Power emergency plan text generation method and device, computer equipment and storage medium

Through the pre-constructed power plan dictionary and knowledge graph, the keywords are identified and the emergency plan text is merged, which solves the problem of insufficient accuracy of emergency plans and improves the accuracy and completeness of emergency plans.

CN119067103APending Publication Date: 2024-12-03GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411001646.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

In the prior art, the accuracy of the emergency plan is insufficient and relies too much on the experience of decision makers, which is prone to omissions, resulting in insufficient accuracy of the emergency plan text in emergencies.

Method used

Through the pre-constructed power plan dictionary and power knowledge graph, we can identify keywords in the target accident information, generate emergency plans, and use triples and associations to merge emergency plan texts to eliminate redundancy and error texts, and improve the integrity of emergency plans. and accuracy.

Benefits of technology

It improves the accuracy and completeness of the emergency plan text, reduces redundancy and error text, and enhances the reliability of the emergency plan.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067103A_ABST
    Figure CN119067103A_ABST
Patent Text Reader

Abstract

The invention relates to a power emergency plan text generation method and device, computer equipment and a storage medium, and relates to the technical field of power safety. The method comprises the following steps: identifying target accident information based on a pre-constructed electric power plan dictionary, and determining keywords contained in the target accident information; generating emergency plans corresponding to the keywords on the basis of triples and the keywords contained in the electric power knowledge graph; the triad is generated based on the entity of each field in the emergency disposal measure library and the incidence relation of each field in the emergency disposal measure library; the emergency disposal measure library is obtained by performing text extraction on historical emergency texts based on an electric power plan dictionary; and according to an association relationship among the keywords, merging the texts of the plurality of emergency plans to obtain an emergency plan text corresponding to the target accident information. By adopting the method, the accuracy of generating the emergency plan text can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of electric power safety technology, and in particular to a method, device, computer equipment, storage medium and computer program product for generating an electric power emergency plan text. Background Art

[0002] The emergency plan is a standardized response procedure for emergency response, the basic rules for emergency handling and the operating guide for emergency response. After an emergency occurs, staff and power equipment can take emergency measures according to the contents of the emergency plan to reduce the losses caused by the emergency.

[0003] However, in actual emergency scenarios, the decision makers usually summarize the existing plans based on their experience to obtain an emergency plan for the current emergency. Due to the complexity of the damage and rescue situations of different accidents, emergency plans that rely too much on experience are prone to omissions, resulting in insufficient accuracy in the generated emergency plan text. Summary of the invention

[0004] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for generating electric power emergency plan text that can improve the accuracy of emergency plan text in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for generating a text of an electric power emergency plan. The method comprises:

[0006] Identify target accident information based on a pre-built power emergency plan dictionary to determine keywords included in the target accident information;

[0007] Based on the triples and keywords contained in the electric power knowledge graph, an emergency plan corresponding to each keyword is generated; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the electric power plan dictionary;

[0008] According to the association relationship between the keywords, the texts of the multiple emergency plans are merged to obtain the emergency plan text corresponding to the target accident information.

[0009] In one embodiment, the method further comprises:

[0010] Perform word segmentation processing on the electric power field text to obtain initial words corresponding to the electric power field text;

[0011] Preprocessing redundant data, missing data, stop word data, and abnormal data in the initial word to obtain a target word;

[0012] A feature value corresponding to the target word is generated, and the power plan dictionary is obtained based on the target word and the feature value of the target word.

[0013] In one embodiment, the method further comprises:

[0014] Perform word segmentation processing on the historical emergency text based on the power plan dictionary to determine the characteristic values ​​of key fields contained in the historical emergency text;

[0015] Classifying the characteristic values ​​of the key fields based on a text recognition algorithm to determine the field type corresponding to each of the key fields;

[0016] For the current field type, determining the measure text corresponding to the key field in the historical emergency text;

[0017] The measure texts corresponding to the multiple key fields are merged to obtain at least one emergency response measure library corresponding to the historical emergency texts.

[0018] In one embodiment, the method further comprises:

[0019] According to a preset association rule mining strategy, rule mining is performed on the key fields contained in the emergency disposal measure library and the measure texts corresponding to the key fields to obtain the association relationship between the key fields.

[0020] In one embodiment, the method further comprises:

[0021] Based on the association relationship between the key fields, performing relationship extraction on the emergency disposal measures library to obtain an initial triple generated by the emergency disposal measures library;

[0022] Entity disambiguation and coreference resolution are performed on the entities in the initial triples to obtain target triples, and multiple target triples are merged to obtain the power knowledge graph.

