Knowledge base construction method and device for power grid dispatching and electronic equipment
The method addresses the challenge of diverse data types in power grid dispatching by using deep learning models to analyze and construct a knowledge base, improving fault diagnosis accuracy.
Patent Information
- Application Number
- CN202411131044.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-12-06
AI Technical Summary
Existing methods for constructing knowledge bases in power grid dispatching fail to accurately extract useful information from diverse data types, including structured, semi-structured, and unstructured data, leading to inaccurate fault diagnosis.
A method involving data division and deep learning-based extraction models (RNN, CRF, BERT) to analyze structured, semi-structured, and unstructured data, respectively, followed by constructing a knowledge base using entity, attribute, and relationship data.
Accurately extracts power grid dispatching data, enhancing fault diagnosis by providing comprehensive and practical knowledge support.
Smart Images

Figure CN119089988A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of data processing and analysis, and in particular to a method for constructing a knowledge base for power grid dispatching, a device for constructing a knowledge base for power grid dispatching, a computer-readable storage medium, and an electronic device. Background Art
[0002] In the power system, grid dispatching is a crucial link, which involves real-time monitoring of the grid operation status, fault prediction, and optimized dispatching. With the rapid development of smart grids, the amount of data in the field of grid dispatching has exploded, and the data types are becoming more and more diverse, including structured data, semi-structured data, and unstructured data. How to effectively manage and utilize this data, extract valuable information, and build a knowledge base to support intelligent decision-making in grid dispatching is an important challenge currently faced. In particular, it is of great significance to monitor the fault status of the grid in real time through the knowledge base, which can timely discover the abnormal conditions and faults of the equipment, and quickly analyze and locate them through the information and rules in the knowledge base. The theoretical knowledge, practical experience, technical specifications, equipment parameters, and other aspects of grid dispatching contained in the knowledge base can provide accurate data support and guidance for fault handling, thereby improving the efficiency and accuracy of fault handling.
[0003] The prior art discloses a method and system for extracting domain knowledge based on self-adjusting parameters, and an electronic device, which includes the following steps: constructing a domain ontology knowledge base through collected domain data, wherein the domain ontology knowledge base includes a domain entity base, a domain relationship base, and a domain attribute base; vectorizing the constructed domain ontology knowledge base and using it as the input of a pre-training model, training the pre-training model, and obtaining a knowledge extraction model; obtaining adjustable parameters in the knowledge extraction model, adjusting the adjustable parameters according to business data, and pre-processing the business data, and then inputting the pre-processed business data and the adjusted adjustable parameters into the knowledge extraction model, thereby obtaining the extraction result of the business data; the extraction result is a triple list of <entity, relationship, attribute>. This solution can complete the knowledge base, uniformly extract entities, relationships, and attributes, and quickly respond to different business needs.
[0004] Although the above patent is applicable to the construction of knowledge base in the field of power grid dispatching, the forms of power grid dispatching data are complex and diverse, including structured data, unstructured data and semi-structured data. The prior art method for constructing a knowledge base in the field of power grid dispatching extracts data, manually annotates entity attribute relationships, and integrates the manually annotated entity attribute relationships to realize knowledge base construction; among them, structured data generally refers to data with a predefined structure and format, semi-structured data has a certain structure, but is not as strict as structured data, and unstructured data has no fixed data structure, but the data volume is large, and valuable content needs to be extracted from a complex format; structured data, unstructured data and semi-structured data contain the same type of information or data elements, and the data complexity is different. The extraction model in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0005] Therefore, a method for accurately extracting power system dispatch data is needed. Summary of the invention
[0006] The main purpose of the present application is to provide a knowledge base construction method for power grid dispatching, a knowledge base construction device for power grid dispatching, a computer-readable storage medium and an electronic device, so as to at least solve the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0007] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for constructing a knowledge base of power grid dispatching is provided, the method comprising: acquiring power grid dispatching data, dividing the power grid dispatching data, and obtaining data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control and regulation of the power grid, and the data division results include at least structured data, semi-structured data and unstructured data; analyzing the power grid dispatching data through a data extraction model according to the data division results to obtain entity data, attribute data and relationship data, wherein the data extraction model is obtained through deep learning training of multiple groups of data, and each group of the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; constructing a knowledge base according to the entity data, the attribute data and the relationship data.
[0008] Optionally, the power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, including: when the power grid dispatching data is the structured data, the structured data is analyzed by a structured data extraction model to obtain first entity data, first attribute data and first relationship data, wherein the structured data extraction model is obtained by deep learning training through multiple groups of first training data, and each group of data in the multiple groups of first training data includes: first entity training data and first entity training data label, first attribute training data and first attribute training data label, first relationship training data and first relationship training data label.
[0009] Optionally, the power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, and also includes: when the power grid dispatching data is the semi-structured data, the semi-structured data is analyzed by a semi-structured data extraction model to obtain second entity data, second attribute data and second relationship data, wherein the semi-structured data extraction model is obtained by deep learning training through multiple groups of second training data, and each group of data in the multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels.
