Tracing method and system based on power enterprise production information

Abstract

The invention discloses a traceability method and system based on power enterprise production information, and the method comprises the steps: carrying out the content recognition of a plurality of production information data matched with a target power enterprise, and obtaining initial information data; processing the initial information data based on a constructed large language model and a classification model in sequence to obtain an information data set corresponding to different information dimensions; traversing the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each information data set, and determining a first relationship result between the information data sets corresponding to different information dimensions; comparing the first relation result with a second relation result obtained by knowledge graph identification to obtain target relation result data; and inputting the target relation result data into a traceability model for processing to obtain each traceability information log of the target power enterprise. According to the method provided by the embodiment of the invention, effective traceability of the production information of the power enterprise is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power enterprise production information data processing, and in particular to a traceability method and system based on power enterprise production information. Background Art

[0002] In the field of power production management, power production involves a large amount of equipment, projects, maintenance records and safety hazard information, and the demand for interaction and query of this data is increasing.

[0003] At present, the traceability method in the field of power construction mainly relies on the clerks of power companies to manually query relevant paper documents in order to track the production information of power companies. However, this method is time-consuming and labor-intensive, and the tracking efficiency is low. The accuracy is also difficult to guarantee. Information is often missed or wrong due to human factors, which cannot meet the needs of modern power companies for efficient and accurate information tracking.

[0004] It can be seen that how to achieve effective traceability of production information of power companies has become a technical problem that technical personnel in this field need to solve urgently. Summary of the invention

[0005] The present invention provides a tracing method and system based on production information of electric power enterprises to solve the current technical problem of low tracking efficiency and to promote the process of intelligent and automated data processing of production information of electric power enterprises.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a traceability method based on production information of an electric power enterprise, comprising: Perform content recognition on a number of production information data matching the target power enterprise to obtain initial information data; The initial information data is processed based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects, and power enterprise equipment; Traversing the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determining a first relationship result between the information data sets corresponding to different information dimensions according to association matching features between the entity information and the attribute information; Comparing and analyzing the first relationship result and the second relationship result obtained by identifying the user-power knowledge graph, to obtain target relationship result data of the target power enterprise; The target relationship result data is input into a traceability model constructed by a hash algorithm for processing to obtain each traceability information log of the target power enterprise.

[0007] As one of the preferred solutions, before performing content recognition on a number of production information data matching the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The constructed neural network recognition model is trained based on the selected traceability sample data to obtain the original production information data; The language representation model is used to perform multi-label classification processing on the original production information data to determine whether the original production information data belongs to a number of production information data matching the target power enterprise; wherein the multi-label classification processing process includes: Construct production information tags based on power companies; Using the language representation model to train the original production information data to obtain production information data to be identified; By using the association between the production information data to be identified and the production information tag, it is determined whether the production information data belongs to a number of production information data that matches the target power enterprise.

[0008] As one of the preferred solutions, the types of the production information data include at least text, image, voice and video; The content recognition of a number of production information data matching the target power enterprise is performed to obtain initial information data, including: Using a Chinese pre-trained model based on the Transformer architecture to perform data cleaning on some of the production information data to obtain noise-free data; Performing word segmentation processing on the noise-free data based on the selected word segmentation method to obtain a word segmentation result; Standardize synonyms on the acquired power enterprise-specific dictionary and project named entity mapping table to obtain the power enterprise mapping relationship; Processing the word segmentation results according to the mapping relationship of the power enterprise to generate several types of standard production information data; Initial information data are obtained by integrating a number of the standard production information data.

[0009] As one of the preferred solutions, the step of processing the initial information data based on the constructed large language model and classification model in sequence to obtain information data sets corresponding to different information dimensions includes: Inputting the initial information data into the constructed large language model to obtain production information data to be classified; The conditional random field classification algorithm is used to perform relationship extraction processing on the production information data to be classified to obtain information data sets corresponding to different information dimensions.

[0010] As one of the preferred solutions, the steps of constructing the traceability model include: Extracting a responsibility tracing label based on the user-power knowledge graph, wherein the responsibility tracing label at least includes user identity, operation time, operation type and operation result; Performing hash calculation on the target relationship result data using a secure hash algorithm to generate an operation log hash tree; Perform label mapping processing on the nodes in the operation log hash tree and the responsibility tracing labels to obtain a label mapping result; A traceability model is constructed based on the operation log hash tree and the label mapping result.

