Information determination method and device of power system and readable storage medium

Abstract

The invention discloses an information determination method and device of a power system and a readable storage medium. The method comprises the steps that an electric power document of an electric power system is acquired, and the electric power document is used for recording basic information of the electric power system; based on the power document, constructing a document graph used for indicating a relationship between a plurality of entities of the power system in the power document and attributes of the power system, and obtaining a business scene of the power system; target information of the electric power system is determined based on the business scene and the document atlas, and the target information is used for indicating basic knowledge related to the electric power system and / or the state of a transformer in the electric power system. The technical problem of low efficiency of determining the basic information of the power system is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of language processing, and in particular to an information determination method, device and readable storage medium of an electric power system. Background Art

[0002] At present, with the acceleration of the digitalization process of the power industry, the digital transformation of power standards has become more urgent. Traditional power standards mainly exist in the form of paper documents or static electronic documents. Therefore, the information of the power system is generally determined based on the documents of the power system.

[0003] In related technologies, the retrieval and application of standard documents rely on manual reading and understanding, which is inefficient and difficult to meet the rapidly changing needs of power systems. In addition, due to the complexity and professionalism of power standard documents, the manual reading execution mode is inefficient, and there is a deviation in the asynchronous execution of associated standard updates. The application effect of existing models in standard document mining is relatively limited, and it is difficult to accurately identify and extract key information in standard documents, and it is also impossible to effectively establish knowledge associations between standard documents. Therefore, there is a technical problem of low efficiency in determining the basic information of the power system.

[0004] With respect to the above-mentioned technical problem of low efficiency in determining basic information of the power system, no effective solution has been proposed so far. Summary of the invention

[0005] Embodiments of the present invention provide a method, device and readable storage medium for determining information of an electric power system, so as to at least solve the technical problem of low efficiency in determining basic information of the electric power system.

[0006] According to one aspect of an embodiment of the present invention, a method for determining information of an electric power system is provided. The method may include: obtaining an electric power document of the electric power system, wherein the electric power document is used to record basic information of the electric power system; constructing a document graph based on the electric power document, and obtaining a business scenario of the electric power system, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; determining target information of the electric power system based on the business scenario and the document graph, wherein the target information is used to indicate basic knowledge related to the electric power system and / or the status of a transformer in the electric power system.

[0007] Optionally, based on the power document, a document graph is constructed, including: preprocessing the power document to obtain a preprocessed power document; disassembling the preprocessed power document to obtain fact tuples and condition tuples, wherein the fact tuple is used to indicate a fact main clause recorded in the power document, and the condition tuple is used to indicate a conditional clause recorded in the power document; integrating the fact tuples and the condition tuples to obtain a document graph.

[0008] Optionally, preprocessing the power document to obtain a preprocessed power document includes: converting the power document into a digital format; and decomposing text information in the converted power document to obtain the preprocessed power document.

[0009] Optionally, the preprocessed power document is disassembled to obtain fact tuples and condition tuples, including: disassembling the preprocessed power document to determine the attributes of multiple sentences in the power document; using a pattern matching strategy to extract entity information in the preprocessed power document, wherein the pattern matching strategy is a rule for extracting entity information in the power document, and the entity information is used to indicate the relationship between multiple entities in the power system; based on the entity information and attributes, determining fact tuples and condition tuples.

[0010] Optionally, based on the business scenario and the document graph, target information of the power system is determined, including: matching the information in the business scenario with the information in the document graph to obtain multiple matching results; determining the information in the document graph with the highest similarity among the multiple matching results as the target information, wherein the similarity is used to indicate the degree of similarity between the information in the business scenario and the information in the document graph.

[0011] Optionally, the information determination method of the power system further includes: storing non-text information in the power document.

[0012] According to another aspect of an embodiment of the present invention, there is also provided an information determination device for an electric power system. The device may include: an acquisition unit, used to acquire an electric power document of the electric power system, wherein the electric power document is used to record basic information of the electric power system; a construction unit, used to construct a document graph based on the electric power document, and acquire a business scenario of the electric power system, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; a determination unit, used to determine target information of the electric power system based on the business scenario and the document graph, wherein the target information is used to indicate basic knowledge related to the electric power system and / or the status of the transformer in the electric power system.

[0013] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is also provided, the computer-readable storage medium including a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the information determination method of the power system in the embodiment of the present invention.

[0014] According to another aspect of an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein when the program is run, the method for determining information of the power system in the embodiment of the present invention is executed.

[0015] According to another aspect of the embodiment of the present invention, a computer program product is also provided, which includes computer instructions, and when the computer instructions are executed by a processor, the method for determining information of the power system in the embodiment of the present invention is implemented.

