Generation method and system of power monitoring work ticket
Through the combination of knowledge graph technology and safety measure library, the inefficiency and error-prone problems of traditional manual compilation of work tickets for power dispatch automation equipment are solved, and the standardization of work tickets is realized, and the safety stability and operation and maintenance efficiency of power facilities are improved. .
Patent Information
- Application Number
- CN202410938461.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-13
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-07-13
AI Technical Summary
The traditional manual power dispatching automation equipment work tickets have problems such as inefficiency, error-prone, and difficulty in adapting to rapid changes. It also lacks systematic and standardized risk identification and response measures, which affects the normal operation of power facilities and personnel safety.
Using knowledge graph technology to build a security measure library and entity relationship extraction method, transform the colloquial intentions of dispatchers into standardized sub-graphs through natural language processing, and combine knowledge graph retrieval to generate work items, risk items and security measures to achieve work tickets Standardized generation.
It significantly improves the standardization and completeness of work tickets, simplifies the compilation complexity, ensures information accuracy and logical clarity, ensures the safety, stability and operation compliance of operation and maintenance operations, and improves the operation and management level of the power industry. and work efficiency.
Smart Images

Figure CN118917656A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric power operation and maintenance, and in particular to a method and system for generating an electric power monitoring work ticket. Background Art
[0002] The traditional manual work ticket preparation model has problems such as low efficiency, easy errors, and difficulty in adapting to the rapid changes in dispatching automation equipment when facing large-scale and complex power dispatching automation equipment operation tasks. In addition, the identification of potential risks and the selection of response measures often rely on the experience and judgment of staff, lacking a systematic and standardized guarantee mechanism. This not only affects the normal operation and maintenance of power facilities, but may also pose a threat to personnel safety, restricting the improvement of the power industry's operational efficiency and service level.
[0003] The power industry has an urgent need for automated and intelligent operation management models. In recent years, knowledge graphs, as a powerful knowledge representation and reasoning tool, have been widely used in various fields, especially in processing complex domain knowledge and realizing intelligent decision-making. Knowledge graphs can organize multiple information such as dispatching automation equipment, operation operations, safety regulations, fault cases, etc. in a structured manner, and form a knowledge network that is easy for machines to understand and process through the modeling of entities, attributes and relationships.
[0004] However, the deep integration and innovative application of knowledge graph technology in the specific scenario of power monitoring work ticket generation is still insufficient. Most existing work ticket generation systems remain at the stage of simple template filling or rule matching, and fail to fully utilize the semantic association capabilities of knowledge graphs to achieve functions such as automatic identification of work content, association of risk factors, and accurate recommendation of safety measures in massive work data. Summary of the invention
[0005] The present invention proposes a method and system for generating a power monitoring work ticket, which utilizes historical data and existing knowledge graphs to accurately generate the work items, risk items, and safety measures items required for the work ticket, thereby solving the problems of traditional manual work ticket compilation being difficult and error-prone, and historical data being difficult to use, and realizing standardized generation of work tickets.
[0006] The present invention adopts the following technical methods.
[0007] A method for generating a power monitoring work ticket comprises the following steps:
[0008] Step S1, building a safety measures library;
[0009] Step S2, extracting entity relationships from the work content to obtain a work content subgraph; the method includes using natural language processing technology to accurately convert the dispatcher's colloquial or informal expression of the inspection or maintenance intention into a standardized subgraph representation;
[0010] Step S3: Search the work content subgraph in the knowledge graph to obtain the associated work item sequence and the risk entity set of the equipment, process, and place associated with the work item;
[0011] Step S4, searching the risk entity in the knowledge graph and the security measure library to obtain a security measure item dictionary;
[0012] Step S5: Fill in the work ticket according to the work item sequence, risk entity set and security item dictionary.