[0010] Optionally, the power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, and also includes: when the power grid dispatching data is the unstructured data, the unstructured data is analyzed by an unstructured data extraction model to obtain third entity data, third attribute data and third relationship data, wherein the unstructured data extraction model is obtained by deep learning training through multiple groups of third training data, and each group of data in the multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels.
[0011] Optionally, the method includes: when the power grid dispatching data is the structured data, training the RNN model through multiple groups of data to obtain the structured data extraction model; when the power grid dispatching data is the semi-structured data, training the CRF model through multiple groups of data to obtain the semi-structured data extraction model; when the power grid dispatching data is the unstructured data, training the BERT model through multiple groups of data to obtain the unstructured data extraction model.
[0012] Optionally, constructing a knowledge base based on the entity data, the attribute data and the relationship data includes: constructing an entity, attribute and relationship one-to-one mapping relationship library; importing the entity data, the attribute data and the relationship data into the entity, attribute and relationship one-to-one mapping relationship library to obtain the knowledge base.
[0013] Optionally, after acquiring the power grid dispatching data, the method further includes: performing data cleaning on the power grid dispatching data; performing data conversion on the power grid dispatching data after data cleaning; and performing data standardization on the power grid dispatching data after data conversion to obtain standardized power grid dispatching data.
[0014] According to another aspect of the present application, a knowledge base construction device for power grid dispatching is provided, and the device includes: a division unit, used to obtain power grid dispatching data, divide the power grid dispatching data, and obtain data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control and regulation of the power grid, and the data division results include at least structured data, semi-structured data and unstructured data; an analysis unit, used to analyze the power grid dispatching data through a data extraction model according to the data division results to obtain entity data, attribute data and relationship data, wherein the data extraction model is obtained through deep learning training with multiple groups of data, and each group of the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; a construction unit, used to construct a knowledge base according to the entity data, the attribute data and the relationship data.
[0015] According to another aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute any one of the knowledge base construction methods for power grid scheduling.
[0016] According to another aspect of the present application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a knowledge base construction method for executing any one of the power grid scheduling methods described.
[0017] By applying the technical solution of the present application, the power grid dispatching data is acquired, the power grid dispatching data is divided, and the data division result is obtained; then, the power grid dispatching data is analyzed through the data extraction model according to the data division result to obtain entity data, attribute data and relationship data; finally, a knowledge base is constructed according to the entity data, attribute data and relationship data. In this way, the data extraction model is used to accurately extract the power grid dispatching data according to the division result of the power grid dispatching data, and the entity, attribute and relationship data are obtained, and a knowledge base is established, so as to achieve the effect of accurately extracting the dispatching data. That is, it solves the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The drawings constituting part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0019] Figure 1 A schematic diagram of a flow chart of a method for constructing a knowledge base for power grid dispatching provided in an embodiment of the present application is shown;
[0020] Figure 2 A schematic diagram of a process of constructing another knowledge base for power grid scheduling provided by an embodiment of the present application is shown;
[0021] Figure 3 A structural block diagram of a knowledge base construction device for power grid dispatching provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0025] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:
[0026] Structured data refers to data that is organized and stored according to certain rules and formats, with clear relationships between fields and values. Structured data usually exists in the form of tables, databases, XML, JSON, etc., which is easy to store, manage and process.
[0027] Semi-structured data is a type of data between structured data and unstructured data. It has a certain structure but no strict mode or format, so it is not suitable for storage and management using traditional relational databases.
[0028] Unstructured data refers to data without a clearly defined structure, usually in the form of natural language, images, audio, video, etc. This type of data is not suitable for direct storage and processing by traditional database management systems and requires special processing to be effectively used.
[0029] Data cleaning is part of data preprocessing, which refers to the operations of screening, organizing, removing errors, filling missing values, etc. on raw data to ensure the accuracy and completeness of the data.
[0030] Data conversion is the process of converting data from one form or format to another. This conversion can involve the conversion of different data types, data structures, or data formats.
[0031] Data standardization refers to the process of converting data into a standard format so that different data can be compared and analyzed. Data standardization can eliminate problems such as unit differences and dimension differences in data and improve the comparability and interpretability of data.
[0032] As introduced in the background technology, the extraction model in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate scheduling data from the knowledge base, which has a certain impact on fault diagnosis. In order to solve the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate scheduling data from the knowledge base, which has a certain impact on fault diagnosis, the embodiments of the present application provide a knowledge base construction method for power grid dispatching, a knowledge base construction device for power grid dispatching, a computer-readable storage medium and an electronic device.
[0033] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.
[0034] In this embodiment, a method for constructing a knowledge base for power grid dispatching is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0035] Figure 1 FIG. 1 is a flow chart of a method for constructing a knowledge base for power grid dispatching according to an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:
[0036] Step S101, acquiring power grid dispatching data, dividing the power grid dispatching data, and obtaining data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data division results at least include structured data, semi-structured data, and unstructured data;
[0037] Specifically, the acquisition of power grid dispatching data can be carried out by using crawler technology through keywords in the field of power grid dispatching. The power grid dispatching data may include equipment operating status, fault records, operation logs, dispatching instructions, etc.
[0038] In some embodiments, after acquiring the power grid dispatching data, the method further includes: performing data cleaning on the power grid dispatching data; performing data conversion on the power grid dispatching data after data cleaning; and performing data standardization on the power grid dispatching data after data conversion to obtain standardized power grid dispatching data.
