault repairing is reduced.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a fault repair method, device, electronic equipment, storage medium and product. Background Art

[0002] The safe and stable operation of the power communication system is of great significance in the power grid field. When a power communication system fails, it needs to be repaired in time to avoid safety accidents.

[0003] The fault repair process of the power communication system includes three parts: fault detection, fault location and fault repair: real-time monitoring of the status data (such as voltage, current, etc.) of all power communication equipment in the power communication system, fault detection based on real-time monitoring data, and judgment of whether there is a fault in the power communication system; if there is a fault in the power communication system, it is necessary to locate the fault, determine the deployment position of the faulty power communication equipment in the power communication system, and repair the fault.

[0004] As the level of intelligence and networking of power communication systems continues to increase, the complexity of power communication systems is also increasing. Since the existing power communication systems are relatively complex, in order to ensure system safety, the fault repair process of the power communication system still requires manual intervention, which is difficult to fully automate and the labor cost of fault repair is high. Summary of the invention

[0005] The present invention provides a fault repair method, device, electronic device, storage medium and product, which are used to solve the defects in the prior art that the fault repair process still requires manual intervention, is difficult to fully realize automation, and has high labor costs for fault repair.

[0006] The present invention provides a fault repair method, comprising: acquiring monitoring data of an electric power communication system; inputting the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; if it is determined based on the fault detection result that a fault exists in the electric power communication system, inputting the monitoring data into a fault location model to obtain a fault location result output by the fault location model; based on the fault location result, performing fault repair on the electric power communication system; wherein the fault detection model is obtained by training based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; and the fault location model is obtained by training based on sample monitoring data and sample fault location result labels corresponding to the sample monitoring data.

[0007] According to a fault repair method provided by the present invention, based on the fault location result, the electric power communication system is repaired, including: based on the fault location result and the business knowledge graph, determining the fault business information of the electric power communication system; based on the fault business information, generating a repair business list; the repair business list includes information of the electric power communication equipment to be repaired; based on the repair business list, the electric power communication system is repaired.

[0008] According to a fault repair method provided by the present invention, before determining the fault business information of the electric power communication system based on the fault location result and the business knowledge graph, it also includes: acquiring business data related to the electric power communication system business; performing knowledge extraction processing on the business data to obtain entity information, entity relationship information and entity attribute information in the business data; performing knowledge fusion processing on the entity information, entity relationship information and entity attribute information to obtain the business knowledge graph.

[0009] According to a fault repair method provided by the present invention, a fault location model includes an input layer, a first network layer, a second network layer, a third network layer and an output layer; before the monitoring data is input into the fault location model and the fault location result output by the fault location model is obtained, it also includes: obtaining sample monitoring data and sample fault location result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data, hybrid training is performed on the first initial model to obtain a fault location model; wherein the input layer is trained based on the error back propagation training method; and the third network layer is trained based on the least squares method.

[0010] According to a fault repair method provided by the present invention, before inputting monitoring data into a fault detection model and obtaining the fault detection result output by the fault detection model, it also includes: obtaining sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault detection result labels corresponding to the sample monitoring data, training a second initial model constructed based on a decision tree algorithm to obtain a fault detection model.

[0011] According to a fault repair method provided by the present invention, monitoring data of an electric power communication system is obtained, including: obtaining original monitoring data of the electric power communication system; preprocessing the original monitoring data to obtain monitoring data; wherein the monitoring data includes real-time status data of each electric power communication device in the electric power communication system, and the real-time status data includes one or more of voltage data, current data, power data, frequency data and temperature data.

[0012] The present invention also provides a fault repair device, including: an acquisition module, used to acquire monitoring data of an electric power communication system; a fault detection module, used to input the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; a fault location module, used to input the monitoring data into a fault location model if it is determined based on the fault detection result that a fault exists in the electric power communication system, to obtain a fault location result output by the fault location model; a repair module, used to repair the electric power communication system based on the fault location result; wherein the fault detection model is trained based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; and the fault location model is trained based on sample monitoring data and sample fault location result labels corresponding to the sample monitoring data.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, any of the above-mentioned fault repair methods is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, any of the above-mentioned fault repair methods is implemented.

[0015] The present invention also provides a computer program product, including a computer program, and when the computer program is executed by a processor, any of the above-mentioned fault repair methods is implemented.