[0013] In step S1, a security measure library is constructed, which specifically includes the following steps:
[0014] Step S1A1, obtaining safety measures corpus, and performing word segmentation on the safety measures corpus;
[0015] Step S1A2: using the LSTM method to determine the first information category of the safety measures corpus; determining whether the word segmentation result of the safety measures corpus contains the vocabulary of the first information;
[0016] Step S1A3: Using the LSTM method, determine the second information category of the safety measures corpus; determine whether the word segmentation result of the safety measures corpus contains the vocabulary of the second information;
[0017] Step S1A4, adding the safety measure corpus to a safety measure library according to the first information category and the second information category;
[0018] In step S1, the security measures library is constructed, and its specific composition is as follows:
[0019] The safety measures library centrally stores the most basic and widely applicable safety measures applicable to power dispatching automation operation scenarios; the library organizes the measures in a structured manner to ensure that they have clear applicability and guiding significance at different time stages and for different types of objects;
[0020] The first information is the time dimension, which is divided into before work, during work, and after work:
[0021] Before work, preventive safety measures are included, including environmental preparation, equipment inspection, notification of the duty desk, and data backup to ensure that all conditions before operation meet safety requirements;
[0022] During work, process control safety measures are included, including operation sequence and isolation verification, to ensure that risks are effectively controlled during the operation;
[0023] After the work, it involves restorative and summary safety measures, including equipment restoration, site cleaning, and equipment verification after the work is completed, to ensure that the system returns to normal after the operation activities are completed.
[0024] The second information is the object type, which is divided into software level, hardware level and general level:
[0025] Software level: covers security measures for business systems, applications, processes, databases, data tables and other software components, including system upgrades and patch management, data backup and recovery, to ensure stable operation and information security of software systems;
[0026] Hardware level: including safety measures for hardware equipment and infrastructure, including personal protective equipment and not touching other running equipment, to ensure the safe and reliable operation of hardware facilities.
[0027] Generic measures: specify common safety requirements applicable to all operational activities, including notification of the duty station, without distinguishing between specific software or hardware objects.
[0028] Preferably, the constructed security measures library can subdivide the software level into business system level, process level, operating system level, database level, and data table level according to object type; subdivide the hardware level into server equipment level, network equipment level, and infrastructure level; and divide the security measures items according to the subdivided levels.
[0029] Each type of safety measure clearly defines its applicable scenarios and content requirements.
[0030] In step S1A2 and step S1A3, the LSTM method is specifically constructed as follows:
[0031] The LSTM method consists of a vector representation layer, an LSTM layer, and a linear layer;
[0032] Input the original security measures corpus Text into the vector representation layer and get
[0033] E=Embedding(Text)=(e1,e2,...,e n ),
[0034] Where n is the length of the corpus; Represents the d-dimensional embedding vector of the i-th word;
[0035] The vector sequence E is input to the LSTM layer, which processes an embedding vector e at each time step t. t , and based on the hidden state h at the previous moment t-1 and cell state C t-1 , update the hidden state h at the current moment t and cell state Ct ; Output the current output direction O t =LSTM(e t ,h t-1 ,C t-1 ) until all sequence information is processed and the output vector O at the last moment is obtained T ;
[0036] Linear layer is used for classification O T As input, we get the probability distribution of different categories.
[0037] In step S2, entity relationships are extracted from the work content to obtain a work content subgraph, which specifically includes the following steps:
[0038] Step S2A1: Use an entity relationship extraction tool to extract entity relationships from the work content to obtain at least 0 triples;
[0039] Step S2A2: when the number of triples is 0, the work content is taken as a single work content entity to form a subgraph with only one entity;
[0040] Step S2A3: When the number of triples is greater than 0, the extracted triples are taken as subgraphs.
[0041] In step S2A1, the entity relationship extraction method is a method based on Transformer and a dictionary rule matching method to obtain entity relationship triples;
[0042] The triplet is specifically constituted as follows:
[0043] The triple is in the form of <h, r, t>, where h represents the head entity, r represents the relationship, and t represents the tail entity;
[0044] Among them, h belongs to one of the work content entity, fault entity, and equipment entity;
[0045] Among them, r belongs to one of the involving relationship and the inducing relationship;
[0046] Among them, t belongs to one of the data table entity, place entity, business system entity, company entity, role entity, process entity, equipment entity, and fault entity.
[0047] In step S3, the work content subgraph is searched in the knowledge graph to obtain the associated work item sequence, which specifically includes the following steps:
[0048] Step S3A1: Retrieve the subgraph in the knowledge graph to locate the fault entity and the work content entity, and locate the work content entity E in the knowledge graph. content ;
[0049] Step S3A2: Retrieve the information related to E in the knowledge graph. content There is a work item entity with a resolution relationship, and its name is stored in the worklist;
[0050] Step S3A3: Retrieve the risk entity names with the current work item as the head entity and involving relationships in the knowledge graph, and store them in the risklist risk entity list;
[0051] Step S3A4: Retrieve work item entities with the current work item as the head entity and a sequential relationship in the knowledge graph, store the entity name in the worklist, and execute step S3A3 until there is no new work item;
[0052] In step S3, the knowledge graph is specifically composed as follows:
[0053] The knowledge graph includes but is not limited to the following entity relationship types:
[0054] Entity types include fault, work content, equipment, business system, process, work item and action entity; relationship types include involving, triggering, resolving and sequential relationships;
[0055] The information corresponding to the entity is stored in the form of entity attributes. The entity attributes at least include entity id and entity name, where entity id is the unique identifier of the entity and entity name stores the content information of the entity. The entity name of the work item entity is the content of the work item.
