Power grid intermediate station data model construction method and device, electronic equipment and storage medium

Abstract

The invention discloses a power grid intermediate station data model construction method and device, electronic equipment and a storage medium, and the method comprises the steps: responding to a model element adding request, adding a corresponding model element in a to-be-constructed data model, and converting a text description word of the model element into a corresponding word vector; calculating a first semantic distance between the text description word of the model element and the text description words of all elements in a preset power grid semantic knowledge base according to the word vector; sorting the first semantic distances, selecting a preset number of corresponding elements as target elements according to a sequence from small to large, and generating an element recommendation list of the model elements according to the target elements, and selecting corresponding elements by a user according to the element recommendation list to construct the power grid intermediate station data model. According to the invention, the construction efficiency of the power grid intermediate station data model can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a method, device, electronic device and storage medium for constructing a power grid middle platform data model. Background Art

[0002] The power grid middle platform is a concept in the enterprise architecture, which is located between the front end (business applications directly facing users) and the back end (the core business systems of the enterprise). In power grid operations, the middle platform system plays a role in connecting various business modules and realizing data sharing and service integration. The power grid middle platform is usually positioned as an enterprise-level multi-type large-data aggregation platform, providing unified data services for front-end applications based on a unified data model and through data tags. The construction of the power grid middle platform is a key step for power grid enterprises to realize data-driven business. It can not only improve the timeliness of data access, enhance data storage and computing capabilities, but also give full play to the value of data through data asset management and support data applications.

[0003] In the process of constructing the power grid middle platform, it is necessary to build a unified data model to ensure the aggregation, integration and sharing of data resources at all levels and in all specialties. The data model includes three parts: the conceptual model, the logical model and the physical model, which together form the basis of the power grid middle platform data architecture. With the progress of information technology, the demand for managing and visualizing the power grid data model is growing continuously. The State Grid unified data model (SG-CIM, Smart Grid Common Information Model) provides a framework for standardizing power grid data.

[0004] However, in the process of constructing the SG-CIM data model of the power grid middle platform, due to the lack of the function of automatically matching and identifying model elements through deep semantic processing in the existing power grid design system, designers need to spend a lot of time searching and connecting relevant elements one by one when conducting model design. Due to the existence of a large number of elements with similar semantics but different expressions, during the model construction process, it may be necessary to combine these elements or select one of them. When combining elements, designers need to spend a lot of time finding relevant elements one by one for combination. If only one is selected, designers may also have the problem of duplicate design due to their different expressions, reducing work efficiency. Summary of the Invention

[0005] The present invention provides a method, device, electronic device and storage medium for constructing a power grid middle platform data model to solve the technical problem that the existing power grid design system lacks the function of automatically matching and identifying model elements through deep semantic processing, and designers need to spend a lot of time searching and connecting relevant elements one by one when conducting model design, reducing work efficiency.

[0006] To solve the above technical problems, an embodiment of the present invention provides a method for constructing a power grid middle - platform data model, including: In response to a model element addition request, add a corresponding model element to the data model to be constructed, and convert the text description word of the model element into a corresponding word vector; wherein, the model elements include: power equipment, equipment attributes of electrical equipment, and equipment relationships between different electrical equipment; the equipment attributes include: rated voltage, rated current, capacity, and impedance; the equipment relationships include: equipment connection relationship and equipment subordination relationship; According to the word vector, calculate the first semantic distance between the text description word of the model element and the text description words of all elements in the preset power grid semantic knowledge base. Sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list for the model element, so that the user can select corresponding elements according to the element recommendation list to construct the power grid middle - platform data model.

[0007] As a preferred solution, the generating the corresponding element recommendation list according to each of the target elements includes: For each target element, obtain the historical addition frequency of the target element and the first weight corresponding to the historical addition frequency; According to the model elements already added in the data model to be constructed, calculate the element correlation between the target element and each of the already added model elements, and obtain the second weight corresponding to the element correlation; Calculate the second semantic distance between the text description word of the target element and the text description words of the already added model elements and the third weight corresponding to the second semantic distance; According to the first weight, second weight, and third weight, perform a weighted sum calculation on the historical addition frequency, element correlation, and second semantic distance to obtain the comprehensive score corresponding to the target element; According to the comprehensive score, select a preset number of target elements in descending order to generate the element recommendation list.

[0008] As a preferred solution, the generation of the power grid semantic knowledge base includes: Obtain the preset power grid text; Perform text segmentation processing, part - of - speech tagging processing, and named entity recognition processing on the power grid text respectively, identify the words in the text and the named entities of each word, and then extract the corresponding keywords from all words according to the named entities; Calculate the semantic similarity between each of the keywords, compare the semantic similarity with a preset similarity threshold, and obtain the correlation between the keywords according to the comparison result; Construct a corresponding power grid domain knowledge graph according to the keywords and the correlation between the keywords, and obtain a corresponding power grid semantic knowledge base according to the power grid domain knowledge graph.

[0009] As a preferred solution, perform word segmentation processing, part-of-speech tagging processing, and named entity recognition processing on the power grid text respectively, identify the words in the text and the named entities of each word, and then extract the corresponding key terms from all the words according to the named entities, including; Obtain a preset power grid professional dictionary, and perform word segmentation processing and part-of-speech tagging processing on the power grid text according to the power grid professional dictionary to obtain the words in the power grid text and the corresponding part-of-speech information of the words; Perform named entity recognition processing on the words according to the words and the corresponding part-of-speech information of the words to identify the named entities of each word; wherein, the named entities include: device entities, parameter entities, and operation entities; Extract the corresponding keywords from all the words according to the named entities.