[0016] The fault repair method, device, electronic device, storage medium and product provided by the present invention obtain the monitoring data of the power communication system; input the monitoring data into the fault detection model to obtain the fault detection result output by the fault detection model; if it is determined based on the fault detection result that the power communication system has a fault, the monitoring data is input into the fault location model to obtain the fault location result output by the fault location model; based on the fault location result, the power communication system is fault repaired. Through the above method, the monitoring data of the power communication system is first input into the fault detection model to obtain the fault detection result. After determining that the power communication system has a fault, the monitoring data is input into the fault location model to obtain the fault location result, and the power communication system is fault repaired based on the fault location result. The automation of fault detection and fault location is realized through the model, avoiding manual intervention in the fault repair process, and reducing the manual cost of fault repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of the fault repair method provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the business knowledge graph ontology provided by the present invention.

[0020] Figure 3 It is a schematic diagram of the process of constructing a business knowledge graph provided by the present invention.

[0021] Figure 4 It is a flowchart of the knowledge fusion processing provided by the present invention.

[0022] Figure 5 It is a structural schematic diagram of the adaptive fuzzy inference system provided by the present invention.

[0023] Figure 6 It is a schematic diagram of a process of optimizing model parameters by using a particle swarm algorithm provided by the present invention.

[0024] Figure 7 It is a tree structure schematic diagram of the decision tree provided by the present invention.

[0025] Figure 8 It is a schematic diagram of the application flow of the decision tree model provided by the present invention.

[0026] Fig. 9 It is a structural schematic diagram of the fault repair device provided by the present invention.

[0027] Fig.10 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] See also Figure 1 , Figure 1: is a flow chart of the fault repair method provided by the present invention. In this embodiment, the fault repair method includes steps S110 to S140, and each step is as follows: S110: Acquire monitoring data of the power communication system.

[0030] Specifically, through advanced sensors and monitoring devices, the power communication equipment in the power communication system is monitored in real time to obtain monitoring data of the power communication system; the monitoring data can be transmitted to the monitoring center through the power communication network for further fault detection and fault location.

[0031] The monitoring data of the power communication system includes the real-time status data of each power communication device in the power communication system, the system topology structure of the power communication system, the fault repair record of the power communication system and other data.

[0032] Optionally, the real-time status data includes one or more of voltage data, current data, power data, frequency data and temperature data.

[0033] S120: Input the monitoring data into the fault detection model to obtain the fault detection result output by the fault detection model.

[0034] Specifically, after receiving the monitoring data, the monitoring center may input the monitoring data into the fault detection model to obtain the fault detection result output by the fault detection model.

[0035] Optionally, the fault detection result includes whether a fault exists in the electric power communication system or whether a fault does not exist in the electric power communication system.

[0036] Optionally, the fault detection result includes the fault degree of the electric power communication system.

[0037] Optionally, the fault detection result includes the fault degree and fault classification of each power communication device in the power communication system.

[0038] S130: If it is determined based on the fault detection result that a fault exists in the electric power communication system, the monitoring data is input into a fault location model to obtain a fault location result output by the fault location model.

[0039] Optionally, the fault location result includes a deployment location of the faulty power communication device (ie, the power communication device to be repaired) in the power communication system.

[0040] S140: Based on the fault location result, the power communication system is repaired.

[0041] Among them, the fault detection model is trained based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data.

[0042] The fault location model is trained based on sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data.

[0043] The fault repair method provided in this embodiment obtains monitoring data of the power communication system; inputs the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; if it is determined based on the fault detection result that the power communication system has a fault, the monitoring data is input into a fault location model to obtain a fault location result output by the fault location model; based on the fault location result, the power communication system is repaired. Through the above method, the monitoring data of the power communication system is first input into the fault detection model to obtain a fault detection result. After determining that the power communication system has a fault, the monitoring data is input into the fault location model to obtain a fault location result, and the power communication system is repaired based on the fault location result. The automation of fault detection and fault location is realized through the model, avoiding manual intervention in the fault repair process and reducing the manual cost of fault repair.

[0044] In some embodiments, based on the fault location result, the power communication system is repaired, including: based on the fault location result and the business knowledge graph, determining the fault business information of the power communication system; based on the fault business information, generating a repair business list; the repair business list includes information of the power communication equipment to be repaired; based on the repair business list, the power communication system is repaired.

[0045] Generally, the fault location result includes the deployment location of the faulty power communication equipment (ie, the power communication equipment to be repaired) in the power communication system.

[0046] In the prior art, after the fault location result is determined, it is necessary to rely on manual fault service analysis to determine which services in the power communication system will be affected by the faulty power communication equipment in order to repair the fault. However, the results of manual analysis are easily affected by human subjective factors, and are prone to problems such as misjudgment due to lack of experience and time-consuming analysis.