[0011] As one of the preferred solutions, after obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: Screening and processing the traceability information log to obtain a fault relationship table; The entropy weight method is used to calculate the weight of the fault relationship table, obtain the fault correlation corresponding to the responsibility tracing label, and determine the responsible object.

[0012] As one of the preferred solutions, after obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The production information integrity check is performed on the traceability information log, wherein the production information integrity check is a hash chain verification, and the hash chain verification process includes calculating a hash value for each piece of production information and comparing it with the hash value stored in the operation log hash tree. If the hash value does not match, the data is determined to be abnormal.

[0013] As one of the preferred solutions, after obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: A local traceability database is constructed based on each traceability information log of the target power enterprise, and a visualization interface of different information dimensions is determined based on the local traceability database.

[0014] As one of the preferred solutions, after obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The traceability information log is processed using a natural language model to generate a natural language report, which is sent to a corresponding client terminal for storage.

[0015] Another embodiment of the present invention provides a traceability system based on production information of an electric power enterprise, including: An identification module is used to perform content identification on a number of production information data matching the target power enterprise to obtain initial information data; A processing module, used to process the initial information data based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects and power enterprise equipment; A matching module, configured to traverse the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determine a first relationship result between the information data sets corresponding to different information dimensions according to the associated matching features between the entity information and the attribute information; An analysis module, configured to compare and analyze the first relationship result and the second relationship result obtained by identifying the user-power knowledge graph, and obtain target relationship result data of the target power enterprise; A generation module is used to input the target relationship result data into a traceability model constructed by a hash algorithm for processing, so as to obtain various traceability information logs of the target power enterprise.

[0016] Compared with the prior art, the embodiments of the present invention have the following advantages: The content of several production information data matching the target power enterprise is identified to obtain initial information data; the initial information data is processed based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects and power enterprise equipment; the information data set is traversed based on the depth-first search algorithm to obtain the entity information and attribute information corresponding to each information data set, and the first relationship result between the information data sets corresponding to different information dimensions is determined according to the association matching characteristics between the entity information and the attribute information; the first relationship result and the second relationship result obtained by the user-power knowledge graph recognition are compared and analyzed to obtain the target relationship result data of the target power enterprise; the target relationship result data is input into the traceability model constructed by the hash algorithm for processing to obtain each traceability information log of the target power enterprise. Compared with the prior art, this method can solve the current technical problem of low tracking efficiency by performing a series of processing processes on several production information data matching the target power enterprise, realize the effective traceability of the production information of the power enterprise, and thus promote the intelligent and automated processing of the production information data of the power enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a method for tracing the source of production information of a power enterprise in one embodiment of the present invention; Figure 2It is a structural schematic diagram of a traceability system based on production information of a power enterprise in one embodiment of the present invention; Reference numerals: Among them, 11, recognition module; 12, processing module; 13, matching module; 14, analysis module; 15, generation module. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] In the description of this application, the terms "first", "second", "third", etc. are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first", "second", "third", etc. may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0020] In the description of the present application, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. The terms "vertical", "horizontal", "left", "right", "upper", "lower" and similar expressions used herein are only for illustrative purposes, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0021] In the description of this application, it should be noted that, unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood by specific circumstances.

[0022] In the field of power production management, power production involves a large amount of equipment, projects, maintenance records and safety hazard information, and the demand for interaction and query of this data is increasing.

[0023] At present, the traceability method in the field of power construction mainly relies on the clerks of power companies to manually query relevant paper documents in order to track the production information of power companies. However, this method is time-consuming and labor-intensive, and the tracking efficiency is low. The accuracy is also difficult to guarantee. Information is often missed or wrong due to human factors, which cannot meet the needs of modern power companies for efficient and accurate information tracking.

[0024] To this end, an embodiment of the present invention provides a method for tracing the source of production information based on power enterprises. For details, see Figure 1 , Figure 1 The figure shows a flow chart of a method for tracing the source of production information of a power enterprise in one embodiment of the present invention, the method comprising: S1: Perform content recognition on a number of production information data matching the target power enterprise to obtain initial information data; S2: Processing the initial information data based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects, and power enterprise equipment; S3: Traversing the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determining a first relationship result between the information data sets corresponding to different information dimensions according to the association matching features between the entity information and the attribute information; S4: Compare and analyze the first relationship result and the second relationship result obtained by identifying the user-power knowledge graph to obtain target relationship result data of the target power enterprise; S5: Input the target relationship result data into the traceability model constructed by the hash algorithm for processing to obtain each traceability information log of the target power enterprise.