[0056] In step S3, the risk entity includes a device entity, a business system entity, a process entity, and a location entity.
[0057] In step S3, the knowledge graph is stored in the graph database for easy updating and searching; the knowledge graph includes the association between the existing ledger information, which is used to locate the inspection or maintenance content, and specifically includes:
[0058] Entity types include people, places, documents, and instruction entities;
[0059] Relationship types include belong to, located in, working member, and responsible for relationships;
[0060] The attributes of the device entity include manufacturer and IP address;
[0061] In step S3, the risk entity includes a device entity, a business system entity, and a process entity; the risk entity also includes a command entity and a location entity.
[0062] In step S4, the specific structure of the security item dictionary is as follows:
[0063] The security item dictionary is stored in the form of {key:value}, where key is one of the three types: before work, during work, and after work, and value is a list type, storing the security item content at different times;
[0064] In step S4, the risk entity is retrieved from the knowledge graph and the security measure library to obtain a list of security measures, which specifically includes the following steps:
[0065] Step S4A1: Retrieve entity types from the risk entities. If there are entities of business system and process types, search for software type and general type security measures in the security measures library and add them to the security measures list. Step S4A2: If there are no entities of business system and process types, search for hardware type and general type security measures in the security measures library and add them to the security measures list.
[0066] Step S4A3: for the risk entities in the risk entity list, searching the knowledge graph for security measures entities with the current risk entity as the last entity and with a relationship involved;
[0067] Step S4A4: for the security measure entity, searching the knowledge graph for a set of risk entities with the current security measure entity as the head entity and with a relationship involved;
[0068] Step S4A5: If the risk entity set is a subset of the risk entities in this work ticket, the security measure entity name is added to the security measure list, and the time is during work.
[0069] A system for generating a power monitoring work ticket, used to implement the above method or steps, includes a memory, a processor, and a computer program instruction stored in the memory and capable of being executed by the processor. The computer program instruction is executed by the processor to implement the method for generating a power monitoring work ticket, specifically including a safety measure library, an intention recognition module, a work item retrieval module, a safety measure retrieval module, and a work ticket generation module; by mining historical data and practical experience, a safety measure library; identifying the maintenance or repair intention of the operator, and obtaining a normalized subgraph through an entity relationship extraction method; using the subgraph to retrieve the knowledge graph, obtaining the work item and the risk item involved; retrieving the risk item, obtaining the safety measure item, and finally generating the work ticket;
[0070] The safety measures library management module is used to manage existing formatted safety measures and to add, delete, modify and check safety measures;
[0071] The knowledge graph management module is composed of the interface provided by the graph database and is used to update the existing knowledge graph;
[0072] The intent recognition module uses an entity relationship extraction algorithm to extract triples from the text input by the scheduler, and inputs the extracted triples into the work item retrieval module;
[0073] The work item retrieval module retrieves the knowledge graph to obtain work items and risk items. The order is determined by the order stored in the knowledge graph. The dispatcher can modify the order by dragging the work item text.
[0074] The safety measures retrieval module retrieves the knowledge graph and safety measures library according to the risk items to obtain the safety measures items. The default order can also be modified by dragging the safety measures item text;
[0075] The work ticket generation module generates a visual work ticket file based on the obtained work items, risk items and safety measures items.
[0076] Compared with the prior art, the present invention has the following beneficial effects:
[0077] The present invention deeply mines historical data and practical experience and systematically constructs a set of standardized safety measures library. The library summarizes and organizes the safety measures that were previously filled in manually, making it a rigorous and unified standard system, which significantly improves the standardization and integrity of the safety measures content.
[0078] Furthermore, the present invention uses natural language processing technology to accurately convert the dispatcher's oral or informal expression of the inspection or maintenance intention into a standardized subgraph representation, and based on the safety measures library and knowledge graph, obtains standardized representations of work items, risk items and safety measures items, and then integrates and generates a complete and standardized work ticket. This intelligent conversion and integration process greatly simplifies the complexity of work ticket compilation and ensures that the information of each task is accurate and logically clear.