[0039] Specifically, data cleaning is used to remove noise, duplicate information and irrelevant information in the crawled data; data conversion is used to convert unstructured data into structured data or semi-structured data format; data standardization is used to unify the units, formats and naming specifications of data to ensure the consistency and comparability of data.
[0040] Step S102, analyzing the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, wherein the above-mentioned data extraction model is obtained through deep learning training of multiple groups of data, and each group of data in the above-mentioned multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels;
[0041] Specifically, obtaining the above-mentioned multiple groups of data includes: defining entities, attributes, and relationships based on the above-mentioned structured data, the above-mentioned semi-structured data, and the above-mentioned unstructured data, respectively, obtaining corresponding entity data, attribute data, and relationship data, respectively, and assigning corresponding entity labels, attribute labels, and relationship labels to the above-mentioned entity data, the above-mentioned attribute data, and the above-mentioned relationship data.
[0042] Furthermore, structured data extraction models, semi-structured data extraction models and unstructured data extraction models were constructed for structured data, semi-structured data and unstructured data respectively, and trained with dedicated data to ensure that as much useful information as possible is extracted from various sources and integrated into the knowledge base, thereby improving the comprehensiveness and practicality of the knowledge base.
[0043] In one embodiment of the present application, in the above step S102, the above-mentioned power grid dispatching data is analyzed by a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, including: when the above-mentioned power grid dispatching data is the above-mentioned structured data, the above-mentioned structured data is analyzed by a structured data extraction model to obtain first entity data, first attribute data and first relationship data, wherein the above-mentioned structured data extraction model is obtained by deep learning training through multiple groups of first training data, and each group of data in the above-mentioned multiple groups of first training data includes: first entity training data and first entity training data label, first attribute training data and first attribute training data label, first relationship training data and first relationship training data label.
[0044] Specifically, the entity data in the above structured data includes: equipment information, fault type, fault event record; the equipment information includes but is not limited to the name, number, and location of the power grid equipment, and the entity label of the equipment information can be {E1, E11, E12, E13}, where E1 represents the power grid type. The power grid equipment type can be classified according to multiple standards, for example, it can be divided into indoor equipment and outdoor equipment according to different installation locations; it can be divided into high-voltage equipment, medium-voltage equipment, and low-voltage equipment according to different voltage levels; it can be divided into primary equipment and secondary equipment according to different electric energy functions; it can be divided into power generation equipment, transmission equipment, and distribution equipment according to different functions; E11 represents the name of the power grid equipment, and E12 represents the number of the power grid equipment, and the specific power grid equipment can be located according to the fault event; E13 represents the location of the power grid equipment, and when a fault occurs, the fault point can be quickly found according to the location of the power grid equipment;
[0045] The attribute data in the above structured data include: equipment status, measurement value, fault level, and fault cause; specifically, the equipment status A1 obtained according to the power grid equipment information includes equipment operation status A11, equipment stop status A12, and equipment fault status A13; the measurement value A2 includes real-time or historical measurement values such as equipment voltage A21, equipment current A22, equipment power A23, and equipment frequency A24; the fault level A3 includes minor fault A31, serious fault A32, and emergency fault A34, etc.; the fault cause A4 includes equipment aging A41, human operation error A42, and natural disaster A43, etc.;
[0046] The relational data in the above structured data include device connection relationships, fault impact range and fault handling process; specifically, device connection relationship R1 includes physical connection relationship R11 between devices and logical connection relationship R12 between devices, etc. For example, the physical connection relationship between devices is the relationship between power grid devices directly connected through cables, wires or other physical media, and the logical connection relationship between devices is that multiple sensors monitor the operating status of a power grid device; the fault impact range R2 includes the device range R21 affected by the fault and the geographical range R22 affected by the fault. Since the power grid equipment and the location of the power grid equipment are entity data, the fault impact range can be used as the relationship between the two entity data; the fault handling process R3 can be used as the relationship between the various links in fault discovery, fault reporting, fault location, fault handling and fault recovery.
[0047] The above structured data extraction model is used to extract entity data, attribute data and relationship data of structured data.
[0048] Each of the above multiple groups of first training data includes: first entity training data and first entity training data labels, first attribute training data and first attribute training data labels, first relationship training data and first relationship training data labels, so as to realize the training of RNN model.
[0049] In one embodiment of the present application, the above-mentioned step S102 includes analyzing the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, and also includes: when the above-mentioned power grid dispatching data is the above-mentioned semi-structured data, analyzing the above-mentioned semi-structured data through a semi-structured data extraction model to obtain second entity data, second attribute data and second relationship data, wherein the above-mentioned semi-structured data extraction model is obtained through deep learning training through multiple groups of second training data, and each group of data in the above-mentioned multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels.
[0050] Specifically, the entity data E1-1 in the above semi-structured data includes: equipment log E1-11 and fault report E1-12. The equipment log records the status changes and operation records of the equipment, and can be regarded as the relevant data of the equipment information entity. The fault report includes key information such as the time, location and type of the fault, and can be regarded as the fault information in the power grid equipment information.