[0025] Before performing content recognition on a number of production information data matching the target power enterprise, it is necessary to judge whether the production information data to be identified belongs to the production information data matching the target power enterprise. If so, the process of tracing the production information begins in step S1. The judging process includes: Step 1: Train the constructed neural network recognition model based on the selected traceability sample data to obtain the original production information data; In this step, according to the characteristics of the selected traceability sample data and task requirements, a suitable neural network architecture can be selected, such as a fully connected neural network, convolutional neural network (CNN), recurrent neural network (RNN), etc., and the traceability sample data can be divided into training set, validation set and test set for model training, verification and testing. After the training is completed, the trained neural network recognition model is used to predict the new production data and obtain the original production information data.

[0026] Step 2: Use a language representation model to perform multi-label classification processing on the original production information data to determine whether the original production information data belongs to a number of production information data that matches the target power enterprise.

[0027] Specifically, in the above step 2, according to the characteristics of the data and the task requirements, a suitable language representation model is selected. Commonly used models include pre-trained language models such as BERT and GPT, as well as fine-tuned versions based on these models. At the same time, before using the language representation model to classify the original production information data, the original production information data needs to be preprocessed. The preprocessing process includes preprocessing steps such as cleaning, word segmentation, and removal of stop words; the language representation model is trained using the preprocessed data and defined labels. During the training process, the model will learn how to map the original production information data to the corresponding labels.

[0028] Specifically, the multi-label classification process includes: Constructing a production information label based on the power enterprise, and using the language representation model to train the original production information data to obtain the production information data to be identified; By using the association between the production information data to be identified and the production information label, it is determined whether it belongs to a number of production information data that matches the target power enterprise. That is to say, the production information data to be identified is input into the trained language representation model. The model will extract the feature vector of the data to be identified and map it to the predefined label. By comparing the similarity between the feature vector of the data to be identified and each label, the model can determine whether the data to be identified belongs to a certain label category. According to the classification result, it is determined whether the production information data to be identified belongs to a number of production information data that matches the target power enterprise. If the data to be identified is classified into the label category related to the target power enterprise, it is considered to be matched; otherwise, it is considered to be unmatched, and the result of the matching judgment is output to the user or system for subsequent processing or decision-making.

[0029] The types of production information data in step S1 include at least text, image, voice and video.

[0030] Specifically, in step S1, the steps of performing content recognition on a number of production information data matching the target power enterprise to obtain initial information data include: Using a Chinese pre-trained model based on the Transformer architecture to perform data cleaning on some of the production information data to obtain noise-free data; Performing word segmentation processing on the noise-free data based on the selected word segmentation method to obtain a word segmentation result; Standardize synonyms on the acquired power enterprise-specific dictionary and project named entity mapping table to obtain the power enterprise mapping relationship; Processing the word segmentation results according to the mapping relationship of the power enterprise to generate several types of standard production information data; Initial information data are obtained by integrating a number of the standard production information data.

[0031] In this process, through the Chinese pre-training model based on the Transformer architecture (such as BERT, ERNIE, etc.), we can use the model's powerful semantic understanding ability to identify and remove noise in production information data, such as invalid characters, redundant information, incorrect formats, etc. The cleaned data is cleaner and more accurate, providing a good foundation for subsequent word segmentation, standardization and other steps.

[0032] The noise-free data is divided into meaningful words or phrases, which is a basic step in natural language processing. The word segmentation results can more clearly show the semantic structure of the data, which is convenient for subsequent tasks such as word frequency statistics, sentiment analysis, and named entity recognition. Integrate and standardize the synonyms in the power company-specific dictionary and project named entity mapping table, and establish specific mapping relationships based on the special needs of power companies, which helps to understand and process related texts more accurately.

[0033] The word segmentation results are mapped and replaced using the mapping relationship of the power enterprise to generate production information data that meets the enterprise standards and specifications; multiple standard production information data are integrated into a complete initial information data set. The integrated initial information data is more comprehensive and systematic, and is easy to use in production management, decision support and other scenarios of power enterprises.

[0034] In step S2, the initial information data obtained in step 1 is processed based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects, and power enterprise equipment, wherein the steps of processing the initial information data based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions include: The initial information data is input into the constructed large language model to obtain production information data to be classified; the conditional random field classification algorithm is used to perform relationship extraction processing on the production information data to be classified to obtain information data sets corresponding to different information dimensions.