[0010] As a preferred solution, the calculating the semantic similarity between each of the keywords, comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between the keywords according to the comparison result includes: Convert each of the keywords into a keyword vector of a preset dimension, and calculate the cosine similarity between the keyword vectors; Identify the specific part-of-speech corresponding to each of the keywords according to a preset specific part-of-speech judgment rule, and obtain the weight corresponding to each of the specific part-of-speeches; wherein, the specific part-of-speeches include: professional terms and general vocabulary; Obtain the keyword weight corresponding to each of the keywords according to the weight corresponding to each of the specific part-of-speeches and the specific part-of-speech corresponding to each of the keywords, take the product of the keyword weight and the cosine similarity as the semantic similarity of the keyword, compare the semantic similarity with a preset similarity threshold, and obtain the correlation between the keywords according to the comparison result.

[0011] As a preferred solution, the constructing a corresponding power grid domain knowledge graph according to the keywords and the correlation between the keywords, and obtaining a corresponding power grid semantic knowledge base according to the power grid domain knowledge graph includes: Extract power grid equipment, equipment attributes of the power grid equipment, and equipment relationships of the power grid equipment from the power grid text, and use the power grid equipment, equipment attributes, and equipment relationships as basic elements; According to the relevance between each keyword, establish associated edges for the relevant keywords, and then construct a corresponding knowledge graph in the power grid field based on the associated edges and the basic elements, and obtain a corresponding power grid semantic knowledge base according to the knowledge graph in the power grid field.

[0012] As a preferred solution, calculate the semantic distance between the text description words of the model elements and the text description words of all elements in the preset power grid semantic knowledge base through the following formula: d(e1, e2) = 1 - cos(v(e1), v(e2)); Where d(e1, e2) is the semantic distance between element e1 and element e2, v(e1) is the semantic vector of element e1, v(e2) is the semantic vector of element e2, and cos is the cosine similarity.

[0013] Based on the above embodiments, another embodiment of the present invention provides a device for constructing a power grid middle platform data model, including: a model element adding module, a semantic distance calculation module, and an element recommendation list generation module; The model element adding module is used to respond to a model element adding request, add corresponding model elements to the data model to be constructed, and convert the text description words of the model elements into corresponding word vectors; wherein, the model elements include: power equipment, equipment attributes of electrical equipment, and equipment relationships between different electrical equipment; the equipment attributes include: rated voltage, rated current, capacity, and impedance; the equipment relationships include: equipment connection relationships and equipment subordination relationships; The semantic distance calculation module is used to calculate the first semantic distance between the text description words of the model elements and the text description words of all elements in the preset power grid semantic knowledge base according to the word vectors; The element recommendation list generation module is used to sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list of the model elements according to each of the target elements, so that the user can select corresponding elements according to the element recommendation list to construct the power grid middle platform data model.

[0014] Based on the above embodiments, another embodiment of the present invention provides an electronic device, the device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the method for constructing a power grid middle platform data model described in the above embodiments of the present invention.

[0015] Based on the above embodiments, another embodiment of the present invention provides a storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the storage medium is located to execute the power grid middle - platform data model construction method described in the above - mentioned embodiments of the present invention.

[0016] Compared with the prior art, the embodiments of the present invention have the following beneficial effects: The present invention provides a method for constructing a power grid middle - platform data model. In response to a model element addition request, corresponding model elements are added to the data model to be constructed, and the text description words of the model elements are converted into corresponding word vectors. Among them, the model elements include: power equipment, device attributes of electrical equipment, and device relationships between different electrical equipment; the device attributes include: rated voltage, rated current, capacity, and impedance; the device relationships include: device connection relationships and device subordination relationships; according to the word vectors, calculate the first semantic distance between the text description words of the model elements and the text description words of all elements in the preset power grid semantic knowledge base; sort the first semantic distances, and select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list for the model elements according to each of the target elements. The elements in the element recommendation list generated by the present invention have similar text descriptions, that is, similar semantics, to the newly added model elements in the data model. When the user newly adds a model element, the present invention selects other elements with similar semantics to the model element from the preset power grid semantic knowledge base to generate an element recommendation list. The user can then combine or select as needed the elements with similar semantics according to the element recommendation list, without manually searching one by one, which can improve the construction efficiency of the power grid middle - platform data model and realize the rapid construction of complex business scenarios of the power grid middle - platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a flowchart of a method for constructing a power grid middle - platform data model provided by an embodiment of the present invention; Figure 2 is an intelligent recommendation block diagram based on a semantic distance algorithm; Figure 3 is an architecture diagram of an intelligent graphic design tool; Figure 4 is a structural diagram of a device for constructing a power grid middle - platform data model provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following will clearly and completely describe the technical solutions in this application in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts fall within the scope of protection of this application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the description of the specification, claims, and above-mentioned drawings of this application are intended to cover non-exclusive inclusion.

[0020] In the description of the embodiments of this application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity, specific order, or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, the meaning of "a plurality" is two or more, unless otherwise specifically defined.