[0047] Based on this, this embodiment introduces knowledge graph technology into the fault repair process, and conducts comprehensive and objective fault analysis based on the knowledge graph to formulate and timely adjust corresponding repair plans to adapt to the complex and changeable communication network environment.

[0048] Specifically, after the fault location result is determined, it is possible to analyze which services in the power communication system will be affected by the faulty power communication equipment based on the fault location result and the business knowledge graph, and determine the fault business information of the power communication system.

[0049] Optionally, the fault service information includes information such as service type, channel terminal, service coordination unit, etc. that will be affected by the faulty power communication equipment.

[0050] Furthermore, based on the fault service information, a repair service list is generated.

[0051] Optionally, the repair service list includes information of power communication equipment to be repaired (ie, information of faulty power communication equipment).

[0052] Furthermore, based on the repair service list, the power communication system is repaired.

[0053] In some embodiments, before determining the fault business information of the power communication system based on the fault location results and the business knowledge graph, it also includes: acquiring business data related to the power communication system business; performing knowledge extraction processing on the business data to obtain entity information, entity relationship information and entity attribute information in the business data; performing knowledge fusion processing on the entity information, entity relationship information and entity attribute information to obtain a business knowledge graph.

[0054] Generally speaking, knowledge graphs can be logically divided into two levels: model layer and data layer.

[0055] Among them, the model layer mainly includes two types: one is "entity-attribute-attribute value"; the other is "entity-relationship-entity"; the data layer is mainly composed of a series of facts, which correspond to the model layer respectively.

[0056] Generally speaking, the construction process of a knowledge graph can be divided into five parts: ontology construction, data source configuration, knowledge extraction, knowledge mapping, and knowledge fusion.

[0057] In this embodiment, before determining the fault service information of the electric power communication system based on the service knowledge graph, it is necessary to first construct a service knowledge graph.

[0058] Specifically, ontology construction is performed first.

[0059] Ontology is the model of the knowledge graph and a constraint on the data that constitutes the knowledge graph. By combing through the knowledge, professional terms, and manual experience in the field of power grids, the construction of the business knowledge graph can be improved in combination with the application scenarios of the knowledge graph, ultimately obtaining entity categories, relationships between categories, and definitions of the attributes contained in the entities. Based on the defined data fields and actual business needs, folders and associations between folders can be established, and the acquired data can be stored in different folders in different categories to preserve the ontology of the business knowledge graph.

[0060] See also Figure 2 , Figure 2 It is a schematic diagram of the business knowledge graph ontology provided by the present invention.

[0061] like Figure 2 As shown, the business knowledge graph ontology of this embodiment includes entities such as equipment, faults, business, management units, equipment categories, fault types, warning phenomena and processing measures, as well as the relationships and attribute information between the entities.

[0062] Furthermore, data source configuration is performed.

[0063] When building a business knowledge graph, it is necessary to configure multiple data sources to collect and integrate various relevant data in the power grid field, such as power communication equipment information, system topology of the power communication system, fault repair records of the power communication system, operation and maintenance records, fault data, etc.

[0064] Generally, for local electronic documents, they need to be classified according to document type and format, parsed and organized into a format that complies with the standards.

[0065] For data from external networks, it is necessary to develop corresponding crawler programs based on the characteristics of the website, crawl relevant data from web pages through the crawler programs, and store them in the local database.

[0066] After completing data collection and data preprocessing, the consistency and availability of the data can be ensured, providing a reliable data foundation for the subsequent construction of business knowledge graphs.

[0067] After completing the ontology construction and data source configuration, you can formally start building the business knowledge graph.

[0068] See also Figure 3 , Figure 3 It is a schematic diagram of the process of constructing a business knowledge graph provided by the present invention.

[0069] like Figure 3 As shown, business data related to the power communication system business can be directly obtained from the configured data source.

[0070] Among them, the business data in the data source includes structured data, semi-structured data and unstructured data.

[0071] Furthermore, knowledge extraction is performed on the business data to obtain entity information, entity relationship information and entity attribute information in the business data.

[0072] It should be noted that knowledge extraction processing refers to extracting structured knowledge with semantic information from raw data for the convenience of storage, management and application.

[0073] In this embodiment, business data can be obtained from multiple data sources, and knowledge extraction processing can be performed on the business data to extract relevant information such as entity information, entity relationship information, and entity attribute information from the business data, such as the system topology structure, operation and maintenance records, fault data, etc. of power communication equipment, power grid or power communication system.

[0074] Specifically, for the information of power communication equipment, basic information of the power communication equipment, such as the type, model, specifications, manufacturer, etc. of the power communication equipment, can be extracted from the technical documents, specification manuals and other documents of the power communication equipment; by parsing the sensor data of the power communication equipment, the real-time status of the power communication equipment can be extracted, such as voltage data, current data, temperature data, etc.