[0051] The attribute data in the above semi-structured data include: device log details A1-1 and fault report details A1-2;
[0052] The relational data in the above semi-structured data include: log association R1-1 and fault propagation R1-2 of different devices. There are correlations between logs of different power grid devices or different time periods. For example, the failure of a certain device may cause abnormal logs of other devices. These correlations can be identified and analyzed through information such as the device number and timestamp in the log.
[0053] The above-mentioned semi-structured data extraction model is used to extract entity data, attribute data and relationship data of semi-structured data.
[0054] Each of the above-mentioned multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels, so as to realize the training of the CRF extraction model; due to the characteristics of the CRF extraction model, it is necessary to convert the multiple groups of second training data into observation sequences, so that the observation sequences correspond to the labels of entities, attributes and relationships one by one, and the observation sequences and the labels of entities, attributes and relationships are constructed into multiple groups of second training data that can be used by the CRF extraction model, and the model is trained. The CRF extraction model can observe the mapping relationship between sequences and labels. During the training process, the model will adjust its parameters to maximize the conditional probability distribution on the training data, so as to make accurate predictions on new unseen data.
[0055] In one embodiment of the present application, the above-mentioned step S102 includes analyzing the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, and also includes: when the above-mentioned power grid dispatching data is the above-mentioned unstructured data, analyzing the above-mentioned unstructured data through an unstructured data extraction model to obtain third entity data, third attribute data and third relationship data, wherein the above-mentioned unstructured data extraction model is obtained by deep learning training through multiple groups of third training data, and each group of data in the above-mentioned multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels.
[0056] The entity data of the above unstructured data includes: equipment data, fault category, fault location and fault time;
[0057] The attribute data of the above unstructured data include: equipment status, fault level, fault cause and fault impact range;
[0058] The relational data of the above unstructured data includes: the causal relationship between equipment and faults, the correlation relationship between equipment and faults, and the temporal relationship between the occurrence of faults.
[0059] The details of entity data and attribute data in unstructured data are the same as those of structured data. Relational data includes causal relationships and association relationships. Causal relationships refer to equipment failures causing other equipment to shut down or power grid fluctuations, and association relationships refer to the association between a certain device and a certain type of failure.
[0060] The above unstructured data extraction model is used to extract entity data, attribute data and relationship data of unstructured data.
[0061] Each of the above-mentioned multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels; implement self-attention-based model BERT training; specifically, according to the entities of unstructured data in the multiple groups of data, traverse all labeled entities, generate all possible entity pairs as potential association relationships, and determine the entity span according to each entity; construct a feature representation according to the potential association relationship of each entity pair, including the characteristics of the entity itself, the text content between the entities, and the semantic representation of the entity and the relationship context; use the potential association relationship of each entity pair as the input of the relationship output layer, and the relationship output layer outputs a relationship type label, such as a causal relationship or an association relationship.
[0062] In the specific implementation process, the above method includes: when the above-mentioned power grid dispatching data is the above-mentioned structured data, training the RNN model through multiple groups of data to obtain the above-mentioned structured data extraction model; when the above-mentioned power grid dispatching data is the above-mentioned semi-structured data, training the CRF model through multiple groups of data to obtain the above-mentioned semi-structured data extraction model; when the above-mentioned power grid dispatching data is the above-mentioned unstructured data, training the BERT model through multiple groups of data to obtain the above-mentioned unstructured data extraction model.
[0063] Specifically, the RNN model is selected as the structured data extraction model. For the recognition of equipment information (E1, E11, E12, E13), the RNN model can learn the patterns in the sequence, thereby identifying the name, number and location of the power grid equipment in the text. For the classification of equipment status (A1), measurement value (A2), fault level (A3) and fault cause (A4), the RNN model can learn the features extracted from the text and map them to the corresponding entity labels; for relational data (R1, R2, R3), the RNN model can learn the entity pairs in the text and the contextual information between them, thereby identifying the relationship between entities;
[0064] The entity data, attribute data and relationship data of semi-structured data are extracted through the CRF extraction model;
[0065] The self-attention-based model BERT is used to extract entity data, attribute data and relationship data from unstructured data. The output layer of the model sets three classification layers to predict the entity labels, attribute categories and relationships between entities in the unstructured data respectively. The entity labels and attribute categories learn the attribute relationships between entities according to the training process of each group of data in multiple groups of data, and output them in the form of labels. For the extraction of relationship data, the entity span of the unstructured data is identified through span classification. The above entity span refers to the starting position and ending position of the entity.
[0066] Furthermore, using specialized models for different types of data can significantly improve the efficiency and accuracy of data processing. This is because each model is designed and optimized for a specific type of data and can better adapt to the characteristics and needs of the data.
[0067] Step S103: construct a knowledge base based on the entity data, the attribute data and the relationship data.