[0016] In an embodiment of the present invention, an electric power document of an electric power system is obtained, wherein the electric power document is used to record basic information of the electric power system; based on the electric power document, a document graph is constructed, and a business scenario of the electric power system is obtained, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; based on the business scenario and the document graph, target information of the electric power system is determined, wherein the target information is used to indicate basic knowledge related to the electric power system, and / or the status of the transformer in the electric power system. That is to say, the embodiment of the present invention solves the technical problem of low efficiency in determining the basic information of the electric power system by processing the electric power document of the electric power system, constructing a document graph, and automatically determining the basic information of the electric power system in combination with the business scenario, thereby achieving the technical effect of improving the efficiency of determining the basic information of the electric power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 is a flow chart of a method for determining information of a power system according to an embodiment of the present invention;

[0019] Figure 2 A flowchart of digitalization of electric power standards based on a conditional event map according to an embodiment of the present invention;

[0020] Figure 3 It is a schematic diagram of an overall framework of digitalization of electric power standards based on a conditional event graph according to an embodiment of the present invention;

[0021] Figure 4 is a flow chart of a knowledge fusion method based on similarity according to an embodiment of the present invention;

[0022] Figure 5 It is a schematic diagram of a digital power standard conditional event map according to an embodiment of the present invention;

[0023] Figure 6 is a schematic diagram of a production rule model according to an embodiment of the present invention;

[0024] Figure 7 is a schematic diagram of an information determination device for a power system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0026] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, functional component or device that includes 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, functional components or devices.

[0027] According to an embodiment of the present invention, an embodiment of a method for determining information of an electric power system 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.

[0028] Figure 1 is a flow chart of a method for determining information of a power system according to an embodiment of the present invention. Figure 1 As shown, the method may include the following steps:

[0029] Step S101, obtaining power documents of the power system.

[0030] In the technical solution provided in the above step S101 of the present invention, the power document is used to record basic information of the power system, wherein the power document may also be referred to as a power standard document or a standard document.

[0031] In this embodiment, the power document of the power system is obtained. The power document may at least include: power standard documents related to transformer status evaluation. For example, power standard documents related to transformer status evaluation include national standards, industry standards and enterprise standards.

[0032] Step S102: construct a document graph based on the power document and obtain the business scenario of the power system.

[0033] In the technical solution provided in the above step S102 of the present invention, the document map is used to indicate the relationship between multiple entities of the power system and the attributes of the power system in the power document. The document map can also be called a conditional reason map or a reason map.

[0034] In this embodiment, after obtaining the power document of the power system in step S101, a document graph is constructed based on the power document, and the business scenario of the power system is obtained. For example, the document graph is constructed by using semantic role labeling technology, sentence decomposition of the power document, etc. This is only an illustrative example and does not limit the specific method of constructing the document graph.

[0035] For example, the entity recognition submodule uses semantic role labeling technology and takes the predicate argument structure as the decomposition center to divide the sentence structure into factual main clauses and conditional subordinate clauses, and identify entities and attributes. The relationship extraction submodule uses pattern matching rules to extract key information from standard sentences and map it to the graph to build the relationship between entities. Through the knowledge fusion submodule, using the pretrained model, the predicate and standard predicate nodes are converted into 300-dimensional word vectors to clean and integrate redundant information. A set of fact tuples and conditional tuples is constructed to form a conditional matter graph.

[0036] Step S103, determining target information of the power system based on the business scenario and the document graph.

[0037] In the technical solution provided in the above step S103 of the present invention, the target information is used to indicate basic knowledge related to the power system and / or the status of the transformer in the power system.

[0038] In this embodiment, after constructing the document graph and acquiring the business scenario of the power system in step S102, the target information of the power system is determined according to the business scenario and the document graph. For example, the target information of the power system is determined by matching the information in the business scenario with the end point information of the document graph. This is only an illustrative example and does not limit the specific method of determining the target information of the power system.

[0039] For example, the similarity calculation unit calculates the similarity between entities or attributes, uses the cosine similarity method to match the obtained word vectors for text similarity, and selects the canonical predicate node with the highest similarity as the matching result. The rule reasoning unit performs reasoning based on preset rules, builds a model using the graph projection rule set, and extracts transformer status evaluation parameters from the database to accurately evaluate the transformer status or determine the basic information of the power system, that is, the target information of the power system.

[0040] It should be noted that the above embodiments can be implemented through a power standard digital system based on a conditional event graph.

[0041] In the above steps S101 to S103 of the present invention, an electric power document of the electric power system is obtained, wherein the electric power document is used to record basic information of the electric power system; based on the electric power document, a document graph is constructed, and a business scenario of the electric power system is obtained, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; based on the business scenario and the document graph, target information of the electric power system is determined, wherein the target information is used to indicate basic knowledge related to the electric power system, and / or the status of the transformer in the electric power system. In other words, the embodiment of the present invention processes the electric power document of the electric power system, constructs a document graph, and automatically determines the basic information of the electric power system in combination with the business scenario, thereby solving the technical problem of low efficiency in determining the basic information of the electric power system and achieving the technical effect of improving the efficiency of determining the basic information of the electric power system.

[0042] The above method of this embodiment is further introduced below.