[0075] For the system topology of the power grid or power communication system, the topology information of the power grid or power communication system can be extracted by analyzing the design drawings, CAD files, etc. of the power grid or power communication system, such as the connection relationship and layout of equipment such as lines, substations, and distribution rooms.

[0076] For operation and maintenance records, change information of the topological structure of the power grid or power communication system can be extracted from historical operation and maintenance records, such as addition, deletion, modification and other operations, to maintain the real-time nature of the business knowledge graph; maintenance records of power grid equipment can be extracted from operation and maintenance record sheets, maintenance reports and other documents, such as maintenance time, maintenance content, maintenance personnel and other information; by parsing inspection and maintenance records, the health status information of power communication equipment in the power grid or power communication system can be extracted, such as equipment aging degree, failure rate, etc.

[0077] For fault data, detailed information on power grid or power communication system faults, such as fault type, fault occurrence time, fault location, fault cause, etc., can be extracted from documents such as fault reports and log records; by parsing sensor data and monitoring system records, real-time fault alarm information and environmental data when the fault occurs, such as temperature, humidity, voltage fluctuations, etc., can be extracted.

[0078] For power grid safety management information, the power grid's safety management policies and regulations, such as safety operation procedures, accident handling procedures, etc., can be extracted from safety management manuals, procedural documents, etc.; by parsing safety event records and safety reports, statistical information on power grid safety events can be extracted, such as the number of accidents, accident cause analysis, etc.

[0079] Furthermore, knowledge mapping and knowledge fusion processing are performed on the entity information, entity relationship information and entity attribute information to obtain a business knowledge graph.

[0080] Knowledge mapping refers to mapping information in different data sources and extracted entity information, entity relationship information, entity attribute information, etc. to corresponding concepts and relationships in the knowledge graph to establish a structured representation of the knowledge graph.

[0081] In this embodiment, the knowledge mapping process in the field of power grid can integrate information from multiple data sources such as equipment information, topology, operation and maintenance records, fault data, etc. into the knowledge graph, thereby forming a complete initial knowledge graph with semantic associations.

[0082] For example, the equipment entities in the equipment information data source are mapped to the equipment nodes in the knowledge graph. The equipment nodes include substations, distribution rooms, lines, switchgear, etc., and the association relationships between the equipment are established, such as the connection relationship between the equipment, upstream and downstream relationships, etc., to complete the mapping of the equipment's technical parameters, model specifications and other attribute information.

[0083] After the knowledge mapping process, a complete initial knowledge graph with semantic association can be formed. The initial knowledge graph is a graph-structured data model used to describe entities, concepts and their relationships. In the field of power grids, the initial knowledge graph can cover entities and relationships in various links such as power generation, transmission, distribution, and power consumption to form a comprehensive knowledge network. However, the initial knowledge graph may contain a large amount of duplicate information, so knowledge fusion processing is required. Knowledge fusion processing refers to the integration, disambiguation, processing, and reasoning verification of information from different knowledge sources to remove duplicate information in order to form a high-quality business knowledge graph and knowledge graph library (i.e., knowledge base).

[0084] Specifically, Figure 3 As shown, after the entity information, entity relationship information and entity attribute information are processed through knowledge mapping and data integration, knowledge fusion processing can be performed through entity alignment and reference resolution to obtain a business knowledge graph.

[0085] Among them, entity alignment and coreference resolution are key technologies in knowledge fusion processing. Entity alignment can be used to solve the ambiguity problem caused by entities with the same name, that is, the same words have different meanings. Coreference resolution can be used to solve the problem that multiple references correspond to the same entity object.

[0086] See also Figure 4 , Figure 4 It is a flowchart of the knowledge fusion processing provided by the present invention.

[0087] like Figure 4 As shown, the process of knowledge fusion processing includes data preprocessing, block processing, record linkage, result evaluation and result output.

[0088] Among them, block processing refers to selecting potential matching record pairs as candidates from all entity pairs in the knowledge base and reducing the size of the candidates as much as possible.

[0089] Understandably, when there is too much data, the time complexity caused by linking one by one is particularly high, so load balancing is required to ensure that the number of entities in all blocks is equal, thereby ensuring that the performance is improved by blocking.

[0090] Record linking means connecting entity information, entity relationship information, and entity attribute information through attribute similarity.

[0091] Result evaluation refers to the running time, accuracy, etc. of the evaluation algorithm.

[0092] It can be understood that the generated business knowledge graph can be stored in the knowledge base.