[0035] In step S2, the integrated initial information data is used as input, and the semantic understanding and generation capabilities of the large language model are used to preliminarily process and parse the information. The output after processing by the large language model is used as production information data to be classified. These data have been reorganized and semantically refined by the large language model, making them easier for subsequent classification and relationship extraction.

[0036] Conditional Random Fields (CRF) is a powerful tool for sequence labeling and segmentation tasks. In step S2, it is used to perform relation extraction on the classified production information data.

[0037] Specifically, the purpose of relationship extraction is to identify entities (such as power company personnel, power company projects, power company equipment, etc.) and the relationships between these entities from the text. Through the CRF algorithm, the entity boundaries and relationship types in the text can be accurately marked.

[0038] After relationship extraction, the original information is decomposed and classified into different information dimensions, forming an information data set corresponding to power enterprise personnel, power enterprise projects, power enterprise equipment, etc.

[0039] In step S3 and step S4, specifically, after obtaining the information data set, the information data set is traversed using a depth-first search algorithm to obtain the entity information and attribute information corresponding to each of the information data sets. The depth-first search (DFS) algorithm is an algorithm for traversing or searching a tree or graph. During the traversal process, the algorithm will identify and extract the entity information (such as power enterprise personnel, projects, equipment, etc.) and attribute information (such as personnel names, project names, equipment types, etc.) in each information data set. By using the association matching features between the entity information and the attribute information (such as the same personnel name appears in different data sets, or the same project has different attribute descriptions at different stages, etc.), the algorithm can identify the potential relationship between the information data sets under different information dimensions. Based on these association matching features, the algorithm will generate a preliminary relationship result, namely the first relationship result, which describes the connection and interaction between different information dimensions.

[0040] Specifically, the user-power knowledge graph is a known template composed of detailed information such as the organizational structure, personnel relationships, project progress, equipment status, etc. of the power company. Comparing and analyzing the first relationship result with the second relationship result can verify and supplement the accuracy of the first relationship result, and help identify relationships that may be omitted or erroneous in the first relationship result.

[0041] Through comparative analysis, the algorithm can integrate the information in the first relationship result and the second relationship result to generate a more comprehensive and accurate target relationship result data. The target relationship result data integrates information from different information dimensions and different data sources, and provides a comprehensive and systematic view to display the internal structure and external relations of the target power enterprise. This information is of great significance for decision support, risk management, project planning and other aspects of the power enterprise.

[0042] The target relationship result data obtained in step S4 is input into the traceability model constructed by the hash algorithm for processing, so as to obtain various traceability information logs of the target power enterprise.

[0043] Specifically, the steps for building the traceability model include: Based on the user-power knowledge graph, responsibility tracing labels are extracted, wherein the responsibility tracing labels at least include user identity (such as user name, user ID), operation time (specific time point for executing the operation), operation type (such as reading data, modifying configuration, etc.) and operation result (whether the operation is successful, the specific impact produced, etc.); these labels are the basis for subsequent traceability analysis.

[0044] A secure hash algorithm (such as SHA-256) is used to perform hash calculation on the target relationship result data to generate an operation log hash tree. Through hash calculation, each log is converted into a unique hash value, and then an operation log hash tree is constructed. The hash tree is a data structure used to efficiently verify the integrity and consistency of data.

[0045] The nodes in the operation log hash tree are subjected to label mapping processing with the responsibility tracing labels to obtain label mapping results, and a traceability model is constructed by combining the structure of the operation log hash tree and the label mapping results. This model can support the rapid search and verification of traceability information for specific operations or events. When it is necessary to trace an operation or event, the corresponding node in the hash tree can be found, and then detailed responsibility tracing information can be obtained according to the label mapping results.

[0046] In this way, efficient and accurate traceability analysis of the target power company's operation logs can be ensured, thereby improving the transparency and security of data management.

[0047] After obtaining each traceability information log of the target power enterprise, the method further includes: The traceability information log is screened to obtain a fault relationship table; the fault relationship table is weighted using an entropy weight method to obtain a fault correlation corresponding to the responsibility tracing label, and the responsible object is determined.