[0021] Referring to "embodiments" herein means that the specific features, structures, or characteristics described in connection with the embodiments may be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those of ordinary skill in the art explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0022] In the description of the embodiments of this application, the term "and / or" is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0023] In the description of the embodiments of this application, the term "a plurality" refers to two or more (including two). Similarly, "multiple groups" refers to two or more groups (including two groups), and "multiple pieces" refers to two or more pieces (including two pieces).

[0024] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installation", "connection", "connection", "fixation" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or integrated; it can also be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and can be the communication inside two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific situations.

[0025] Embodiment 1 Please refer to Figure 1 , which is a schematic flowchart of a method for constructing a data model of a power grid middle platform provided by an embodiment of the present invention, including the following specific steps: S1. Respond to the model element addition request, add the corresponding model element to the data model to be constructed, and convert the text description word of the model element into the corresponding word vector; wherein, the model element includes: device attributes of power equipment and electrical equipment, and device relationships between different electrical equipment; the device attributes include: rated voltage, rated current, capacity, and impedance; the device relationships include: device connection relationship and device subordination relationship; Specifically, the present invention details the method of designing the State Grid unified data model by the semantic distance algorithm. By applying natural language processing technologies such as the semantic distance algorithm, automatic matching and intelligent recommendation of model elements are realized, effectively reducing the repetitive work of designers.

[0026] Please refer to Figure 2 , which is an intelligent recommendation block diagram based on the semantic distance algorithm. By using natural language processing (NLP) technology, professional terms and concepts in the power grid field are analyzed to construct a domain-specific semantic knowledge base. Through NLP technology, the system can extract keywords and terms from these texts and establish semantic relationships between the terms. Finally, these processed information is used to construct a semantic knowledge base specific to the power grid industry, providing a basis for subsequent semantic analysis.

[0027] Based on this knowledge base, the semantic similarity between different elements (such as entities, attributes, relationships, etc.) in the calculation model is calculated to form a semantic distance matrix, providing important data support for subsequent intelligent recommendation and model matching. During the model design process, a semantic distance algorithm is used for intelligent recommendation. For example, when a designer adds a new entity, the system will automatically recommend semantically similar attributes and relationships. When performing model matching and searching, the semantic distance algorithm is used to improve accuracy. The system can identify elements that are semantically similar but have different expressions, avoiding duplicate design. In addition, the present invention also applies the semantic distance algorithm to the intelligent matching and recommendation of business middle platform services. Based on historical usage data and semantic similarity, the system can recommend appropriate business middle platform services to users and provide service combination suggestions to support the rapid construction of complex business scenarios.

[0028] The specific implementation steps are as follows: When a user adds a new model element e, the system responds to the model element addition request, adds the corresponding model element e to the data model to be constructed, and converts the text description word of the model element e into a corresponding word vector.

[0029] S2. According to the word vector, calculate the first semantic distance between the text description word of the model element and the text description words of all elements in the preset power grid semantic knowledge base; Preferably, the semantic distance between the text description word of the model element and the text description words of all elements in the preset power grid semantic knowledge base is calculated by the following formula: d(e1, e2) = 1 - cos(v(e1), v(e2)); Where d(e1, e2) is the semantic distance between element e1 and element e2, v(e1) is the semantic vector of element e1, v(e2) is the semantic vector of element e2, and cos is the cosine similarity.

[0030] Preferably, the generation of the power grid semantic knowledge base includes: obtaining the preset power grid text; performing text segmentation processing, part-of-speech tagging processing, and named entity recognition processing on the power grid text respectively to identify the words in the text and the named entities of each word, and then extracting the corresponding keywords from all the words according to the named entities; calculating the semantic similarity between each keyword, comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between each keyword according to the comparison result; constructing a corresponding power grid domain knowledge graph according to the keywords and the correlation between each keyword, and obtaining the corresponding power grid semantic knowledge base according to the power grid domain knowledge graph.

[0031] Preferably, the power grid text is respectively subjected to text word segmentation processing, part-of-speech tagging processing, and named entity recognition processing to identify the words in the text and the named entities of each word, and then corresponding key terms are extracted from all the words according to the named entities, including: obtaining a preset power grid professional dictionary, and performing text word segmentation processing and part-of-speech tagging processing on the power grid text according to the power grid professional dictionary to obtain the words in the power grid text and the part-of-speech information corresponding to the words; performing named entity recognition processing on the words according to the words and the part-of-speech information corresponding to the words to identify the named entities of each word; wherein, the named entities include: device entities, parameter entities, and operation entities; extracting corresponding keywords from all the words according to the named entities.

[0032] Preferably, calculating the semantic similarity between each of the keywords, and comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between each keyword according to the comparison result, including: converting each of the keywords into a keyword vector in a preset dimension, and calculating the cosine similarity between the keyword vectors; according to a preset specific part-of-speech judgment rule, identifying the specific part-of-speech corresponding to each of the keywords, and obtaining the weight corresponding to each specific part-of-speech; wherein, the specific part-of-speech includes: professional terms and general vocabulary; obtaining the keyword weight corresponding to each of the keywords according to the weight corresponding to each specific part-of-speech and the specific part-of-speech corresponding to each of the keywords, taking the product of the keyword weight and the cosine similarity as the semantic similarity of the keyword, and comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between each keyword according to the comparison result.