[0079] The present invention can provide fast, convenient, accurate and effective knowledge support for personnel in power dispatching automation operation and maintenance, inspection, technical management, etc., convert the inspection or maintenance intention into a standardized subgraph, which can be used for knowledge graph and safety measures library retrieval to obtain a standardized work ticket. It effectively solves the problem of difficulty in obtaining information, reliance on personal experience and prone to omissions in the traditional manual filling method, effectively guarantees the safety and stability of operation and inspection operations, as well as the operational compliance of relevant personnel, and is of great significance to improving the operation management level and work efficiency of the entire power industry.
[0080] A method for generating an electric power monitoring work ticket of the present invention specifically includes constructing a safety measure library, an intention recognition module, a work item retrieval module, a safety measure retrieval module and a work ticket generation module; mining historical data and practical experience, a safety measure library; identifying the inspection or maintenance intention of the operator, and obtaining a normalized subgraph through an entity relationship extraction method; using the subgraph to retrieve a knowledge graph to obtain work items and the risk items involved; retrieving the risk items to obtain safety measure items, and finally generating a work ticket.
[0081] The present invention utilizes a large amount of historical data and experience information to automatically generate work tickets, effectively solving the problems of difficulty in obtaining information, reliance on personal experience and prone to omissions under the traditional manual filling method, and effectively ensuring the safety and stability of power facility operations and the operational compliance of relevant personnel. BRIEF DESCRIPTION OF THE DRAWINGS
[0082] The present invention is further described in detail below with reference to the accompanying drawings and specific embodiments:
[0083] Attached Figure 1 It is a schematic diagram of the principle framework of the generation method described in the present invention. DETAILED DESCRIPTION
[0084] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings.
[0085] As shown in the figure, a method for generating a power monitoring work ticket includes the following steps:
[0086] Step S1, building a safety measures library;
[0087] Step S2, extracting entity relationships from the work content to obtain a work content subgraph; the method includes using natural language processing technology to accurately convert the dispatcher's colloquial or informal expression of the inspection or maintenance intention into a standardized subgraph representation;
[0088] Step S3: Search the work content subgraph in the knowledge graph to obtain the associated work item sequence and the risk entity set of the equipment, process, and place associated with the work item;
[0089] Step S4, searching the risk entity in the knowledge graph and the security measure library to obtain a security measure item dictionary;
[0090] Step S5: Fill in the work ticket according to the work item sequence, risk entity set and security item dictionary.
[0091] In step S1, a security measure library is constructed, which specifically includes the following steps:
[0092] Step S1A1, obtaining safety measures corpus, and performing word segmentation on the safety measures corpus;
[0093] Step S1A2: using the LSTM method to determine the first information category of the safety measures corpus; determining whether the word segmentation result of the safety measures corpus contains the vocabulary of the first information;
[0094] Step S1A3: Using the LSTM method, determine the second information category of the safety measures corpus; determine whether the word segmentation result of the safety measures corpus contains the vocabulary of the second information;
[0095] Step S1A4, adding the safety measure corpus to a safety measure library according to the first information category and the second information category;
[0096] In step S1, the security measures library is constructed, and its specific composition is as follows:
[0097] The safety measures library centrally stores the most basic and widely applicable safety measures applicable to power dispatching automation operation scenarios; the library organizes the measures in a structured manner to ensure that they have clear applicability and guiding significance at different time stages and for different types of objects;
[0098] The first information is the time dimension, which is divided into before work, during work, and after work:
[0099] Before work, preventive safety measures are included, including environmental preparation, equipment inspection, notification of the duty desk, and data backup to ensure that all conditions before operation meet safety requirements;
[0100] During work, process control safety measures are included, including operation sequence and isolation verification, to ensure that risks are effectively controlled during the operation;
[0101] After the work, it involves restorative and summary safety measures, including equipment restoration, site cleaning, and equipment verification after the work is completed, to ensure that the system returns to normal after the operation activities are completed.
[0102] The second information is the object type, which is divided into software level, hardware level and general level:
[0103] Software level: covers security measures for business systems, applications, processes, databases, data tables and other software components, including system upgrades and patch management, data backup and recovery, to ensure stable operation and information security of software systems;
[0104] Hardware level: including safety measures for hardware equipment and infrastructure, including personal protective equipment and not touching other running equipment, to ensure the safe and reliable operation of hardware facilities.