[0023] In one embodiment, the generating of emergency plans corresponding to the keywords based on the triples and keywords included in the electric power knowledge graph includes:

[0024] Searching the electric power knowledge graph according to the keyword to obtain a key triple corresponding to the keyword;

[0025] Based on the association relationship corresponding to each entity in the key triple, obtaining an associated triple adjacent to the key triple;

[0026] A first emergency plan included in the key triple is obtained, and a second emergency plan included in the associated triple is obtained, and the first emergency plan and the second emergency plan are determined to be emergency plans corresponding to the keyword.

[0027] In one embodiment, the texts of the multiple emergency plans are merged according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information, including:

[0028] If the association relationship is a superior-subordinate relationship, the text of the emergency plan is merged according to the superior-subordinate order between the keywords to obtain the emergency plan text corresponding to the target accident information;

[0029] If the correlation relationship between the keywords is no relationship, the emergency plans corresponding to the keywords with no correlation relationship are deleted;

[0030] If the association relationship is a type duplication relationship, the emergency plan corresponding to the low-weight keyword is deleted; the low-weight keyword is determined based on the weight values ​​of two keywords of the same type in the target accident information.

[0031] In a second aspect, the present application also provides a device for generating a text of an electric power emergency plan. The device comprises:

[0032] A keyword recognition module, used to recognize target accident information based on a pre-built power emergency plan dictionary, and determine keywords contained in the target accident information;

[0033] A plan generation module is used to generate emergency plans corresponding to each keyword based on the triples and keywords contained in the power knowledge graph; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the power plan dictionary;

[0034] The text merging module is used to merge the texts of multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information.

[0035] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the following steps are implemented:

[0036] Identify target accident information based on a pre-built power emergency plan dictionary to determine keywords included in the target accident information;

[0037] Based on the triples and keywords contained in the electric power knowledge graph, an emergency plan corresponding to each keyword is generated; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the electric power plan dictionary;

[0038] According to the association relationship between the keywords, the texts of the multiple emergency plans are merged to obtain the emergency plan text corresponding to the target accident information.

[0039] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0040] Identify target accident information based on a pre-built power emergency plan dictionary to determine keywords included in the target accident information;

[0041] Based on the triples and keywords contained in the electric power knowledge graph, an emergency plan corresponding to each keyword is generated; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the electric power plan dictionary;

[0042] According to the association relationship between the keywords, the texts of the multiple emergency plans are merged to obtain the emergency plan text corresponding to the target accident information.

[0043] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0044] Identify target accident information based on a pre-built power emergency plan dictionary to determine keywords included in the target accident information;

[0045] Based on the triples and keywords contained in the electric power knowledge graph, an emergency plan corresponding to each keyword is generated; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the electric power plan dictionary;

[0046] According to the association relationship between the keywords, the texts of the multiple emergency plans are merged to obtain the emergency plan text corresponding to the target accident information.

[0047] The above-mentioned method, device, computer equipment, storage medium and computer program product for generating the text of the power emergency plan identify the target accident information through the pre-constructed power plan dictionary, obtain multiple keywords contained in the target accident information, and generate the emergency plan corresponding to each keyword based on the keywords and the triples contained in the power knowledge graph. The triples are generated by the entities of each field in the emergency disposal measures library and the association relationship between each field. The emergency disposal measures library is obtained by extracting the text of the historical emergency text through the power plan dictionary. Based on this, the server can merge the text of multiple emergency plans for each keyword according to the association relationship between the keywords, so as to obtain the emergency plan text corresponding to the target accident information. Since the keywords are extracted from the target accident information, and the keywords can be used as the retrieval starting point of the power knowledge graph, the triples associated with the keywords are found from the power knowledge graph, and the emergency plan corresponding to the triples is obtained. The triples in the power knowledge graph are derived from the emergency response measures library, and the fields in the emergency response measures library are linked through association relationships. Therefore, the emergency plan obtained is the emergency plan text of each keyword in the emergency response measures library, and also includes the emergency plan text associated with the keywords, thereby improving the completeness and accuracy of the emergency plan text. In addition, by merging the text of the emergency plan through the association relationship of keywords, redundant text and erroneous text can be eliminated, and the accuracy of the emergency plan text corresponding to the target accident information can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0049] FIG1 is an application environment diagram of a method for generating a text of an electric power emergency plan in one embodiment;