[0033] Preferably, constructing a corresponding power grid domain knowledge graph according to the keywords and the correlation between each keyword, and obtaining a corresponding power grid semantic knowledge base according to the power grid domain knowledge graph, including: extracting power grid equipment, equipment attributes of the power grid equipment, and equipment relationships of the power grid equipment in the power grid text, and taking the power grid equipment, equipment attributes, and equipment relationships as basic elements; establishing association edges for related keywords according to the correlation between each keyword, and then constructing a corresponding power grid domain knowledge graph according to the association edges and the basic elements, and obtaining a corresponding power grid semantic knowledge base according to the power grid domain knowledge graph.

[0034] Specifically, the system calculates the first semantic distance between the text description words of the model element e and the text description words of all elements in the preset power grid semantic knowledge base according to the word vector.

[0035] Among them, the construction process of the power grid semantic knowledge base is as follows: (1) Collect power grid texts such as power grid domain professional literature, standards and specifications, and design documents.

[0036] (2) Use natural language processing techniques, such as word segmentation, part-of-speech tagging, named entity recognition, etc., to extract keywords and concepts from the power grid text. The specific implementation process is as follows: First, perform text word segmentation through the jieba word segmentation tool in combination with a power grid professional dictionary. This professional dictionary systematically integrates the core terms in the power grid field, including equipment terms such as transformers, circuit breakers, capacitors, instrument transformers, arresters, etc., technical parameters such as rated voltage, rated current, capacity, impedance, etc., and operation terms such as grid connection, disconnection, commissioning, maintenance, etc. The system imports these professional dictionaries through the jieba.load_userdict() method, significantly improving the word segmentation accuracy of professional terms.

[0037] Then, use the HMM model for part-of-speech tagging to establish a complete power grid-specific part-of-speech rule system. This system includes equipment category words (EQP, such as transformers, circuit breakers), parameter category words (PAR, such as voltage, current), operation category words (OPR, such as grid connection, disconnection), and location category words (LOC, such as substations, distribution rooms). The system constructs a state transition matrix and an emission matrix by training the labeled power grid field corpus to achieve accurate tagging of professional terms.

[0038] Next, adopt the BiLSTM-CRF model for named entity recognition. This model is designed with a three-layer architecture. The input layer converts words into word vector representations, the BiLSTM layer is responsible for capturing context feature information, and the CRF layer optimizes the tagging sequence. The system supports the recognition of three types of entities: equipment entities (B-EQP, I-EQP), parameter entities (B-PAR, I-PAR), and operation entities (B-OPR, I-OPR), effectively improving the accuracy of entity recognition.

[0039] Finally, the system combines the TF-IDF algorithm and the TextRank algorithm to extract keywords. TF-IDF calculates the product of the term frequency (the number of times a word appears in a document / the total number of words in the document) and the inverse document frequency (the logarithm of the ratio of the total number of documents to the number of documents containing the word) to identify keywords in the document, effectively filtering out common words and highlighting keywords. At the same time, the system sets a sliding window of 5 words, counts the number of times word pairs co-occur within the window, constructs an undirected weighted graph, and uses it to discover semantic associations between words and identify common phrases and collocations.

[0040] Through this processed information, the system constructs a complete power grid industry semantic knowledge base. This knowledge base contains the association relationships between equipment and its parameters, operations, and locations, the definitions of units, related parameters, and usage scenarios of keywords, and different expressions of keywords such as equipment and operations. This knowledge base provides solid basic support for subsequent semantic analysis and significantly improves the model element matching efficiency.

[0041] (3) Use the word vector model BERT to calculate the semantic similarity between keywords. To improve the design efficiency of the unified power grid data model, the BERT pre-trained model with strong context understanding ability is selected as the word vector model.

[0042] First, use the constructed semantic knowledge base as the training corpus, and collect power grid domain corpora such as the State Grid unified data model (SG-CIM) specification documents, power grid design standard specifications, and professional literature to perform domain adaptation fine-tuning on the pre-trained BERT model. The input includes power grid keywords extracted from the semantic knowledge base, the context information of the keywords (from SG-CIM specification documents and design standards), and the association relationships between the keywords (from the previously constructed knowledge graph). Convert the keywords into 768-dimensional word vector representations, that is, use vectors composed of 768 numerical values to comprehensively describe the semantic features of each keyword, including information in multiple dimensions such as part of speech, context relationship, and professional attributes.

[0043] Then, use the improved cosine similarity formula to calculate the semantic similarity between keywords: sim(A,B) = cos(θ) = (A·B) / (||A||·||B||), where A and B are the word vectors corresponding to the keywords. To highlight the importance of specific terms in the power grid domain and improve the matching accuracy, the system identifies professional terms and general vocabulary through three levels: comparison based on the previously established power grid professional dictionary, judgment using the set power grid specific part-of-speech rules (such as EQP, PAR, OPR, etc.), and verification in combination with the entity types in the knowledge graph. Assign a weight of 1.2 to the identified professional terms (such as equipment terms like transformers and circuit breakers), and a weight of 0.8 to the general vocabulary.