[0105] Generic measures: specify common safety requirements applicable to all operational activities, including notification of the duty station, without distinguishing between specific software or hardware objects.
[0106] Each type of safety measure clearly defines its applicable scenarios and content requirements.
[0107] In step S1A2 and step S1A3, the LSTM method is specifically constructed as follows:
[0108] The LSTM method consists of a vector representation layer, an LSTM layer, and a linear layer;
[0109] Input the original security measures corpus Text into the vector representation layer and get
[0110] E=Embedding(Text)=(e1,e2,...,e n ),
[0111] Where n is the length of the corpus; Represents the d-dimensional embedding vector of the i-th word;
[0112] The vector sequence E is input to the LSTM layer, which processes an embedding vector e at each time step t. t , and based on the hidden state h at the previous moment t-1 and cell state C t-1 , update the hidden state h at the current moment t and cell state C t ; Output the current output direction O t =LSTM(e t ,h t-1 ,C t-1 ) until all sequence information is processed and the output vector O at the last moment is obtained T ;
[0113] Linear layer is used for classification O T As input, we get the probability distribution of different categories.
[0114] In step S2, entity relationships are extracted from the work content to obtain a work content subgraph, which specifically includes the following steps:
[0115] Step S2A1: Use an entity relationship extraction tool to extract entity relationships from the work content to obtain at least 0 triples;
[0116] Step S2A2: when the number of triples is 0, the work content is taken as a single work content entity to form a subgraph with only one entity;
[0117] Step S2A3: When the number of triples is greater than 0, the extracted triples are taken as subgraphs.
[0118] In step S2A1, the entity relationship extraction method is a method based on Transformer and a dictionary rule matching method to obtain entity relationship triples;
[0119] The triplet is specifically constituted as follows:
[0120] The triple is in the form of <h, r, t>, where h represents the head entity, r represents the relationship, and t represents the tail entity;
[0121] Among them, h belongs to one of the work content entity, fault entity, and equipment entity;
[0122] Among them, r belongs to one of the involving relationship and the inducing relationship;
[0123] Among them, t belongs to one of the data table entity, place entity, business system entity, company entity, role entity, process entity, equipment entity, and fault entity.
[0124] In step S3, the work content subgraph is searched in the knowledge graph to obtain the associated work item sequence, which specifically includes the following steps:
[0125] Step S3A1: Retrieve the subgraph in the knowledge graph to locate the fault entity and the work content entity, and locate the work content entity E in the knowledge graph. content ;
[0126] Step S3A2: Retrieve the information related to E in the knowledge graph. content There is a work item entity with a resolution relationship, and its name is stored in the worklist;
[0127] Step S3A3: Retrieve the risk entity names with the current work item as the head entity and involving relationships in the knowledge graph, and store them in the risklist risk entity list;
[0128] Step S3A4: Retrieve work item entities with the current work item as the head entity and a sequential relationship in the knowledge graph, store the entity name in the worklist, and execute step S3A3 until there is no new work item;
[0129] In step S3, the knowledge graph is specifically composed as follows:
[0130] The knowledge graph includes but is not limited to the following entity relationship types:
[0131] Entity types include fault, work content, equipment, business system, process, work item and action entity; relationship types include involving, triggering, resolving and sequential relationships;
[0132] The information corresponding to the entity is stored in the form of entity attributes. The entity attributes at least include entity id and entity name, where entity id is the unique identifier of the entity and entity name stores the content information of the entity. The entity name of the work item entity is the content of the work item.
[0133] In step S3, the risk entity includes a device entity, a business system entity, a process entity, and a location entity.
[0134] In step S3, the knowledge graph is stored in the graph database for easy updating and searching; the knowledge graph includes the association between the existing ledger information, which is used to locate the inspection or maintenance content, and specifically includes:
[0135] Entity types include people, places, documents, and instruction entities;
[0136] Relationship types include belong to, located in, working member, and responsible for relationships;
[0137] The attributes of the device entity include manufacturer and IP address;
[0138] In step S3, the risk entity includes a device entity, a business system entity, and a process entity; the risk entity also includes a command entity and a location entity.