[0043] As an optional implementation method, a document graph is constructed based on an electric power document, including: preprocessing the electric power document to obtain a preprocessed electric power document; disassembling the preprocessed electric power document to obtain fact tuples and condition tuples, wherein the fact tuple is used to indicate a fact main clause recorded in the electric power document, and the condition tuple is used to indicate a conditional clause recorded in the electric power document; integrating the fact tuples and the condition tuples to obtain a document graph.

[0044] In this embodiment, the power document is preprocessed to obtain a preprocessed power document, wherein the preprocessing may include: converting the power document into a digital format, optically decomposing the text information in the power document, converting the text in the image in the power document into editable text data, etc.

[0045] For example, these standard documents are converted into digital formats, such as PDF or image scans, for input into the system. Preprocessing and input: Use the Python-tesseract library to optically disassemble the text information in the PDF document and convert the text in the image into editable text data. Manually check the disassembled text to ensure the accuracy of the conversion, and use the sentence segmentation method to collect the data set. Non-text objects in the document, such as pictures, formulas, tables, etc., are stored in a standardized manner so that they can be easily retrieved in a relational database.

[0046] Optionally, after preprocessing the power document, the preprocessed power document is disassembled to obtain fact tuples and condition tuples, and the fact tuples and condition tuples are integrated to obtain a document graph. For example, the preprocessed power document is disassembled using a pretrained model, pattern matching rules, etc. to obtain fact tuples and condition tuples.

[0047] As an optional implementation method, preprocessing the power document to obtain the preprocessed power document includes: converting the power document into a digital format; and decomposing text information in the converted power document to obtain the preprocessed power document.

[0048] In this embodiment, the power document is converted into a digital format, and the text information in the power document is decomposed to obtain a preprocessed power document. For example, the Python-tesseract library is used to optically decompose the text information in the PDF document and convert the text in the image into editable text data.

[0049] Optionally, by preprocessing the power document, the comprehensiveness of the information in the power document is ensured, thereby improving the efficiency of determining the basic information of the power system.

[0050] As an optional implementation method, the preprocessed power document is disassembled to obtain fact tuples and condition tuples, including: disassembling the preprocessed power document to determine the attributes of multiple sentences in the power document; using a pattern matching strategy to extract entity information in the preprocessed power document, wherein the pattern matching strategy is a rule for extracting entity information in the power document, and the entity information is used to indicate the relationship between multiple entities in the power system; based on the entity information and attributes, determining fact tuples and conditional tuples.

[0051] In this embodiment, the preprocessed power document is disassembled to determine entities and attributes of multiple sentences in the power document. For example, the power document is disassembled by semantic role labeling technology to determine entities and attributes of multiple sentences in the power document.

[0052] For example, by using the entity recognition submodule and semantic role labeling technology, the sentence structure is divided into factual clauses and conditional clauses with the predicate argument structure as the decomposition center, and entities and attributes are identified.

[0053] Optionally, after determining the entities and attributes of multiple sentences in the power document, a pattern matching strategy is used to extract entity information in the preprocessed power document. For example, a relation extraction submodule is used to extract key information from standard sentences based on pattern matching rules, and map it to a graph to construct relationships between entities.

[0054] Optionally, fact tuples and condition tuples are constructed by integrating entity information and attributes.

[0055] As an optional implementation method, based on business scenarios and document graphs, target information of the power system is determined, including: matching information in the business scenario with information in the document graph to obtain multiple matching results; determining the information in the document graph with the highest similarity among the multiple matching results as the target information, wherein the similarity is used to indicate the degree of similarity between the information in the business scenario and the information in the document graph.

[0056] In this embodiment, the information in the business scenario is matched with the information in the document graph to obtain multiple matching results. For example, the similarity between entities or attributes is calculated by a similarity calculation unit, and the obtained word vectors are matched for text similarity using a cosine similarity method.

[0057] Optionally, information in a document graph with the highest similarity among multiple matching results is determined as target information, wherein the target information may be a transformer state evaluation parameter or an answer to a question raised by a user related to the power system.

[0058] As an optional implementation manner, a method for determining information of an electric power system includes: storing non-text information in an electric power document.

[0059] In this embodiment, non-text information in the power document is stored, wherein the non-text information may at least include: pictures, formulas, tables and other information.

[0060] Optionally, non-text information in the power document is stored in a standardized manner so as to be easily retrieved in a relational database.

[0061] It should be noted that the above embodiments can be implemented through a power standard digital system based on a conditional event graph.

[0062] In this embodiment, an electric power document of the electric power system is obtained, wherein the electric power document is used to record basic information of the electric power system; based on the electric power document, a document graph is constructed, and a business scenario of the electric power system is obtained, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; based on the business scenario and the document graph, target information of the electric power system is determined, wherein the target information is used to indicate basic knowledge related to the electric power system, and / or the status of the transformer in the electric power system. That is to say, the embodiment of the present invention solves the technical problem of low efficiency in determining the basic information of the electric power system by processing the electric power document of the electric power system, constructing a document graph, and automatically determining the basic information of the electric power system in combination with the business scenario, thereby achieving the technical effect of improving the efficiency of determining the basic information of the electric power system.