[0044] The system sets 0.75 as the basic similarity threshold for preliminary screening, and the final similarity is calculated by multiplying the original similarity by the weight. For example, when the original similarity between two keywords is 0.7, if both are professional terms, the final similarity is 0.7×1.2 = 0.84, exceeding the threshold, and it is determined to be relevant; if both are general vocabulary, the final similarity is 0.7×0.8 = 0.56, lower than the threshold, and it is determined to be irrelevant. Through this multi-level screening and verification mechanism, the system pre-computes and stores the weighted similarities between keywords, avoiding real-time calculation. At the same time, the verification mechanism based on the knowledge graph improves the matching accuracy, can quickly locate keywords with similar semantics, reduces the search scope, and thus realizes the precise matching of keywords in the power grid domain, significantly improving the model element matching efficiency.

[0045] (4)Construct a knowledge graph for the power grid domain to represent the semantic relationships between keywords. When constructing the knowledge graph, first, based on the SG-CIM standard specification and power grid design documents, extract entities (such as equipment like transformers, circuit breakers, etc.), attributes (such as parameters like rated voltage, rated current, etc.), and relationships (such as connection relationships, subordination relationships between equipment, etc.) in the power grid domain as the basic elements of the knowledge graph; then, using the obtained word vector representations and semantic similarity calculation results, establish association edges for keywords with similar semantics, and assign different weights to different types of relationships (weight of 1.5 for physical connection relationships between equipment, weight of 1.2 for parameter subordination relationships); finally, store these nodes and relationships through a graph database (such as Neo4j) to form a complete knowledge graph for the power grid domain.

[0046] Through the stored physical connection relationships and parameter subordination relationships between equipment, this knowledge graph provides context information support for the intelligent recommendation system, helping to analyze equipment types, parameter configurations, and topological relationships; at the same time, its relationship network can be used to verify the rationality of the recommendation results, and through the fast query ability of the graph database, provide comprehensive knowledge support for model matching and recommendation decision-making, complementing the semantic distance algorithm, and providing a knowledge basis for subsequent intelligent recommendation and model matching.

[0047] S3. Sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list for the model elements based on each of the target elements, so that users can select corresponding elements according to the element recommendation list to construct the power grid middle platform data model.

[0048] Preferably, generating the corresponding element recommendation list based on each of the target elements includes: for each target element, obtain the historical addition frequency of the target element and the first weight corresponding to the historical addition frequency; calculate the element correlation between the target element and each of the added model elements in the data model to be constructed, and obtain the second weight corresponding to the element correlation; calculate the second semantic distance between the text description words of the target element and the text description words of the added model elements and the third weight corresponding to the second semantic distance; perform a weighted summation calculation on the historical addition frequency, element correlation, and second semantic distance according to the first weight, second weight, and third weight to obtain the comprehensive score corresponding to the target element; select a preset number of target elements in descending order according to the comprehensive score to generate the element recommendation list.

[0049] Specifically, sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and then combine the historical usage frequency and context information to sort each of the target elements to generate an element recommendation list for the model elements, so that the user can select corresponding elements according to the element recommendation list to construct the power grid middle platform data model.

[0050] When generating the final element recommendation list, a multi-dimensional intelligent weighted sorting method for the power grid field is adopted: First, count the usage frequency of target elements (such as equipment like transformers and circuit breakers) in historical power grid designs, and calculate the usage frequency score (accounting for 30%); then analyze the current design context, including information such as the types of equipment already selected, electrical parameter configurations, and topological connection relationships, and calculate the context relevance score (accounting for 40%); finally, combine the semantic distance scores between elements calculated based on the BERT model (accounting for 30%), and obtain the comprehensive score through weighted summation. The system sorts the target elements in descending order according to the comprehensive score, selects the top N target elements with the highest scores to generate the final element recommendation list, and dynamically adjusts the weights of each dimension based on the feedback of actual engineering designs to continuously optimize the power grid model design recommendation and further improve the design efficiency and accuracy.

[0051] In another specific embodiment, the present invention also designs a real-time collaborative design platform, which is an important infrastructure for supporting the application of the semantic distance algorithm. Through the organic collaboration of five core functional modules, the platform realizes the efficient execution of semantic knowledge base construction, semantic distance calculation, and intelligent recommendation algorithms. The platform adopts a layered architecture design. The business middle platform management module serves as the top-level control center to coordinate other modules. The access control module serves as the basic security layer, and the real-time data synchronization module, version control module, and collaborative communication module constitute the core functional layer. Each module communicates with each other through standardized interfaces to form a complete data processing link during the execution of the semantic algorithm.

[0052] Among them, the business middle platform management module, as the top-level control center of the platform, is designed with a microservices architecture and is responsible for the configuration management and scheduling execution of the semantic algorithm. This module realizes service registration and discovery through Eureka, conducts unified configuration management through Apollo, and is equipped with a monitoring collector to track indicators such as the progress of semantic knowledge base construction and calculation performance in real time. The system status is displayed through a visual operation dashboard, and tasks such as knowledge base update and algorithm optimization are uniformly processed by the lifecycle manager. This centralized management mechanism significantly improves the efficiency of semantic processing and multi-party collaboration.

[0053] As the basic security layer of the platform, the access control module implements the permission management system based on RBAC (Role-Based Access Control). The module includes a role definition engine to support custom permission rules, a permission validator using JWT tokens to implement identity authentication, and an operation audit recorder to record operation logs such as semantic knowledge base access and algorithm calls. The dynamic permission manager is used to adjust permissions in real time to ensure the security and controllability of semantic data and algorithms.