[0139] In step S4, the specific structure of the security item dictionary is as follows:
[0140] The security item dictionary is stored in the form of {key:value}, where key is one of the three types: before work, during work, and after work, and value is a list type, storing the security item content at different times;
[0141] In step S4, the risk entity is retrieved from the knowledge graph and the security measure library to obtain a list of security measures, which specifically includes the following steps:
[0142] Step S4A1: Retrieve entity types from the risk entities. If there are entities of business system and process types, search for software type and general type security measures in the security measures library and add them to the security measures list. Step S4A2: If there are no entities of business system and process types, search for hardware type and general type security measures in the security measures library and add them to the security measures list.
[0143] Step S4A3: for the risk entities in the risk entity list, searching the knowledge graph for security measures entities with the current risk entity as the last entity and with a relationship involved;
[0144] Step S4A4: for the security measure entity, searching the knowledge graph for a set of risk entities with the current security measure entity as the head entity and with a relationship involved;
[0145] Step S4A5: If the risk entity set is a subset of the risk entities in this work ticket, the security measure entity name is added to the security measure list, and the time is during work.
[0146] A system for generating a power monitoring work ticket, used to implement the above method or steps, includes a memory, a processor, and a computer program instruction stored in the memory and capable of being executed by the processor. The computer program instruction is executed by the processor to implement the method for generating a power monitoring work ticket, specifically including a safety measure library, an intention recognition module, a work item retrieval module, a safety measure retrieval module, and a work ticket generation module; by mining historical data and practical experience, a safety measure library; identifying the maintenance or repair intention of the operator, and obtaining a normalized subgraph through an entity relationship extraction method; using the subgraph to retrieve the knowledge graph, obtaining the work item and the risk item involved; retrieving the risk item, obtaining the safety measure item, and finally generating the work ticket;
[0147] The safety measures library management module is used to manage existing formatted safety measures and to add, delete, modify and check safety measures;
[0148] The knowledge graph management module is composed of the interface provided by the graph database and is used to update the existing knowledge graph;
[0149] The intent recognition module uses an entity relationship extraction algorithm to extract triples from the text input by the scheduler, and inputs the extracted triples into the work item retrieval module;
[0150] The work item retrieval module retrieves the knowledge graph to obtain work items and risk items. The order is determined by the order stored in the knowledge graph. The dispatcher can modify the order by dragging the work item text.
[0151] The safety measures retrieval module retrieves the knowledge graph and safety measures library according to the risk items to obtain the safety measures items. The default order can also be modified by dragging the safety measures item text;
[0152] The work ticket generation module generates a visual work ticket file based on the obtained work items, risk items and safety measures items.
[0153] This embodiment implements the generation of work items, risk items and safety measures items in work tickets based on the characteristics of power monitoring work tickets. First, safety measures are summarized and summarized based on historical work tickets and historical experience to form a standardized safety measures library; entity relationships are extracted from the inspection and maintenance intentions of dispatchers to obtain subgraphs; work item sequences and risk item sets are retrieved in the knowledge graph based on the subgraphs, and safety measures sequences are obtained by searching the knowledge graph and safety measures library based on the risk item set, thus realizing the automatic generation of power monitoring work tickets. This realizes the effective matching of work sequence arrangement, accurate risk identification and safety measures for power dispatching automation operations, and effectively promotes the information and intelligent upgrade of dispatching operations.
[0154] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0155] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0156] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0157] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0158] The above is only a preferred embodiment of the present invention, and does not limit the present invention in other forms. Any technician familiar with the profession may use the above disclosed technical content to change or modify it into an equivalent embodiment with equivalent changes. However, any simple modification, equivalent change and modification made to the above embodiment according to the technical essence of the present invention without departing from the technical solution of the present invention still belongs to the protection scope of the technical solution of the present invention.
Claims
1. A method for generating a power monitoring work ticket, characterized in that: The steps include: Step S1, building a safety measures library; Step S2, extracting entity relationships from the work content to obtain a work content subgraph; the method includes using natural language processing technology to accurately convert the dispatcher's colloquial or informal expression of the inspection or maintenance intention into a standardized subgraph representation; Step S3: Search the work content subgraph in the knowledge graph to obtain the associated work item sequence and the risk entity set of the equipment, process, and place associated with the work item; Step S4, searching the risk entity in the knowledge graph and the security measure library to obtain a security measure item dictionary; Step S5: Fill in the work ticket according to the work item sequence, risk entity set and security item dictionary.