[0063] The technical solution of the embodiment of the present invention is illustrated below in conjunction with preferred implementation modes.

[0064] At present, with the acceleration of the digitalization process of the power industry, the digital transformation of power standards has become more urgent. Traditional power standards mainly exist in the form of paper documents or static electronic documents. Therefore, the information of the power system is generally determined based on the documents of the power system.

[0065] In the related technologies, the retrieval and application of standard documents rely on manual reading and understanding, which is inefficient and difficult to meet the rapidly changing needs of power systems. In addition, due to the complexity and professionalism of power standard documents, the manual reading execution mode is inefficient, and there are deviations in the asynchronous execution of related standard updates. The application effect of existing models in standard document mining is relatively limited, and it is difficult to accurately identify and extract key information in standard documents, and it is also impossible to effectively establish knowledge associations between standard documents. Therefore, there is a technical problem of low efficiency in determining the basic information of the power system. At present, no effective solution has been proposed for the above-mentioned technical problem of low efficiency in determining the basic information of the power system.

[0066] However, the embodiment of the present invention proposes a method for digitizing power standards based on conditional reasoning graphs, which collects and organizes power standard documents related to transformer status evaluation, converts these standard documents into digital format, and optically disassembles the text information in the documents; through semantic role labeling technology, with the predicate argument structure as the disassembly center, the sentence structure is divided into factual main clauses and conditional clauses, and entities and attributes are identified; based on pattern matching rules, key information is extracted from standard sentences and mapped to the graph to build relationships between entities; using a pre-trained model, predicates and standard predicate nodes are converted into 300-dimensional word vectors to clean and integrate redundant information. A set of fact tuples and conditional tuples is constructed to form a conditional reasoning graph, and the similarity between entities or attributes is calculated by a similarity calculation unit, and the standard predicate node with the highest similarity is selected as the matching result. Reasoning is performed according to preset rules, and a model is established using a graph projection rule set to extract transformer status evaluation parameters from a database to accurately evaluate the transformer status, thereby solving the technical problem of low efficiency in determining the basic information of the power system and achieving the technical effect of improving the efficiency of determining the basic information of the power system.

[0067] The embodiments of the present invention are further described below.

[0068] Figure 2 The present invention is a flowchart of a digital power standard based on a conditional event map according to an embodiment of the present invention. The analysis method includes the following steps:

[0069] Step S201, collecting and organizing power standard documents related to transformer status evaluation and preprocessing the power standard documents.

[0070] In this embodiment, power standard documents related to transformer status evaluation are collected and organized, including national standards, industry standards, and enterprise standards. These standard documents are converted into digital formats, such as PDF or image scans, for input into the system. Preprocessing and input: Use the Python-tesseract library to optically disassemble the text information in the PDF document and convert the text in the image into editable text data. The disassembled text is manually verified to ensure the accuracy of the conversion, and the data set is collected using the sentence segmentation method. Non-text objects in the document, such as pictures, formulas, tables, etc., are stored in a standardized manner so that they can be easily retrieved in a relational database.

[0071] Step S202, constructing a matter diagram according to the electric power standard document.

[0072] In this embodiment, the entity recognition submodule is used to divide the sentence structure into factual main clauses and conditional clauses by semantic role labeling technology, with the predicate argument structure as the disassembly center, and the entities and attributes are identified. The relationship extraction submodule is used to extract key information from standard sentences based on pattern matching rules, and map it to the graph to build the relationship between entities. Through the knowledge fusion submodule, the predicate and standard predicate nodes are converted into 300-dimensional word vectors using a pre-trained model to clean and integrate redundant information. A set of fact tuples and conditional tuples is constructed to form a conditional matter graph, which is stored in the Neo4j graph database.

[0073] Optionally, in a preferred implementation of the method, the method for extracting entities and their attributes is specifically:

[0074] Standard structure information entities are the main components of the SL structure layer, including: a sn ={standard document: standard name, standard code (such as "GB"), standard number, standard level or category, International Standard Classification (ICS) number, Chinese Standard Document Classification (CCS) number, release date, implementation date, release agency, standard status, replacement standard, replaced standard, drafting unit, competent department, centralized unit, drafter}; a ct ={Chapter title: title name, title number}; a st ={sentence: content, sentence number, link title, link title number}, in particular, the sentence number is represented by s i , where i is numbered based on the entire article.

[0075] The standard content information entity is a further decomposition of the standard content and an in-depth analysis of the standard content semantics and standard rule knowledge. The standard content information entity includes: a pr ={predicate: predicate relationship, belonging sentence}, e={entity: entity category, corresponding attribute}. Among them, Table 1 is a table of standard content entity categories and their corresponding attributes, and specific entity categories and their corresponding attributes and examples are partially shown in Table 1.