[0068] In the above steps, the power grid dispatching data is obtained and divided to obtain the data division results; then, the power grid dispatching data is analyzed by the data extraction model according to the data division results to obtain entity data, attribute data and relationship data; finally, a knowledge base is constructed according to the entity data, attribute data and relationship data. In this way, the data extraction model is used to accurately extract the power grid dispatching data according to the division results of the power grid dispatching data, to obtain entity, attribute and relationship data, and to establish a knowledge base, so as to achieve the effect of accurately extracting the dispatching data. That is, it solves the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0069] Specifically, after building the knowledge base, when a power grid failure occurs, structured data can be used to quickly locate fault-related equipment, parameters, or events by directly querying the knowledge base. For semi-structured and unstructured data, the entities, attributes, and relationships extracted by the model help to quickly identify key information related to the fault.
[0070] During the specific implementation process, the above-mentioned step S103 constructs a knowledge base based on the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data, including: constructing an entity, attribute and relationship one-to-one mapping relationship library; importing the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data into the above-mentioned entity, attribute and relationship one-to-one mapping relationship library to obtain the above-mentioned knowledge base.
[0071] Specifically, a corresponding extraction model is selected according to the type of data to be extracted, and the data to be extracted is input into the trained corresponding extraction model to output the extraction result of the data to be extracted; the entity, attribute and relationship mapping relationship library obtained according to the extraction result is used to realize the knowledge base design, and an entry is constructed for each entity in the knowledge base, and the entry name and description information are stored, and a table is created for each relationship type, which contains references to the starting entity, relationship type and end entity, and attribute information is added for each entity, and finally the extraction result is imported. Specifically, the entities, relationships and attributes in the extraction result are mapped with the patterns in the knowledge base and imported into the entity, attribute and relationship mapping relationship library to realize the knowledge base construction.
[0072] This embodiment studies knowledge extraction methods suitable for structured, semi-structured and unstructured data in view of the complex and diverse forms of regulatory data. Through targeted knowledge extraction methods, useful information in these data can be effectively extracted, thereby improving the efficiency of data utilization. For different types of data forms, corresponding knowledge extraction methods are adopted to ensure the flexibility and adaptability of the method. This means that even in the face of new data types or changes in data formats, the knowledge extraction method can be quickly adjusted and optimized to ensure the continuous use of data and maximize its value.
[0073] By constructing knowledge of power grid operation data, equipment data, regulations, accident plans and reports in the field of power dispatching, a basic database in the field of power dispatching is constructed. On this basis, text mining, association analysis, entity extraction, relationship extraction and other technologies are used to carry out dynamic extraction and knowledge mapping research on power grid dispatching knowledge elements, establish a complete knowledge graph in the field of power dispatching, realize the integration and precipitation of dispatching knowledge, and realize knowledge discovery in the dispatching field by constructing a knowledge graph in the dispatching field, providing knowledge enhancement for large models.
[0074] In order to enable those skilled in the art to more clearly understand the technical solution of the present application, the implementation process of the knowledge base construction method for power grid scheduling of the present application will be described in detail below in combination with specific embodiments.
[0075] This embodiment relates to a specific knowledge base construction method for power grid dispatching, such as Figure 2 As shown, including:
[0076] Step S1: acquiring power grid dispatching data, performing data cleaning on the power grid dispatching data; performing data conversion on the cleaned power grid dispatching data; performing data standardization on the converted power grid dispatching data to obtain standardized power grid dispatching data;
[0077] Step S2: dividing the power grid dispatching data to obtain a data division result, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data division result includes at least structured data, semi-structured data and unstructured data;
[0078] Step S3: When the power grid dispatching data is structured data, the RNN model is trained through multiple sets of data to obtain a structured data extraction model; when the power grid dispatching data is semi-structured data, the CRF model is trained through multiple sets of data to obtain a semi-structured data extraction model; when the power grid dispatching data is unstructured data, the BERT model is trained through multiple sets of data to obtain an unstructured data extraction model;
[0079] Step S4: Analyze the power grid dispatching data through the data extraction model according to the data division result to obtain entity data, attribute data and relationship data, wherein the data extraction model is obtained through deep learning training of multiple groups of data, and each group of data in the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels;
[0080] Step S5: Construct a knowledge base based on entity data, attribute data and relationship data.
[0081] The embodiment of the present application also provides a knowledge base construction device for power grid dispatching. It should be noted that the knowledge base construction device for power grid dispatching in the embodiment of the present application can be used to execute the knowledge base construction method for power grid dispatching provided in the embodiment of the present application. The device is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware for a predetermined function. Although the device described in the following embodiments is preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceived.
[0082] The following is an introduction to the knowledge base construction device for power grid dispatching provided in an embodiment of the present application.
[0083] Figure 3 1 is a structural block diagram of a knowledge base construction device for power grid dispatching provided in an embodiment of the present application. Figure 3 As shown, the device comprises:
[0084] A partitioning unit 10 is used to obtain power grid dispatching data, partition the power grid dispatching data, and obtain a data partitioning result, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data partitioning result includes at least structured data, semi-structured data, and unstructured data;
[0085] The analysis unit 20 is used to analyze the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, wherein the above-mentioned data extraction model is obtained through deep learning training of multiple groups of data, and each group of data in the above-mentioned multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels;
[0086] The construction unit 30 is used to construct a knowledge base according to the above entity data, the above attribute data and the above relationship data.