[0093] Preferably, knowledge processing is performed on the business knowledge graph in the knowledge base, such as quality assessment and knowledge reasoning on the business knowledge graph or the knowledge base, and the knowledge graph is updated and maintained in a timely manner.

[0094] Preferably, the business knowledge graph in the knowledge base is updated regularly to adapt to the complex and changing business environment.

[0095] The fault repair method provided in this embodiment introduces knowledge graph technology into the fault repair process. Knowledge graph technology can solve the problems of low efficiency of manual fault analysis and imperfect business analysis, and can greatly improve the efficiency and accuracy of fault repair. At the same time, knowledge graph technology can comprehensively consider various information, including system topology, business characteristics, equipment status, etc., so as to analyze the impact of faults more comprehensively and objectively to formulate corresponding fault repair plans. Knowledge graph technology can update and maintain business knowledge graphs in real time, and adjust plans in time according to real-time monitoring data and fault location results to adapt to complex and changeable communication environments.

[0096] In some embodiments, the fault localization model includes an input layer, a first network layer, a second network layer, a third network layer, and an output layer.

[0097] Before inputting the monitoring data into the fault location model and obtaining the fault location result output by the fault location model, it also includes: obtaining sample monitoring data and sample fault location result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data, performing hybrid training on the first initial model to obtain the fault location model; wherein the input layer is trained based on the error back propagation training method; and the third network layer is trained based on the least squares method.

[0098] Specifically, the fault location model is an adaptive fuzzy reasoning system, which can make inferences and decisions on complex systems by combining fuzzy logic and adaptive algorithms, and is particularly suitable for fault location. The three basic processes of fuzzification, fuzzy reasoning and defuzzification of fuzzy control can all be realized through neural networks.

[0099] See also Figure 5 , Figure 5 It is a structural schematic diagram of the adaptive fuzzy inference system provided by the present invention.

[0100] like Figure 5 As shown, the adaptive fuzzy reasoning system (ie, the fault location model) includes an input layer, a first network layer, a second network layer, a third network layer and an output layer; wherein the square nodes represent nodes with adjustable parameters.

[0101] Specifically, the input layer includes input nodes, such as Figure 5 A1, A2, B1, and B2 are the four input nodes respectively.

[0102] The input layer uses the membership function to determine the membership value of the input node to various fuzzy intervals. The calculation formula of the membership value is as follows:

[0103] Among them, x, y are the input data of node i; A i , B i is a fuzzy set; Yes A i The membership value is used to indicate that x, y belongs to A i the extent of; It is B i The membership value is used to indicate that x, y belongs to B i degree; μ is a parameter.

[0104] It should be noted that, in this embodiment, the input data x, y are monitoring data of the power communication system, including real-time status data of each power communication device in the power communication system, system topology structure of the power communication system, fault repair records of the power communication system and other data.

[0105] The first network layer is the second layer of the adaptive fuzzy inference system, which contains multiple nodes, e.g. Figure 5 The Π in represents a node. Each node in this layer is used to calculate the credibility of the rule. The calculation formula of the credibility is as follows: ; Among them, W i is the credibility of the ith node.

[0106] The second network layer is the third layer of the adaptive fuzzy inference system, which contains multiple nodes, e.g. Figure 5 N in represents a node, which is used to normalize the strength of all rules. The calculation formula for normalization is as follows: ; Among them, the normalized processing result O 3i is the output of the second network layer.

[0107] The third network layer is the fourth layer of the adaptive fuzzy inference system, which includes multiple nodes, such as Figure 5 In the above equation, f1 and f2 are two nodes, which are used to calculate the output of fuzzy rules. Each node i in this layer is an adaptive node, and the calculation formula of its output data is as follows: ; in, is the output data O of the second network layer 3i ; is the parameter set of node i.

[0108] The output layer is the fifth layer of the adaptive fuzzy inference system, which is used to generate the final output data. The calculation formula of the output data is as follows: .

[0109] It should be noted that for adaptive fuzzy inference systems, hybrid training is one of the most effective methods for model training.

[0110] Specifically, sample monitoring data and sample fault location result labels corresponding to the sample monitoring data are obtained; based on the sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data, a first initial model is hybrid trained to obtain a fault location model.

[0111] Among them, the input layer is trained based on the error back propagation training method; the third network layer is trained based on the least squares method.

[0112] Under this training method, the adaptive fuzzy inference system can obtain characteristic information from the input data (such as line voltage frequency signal, current data, etc.) as input variables, and its output is the fault location result to realize the detection of the fault location.