[0048] Specifically, the traceability information log is filtered according to preset filtering criteria or rules, such as keywords or patterns such as error codes, abnormal status, operation failures, etc. recorded in the log. The filtering scope can also be further narrowed by considering conditions such as time range and operation type.

[0049] The selected log entries will be organized into a fault relationship table. This table usually contains the time when the fault occurred, the fault type, the equipment or personnel involved, and the relevant responsibility traceability tags (such as operator, operation time, etc.). Establishing a fault relationship table helps to clearly show the relationship between the fault and each responsibility traceability tag.

[0050] Based on the calculated fault correlation, the responsibility tracing tags most relevant to the occurrence of the fault can be identified. Through these tags, the responsible objects can be further determined, such as specific personnel, departments or equipment. The determination of the responsible objects will help with subsequent fault handling, accountability and the formulation of improvement measures.

[0051] After obtaining each traceability information log of the target power enterprise, the method further includes: The production information integrity check is performed on the traceability information log, wherein the production information integrity check is a hash chain verification, and the hash chain verification process includes calculating a hash value for each piece of production information and comparing it with the hash value stored in the operation log hash tree. If the hash value does not match, the data is determined to be abnormal.

[0052] That is to say, in the traceability information log, for each entry containing production information, the above-mentioned hash chain verification process is applied. If data anomalies are found, further measures can be taken, such as re-obtaining production information, marking abnormal data, triggering alarms, etc. Through hash chain verification, it can be ensured that the production information in the traceability information log has not been tampered with or damaged during transmission and storage, thereby ensuring the integrity and accuracy of the data.

[0053] After obtaining each traceability information log of the target power enterprise, the method further includes: A local traceability database is constructed based on each traceability information log of the target power enterprise, and a visualization interface of different information dimensions is determined based on the local traceability database.

[0054] After obtaining each traceability information log of the target power enterprise, the method further includes: The traceability information log is processed using a natural language model to generate a natural language report, which is sent to a corresponding client terminal for storage.

[0055] That is to say, the preprocessed logs are processed using a natural language model. During the processing, the natural language model will extract key information from the logs, such as fault descriptions, responsibility tracing labels, timestamps, etc., and integrate this information into a coherent narrative to generate a detailed natural language report. This report presents the core content of the traceability information log in a human-readable manner. The information in the report is directly derived from the traceability information log, ensuring the traceability and accuracy of the information.

[0056] Another embodiment of the present invention provides a traceability system based on production information of power enterprises. For details, see Figure 2 , Figure 2 The structure diagram of the traceability system based on production information of power enterprises in one embodiment of the present invention is shown, and the structure includes: The identification module 11 is used to perform content identification on a number of production information data matching the target power enterprise to obtain initial information data; The processing module 12 is used to process the initial information data based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects and power enterprise equipment; A matching module 13 is used to traverse the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determine a first relationship result between the information data sets corresponding to different information dimensions according to the associated matching features between the entity information and the attribute information; An analysis module 14 is used to compare and analyze the first relationship result and the second relationship result obtained by user-power knowledge graph identification to obtain target relationship result data of the target power enterprise; The generating module 15 is used to input the target relationship result data into the traceability model constructed by the hash algorithm for processing, so as to obtain each traceability information log of the target power enterprise.

[0057] Compared with the prior art, the embodiments of the present invention have the following advantages: The content of several production information data matching the target power enterprise is identified to obtain initial information data; the initial information data is processed based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects and power enterprise equipment; the information data set is traversed based on the depth-first search algorithm to obtain the entity information and attribute information corresponding to each information data set, and the first relationship result between the information data sets corresponding to different information dimensions is determined according to the association matching characteristics between the entity information and the attribute information; the first relationship result and the second relationship result obtained by the user-power knowledge graph recognition are compared and analyzed to obtain the target relationship result data of the target power enterprise; the target relationship result data is input into the traceability model constructed by the hash algorithm for processing to obtain each traceability information log of the target power enterprise. Compared with the prior art, this method can solve the current technical problem of low tracking efficiency by performing a series of processing processes on several production information data matching the target power enterprise, realize the effective traceability of the production information of the power enterprise, and thus promote the intelligent and automated processing of the production information data of the power enterprise.