2. A method and system for generating a power monitoring work ticket according to claim 1, characterized in that: In step S1, a security measure library is constructed, which specifically includes the following steps: Step S1A1, obtaining safety measures corpus, and performing word segmentation on the safety measures corpus; Step S1A2: using the LSTM method to determine the first information category of the safety measures corpus; determining whether the word segmentation result of the safety measures corpus contains the vocabulary of the first information; Step S1A3: Using the LSTM method, determine the second information category of the safety measures corpus; determine whether the word segmentation result of the safety measures corpus contains the vocabulary of the second information; Step S1A4: Add the safety measure corpus to a safety measure library according to the first information category and the second information category.
3. The method for generating a power monitoring work ticket according to claim 2, characterized in that: In step S1, the security measures library is constructed, and its specific composition is as follows: The safety measures library centrally stores the most basic and widely applicable safety measures applicable to power dispatching automation operation scenarios; the library organizes the measures in a structured manner to ensure that they have clear applicability and guiding significance at different time stages and for different types of objects; The first information is the time dimension, which is divided into before work, during work, and after work: Before work, preventive safety measures are included, including environmental preparation, equipment inspection, notification of the duty desk, and data backup to ensure that all conditions before operation meet safety requirements; During work, process control safety measures are included, including operation sequence and isolation verification, to ensure that risks are effectively controlled during the operation; After the work, it involves restorative and summary safety measures, including equipment restoration, site cleaning, and equipment verification after the work is completed, to ensure that the system returns to normal after the operation activities are completed. The second information is the object type, which is divided into software level, hardware level and general level: Software level: covers security measures for business systems, applications, processes, databases, data tables and other software components, including system upgrades and patch management, data backup and recovery, to ensure stable operation and information security of software systems; Hardware level: including safety measures for hardware equipment and infrastructure, including personal protective equipment and not touching other running equipment, to ensure the safe and reliable operation of hardware facilities. Generic measures: specify common safety requirements applicable to all operational activities, including notification of the duty station, without distinguishing between specific software or hardware objects. Each type of safety measure clearly defines its applicable scenarios and content requirements.
4. The method for generating a power monitoring work ticket according to claim 2, characterized in that: In step S1A2 and step S1A3, the LSTM method is specifically constructed as follows: The LSTM method consists of a vector representation layer, an LSTM layer, and a linear layer; Input the original security measures corpus Text into the vector representation layer and get E=Embedding(Text)=(e1,e2,...,e n ), Where n is the length of the corpus; Represents the d-dimensional embedding vector of the i-th word; The vector sequence E is input to the LSTM layer, which processes an embedding vector e at each time step t. t , and based on the hidden state h at the previous moment t-1 and cell state C t-1 , update the hidden state h at the current moment t and cell state C t ; Output the current output direction O t =LSTM(e t ,h t-1 ,C t-1 ) until all sequence information is processed and the output vector O at the last moment is obtained T ; Linear layer is used for classification O T As input, we get the probability distribution of different categories.
5. The method for generating a power monitoring work ticket according to claim 2, characterized in that: In step S2, entity relations are extracted from the work content to obtain a work content subgraph, which specifically includes the following steps: step S2A1, using an entity relation extraction tool to extract entity relations from the work content to obtain at least 0 triples; Step S2A2: when the number of triples is 0, the work content is taken as a single work content entity to form a subgraph with only one entity; Step S2A3: When the number of triples is greater than 0, the extracted triples are taken as subgraphs.
6. A method for generating a power monitoring work ticket according to claim 2, characterized in that: In step S2A1, the entity relationship extraction method is to obtain entity relationship triples by a Transformer-based method and a dictionary rule matching method; The triplet is specifically constituted as follows: The triplet is in the form of <h, r, t>, where h represents the head entity, r represents the relationship, and t represents the tail entity; Among them, h belongs to one of the work content entity, fault entity, and equipment entity; Among them, r belongs to one of the involving relationship and the inducing relationship; Among them, t belongs to one of the data table entity, place entity, business system entity, company entity, role entity, process entity, equipment entity, and fault entity.