[0076] Table 1 Standard content entity categories and their corresponding attributes

[0077]

[0078] Optionally, the semantic role labeling (Semantic Role Labeling, SRL) is specifically: taking the predicate argument structure as the decomposition center, the sentence structure is divided into factual main clauses and conditional clauses. SRL is a key technology in natural language processing, and its goal is to map words, phrases or clauses in a sentence to specific semantic roles. The predicate usually refers to the core verb in a sentence, and the predicate argument structure is centered on the predicate to analyze the relationship between each component in the sentence and the predicate. The extracted predicate is used as a pr The nodes constitute the tuple layer TL, and the entity e associated with this predicate constitutes the entity layer EL. First, Table 2 is a dictionary table in the power field. As shown in Table 2, the standard content information needs to be segmented in units of sentences through the self-built power field dictionary. The word segmentation uses the HanLP Chinese word segmentation tool, and the introduction of the power field dictionary is taken from the "Electric Power Terminology (Third Edition)" Some nouns are shown in Table 2.

[0079] Table 2 Dictionary table of power industry

[0080]

[0081] Optionally, SRL is performed on the sentence after word segmentation. Table 3 is the Sexp predicate argument structure table. Taking the sentence Sexp "When the single deduction points of any state quantity and the total deduction points of the components meet the requirements of Table 3 at the same time, the evaluation is normal." as an example, the predicate argument structure of this sentence is Table 3, and the predicates that can be extracted are "satisfy" and "evaluate".

[0082] Table 3 Sexp predicate argument structure

[0083]

[0084] Optionally, Table 4 is a relation extraction rule table, and the method for extracting the relation and its attributes is specifically: a relation extraction method based on pattern matching rules is selected, and the relation extraction rules are shown in Table 4. This method can effectively extract key information from standard sentences, and accurately map the relationship between entities to the graph according to the preset pattern, thereby realizing the effective expression and presentation of things.

[0085] Table 4 Relationship extraction rules table

[0086]

[0087] Optionally, Table 5 is a standard predicate node table, and the standard predicate nodes are shown in Table 5. The extracted predicates and standard predicate nodes are converted into word vectors using word2vec. The word vector conversion model is pre-trained for the Chinese People's Daily corpus, the vector dimension is 300, and the training granularity includes characters and words. The obtained word vectors are matched for text similarity, and the cosine similarity is selected for similarity calculation, as shown in the following formula, and the standard predicate node with the highest similarity is selected as the matching result.

[0088]

[0089] Optionally, taking the sentence Sexp as an example, the extracted predicates are "satisfy" and "evaluate". "Satisfy" can take the canonical predicate nodes as "satisfy", and the similarity matching result is 1; "evaluate" can take the canonical predicate nodes as "assess", and the similarity matching result is.

[0090] Table 5 Standard predicate node table

[0091]

[0092] Optionally, the knowledge storage and update method is specifically as follows: using a Neo4j graph database to store the constructed conditional event graph for the equipment status evaluation standard. Neo4j is a native graph database engine that makes full use of the natural extension characteristics of the graph structure to optimize the query algorithm, thereby achieving efficient query performance, and can easily store and query any type of data and provide visualization tools to browse and visualize graph data intuitively and conveniently. In the conditional event graph for the equipment status evaluation standard, this flexibility enables the database to adapt to complex and changeable data structures and relationships, achieve accurate data analysis, and quickly respond to equipment status evaluation tasks, thereby more comprehensively reflecting the status and evaluation standards of the equipment. As long-term high-reliability knowledge, power standards have the ability to add new knowledge to the Neo4j database when the standards are updated, replaced, or added, and ensure the accuracy and completeness of the knowledge. Cypher is the query language of Neo4j, which can also be used to update data. By writing Cypher statements, nodes and relationships can be added, modified, or deleted. At the same time, Neo4j supports transaction processing, which means that multiple update operations can be completed in one transaction to ensure data consistency.

[0093] Step S203: construct a knowledge reasoning question-answering system.

[0094] In this embodiment, the similarity between entities or attributes is calculated by the similarity calculation unit, and the obtained word vectors are matched for text similarity using the cosine similarity method, and the canonical predicate node with the highest similarity is selected as the matching result. The rule reasoning unit performs reasoning according to the preset rules, establishes a model using the graph projection rule set, and extracts transformer status evaluation parameters from the database to accurately evaluate the transformer status.

[0095] Optionally, the specific steps of building the question-answering system are as follows: Data modeling and import: First, the standard knowledge of power equipment needs to be modeled. Based on the digital power standard event map established in this example, 1219 related nodes and 1610 relationships are screened for the transformer status evaluation field. A map projection is established for the screened map data to optimize query efficiency and quality.

[0096] Optionally, build a question-answering system: To implement natural language question-answering, natural language processing technology is needed to analyze the questions asked by users. Based on the user's questions, knowledge extraction technology is used to extract entities and related predicates in order to accurately understand the user's query intent. After that, the entities and related predicates are converted into Cypher query language for querying in the Neo4j database.