[0054] The real-time data synchronization module is the core supporting module for the execution of semantic algorithms. Based on the distributed architecture design, this module is responsible for tasks such as incremental update synchronization of the semantic knowledge base, real-time push of model element changes, and distribution of semantic calculation results. Local modifications are captured through data change listeners, Redis message queues store events to be synchronized, and updates are pushed based on the WebSocket protocol. Distributed locks and version number mechanisms are used to resolve concurrency conflicts and ensure the consistency of semantic data.

[0055] The version control module is responsible for the version management of the semantic knowledge base and algorithm model. It adopts a distributed version control system architecture, uses the Myers difference algorithm to calculate the knowledge base version changes through the difference calculation engine, maintains the evolution history of the semantic relationship, and the change log recorder saves detailed algorithm optimization records. This module enables the design team to track the complete evolution process of the semantic model and supports version rollback and comparison.

[0056] The collaborative communication module provides real-time communication support for the collaborative application of semantic algorithms. It builds a P2P communication network based on WebRTC technology, manages client connections through a signaling server, is equipped with a data channel manager to maintain point-to-point transmission of semantic data, and a session state manager to track user online status. It supports the design team to discuss semantic matching results and recommended solutions in real time, improving collaboration efficiency.

[0057] In actual operation, data processing related to semantic algorithms follows a complete flow chain: first, the business middle-end management module starts the relevant tasks, and after the access control module verifies the permissions, the real-time data synchronization module performs specific data processing and distribution, the version control module records the processing process and results, and the collaborative communication module supports team communication. Through this close collaboration mechanism, the platform realizes the efficient execution of semantic algorithms and multi-user collaboration, providing strong support for the intelligent design of the unified data model of the power grid.

[0058] In another specific embodiment, the present invention also designs a visualization carrier for implementing the semantic distance algorithm and the real-time collaborative design platform, and presents the semantic analysis ability and the collaborative design function to the user completely through a graphical interface. This tool is based on deep learning technology and integrates advanced technologies such as the double-layer rendering technology of HTML5 Canvas and SVG, the deep learning algorithm cluster, and multimodal intelligent search. Through the five core modules of the interactive graphical interface, machine learning recommendation, intelligent error detection, semantic search, and business middle platform management, the intelligent upgrade of the entire process of model design is realized. Innovatively applying deep learning technology to the field of power grid model design significantly improves the design efficiency and accuracy, and provides strong support for the standardized and intelligent design of the unified data model of the power grid. Please refer to Figure 3 , which is the architecture diagram of the intelligent graphic design tool.

[0059] As the visualization implementation layer of the semantic algorithm, the interactive graphical interface engine constructs a double-layer rendering architecture based on HTML5 Canvas and SVG technologies. The underlying SVG processes the basic graphic drawing, and the upper-layer Canvas processes the dynamic interaction effects. The user operations are captured through a custom event processor, and the graphical operations of the user are converted into the input data required for semantic analysis, and the analysis results of the semantic algorithm are presented in a visual manner. At the same time, through the data synchronization mechanism of the real-time collaborative design platform, the engine ensures that the graphic operations among multiple users can be synchronized in real time.

[0060] The machine learning intelligent recommendation system directly integrates the core functions of the semantic distance algorithm and constructs a parameter intelligent adjustment engine using a deep learning model. The system includes a data collection layer that records the historical operation data of designers, a feature engineering layer that extracts the operation sequence features, a model training layer that uses an LSTM network to learn the time series features, and a recommendation generation layer that predicts the optimal parameter configuration in real time based on the current design context and the semantic analysis results. The system records the parameter adjustment history through the version control module of the collaborative platform to realize the knowledge sharing among multiple users.

[0061] The intelligent error detection and correction system constructs a multi-level verification framework based on the semantic knowledge base, including a rule engine based on the expert knowledge base to define the power grid design specifications, a graph structure analyzer using graph theory algorithms to verify the correctness of the network topology, and a parameter validator using the constraint propagation algorithm to ensure the parameter consistency. Through the real-time data synchronization module of the collaborative platform, the system ensures that the error detection results can be pushed to all relevant users in a timely manner.

[0062] The semantic intelligent search engine directly reuses the core technology of the semantic distance algorithm to implement a multi-modal search system based on deep learning. It includes a text understanding module using the BERT model, a graphical feature extractor adopting a graph neural network, and a multi-modal fusion layer combining text and graphical features. The engine ensures the compliance of search permissions through the access control module of the collaboration platform and realizes the real-time sharing of search results through the data synchronization module.

[0063] The business middle platform integration system is closely integrated with the real-time collaborative design platform to build a hierarchical service management architecture. The system uniformly schedules service resources through the business middle platform management module of the collaboration platform, tracks the service change history through the version control module, and ensures the consistency of service status through the real-time data synchronization module. At the same time, the system also provides the necessary computing resources and operating environment for the semantic algorithm.

[0064] Through the above design, the intelligent graphic design tool realizes the deep integration with the semantic distance algorithm and the real-time collaborative design platform, and the three form an organic whole: the semantic algorithm provides intelligent analysis capabilities, the collaborative platform provides multi-user collaboration support, and the graphic tool provides an intuitive operation interface. This close integration not only ensures the unity of the patent but also significantly improves the efficiency and accuracy of the power grid unified data model design.