[0058] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A tracing method based on production information of electric power enterprises, characterized in that: include: Perform content recognition on a number of production information data matching the target power enterprise to obtain initial information data; The initial information data is processed based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects, and power enterprise equipment; Traversing the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determining a first relationship result between the information data sets corresponding to different information dimensions according to association matching features between the entity information and the attribute information; Comparing and analyzing the first relationship result and the second relationship result obtained by identifying the user-power knowledge graph, to obtain target relationship result data of the target power enterprise; The target relationship result data is input into a traceability model constructed by a hash algorithm for processing to obtain each traceability information log of the target power enterprise.

2. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: Before performing content recognition on a number of production information data matching the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The constructed neural network recognition model is trained based on the selected traceability sample data to obtain the original production information data; The language representation model is used to perform multi-label classification processing on the original production information data to determine whether the original production information data belongs to a number of production information data matching the target power enterprise; wherein the multi-label classification processing process includes: Construct production information tags based on power companies; Using the language representation model to train the original production information data to obtain production information data to be identified; By using the association between the production information data to be identified and the production information tag, it is determined whether the production information data belongs to a number of production information data that matches the target power enterprise.

3. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: The types of the production information data include at least text, image, voice and video; The content recognition of a number of production information data matching the target power enterprise is performed to obtain initial information data, including: Using a Chinese pre-trained model based on the Transformer architecture to perform data cleaning on some of the production information data to obtain noise-free data; Performing word segmentation processing on the noise-free data based on the selected word segmentation method to obtain a word segmentation result; Standardize synonyms on the acquired power enterprise-specific dictionary and project named entity mapping table to obtain the power enterprise mapping relationship; Processing the word segmentation results according to the mapping relationship of the power enterprise to generate several types of standard production information data; Initial information data are obtained by integrating a number of the standard production information data.

4. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: The step of sequentially processing the initial information data based on the constructed large language model and classification model to obtain information data sets corresponding to different information dimensions includes: Inputting the initial information data into the constructed large language model to obtain production information data to be classified; The conditional random field classification algorithm is used to perform relationship extraction processing on the production information data to be classified to obtain information data sets corresponding to different information dimensions.

5. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: The steps of constructing the traceability model include: Extracting a responsibility tracing label based on the user-power knowledge graph, wherein the responsibility tracing label at least includes user identity, operation time, operation type and operation result; Performing hash calculation on the target relationship result data using a secure hash algorithm to generate an operation log hash tree; Perform label mapping processing on the nodes in the operation log hash tree and the responsibility tracing labels to obtain a label mapping result; A traceability model is constructed based on the operation log hash tree and the label mapping result.

6. The method for tracing the source of production information based on power enterprises according to claim 5, characterized in that: After obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: Screening and processing the traceability information log to obtain a fault relationship table; The entropy weight method is used to calculate the weight of the fault relationship table, obtain the fault correlation corresponding to the responsibility tracing label, and determine the responsible object.

7. The method for tracing the source of production information based on power enterprises according to claim 5, characterized in that: After obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The production information integrity check is performed on the traceability information log, wherein the production information integrity check is a hash chain verification, and the hash chain verification process includes calculating a hash value for each piece of production information and comparing it with the hash value stored in the operation log hash tree. If the hash value does not match, the data is determined to be abnormal.

8. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: After obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: A local traceability database is constructed based on each traceability information log of the target power enterprise, and a visualization interface of different information dimensions is determined based on the local traceability database.

9. The method for tracing the source of production information based on power enterprises according to claim 1, characterized in that: After obtaining each traceability information log of the target power enterprise, the traceability method based on the production information of the power enterprise further includes: The traceability information log is processed using a natural language model to generate a natural language report, which is sent to a corresponding client terminal for storage.

10. A traceability system based on production information of power enterprises, characterized in that: include: An identification module is used to perform content identification on a number of production information data matching the target power enterprise to obtain initial information data; A processing module, used to process the initial information data based on the constructed large language model and classification model in turn to obtain information data sets corresponding to different information dimensions, wherein the information dimensions at least include power enterprise personnel, power enterprise projects and power enterprise equipment; A matching module, configured to traverse the information data sets based on a depth-first search algorithm to obtain entity information and attribute information corresponding to each of the information data sets, and determine a first relationship result between the information data sets corresponding to different information dimensions according to the associated matching features between the entity information and the attribute information; An analysis module, configured to compare and analyze the first relationship result and the second relationship result obtained by identifying the user-power knowledge graph, and obtain target relationship result data of the target power enterprise; A generation module is used to input the target relationship result data into a traceability model constructed by a hash algorithm for processing, so as to obtain various traceability information logs of the target power enterprise.

Images