[0050] FIG2 is a schematic diagram of a flow chart of a method for generating a text of an electric power emergency plan in one embodiment;

[0051] FIG3 is a schematic flow diagram of the steps of generating an emergency plan for keywords in one embodiment;

[0052] FIG4 is a flow chart of a method for generating a text of an electric power emergency plan in another embodiment;

[0053] FIG5 is a structural block diagram of a device for generating a text of an electric power emergency plan in one embodiment;

[0054] FIG. 6 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] The method for generating the text of the power emergency plan provided in the embodiment of the present application can be applied to the application environment shown in FIG1. ​​Among them, the terminal 102 communicates with the server 104 through the network. The terminal 102 can send the acquired target accident information to the server 104, and the server 104 can parse the target accident information to obtain keywords. And through the triples contained in the keywords and the power knowledge graph, the emergency plan corresponding to each keyword is generated, and multiple emergency plans can be merged through the association relationship between the keywords to obtain the emergency plan text of the target accident information. Among them, the server 104 can pre-construct the power plan dictionary and the power knowledge graph, and the server 104 can use each field in the emergency disposal measures library as a component of the power knowledge graph. The server 104 can extract the data in the historical emergency text through the power plan dictionary, thereby constructing the emergency disposal measures library. The data storage system can store the data that the server 104 needs to process, such as target accident information, power plan dictionary, emergency disposal measures library and other data. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The head-mounted devices may be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0057] In an exemplary embodiment, as shown in FIG2 , a method for generating a text of an electric power emergency plan is provided, and the method is described by taking the application of the method to the server in FIG1 as an example, and includes the following steps S202 to S206. Among them:

[0058] Step S202: identifying the target accident information based on the pre-constructed power emergency plan dictionary, and determining the keywords included in the target accident information.

[0059] Among them, the power plan dictionary is used to identify professional terms in the power field and convert each professional term in the power field into a corresponding feature value. For example, the feature value can be a feature vector. The target accident information is the information describing the event when an emergency occurs. For example, the target accident information can include information such as the content of the accident, the cause of the accident, the type of accident, the scope of the accident, the address of the accident, and influencing factors. The target accident information can be an unstructured text, such as a text description of the accident by the staff, and the text can be a text converted from speech. The target accident information can be a structured text, such as a structured document after the staff fills in the document according to the template for the accident.

[0060] Specifically, the server obtains the target accident information, and segments the target accident information through a pre-built power plan dictionary to obtain multiple words. The server can perform statistical analysis on each segmented word in the target accident information, and determine the keyword corresponding to the target accident information from the multiple words.

[0061] In one example, the server may use the power plan dictionary and word segmentation software to perform multiple rounds of recognition on the target accident information, thereby completing word segmentation and obtaining words that only appear in the power plan dictionary.

[0062] In one example, the server can perform statistical analysis on multiple words after word segmentation through a word frequency analysis algorithm, and obtain keywords based on the analysis results. For example, the TF-IDF (term frequency–inverse document frequency) algorithm. The server can also add weights to each word through the power plan dictionary and combine it with the above-mentioned word frequency analysis algorithm to comprehensively determine keywords based on the weight of each word and the word frequency analysis results.

[0063] Step S204: Generate emergency plans corresponding to each keyword based on the triples and keywords contained in the power knowledge graph.

[0064] Among them, the triples are generated based on the entities of each field in the emergency response measures library and the association relationship between each field in the emergency response measures library; the emergency response measures library is obtained by extracting historical emergency texts based on the power plan dictionary. Historical emergency texts are plan documents corresponding to various situations and accidents formulated in the power field. For example, there are various types of document materials such as plan standard documents, historical real emergency measures documents, special emergency plans, on-site disposal plans, etc. The emergency response measures library contains multiple fields, and each field has corresponding attribute values. For example, the fields of the emergency response measures library can include area, title, event type, event level, response level, reference document, and measure description.

[0065] Specifically, the server traverses each keyword, and for the current keyword, the server finds the entity corresponding to the keyword in the electric power knowledge graph, as well as the triples related to the entity, and obtains multiple triples corresponding to the keyword. According to the attribute values ​​of other entities in the triple, the emergency plan for the keyword is obtained. In an example, the keyword is transformer failure, then multiple entities of transformer failure are retrieved in the electric power knowledge graph, and based on the association relationship between the entities in the electric power knowledge graph, multiple entities related to transformer failure are determined, and corresponding attribute values ​​are obtained from the related entities. The attribute value can be a hyperlink to a reference document, a measure description text, an accident level, a corresponding level, an accident scope, etc., which is not specifically limited in the embodiments of the present application. The attribute value can be the value of a field in the emergency response measures library.