[0065] It can be seen that the present invention provides a method for constructing a power grid middle platform data model. The present invention can bring important improvements to the power grid industry and effectively meet the industry's demand for efficient design tools. By integrating natural language processing technology, it realizes the automatic matching and intelligent recommendation of model elements, significantly reducing the workload of designers. Version management and real-time multi-user collaboration effectively reduce human errors and greatly improve the design quality. By integrating the business middle platform operation management model, the present invention realizes the full life cycle management of business middle platform services, significantly improving the availability and maintainability of services. The service matching and recommendation mechanism based on the semantic distance algorithm enables users to find the required services more quickly and accurately, greatly enhancing the usage efficiency of the business middle platform. Specifically, the following beneficial effects can be achieved through the present invention: (1) Improve design efficiency: Through the semantic distance algorithm and intelligent recommendation mechanism based on deep learning, combined with the double-layer rendering technology of HTML5 Canvas and SVG, the model design efficiency is significantly improved. When the system processes large-scale complex models, the rendering performance is improved by more than 60%, and the user interaction response time is shortened by 70%. In specific applications, the model element matching time is reduced by more than 50%, the parameter adjustment efficiency is increased by more than 50%, and the overall design process efficiency is increased by more than 40%.

[0066] (2) Reduce error rate: By adopting a multi-level verification framework, including a rule engine based on an expert knowledge base, topological analysis using graph theory algorithms, and parameter verification using constraint propagation algorithms, real-time error detection and intelligent correction during the design process are achieved. Through this mechanism, the design error rate is reduced by more than 30%, parameter configuration errors are reduced by 45%, and permission-related accidents are reduced by more than 80%.

[0067] (3) Improve service quality: Based on an improved semantic distance algorithm, an intelligent service matching and recommendation mechanism is implemented, combined with a multi-modal search system using deep learning, to increase the service matching accuracy rate to more than 90%. At the same time, the user's service search time is reduced by 65%, and the service operation and maintenance efficiency is increased by more than 40%, effectively ensuring the stability and availability of the system.

[0068] (4) Optimize project management: Through a hierarchical service management architecture and a business middle platform integration system, intelligent management of the entire service life cycle is achieved. Combined with visual dependency analysis and performance monitoring, the overall project cycle is shortened by 20% - 30%, and the multi-user collaboration efficiency is increased by 55%, significantly improving the project delivery efficiency.

[0069] Embodiment 2 Please refer to Figure 4 , which is a schematic structural diagram of a power grid middle platform data model construction device provided by an embodiment of the present invention. The device includes: a model element adding module, a semantic distance calculation module, and an element recommendation list generation module; The model element adding module is used to respond to a model element adding request, add corresponding model elements to the data model to be constructed, and convert the text description words of the model elements into corresponding word vectors; wherein, the model elements include: power equipment, equipment attributes of electrical equipment, and equipment relationships between different electrical equipment; the equipment attributes include: rated voltage, rated current, capacity, and impedance; the equipment relationships include: equipment connection relationships and equipment subordination relationships; The semantic distance calculation module is used to calculate a first semantic distance between the text description words of the model elements and the text description words of all elements in a preset power grid semantic knowledge base according to the word vectors; The element recommendation list generation module is used to sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list of the model elements according to each of the target elements, so that users can select corresponding elements according to the element recommendation list to construct the power grid middle platform data model.

[0070] It should be noted that the device embodiments described above are merely illustrative. 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 to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement without creative efforts.

[0071] Those skilled in the art can clearly understand that for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the foregoing method embodiments, and will not be elaborated herein.

[0072] Embodiment 3 Correspondingly, an embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the method for constructing the power grid middle platform data model described in the foregoing embodiments of the present invention.

[0073] The electronic device may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The device may include, but is not limited to, a processor and a memory.

[0074] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the device, and connects various parts of the entire device through various interfaces and lines.

[0075] Embodiment 4 Accordingly, an embodiment of the present invention provides a storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the storage medium is located to execute the power grid middle platform data model construction method described in the above embodiments of the present invention.

[0076] The memory can be used to store the computer program. By running or executing the computer program stored in the memory and calling the data stored in the memory, the various functions of the device can be realized. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card (FlashCard), at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0077] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by a processor, the steps of the above various method embodiments can be realized. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0078] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A method for constructing a data model in a power grid, characterized in that: include: In response to a model element addition request, a corresponding model element is added to the data model to be constructed, and a text description word of the model element is converted into a corresponding word vector; wherein the model element includes: power equipment, equipment attributes of electrical equipment and equipment relationships between different electrical equipment; the equipment attributes include: rated voltage, rated current, capacity and impedance; the equipment relationship includes: equipment connection relationship and equipment subordination relationship; Calculating, based on the word vector, a first semantic distance between the text description word of the model element and the text description words of all elements in a preset power grid semantic knowledge base; The first semantic distances are sorted, and a preset number of corresponding elements are selected as target elements in order from small to large, and an element recommendation list of the model elements is generated according to each of the target elements, so that the user can select the corresponding element according to the element recommendation list to construct the power grid middle station data model.