[0097] Optionally, query execution and result return: execute the converted Cypher query and get the result from the Neo4j database. The following will show some common problems and the corresponding return results.

[0098] Optionally, the rule-based reasoning device state evaluation method is specifically as follows: production rules are a form of knowledge with an associative relationship, which is used to characterize procedural knowledge or rules in artificial intelligence systems. Each production rule consists of two parts: a condition and an action. This rule can be expressed in the form of "if the condition is met, then perform the action", which is the so-called condition-action rule (CA rule). In the field of artificial intelligence, production rules are widely used in expert systems, natural language processing, intelligent control and other fields. By simulating human thinking processes and reasoning patterns, production rules can help machines make more intelligent decisions and judgments.

[0099] Step S204: Use the knowledge reasoning question answering system to output the query results or the analyzed transformer status.

[0100] In this embodiment, the user inputs a natural language query request through the query interface, and the system converts the natural language analysis into a structured query. The result display unit displays the query results or the analyzed transformer status to the user in a table form, providing intuitive data display.

[0101] Optionally, in the preferred implementation of the method, the specific query request and query result are: on the basis of the digital power standard based on the conditional causal graph, combined with the Cy language and graph algorithm, auxiliary knowledge question and answer and status assessment tasks for power equipment are implemented. This example will take the transformer status evaluation as an example to demonstrate the standard knowledge question and answer and status assessment business example of power equipment. Table 6 is a standard knowledge question and answer table for power equipment, and the specific results are shown as follows.

[0102] Table 6 Questions and answers on standard knowledge of power equipment

[0103]

[0104] Optionally, the result of the equipment status evaluation is specifically as follows: This example uses the example in "DL / T 1685-2017 Guidelines for Condition Evaluation of Oil-Immersed Transformers (Reactors)" (hereinafter referred to as the Guidelines) to verify the effectiveness of the digital power standards constructed in this article in equipment status evaluation business applications.

[0105] Optionally, the guideline case provides equipment status information as follows: "A SFPSZ9-150000 / 220 transformer was manufactured in March 2003 and put into operation in March 2004. No abnormalities were found in the routine test on March 11, 2005. On May 19, 2005, the acetylene and total hydrocarbon contents of the transformer were found to have increased sharply to 2.45uL / L and 348.55uL / L respectively during the routine oil chromatography test. Other oil chromatographic dissolved gas components also increased. After that, long-term follow-up detection was carried out. The test data is shown in Table 7. There are no abnormalities in other status quantities." Table 7 is the oil chromatography test data table of a transformer in the guideline, as shown in Table 7.

[0106] Table 7 Guidelines for a certain transformer oil chromatography test data table

[0107]

[0108] Optionally, Table 8 is a key information return result table, and knowledge extraction is performed on the above-mentioned equipment status information. For the key information of "transformer" and "dissolved gas in oil", the returned results are shown in Table 8.

[0109] Optionally, it can be seen from Table 8 that the physical parameters of "total hydrocarbons", "acetylene", "H2", "CO", "relative gas production rate" and "acetylene growth trend" still need to be added.

[0110] Table 8 Key information return result table

[0111]

[0112] Optionally, Table 9 is the status evaluation return result table. Physical parameters need to be added as attributes, and nodes are established using the data in Table 7 as corresponding attribute values. The label is data, and the graph is input and calculated and analyzed according to the judgment method and threshold. The returned result is Table 9.

[0113] Table 9 Status evaluation return result table

[0114]

[0115] Optionally, it can be concluded that the transformer status is abnormal, which is the same as the guideline case. The system proposed in the present invention only takes 13 milliseconds in the retrieval process, showing extremely high efficiency. On the basis of systematic data entry, the data call speed can be further optimized, thereby significantly shortening the time required for manual search and review of standards while ensuring the accuracy of retrieval. The digital power standard proposed in the present invention greatly improves the retrieval efficiency of standard knowledge. Through the verification of actual cases, the effectiveness of this method in retrieval and result judgment has been fully demonstrated.

[0116] Figure 3is a schematic diagram of an overall framework of digitalization of power standards based on a conditional event graph according to an embodiment of the present invention, such as Figure 3 As shown, the overall framework of the digitalization of power standards based on the conditional event graph includes a standard input layer, an event graph layer, a knowledge reasoning layer, and a knowledge application layer.

[0117] Figure 4 is a flow chart of a knowledge fusion method based on similarity according to an embodiment of the present invention. Figure 4 As shown, the knowledge fusion method includes the following steps:

[0118] Step S401, extracting predicates.

[0119] In this embodiment, predicates in the power standard document are extracted.

[0120] Step S402: pre-training corpus.

[0121] In this embodiment, the power standard document pre-training corpus is utilized.

[0122] Step S403: word vector conversion.