[0066] Step S206, merging the texts of multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information.

[0067] Among them, the association relationship between keywords can be pre-configured through the power plan dictionary. Keywords can be descriptions of emergencies, the content of emergency management models, the organizational system and business processes of emergency management, and the influencing factors of emergencies. There can be pre-configured association relationships within the above content, and the association relationship of each keyword is recorded in the power plan dictionary. There may be an association relationship between keywords, or there may not be an association relationship. The association relationship between multiple keywords can be represented by a graph network structure, and the target accident information can include a graph network structure composed of one or more keywords.

[0068] Specifically, the server can treat the keywords in the same graph network structure as a category of keywords, and only merge the corresponding emergency plans in the same graph network structure to obtain the emergency plan text corresponding to the graph network structure. In addition, the emergency plan contains various types of text, such as blank text, unstructured text, structured text, and link text. For unstructured text and structured text, the field corresponding to the text in the emergency disposal measures library can be obtained, and the field and the text are used as a set of text data. For link-type text, it is necessary to obtain the text data corresponding to the link, and the field corresponding to the text in the emergency disposal measures library and the text data corresponding to the link are used as a set of text data. Based on this, the server can convert various types of text in the emergency plan into text data corresponding to the emergency disposal measures library. And merge multiple emergency plans contained in the same graph network structure to obtain the emergency plan text, and use it as the emergency plan text corresponding to the target accident information.

[0069] Optionally, the keywords of the target accident information may constitute multiple graph network structures, and the server may perform statistical analysis on the multiple graph network structures to determine the graph network structure that best matches the target view information. The server deletes the emergency plans corresponding to other graph network structures, merges the texts of the multiple emergency plans of the graph network structure that best matches the target view information, and uses the merged emergency plan text as the emergency plan text corresponding to the target accident information.

[0070] In the above-mentioned method for generating the text of the power emergency plan, the target accident information is identified by a pre-constructed power plan dictionary to obtain multiple keywords contained in the target accident information, and based on the keywords and the triples contained in the power knowledge graph, an emergency plan corresponding to each keyword is generated. The triples are generated by the entities of each field in the emergency disposal measures library and the association relationship between each field. The emergency disposal measures library is obtained by extracting the text of the historical emergency text through the power plan dictionary. Based on this, the server can merge the text of multiple emergency plans for each keyword according to the association relationship between the keywords, so as to obtain the emergency plan text corresponding to the target accident information. Since the keywords are extracted from the target accident information, and the keywords can be used as the retrieval starting point of the power knowledge graph, the triples associated with the keywords are found from the power knowledge graph, and the emergency plan corresponding to the triples is obtained. The triples in the power knowledge graph are derived from the emergency response measures library, and the fields in the emergency response measures library are linked through association relationships. Therefore, the emergency plan obtained is the emergency plan text of each keyword in the emergency response measures library, and also includes the emergency plan text associated with the keywords, thereby improving the completeness and accuracy of the emergency plan text. In addition, by merging the text of the emergency plan through the association relationship of keywords, redundant text and erroneous text can be eliminated, and the accuracy of the emergency plan text corresponding to the target accident information can be improved.

[0071] In an exemplary embodiment, the method for generating a power emergency plan text further includes:

[0072] The electric power field text is segmented to obtain the initial words corresponding to the electric power field text; the redundant data, missing data, stop word data, and abnormal data in the initial words are preprocessed to obtain the target words; the feature values ​​corresponding to the target words are generated, and the electric power plan dictionary is obtained based on the target words and the feature values ​​of the target words.

[0073] The power field texts can be all standard documents in the power field, such as standard documents corresponding to multiple power sub-fields such as power infrastructure, power transmission, power maintenance, power failure, etc. The power field texts can be professional books, papers and other texts published by professionals.

[0074] Specifically, the server can segment each power field text through the word segmentation software to obtain all the initial words contained in the power field text. Due to the defects of the word segmentation software itself and the text quality of the power field text, the initial words may contain redundant data, missing data, stop word data, abnormal data and other data. The server can perform deduplication processing on redundant data to remove duplicate data. The server can delete or supplement the actual data, delete the missing data, or generate a preset value to assign a value to the missing data. The server can perform deletion processing on the stop word data and delete the stop words that have no actual semantics. The server can delete the value of the abnormal data for the abnormal data, or use the normal value corresponding to the abnormal data to cover the abnormal data. For example, the abnormal data is 100000, but the value range of the field is 0 to 1000, then the average value of 500 can be taken to replace 100000. Alternatively, the abnormal data 100000 can be deleted directly.