2. The method for constructing a data model in a power grid according to claim 1, characterized in that: The generating a corresponding element recommendation list according to each of the target elements includes: For each target element, obtain the historical adding frequency of the target element and the first weight corresponding to the historical adding frequency; According to the added model elements in the data model to be constructed, calculating the element correlation between the target element and each of the added model elements, and obtaining a second weight corresponding to the element correlation; Calculating a second semantic distance between the text description word of the target element and the text description word of the added model element and a third weight corresponding to the second semantic distance; According to the first weight, the second weight and the third weight, a weighted sum calculation is performed on the historical adding frequency, the element relevance and the second semantic distance to obtain a comprehensive score corresponding to the target element; According to the comprehensive score, a preset number of target elements are selected in descending order to generate the element recommendation list.

3. The method for constructing a data model in a power grid according to claim 1, characterized in that: The generation of the power grid semantic knowledge base includes: Get the preset grid text; Performing text segmentation, part-of-speech tagging, and named entity recognition on the power grid text, identifying the words in the text and the named entities of each word, and then extracting corresponding keywords from all the words according to the named entities; Calculating the semantic similarity between the keywords, and comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between the keywords according to the comparison result; According to the keyword and the correlation between each keyword, a corresponding power grid domain knowledge graph is constructed, and according to the power grid domain knowledge graph, a corresponding power grid semantic knowledge base is obtained.

4. The method for constructing a data model in a power grid according to claim 3, characterized in that: The power grid text is subjected to text segmentation processing, part-of-speech tagging processing and named entity recognition processing respectively, the words in the text and the named entities of each word are identified, and then the corresponding key terms are extracted from all the words according to the named entities, including: Obtaining a preset power grid professional dictionary, and performing text segmentation processing and part-of-speech tagging processing on the power grid text according to the power grid professional dictionary, to obtain words in the power grid text and part-of-speech information corresponding to the words; According to the words and the part-of-speech information corresponding to the words, the words are subjected to named entity recognition processing to identify the named entities of each word; wherein the named entities include: device entities, parameter entities and operation entities; Corresponding keywords are extracted from all words according to the named entity.

5. The method for constructing a data model in a power grid according to claim 3, characterized in that: The calculating of the semantic similarity between the keywords, comparing the semantic similarity with a preset similarity threshold, and obtaining the correlation between the keywords according to the comparison result includes: Convert each of the keywords into a keyword vector of a preset dimension, and calculate the cosine similarity between the keyword vectors; According to the preset specific part-of-speech judgment rules, the specific part-of-speech corresponding to each of the keywords is identified, and the weight corresponding to each of the specific parts-of-speech is obtained; wherein the specific parts-of-speech include: professional terms and general terms; According to the weight corresponding to each of the specific parts of speech and the specific parts of speech corresponding to each of the keywords, the keyword weight corresponding to each of the keywords is obtained, the product of the keyword weight and the cosine similarity is used as the semantic similarity of the keyword, and the semantic similarity is compared with a preset similarity threshold, and the correlation between the keywords is obtained according to the comparison result.

6. The method for constructing a data model in a power grid according to claim 3, characterized in that: The method of constructing a corresponding power grid domain knowledge graph based on the keyword and the correlation between the keywords, and obtaining a corresponding power grid semantic knowledge base based on the power grid domain knowledge graph includes: Extracting power equipment, equipment attributes of power equipment and equipment relationships of power equipment in the power grid field from the power grid text, and taking the power equipment, equipment attributes and equipment relationships as basic elements; According to the correlation between the keywords, associated edges are established for the related keywords, and then the corresponding power grid domain knowledge graph is constructed according to the associated edges and the basic elements, and the corresponding power grid semantic knowledge base is obtained according to the power grid domain knowledge graph.

7. The method for constructing a data model in a power grid according to claim 1, characterized in that: The semantic distance between the text description words of the model element and the text description words of all elements in the preset power grid semantic knowledge base is calculated by the following formula: d(e1, e2) = 1 - cos(v(e1), v(e2)); Among them, d(e1, e2) is the semantic distance between element e1 and element e2, v(e1) is the semantic vector of element e1, v(e2) is the semantic vector of element e2, and cos is the cosine similarity.

8. A device for constructing a data model in a power grid, characterized in that: include: Model element adding module, semantic distance calculation module and element recommendation list generation module; The model element adding module is used to respond to the model element adding request, add the corresponding model element in the data model to be constructed, and convert the text description words of the model element into the corresponding word vector; wherein the model elements include: power equipment, equipment attributes of electrical equipment and equipment relations between different electrical equipment; the equipment attributes include: rated voltage, rated current, capacity and impedance; the equipment relations include: equipment connection relations and equipment subordination relations; The semantic distance calculation module is used to calculate the first semantic distance between the text description word of the model element and the text description words of all elements in the preset power grid semantic knowledge base according to the word vector; The element recommendation list generation module is used to sort the first semantic distances, select a preset number of corresponding elements as target elements in ascending order, and generate an element recommendation list of the model elements based on each of the target elements, so that the user can select the corresponding elements according to the element recommendation list to construct the power grid middle station data model.

9. An electronic device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements the method for constructing a data model of a power grid station as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the method for constructing a data model in a power grid station as described in any one of claims 1 to 7.

Images