[0123] In this embodiment, word2vec is used to convert the extracted predicates and canonical predicate nodes into word vectors.

[0124] Step S404: calculating cosine similarity.

[0125] In this embodiment, text similarity matching is performed on the obtained word vectors, and cosine similarity is selected for similarity calculation.

[0126] Step S405: whether it is a maximum value.

[0127] In this embodiment, it is determined whether the current similarity is a maximum value. If it is a maximum value, step S406 is executed; if it is not a maximum value, step S408 is executed.

[0128] Step S406: whether the similarity is greater than 0.5.

[0129] In this embodiment, it is determined whether the current similarity is greater than 0.5. If it is greater than 0.5, step S407 is executed; if it is not greater than 0.5, step S409 is executed.

[0130] Step S407, outputting the similarity and its canonical predicate node.

[0131] In this embodiment, the calculated similarity and its canonical predicate node are output.

[0132] Step S408, standardizing predicate nodes.

[0133] In this embodiment, the predicate nodes are re-normalized and step S402 is re-executed.

[0134] Step S409: manual review.

[0135] In this embodiment, the similarity is manually reviewed.

[0136] Figure 5 is a schematic diagram of a digital power standard conditional event map according to an embodiment of the present invention, such as Figure 5 As shown, it contains more than 6,000 entities and 8,000 relationships, fully realizing the standard multi-granularity decomposition and the maximum degree of semantic relationship retention. Figure 5 The visualization effect of some knowledge graphs is shown. Through intuitive graphical display, we can understand the conditions and logical relationships between equipment status, evaluation criteria, physical parameters, and their positions in the overall knowledge network. Such visualization is not only intuitive, but also helps to deeply explore and analyze the complex associations in the graph, providing strong data support for equipment maintenance and management.

[0137] Figure 6 is a schematic diagram of a production rule model according to an embodiment of the present invention, such as Figure 6 As shown, the conditional event graph is projected, the standard names are split, and reasoning is performed based on the rule base to obtain parameters related to transformer status evaluation from the database.

[0138] In this embodiment, by collecting and organizing the power standard documents related to the transformer status evaluation, these standard documents are converted into digital format, and the text information in the documents is optically disassembled; by using the semantic role labeling technology, the sentence structure is divided into factual main clauses and conditional clauses with the predicate argument structure as the disassembly center, and the entities and attributes are identified; based on the pattern matching rules, key information is extracted from the standard sentences and mapped to the graph to build the relationship between entities; using the pre-trained model, the predicate and the standard predicate nodes are converted into 300-dimensional word vectors, and the redundant information is cleaned and integrated. A set containing fact tuples and condition tuples is constructed to form a conditional reasoning graph, and the similarity between entities or attributes is calculated by the similarity calculation unit, and the standard predicate node with the highest similarity is selected as the matching result. Reasoning is performed according to the preset rules, and a model is established using the graph projection rule set to extract transformer status evaluation parameters from the database to accurately evaluate the transformer status, thereby solving the technical problem of low efficiency in determining the basic information of the power system and achieving the technical effect of improving the efficiency of determining the basic information of the power system.

[0139] According to an embodiment of the present invention, there is also provided an information determination device for a power system. It should be noted that the information determination device for a power system can be used to execute the information determination method for a power system in the method embodiment.

[0140] Figure 7 is a schematic diagram of an information determination device for a power system according to an embodiment of the present invention. Figure 7 As shown, the power system information determination device 700 may include: an acquisition unit 701 , a construction unit 702 and a determination unit 703 .

[0141] The acquisition unit 701 is used to acquire a power document of the power system, wherein the power document is used to record basic information of the power system.

[0142] The construction unit 702 is used to construct a document graph based on the power document and obtain the business scenario of the power system, wherein the document graph is used to indicate the relationship between multiple entities of the power system and the attributes of the power system in the power document.

[0143] The determination unit 703 is used to determine target information of the power system based on the business scenario and the document map, wherein the target information is used to indicate basic knowledge related to the power system and / or the status of the transformer in the power system.

[0144] Optionally, construction unit 702 may include: a preprocessing module, used to preprocess the power document to obtain a preprocessed power document; a disassembly module, used to disassemble the preprocessed power document to obtain fact tuples and condition tuples, wherein the fact tuple is used to indicate the fact main clause recorded in the power document, and the condition tuple is used to indicate the conditional clause recorded in the power document; an integration module, used to integrate the fact tuples and condition tuples to obtain a document graph.

[0145] Optionally, the preprocessing module may include: a conversion submodule, used to convert the power document into a digital format; and a disassembly submodule, used to disassemble the text information in the converted power document to obtain a preprocessed power document.

[0146] Optionally, the disassembly module may include: a disassembly sub-module, used to disassemble the preprocessed power document and determine the attributes of multiple sentences in the power document; an extraction sub-module, used to extract entity information in the preprocessed power document using a pattern matching strategy, wherein the pattern matching strategy is a rule for extracting entity information in the power document, and the entity information is used to indicate the relationship between multiple entities in the power system; a determination sub-module, used to determine fact tuples and conditional tuples based on the entity information and attributes.