[0087] In the above device, the power grid dispatching data is obtained, the power grid dispatching data is divided, and the data division result is obtained; then, the power grid dispatching data is analyzed by the data extraction model according to the data division result to obtain entity data, attribute data and relationship data; finally, a knowledge base is constructed according to the entity data, attribute data and relationship data. In this way, the data extraction model is used to accurately extract the power grid dispatching data according to the division result of the power grid dispatching data, and the entity, attribute and relationship data are obtained, and a knowledge base is established, so as to achieve the effect of accurately extracting the dispatching data. That is, it solves the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0088] In one embodiment of the present application, the analysis unit includes a first analysis module, which is used to analyze the above-mentioned structured data through a structured data extraction model when the above-mentioned power grid dispatching data is the above-mentioned structured data, so as to obtain first entity data, first attribute data and first relationship data, wherein the above-mentioned structured data extraction model is obtained by deep learning training through multiple groups of first training data, and each group of data in the above-mentioned multiple groups of first training data includes: first entity training data and first entity training data label, first attribute training data and first attribute training data label, first relationship training data and first relationship training data label.
[0089] In one embodiment of the present application, the analysis unit includes a second analysis module, which is used to analyze the above-mentioned semi-structured data through a semi-structured data extraction model when the above-mentioned power grid dispatching data is the above-mentioned semi-structured data, so as to obtain second entity data, second attribute data and second relationship data, wherein the above-mentioned semi-structured data extraction model is obtained by deep learning training through multiple groups of second training data, and each group of data in the above-mentioned multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels.
[0090] In one embodiment of the present application, the analysis unit includes a third analysis module, which is used to analyze the above-mentioned unstructured data through an unstructured data extraction model when the above-mentioned power grid dispatching data is the above-mentioned unstructured data, so as to obtain third entity data, third attribute data and third relationship data, wherein the above-mentioned unstructured data extraction model is obtained by deep learning training through multiple groups of third training data, and each group of data in the above-mentioned multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels.
[0091] In one embodiment of the present application, the above-mentioned device also includes a first training unit, a second training unit and a third training unit. The first training unit is used to train the RNN model through multiple groups of data when the above-mentioned power grid dispatching data is the above-mentioned structured data to obtain the above-mentioned structured data extraction model; the second training unit is used to train the CRF model through multiple groups of data when the above-mentioned power grid dispatching data is the above-mentioned semi-structured data to obtain the above-mentioned semi-structured data extraction model; the third training unit is used to train the BERT model through multiple groups of data when the above-mentioned power grid dispatching data is the above-mentioned unstructured data to obtain the above-mentioned unstructured data extraction model.
[0092] In one embodiment of the present application, the construction unit includes a construction module and an import module. The construction module is used to construct an entity, attribute and relationship one-to-one mapping relationship library; the import module is used to import the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data into the above-mentioned entity, attribute and relationship one-to-one mapping relationship library to obtain the above-mentioned knowledge base.
[0093] In one embodiment of the present application, the above-mentioned device also includes a first processing unit, a second processing unit and a third processing unit. After acquiring the power grid dispatching data, the first processing unit is used to perform data cleaning on the above-mentioned power grid dispatching data; the second processing unit is used to perform data conversion on the above-mentioned power grid dispatching data after data cleaning; and the third processing unit is used to perform data standardization on the above-mentioned power grid dispatching data after data conversion to obtain standardized above-mentioned power grid dispatching data.
[0094] The above-mentioned knowledge base construction device for power grid dispatching includes a processor and a memory, and the above-mentioned division unit, analysis unit and construction unit are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions. The above-mentioned modules are all located in the same processor; or, the above-mentioned modules are located in different processors in the form of any combination.
[0095] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters can be adjusted to solve the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0096] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0097] An embodiment of the present invention provides a computer-readable storage medium, which includes a stored program, wherein when the program is running, the device where the computer-readable storage medium is located is controlled to execute the knowledge base construction method for power grid scheduling.
[0098] Specifically, the knowledge base construction method for power grid dispatching includes:
[0099] Step S101, acquiring power grid dispatching data, dividing the power grid dispatching data, and obtaining data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data division results at least include structured data, semi-structured data, and unstructured data;
[0100] Specifically, the acquisition of power grid dispatching data can be carried out by using crawler technology through keywords in the field of power grid dispatching. The power grid dispatching data may include equipment operating status, fault records, operation logs, dispatching instructions, etc.
[0101] Step S102, analyzing the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division result to obtain entity data, attribute data and relationship data, wherein the above-mentioned data extraction model is obtained through deep learning training of multiple groups of data, and each group of data in the above-mentioned multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels;
[0102] Specifically, obtaining the above-mentioned multiple groups of data includes: defining entities, attributes, and relationships based on the above-mentioned structured data, the above-mentioned semi-structured data, and the above-mentioned unstructured data, respectively, obtaining corresponding entity data, attribute data, and relationship data, respectively, and assigning corresponding entity labels, attribute labels, and relationship labels to the above-mentioned entity data, the above-mentioned attribute data, and the above-mentioned relationship data.
[0103] Step S103: construct a knowledge base based on the entity data, the attribute data and the relationship data.