[0113] It should be noted that when there is a fault in the power communication system, the next step is to determine the exact location of the faulty power communication equipment in the power communication system line. Therefore, the adaptive fuzzy inference system can be trained by extracting the frequency deviation characteristics of each fault point in the power communication system line within the time range of the fault occurrence in the input data.

[0114] Preferably, during the training process of the adaptive fuzzy inference system, the fault location can be optimized based on the line length, and the feature weights in the adaptive fuzzy inference system can be adjusted by a particle swarm algorithm to obtain the optimal system response.

[0115] See also Figure 6 , Figure 6 It is a schematic diagram of a process of optimizing model parameters by using a particle swarm algorithm provided by the present invention.

[0116] like Figure 6 As shown, sample monitoring data and sample fault location result labels corresponding to the sample monitoring data are obtained as blurred training data; based on the training data, hybrid training is performed on the first initial model to obtain a fault location model.

[0117] Among them, the training data includes historical fault data of the power communication system, sensor data, system topology, fault repair records, etc.

[0118] By comprehensively utilizing the various types of data mentioned above, a comprehensive and rich training set can be constructed, which can effectively train the adaptive fuzzy inference system so that it can accurately locate faults and perform well in actual operation.

[0119] N represents the number of particles. Max Indicates the maximum number of particles.

[0120] During the model training process, the premise parameters of the model are divided into several sets, and a fitness function value is assigned to each particle; the fitness function value is updated when the particle moves until the optimal solution is generated and the optimized premise parameters are obtained; the optimized premise parameters are used as initial parameters and enter the adaptive fuzzy inference system to train the adaptive fuzzy inference system.

[0121] The fault repair method provided in this embodiment introduces an adaptive neural fuzzy reasoning system into the fault repair process. The adaptive neural fuzzy reasoning system is an intelligent technology that can make inferences and decisions on complex systems by combining fuzzy logic and adaptive algorithms, and is particularly suitable for fault location. The power communication system involves numerous power communication devices and complex circuit structures, and environmental factors and external interference may cause system failures. The adaptive fuzzy reasoning system can adapt to the complexity and uncertainty of the power communication system, process uncertainty information through fuzzy logic, and dynamically adjust the system using adaptive algorithms, so that it can better adapt to the complex changes of the power communication system and improve the accuracy of fault location. Accuracy and stability; secondly, the adaptive fuzzy reasoning system can use multi-source data to locate faults. In the power communication system, there are usually data from various sensors and monitoring equipment, including current, voltage, temperature and other information. The adaptive fuzzy reasoning system can fuse these multi-source data and conduct comprehensive analysis through fuzzy reasoning, so as to have a more comprehensive understanding of the operating status of the power communication system and improve the accuracy of fault location; in addition, the adaptive fuzzy reasoning system also has self-learning and optimization capabilities. By constantly interacting with the actual power communication system, the adaptive fuzzy reasoning system can learn from historical experience, continuously optimize the reasoning rules and algorithms, and improve the efficiency and accuracy of fault location.

[0122] In some embodiments, before the monitoring data is input into the fault detection model and the fault detection result output by the fault detection model is obtained, it also includes: obtaining sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault detection result labels corresponding to the sample monitoring data, training a second initial model constructed based on the decision tree algorithm to obtain a fault detection model.

[0123] Traditional fault detection methods have low detection accuracy and problems such as untimely analysis of real-time monitoring data. Based on this, fault detection can be performed by applying a decision tree model.

[0124] The decision tree algorithm is a classification algorithm based on a tree structure. It can automatically build a classification model according to the characteristics of the monitoring data and is used to determine whether the power communication equipment in the power communication system is abnormal or faulty. Through continuous learning and optimization, the fault detection model can continuously improve the accuracy and reliability of fault detection.

[0125] See also Figure 7 , Figure 7 It is a tree structure schematic diagram of the decision tree provided by the present invention.

[0126] like Figure 7 As shown in the figure, in the decision tree, the attributes and categories of the data are stored in a tree structure.

[0127] The sample monitoring data and the sample fault detection result labels corresponding to the sample monitoring data are used as training data samples of the power communication system.

[0128] Assume that the training data sample is , the classes in the training data samples are respectively Indicates that the training data sample set is divided into N subsets, using If , then the average amount of information of the data set can be expressed as: ; in, represents the probability of the i-th subset, Indicates the number of data samples.

[0129] Assumptions There are x values ,in ,according to Will Subsets are divided into subset; when specified When The subset is classified into the i-th data set, and the average information content of the data set can be expressed as: ; in, , .