[0147] Optionally, the determination unit 703 may include: a matching module, used to match the information in the business scenario with the information in the document graph to obtain multiple matching results; a determination module, used to determine the information in the document graph with the highest similarity among the multiple matching results as the target information, wherein the similarity is used to indicate the degree of similarity between the information in the business scenario and the information in the document graph.

[0148] Optionally, the information determination device 700 of the power system may further include: a storage unit, configured to store non-text information in the power document.

[0149] In this embodiment, an electric power document of the electric power system is obtained, wherein the electric power document is used to record basic information of the electric power system; based on the electric power document, a document graph is constructed, and a business scenario of the electric power system is obtained, wherein the document graph is used to indicate the relationship between multiple entities of the electric power system and the attributes of the electric power system in the electric power document; based on the business scenario and the document graph, target information of the electric power system is determined, wherein the target information is used to indicate basic knowledge related to the electric power system, and / or the status of the transformer in the electric power system. That is to say, the embodiment of the present invention solves the technical problem of low efficiency in determining the basic information of the electric power system by processing the electric power document of the electric power system, constructing a document graph, and automatically determining the basic information of the electric power system in combination with the business scenario, thereby achieving the technical effect of improving the efficiency of determining the basic information of the electric power system.

[0150] According to an embodiment of the present invention, a computer-readable storage medium is further provided. The storage medium includes a stored program, wherein the program executes the method for determining information of the power system in the method embodiment.

[0151] According to an embodiment of the present invention, a processor is further provided. The processor is used to run a program, wherein the method for determining information of a power system in the method embodiment is executed when the program is run.

[0152] According to an embodiment of the present invention, a computer program product is further provided. The computer program product includes computer instructions. When the computer instructions are executed by a processor, the method for determining information of the power system in the method embodiment is implemented.

[0153] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0154] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0155] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0156] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed over multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0157] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0158] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent functional component, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software functional component, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store program codes.

[0159] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining information of a power system, characterized in that: include: Acquire a power document of the power system, wherein the power document is used to record basic information of the power system; Based on the power document, a document graph is constructed, and a business scenario of the power system is obtained, wherein the document graph is used to indicate the relationship between multiple entities of the power system in the power document and the attributes of the power system; Based on the business scenario and the document graph, target information of the power system is determined, wherein the target information is used to indicate basic knowledge related to the power system and / or a status of a transformer in the power system.

2. The method according to claim 1, characterized in that Based on the power document, a document graph is constructed, including: Preprocessing the power document to obtain the preprocessed power document; Decomposing the preprocessed power document to obtain a fact tuple and a condition tuple, wherein the fact tuple is used to indicate a fact main clause recorded in the power document, and the condition tuple is used to indicate a conditional subordinate clause recorded in the power document; The fact tuple and the condition tuple are integrated to obtain the document graph.

3. The method according to claim 2, characterized in that Preprocessing the power document to obtain the preprocessed power document includes: converting said electrical documentation into a digital format; The text information in the converted power document is disassembled to obtain the preprocessed power document.

4. The method according to claim 2, characterized in that: The preprocessed power document is disassembled to obtain fact tuples and condition tuples, including: Decomposing the preprocessed power document to determine the attributes of multiple sentences in the power document; Extracting entity information from the preprocessed power document using a pattern matching strategy, wherein the pattern matching strategy is a rule for extracting the entity information from the power document, and the entity information is used to indicate a relationship between multiple entities in the power system; Based on the entity information and the attributes, the fact tuple and the condition tuple are determined.

5. The method according to claim 1, characterized in that Determining target information of the power system based on the business scenario and the document graph includes: Matching the information in the business scenario with the information in the document graph to obtain multiple matching results; The information in the document graph with the highest similarity among the multiple matching results is determined as the target information, wherein the similarity is used to indicate the degree of similarity between the information in the business scenario and the information in the document graph.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: The non-text information in the power document is stored.

7. An information determination device for a power system, characterized in that: include: An acquisition unit, used for acquiring a power document of a power system, wherein the power document is used for recording basic information of the power system; A construction unit, configured to construct a document graph based on the power document and obtain a business scenario of the power system, wherein the document graph is used to indicate the relationship between multiple entities of the power system in the power document and attributes of the power system; A determination unit is used to determine target information of the power system based on the business scenario and the document map, wherein the target information is used to indicate basic knowledge related to the power system and / or the status of a transformer in the power system.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein when the program is executed by a processor, the device where the storage medium is located is controlled to execute the method according to any one of claims 1 to 6.

9. A processor, characterized in that: The processor is used to run a program, wherein the program executes the method according to any one of claims 1 to 6 when running.

10. A computer program product, characterized in that The computer program product comprises computer instructions, which implement the method according to any one of claims 1 to 6 when executed by a processor.

Images