[0104] Specifically, after building the knowledge base, when a power grid failure occurs, structured data can be used to quickly locate fault-related equipment, parameters, or events by directly querying the knowledge base. For semi-structured and unstructured data, the entities, attributes, and relationships extracted by the model help to quickly identify key information related to the fault.
[0105] Optionally, the power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, including: when the power grid dispatching data is the structured data, the structured data is analyzed by a structured data extraction model to obtain first entity data, first attribute data and first relationship data, wherein the structured data extraction model is obtained by deep learning training through multiple groups of first training data, and each group of data in the multiple groups of first training data includes: first entity training data and first entity training data label, first attribute training data and first attribute training data label, first relationship training data and first relationship training data label.
[0106] Optionally, the power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, and also includes: when the power grid dispatching data is the semi-structured data, the semi-structured data is analyzed by a semi-structured data extraction model to obtain second entity data, second attribute data and second relationship data, wherein the semi-structured data extraction model is obtained by deep learning training through multiple groups of second training data, and each group of data in the multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels.
[0107] Optionally, the power grid dispatching data is analyzed through a data extraction model according to the above data division results to obtain entity data, attribute data and relationship data, and also includes: when the above power grid dispatching data is the above unstructured data, the unstructured data is analyzed through an unstructured data extraction model to obtain third entity data, third attribute data and third relationship data, wherein the above unstructured data extraction model is obtained through deep learning training of multiple groups of third training data, and each group of data in the above multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels.
[0108] Optionally, the method includes: when the power grid dispatching data is the structured data, training the RNN model through multiple sets of data to obtain the structured data extraction model; when the power grid dispatching data is the semi-structured data, training the CRF model through multiple sets of data to obtain the semi-structured data extraction model; when the power grid dispatching data is the unstructured data, training the BERT model through multiple sets of data to obtain the unstructured data extraction model.
[0109] Optionally, constructing a knowledge base based on the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data includes: constructing an entity, attribute and relationship one-to-one mapping relationship library; importing the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data into the above-mentioned entity, attribute and relationship one-to-one mapping relationship library to obtain the above-mentioned knowledge base.
[0110] Optionally, after acquiring the power grid dispatching data, the method further includes: performing data cleaning on the power grid dispatching data; performing data conversion on the power grid dispatching data after data cleaning; and performing data standardization on the power grid dispatching data after data conversion to obtain standardized power grid dispatching data.
[0111] An embodiment of the present invention provides an electronic device, including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, at least the following steps are implemented: obtaining power grid dispatching data, dividing the power grid dispatching data, and obtaining data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data division results at least include structured data, semi-structured data, and unstructured data; analyzing the power grid dispatching data through a data extraction model according to the data division results to obtain entity data, attribute data, and relationship data, wherein the data extraction model is obtained through deep learning training of multiple groups of data, and each group of the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; building a knowledge base based on the entity data, the attribute data, and the relationship data. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.
[0112] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the above-mentioned method in each embodiment of the present application: obtaining power grid dispatching data, dividing the above-mentioned power grid dispatching data, and obtaining data division results, wherein the above-mentioned power grid dispatching data is data that characterizes the operating conditions and control and regulation of the power grid, and the above-mentioned data division results at least include structured data, semi-structured data and unstructured data; analyzing the above-mentioned power grid dispatching data through a data extraction model according to the above-mentioned data division results to obtain entity data, attribute data and relationship data, wherein the above-mentioned data extraction model is obtained through deep learning training of multiple groups of data, and each group of the above-mentioned multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; building a knowledge base according to the above-mentioned entity data, the above-mentioned attribute data and the above-mentioned relationship data.
[0113] By acquiring the power grid dispatching data, the power grid dispatching data is divided to obtain the data division results; then the power grid dispatching data is analyzed through the data extraction model according to the data division results to obtain entity data, attribute data and relationship data; finally, a knowledge base is constructed based on the entity data, attribute data and relationship data. In this way, according to the division results of the power grid dispatching data, the data extraction model is used to accurately extract the power grid dispatching data, obtain entity, attribute and relationship data, and establish a knowledge base, so as to achieve the effect of accurately extracting the dispatching data. That is, it solves the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0114] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.
[0115] 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.
[0116] 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.
[0117] 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.
[0118] 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 The steps for the functions specified in one or more boxes.
[0119] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0120] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0121] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0122] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0123] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:
[0124] 1) The knowledge base construction method of the power grid dispatching of the present application obtains the power grid dispatching data, divides the power grid dispatching data, and obtains the data division result; then, the power grid dispatching data is analyzed by the data extraction model according to the data division result to obtain entity data, attribute data and relationship data; finally, the knowledge base is constructed according to the entity data, attribute data and relationship data. In this way, the data extraction model is used to accurately extract the power grid dispatching data according to the division result of the power grid dispatching data, and the entity, attribute and relationship data are obtained, and a knowledge base is established, so as to achieve the effect of accurately extracting the dispatching data. That is, it solves the problem that the information extraction model of the power system in the prior art cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0125] 2) The knowledge base construction device for power grid dispatching of the present application obtains power grid dispatching data, divides the power grid dispatching data, and obtains data division results; then analyzes the power grid dispatching data through a data extraction model according to the data division results to obtain entity data, attribute data, and relationship data; finally, a knowledge base is constructed according to the entity data, attribute data, and relationship data. In this way, the data extraction model is used to accurately extract data from the power grid dispatching data according to the division results of the power grid dispatching data, and entity, attribute, and relationship data are obtained, and a knowledge base is established, achieving the effect of accurately extracting dispatching data. That is, it solves the problem in the prior art that the information extraction model of the power system cannot accurately extract more useful information, resulting in the inability to obtain accurate dispatching data from the knowledge base, which has a certain impact on fault diagnosis.