[0130] At this time, the information gain of the data set D can be obtained by the above formula, and the information gain of the data set D can be expressed as: ; The information gain is used as an indicator for classifying the training data sample set, that is, the attribute with the largest information gain is selected as the indicator for classifying the training data sample set. The decision tree construction process includes the following steps: (1) Select the optimal partitioning attribute for the current node, that is, select the attribute with the largest information gain. If it does not exist, the node is determined to be a leaf node.

[0131] (2) For each value of the selected optimal attribute, construct a child node and assign sample points to these child nodes.

[0132] (3) For each child node, recursively execute steps ab until the termination condition is met, that is, a leaf node is reached or further partitioning is impossible.

[0133] (4) After building the decision tree, the decision tree model can be used to perform classification predictions on test data.

[0134] See also Figure 8 , Figure 8 It is a schematic diagram of the application flow of the decision tree model provided by the present invention.

[0135] like Figure 8 As shown in the figure, before applying the decision tree model, data preprocessing is required and sample data is used for training; the decision tree model can extract features and labels from sample data and generate a decision tree model (i.e., a fault detection model) through training.

[0136] It can be understood that the trained decision tree model is cross-validated through the test set data, and tested through or real-time data to determine whether the decision tree model is trained; the trained decision tree model can be officially put into use.

[0137] By analyzing historical data (sample data), the decision tree model can summarize the characteristic patterns of different types of faults, such as abnormal voltage and excessive load under specific conditions; a set of judgment rules can be constructed based on these characteristic patterns. When new monitoring data is input, fast and accurate fault detection can be performed based on these rules; secondly, the decision tree can also use real-time sensor data for fault diagnosis. Power communication systems are usually equipped with a large number of sensors for real-time monitoring of parameters such as voltage, current, and load. The decision tree can combine these real-time data to promptly detect abnormal conditions in the power communication system and determine whether there is a fault.

[0138] The fault repair method provided in this embodiment introduces decision tree technology in the fault repair process. Decision tree is a machine learning technology that can be used in power grid environments to support the decision-making process. The decision tree is constructed based on the voltage, current, power and other characteristics of the data, and can be used for fault detection in power communication systems. One of the main advantages of decision trees is that they can process continuous and discrete data, which makes the model particularly useful in practical applications; in addition, decision trees are easy to visualize, which can help power communication system operators understand the decision-making process and improve the transparency of the system; decision trees can also be used for fault detection and diagnosis of power communication systems. By analyzing the patterns of voltage, current and power, decision trees can identify the location of faults and classify them according to their severity. This information can be used to decide how to respond to faults, such as shutting down certain parts of the power communication system or rerouting power to other areas; another potential use of decision trees in the field of power grids is to predict the possibility of future power outages. By analyzing the historical data of power outages and the conditions that led to power outages, decision trees can be used to identify patterns and predict the probability of future power outages.

[0139] In some embodiments, obtaining monitoring data of the power communication system includes: obtaining original monitoring data of the power communication system; preprocessing the original monitoring data to obtain monitoring data; wherein the monitoring data includes real-time status data of each power communication device in the power communication system, and the real-time status data includes one or more of voltage data, current data, power data, frequency data and temperature data.

[0140] The present invention also provides a fault repair device. Fig. 9 , Fig. 9 9 is a schematic diagram of the structure of the fault repair device provided by the present invention. In this embodiment, the fault repair device includes an acquisition module 910 , a fault detection module 920 , a fault location module 930 and a repair module 940 .

[0141] The acquisition module 910 is used to acquire monitoring data of the power communication system.

[0142] The fault detection module 920 is used to input the monitoring data into the fault detection model to obtain the fault detection result output by the fault detection model.

[0143] The fault location module 930 is used to input the monitoring data into the fault location model to obtain the fault location result output by the fault location model if it is determined that there is a fault in the power communication system based on the fault detection result.

[0144] The repair module 940 is used to repair the fault of the power communication system based on the fault location result.

[0145] Among them, the fault detection model is trained based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data.

[0146] The fault location model is trained based on sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data.

[0147] In some embodiments, the repair module 940 is used to determine the fault business information of the power communication system based on the fault location results and the business knowledge graph; generate a repair business list based on the fault business information; the repair business list includes information on the power communication equipment to be repaired; and perform fault repair on the power communication system based on the repair business list.

[0148] In some embodiments, the repair module 940 is used to obtain business data related to the power communication system business; perform knowledge extraction processing on the business data to obtain entity information, entity relationship information and entity attribute information in the business data; perform knowledge fusion processing on the entity information, entity relationship information and entity attribute information to obtain a business knowledge graph.

[0149] In some embodiments, the fault localization model includes an input layer, a first network layer, a second network layer, a third network layer, and an output layer.