7. The method for generating a power monitoring work ticket according to claim 1, characterized in that: In step S3, the work content subgraph is searched in the knowledge graph to obtain the associated work item sequence, which specifically includes the following steps: Step S3A1: Retrieve the subgraph in the knowledge graph to locate the fault entity and the work content entity, and locate the work content entity E in the knowledge graph. content ; Step S3A2: Retrieve the information related to E in the knowledge graph. content There is a work item entity with a resolution relationship, and its name is stored in the worklist; Step S3A3: Retrieve the risk entity names with the current work item as the head entity and involving relationships in the knowledge graph, and store them in the risklist risk entity list; Step S3A4: Retrieve work item entities with the current work item as the head entity and a sequential relationship in the knowledge graph, store the entity name in the worklist, and execute step S3A3 until there is no new work item; In step S3, the knowledge graph is specifically constructed as follows: The knowledge graph includes but is not limited to the following entity relationship types: Entity types include fault, work content, equipment, business system, process, work item and security measure entity; Relationship types include involving, initiating, resolving, and sequential relationships; The information corresponding to the entity is stored in the form of entity attributes. The entity attributes at least include entity id and entity name, where entity id is the unique identifier of the entity and entity name stores the content information of the entity. The entity name of the work item entity is the content of the work item. In step S3, the risk entity includes a device entity, a business system entity, a process entity, and a location entity.
8. A method for generating a power monitoring work ticket according to claim 7, characterized in that: In step S3, the knowledge graph is stored in the graph database for easy updating and searching; the knowledge graph includes the association between the existing ledger information, which is used to locate the inspection or maintenance content, and specifically includes: Entity types include people, places, documents, and instruction entities; Relationship types include belong to, located within, working member, and responsible for relationships; The attributes of the device entity include manufacturer and IP address; In step S3, the risk entity includes a device entity, a business system entity, and a process entity; the risk entity also includes a command entity and a location entity.
9. The method for generating a power monitoring work ticket according to claim 1, characterized in that: In step S4, the specific structure of the security item dictionary is as follows: The security item dictionary is stored in the form of {key:value}, where key is one of the three types: before work, during work, and after work, and value is a list type, storing the security item content at different times; In step S4, the risk entity is retrieved from the knowledge graph and the security measure library to obtain a list of security measures, which specifically includes the following steps: Step S4A1: Retrieve entity types from the risk entities. If there are entities of business system and process types, search for software type and general type security measures in the security measures library and add them to the security measures list. Step S4A2: If there are no entities of business system and process types, search for hardware type and general type security measures in the security measures library and add them to the security measures list. Step S4A3: for the risk entities in the risk entity list, searching the knowledge graph for security measures entities with the current risk entity as the last entity and with a relationship involved; Step S4A4: for the security measure entity, searching the knowledge graph for a set of risk entities with the current security measure entity as the head entity and having a relationship involved; Step S4A5: If the risk entity set is a subset of the risk entities in this work ticket, the security measure entity name is added to the security measure list, and the time is during work.
10. A system for generating a power monitoring work ticket, used to implement the method or step in any one of claims 1 to 9, characterized in that: The invention comprises a memory, a processor and a computer program instruction stored in the memory and capable of being executed by the processor. The computer program instruction is executed by the processor to realize a method for generating a power monitoring work ticket, and specifically comprises a safety measure library, an intention recognition module, a work item retrieval module, a safety measure retrieval module and a work ticket generation module; by mining historical data and practical experience, a safety measure library is used; the maintenance or repair intention of the operator is identified, and a normalized subgraph is obtained by an entity relationship extraction method; the knowledge graph is retrieved using the subgraph to obtain the work item and the risk item involved; the risk item is retrieved to obtain the safety measure item, and finally a work ticket is generated; The safety measures library management module is used to manage existing formatted safety measures and to add, delete, modify and check safety measures; The knowledge graph management module is composed of the interface provided by the graph database and is used to update the existing knowledge graph; The intent recognition module uses an entity relationship extraction algorithm to extract triples from the text input by the scheduler, and inputs the extracted triples into the work item retrieval module; The work item retrieval module retrieves the knowledge graph to obtain work items and risk items. The order is determined by the order stored in the knowledge graph. The dispatcher can modify the order by dragging the work item text. The safety measures retrieval module retrieves the knowledge graph and safety measures library according to the risk items to obtain the safety measures items. The default order can also be modified by dragging the safety measures item text; The work ticket generation module generates a visual work ticket file based on the obtained work items, risk items and safety measures items.
Citation Information
Patent Citations
Intelligent management method and device for work tickets
CN115062795A
Power work ticket generation method and system based on relational graph convolutional neural network
CN116628220A
System for on-site management of an event
US8397984B1
Knowledge graph question-answer method and apparatus based on deep learning technology, and device
WO2021139283A1
Cited By
Electrical work ticket intelligent generation method and device, medium and equipment
CN120911429A
Generation method of power grid equipment maintenance work ticket and electronic equipment
CN121480456A
Safety supervision work ticket intelligent generation method based on ReAct intelligent agent
CN121503703A