[0126] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.
Claims
1. A knowledge base construction method for power grid dispatching, It is characterized in that include: Acquire power grid dispatching data, divide the power grid dispatching data, and obtain data division results, wherein the power grid dispatching data is data characterizing the operating conditions and control regulation of the power grid, and the data division results at least include structured data, semi-structured data, and unstructured data; According to the data division result, the power grid dispatching data is analyzed by a data extraction model to obtain entity data, attribute data and relationship data, wherein the data extraction model is obtained by deep learning training through multiple groups of data, and each group of data in the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; A knowledge base is constructed according to the entity data, the attribute data and the relationship data.
2. The knowledge base construction method according to claim 1, It is characterized in that The power grid dispatching data is analyzed by a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, including: In the case where the power grid dispatching data is the structured data, the structured data is analyzed by a structured data extraction model to obtain first entity data, first attribute data and first relationship data, wherein the structured data extraction model is obtained by deep learning training through multiple groups of first training data, and each group of data in the multiple groups of first training data includes: first entity training data and first entity training data labels, first attribute training data and first attribute training data labels, first relationship training data and first relationship training data labels.
3. The knowledge base construction method according to claim 1, It is characterized in that According to the data division result, the power grid dispatching data is analyzed by a data extraction model to obtain entity data, attribute data and relationship data, and also includes: In the case where the power grid dispatching data is the semi-structured data, the semi-structured data is analyzed by a semi-structured data extraction model to obtain second entity data, second attribute data and second relationship data, wherein the semi-structured data extraction model is obtained by deep learning training through multiple groups of second training data, and each group of data in the multiple groups of second training data includes: second entity training data and second entity training data labels, second attribute training data and second attribute training data labels, second relationship training data and second relationship training data labels.
4. The knowledge base construction method according to claim 1, It is characterized in that According to the data division result, the power grid dispatching data is analyzed by a data extraction model to obtain entity data, attribute data and relationship data, and also includes: In the case where the power grid dispatching data is the unstructured data, the unstructured data is analyzed by an unstructured data extraction model to obtain third entity data, third attribute data and third relationship data, wherein the unstructured data extraction model is obtained by deep learning training through multiple groups of third training data, and each group of data in the multiple groups of third training data includes: third entity training data and third entity training data labels, third attribute training data and third attribute training data labels, third relationship training data and third relationship training data labels.
5. The knowledge base construction method according to claim 1, It is characterized in that The method comprises: In the case where the power grid dispatching data is the structured data, the RNN model is trained by using multiple groups of data to obtain the structured data extraction model; In the case where the power grid dispatching data is the semi-structured data, the CRF model is trained by using multiple groups of data to obtain the semi-structured data extraction model; In the case where the power grid dispatching data is the unstructured data, the BERT model is trained using multiple groups of data to obtain the unstructured data extraction model.
6. The knowledge base construction method according to claim 1, It is characterized in that Building a knowledge base according to the entity data, the attribute data and the relationship data includes: Build a one-to-one mapping relationship library among entities, attributes and relationships; The entity data, the attribute data and the relationship data are imported into the entity, attribute and relationship one-to-one mapping relationship library to obtain the knowledge base.
7. The knowledge base construction method according to claim 1, It is characterized in that After acquiring the power grid dispatching data, the method further includes: Performing data cleaning on the power grid dispatching data; Performing data conversion on the power grid dispatching data after data cleaning; The power grid dispatching data after data conversion is subjected to data standardization to obtain standardized power grid dispatching data.
8. A knowledge base construction device for power grid dispatching, It is characterized in that include: A partitioning unit, used to obtain power grid dispatching data, partition the power grid dispatching data, and obtain a data partitioning result, wherein the power grid dispatching data is data characterizing an operating condition and control regulation of a power grid, and the data partitioning result includes at least structured data, semi-structured data, and unstructured data; An analysis unit, configured to analyze the power grid dispatching data through a data extraction model according to the data division result to obtain entity data, attribute data and relationship data, wherein the data extraction model is obtained through deep learning training of multiple groups of data, and each group of data in the multiple groups of data includes: entity training data and entity training data labels, attribute training data and attribute training data labels, relationship training data and relationship training data labels; A construction unit is used to construct a knowledge base according to the entity data, the attribute data and the relationship data.
9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium includes a stored program, wherein when the program is executed, the device where the computer-readable storage medium is located is controlled to execute the knowledge base construction method for power grid dispatching according to any one of claims 1 to 7.
10. An electronic device, It is characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include a knowledge base construction method for executing the power grid dispatch described in any one of claims 1 to 7.
Citation Information
Patent Citations
Fan standard operation knowledge graph construction method and device and operation machine
CN115391567A