[0150] The fault location module 930 is used to obtain sample monitoring data and sample fault location result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data, the first initial model is hybrid trained to obtain a fault location model; wherein the input layer is trained based on the error back propagation training method; and the third network layer is trained based on the least squares method.

[0151] In some embodiments, the fault detection module 920 is used to obtain sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; based on the sample monitoring data and the sample fault detection result labels corresponding to the sample monitoring data, a second initial model constructed based on the decision tree algorithm is trained to obtain a fault detection model.

[0152] In some embodiments, the acquisition module 910 is used to acquire original monitoring data of the power communication system; preprocess the original monitoring data to obtain monitoring data; wherein the monitoring data includes real-time status data of each power communication device in the power communication system, and the real-time status data includes one or more of voltage data, current data, power data, frequency data and temperature data.

[0153] The present invention also provides an electronic device, Fig.10 is a schematic diagram of the structure of the electronic device provided by the present invention, such as Fig.10 As shown, the electronic device may include: a processor 1010, a communications interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communications interface 1020 and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call the logic instructions in the memory 1030 to execute the fault repair method.

[0154] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0155] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the fault repair method provided by the above methods is implemented.

[0156] The present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the fault repair methods provided by the above methods.

[0157] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0158] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art 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 ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fault repair method, characterized in that: include: Obtain monitoring data of the power communication system; Inputting the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; If it is determined that a fault exists in the power communication system based on the fault detection result, the monitoring data is input into a fault location model to obtain a fault location result output by the fault location model; Based on the fault location result, repair the fault of the power communication system; Wherein, the fault detection model is obtained by training based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; The fault location model is trained based on sample monitoring data and sample fault location result labels corresponding to the sample monitoring data.

2. The fault repair method according to claim 1, characterized in that: The performing fault repair on the power communication system based on the fault location result includes: Determine fault service information of the electric power communication system based on the fault location result and the service knowledge graph; Based on the fault service information, a repair service list is generated; the repair service list includes information of the power communication equipment to be repaired; Based on the repair service list, the power communication system is repaired.

3. The fault repair method according to claim 2, characterized in that: Before determining the fault service information of the power communication system based on the fault location result and the service knowledge graph, the method further includes: Acquire business data related to power communication system business; Performing knowledge extraction processing on the business data to obtain entity information, entity relationship information and entity attribute information in the business data; The entity information, the entity relationship information and the entity attribute information are subjected to knowledge fusion processing to obtain the business knowledge graph.

4. The fault repair method according to claim 1, characterized in that: The fault location model includes an input layer, a first network layer, a second network layer, a third network layer and an output layer; Before inputting the monitoring data into the fault location model to obtain the fault location result output by the fault location model, the method further includes: Acquire the sample monitoring data and a sample fault location result label corresponding to the sample monitoring data; Based on the sample monitoring data and the sample fault location result labels corresponding to the sample monitoring data, hybrid training is performed on the first initial model to obtain the fault location model; Among them, the input layer is trained based on the error back propagation training method; and the third network layer is trained based on the least squares method.

5. The fault repair method according to claim 1, characterized in that: Before inputting the monitoring data into the fault detection model to obtain the fault detection result output by the fault detection model, the method further includes: Acquire the sample monitoring data and a sample fault detection result label corresponding to the sample monitoring data; Based on the sample monitoring data and the sample fault detection result labels corresponding to the sample monitoring data, a second initial model constructed based on a decision tree algorithm is trained to obtain the fault detection model.

6. The fault repair method according to claim 1, characterized in that: The obtaining of monitoring data of the power communication system includes: Acquiring original monitoring data of the power communication system; Preprocessing the original monitoring data to obtain the monitoring data; The monitoring data includes real-time status data of each power communication device in the power communication system, and the real-time status data includes one or more of voltage data, current data, power data, frequency data and temperature data.

7. A fault repair device, characterized in that: include: An acquisition module, used for acquiring monitoring data of the power communication system; A fault detection module, used to input the monitoring data into a fault detection model to obtain a fault detection result output by the fault detection model; A fault location module, for inputting the monitoring data into a fault location model to obtain a fault location result output by the fault location model if it is determined that the power communication system has a fault based on the fault detection result; A repair module, used for repairing the fault of the power communication system based on the fault location result; Wherein, the fault detection model is obtained by training based on sample monitoring data and sample fault detection result labels corresponding to the sample monitoring data; The fault location model is trained based on sample monitoring data and sample fault location result labels corresponding to the sample monitoring data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the fault recovery method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the fault recovery method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the fault recovery method according to any one of claims 1 to 6 is implemented.

Images