[0075] Based on this, the server performs data preprocessing on all the initial words to obtain the target words. The server can establish a feature vector for each target word based on the word vector model and use it as the feature value corresponding to the target word, so as to uniquely represent each target word through the feature value, and can also determine the similarity between each target word through the feature vector. The server takes the target word and the feature value of the target word as a set of data and stores it in the power plan dictionary, thereby obtaining a power plan dictionary composed of all the target words.

[0076] Optionally, the server may receive the association relationship between preset target words and the weight coefficient of the target word itself, and store them in the power plan dictionary.

[0077] Optionally, the server can perform word segmentation on the new power field text to obtain the initial words corresponding to the new power field text, and pre-process the redundant data, missing data, stop word data, and abnormal data in the initial words to obtain the target words, and update the target words and the feature values ​​of the target words to the existing power plan dictionary.

[0078] In this embodiment, the power field text is segmented and the initial words after segmentation are preprocessed to obtain the target words. Finally, the target words and the feature values ​​of the target words are stored as a set of data, thereby obtaining a power plan dictionary composed of all target words, which can improve the accuracy of the power plan dictionary.

[0079] In an exemplary embodiment, the method for generating a power emergency plan text further includes:

[0080] Based on the power plan dictionary, the historical emergency text is segmented to determine the characteristic values ​​of the key fields contained in the historical emergency text; based on the text recognition algorithm, the characteristic values ​​of the key fields are classified to determine the field type corresponding to each key field; according to the current field type, the measure text corresponding to the key field in the historical emergency text is determined; the measure texts corresponding to multiple key fields are merged to obtain at least one emergency response measure library corresponding to the historical emergency text.

[0081] Among them, the text recognition algorithm can be a Bayesian classifier, a support vector machine, a logistic regression, a decision tree, a deep learning model or other algorithms.

[0082] Specifically, the server can perform word segmentation processing on each historical emergency text through the power plan dictionary established above, so as to obtain the key fields contained in the historical emergency text and the characteristic values ​​corresponding to the key fields. The server can input the characteristic values ​​of the key fields into the text recognition algorithm, and the text recognition algorithm processes the characteristic values ​​and obtains the classification results corresponding to the characteristic values, so as to obtain the field type of each key field. If the historical emergency text is a structured text, the server can obtain the measure text corresponding to the key field. If the historical emergency text is an unstructured text, the server can extract the measure text corresponding to the key field in the historical emergency text through a preset text extraction algorithm to obtain the measure text.

[0083] Based on this, the server merges multiple sets of key fields and measure texts to obtain the emergency response measure library contained in each historical emergency text. The server can divide the measure text into different emergency response measure libraries according to the accident scenario and response level. An emergency response measure library can be a library of emergency response measures for a certain level under a certain scenario. A historical emergency text can contain multiple accident scenarios and multiple response levels, so a historical emergency text can generate multiple emergency response measure libraries.

[0084] In this embodiment, historical emergency texts are classified through the power plan dictionary, the field types of key fields are determined, and the measure texts corresponding to the key fields are obtained. Finally, the key fields and the measure texts are taken as a set of data to obtain an emergency response measures library composed of multiple sets of data. This can generate an emergency response measures library for various scenarios, thereby improving the accuracy and richness of the emergency response measures library.

[0085] In an exemplary embodiment, the method for generating a power emergency plan text further includes:

[0086] According to the preset association rule mining strategy, rule mining is performed on the key fields contained in the emergency response measures library and the measure texts corresponding to the key fields to obtain the association relationship between the key fields.

[0087] Among them, the association rule mining strategy can process the existing text data to obtain the implicit relationship between the text data. The association rule mining strategy can include the Apriori algorithm, FP-Growth algorithm, Eclat algorithm and other strategies.

[0088] Specifically, the server processes the key fields and the measure texts of the key fields contained in all emergency response measures libraries through a preset association rule mining strategy, and determines the implicit relationship between the key fields or between the measure texts based on the existing correspondence between the key fields and the measure texts. The server can pre-process multiple associations generated by the association rule mining strategy, merge multiple associations with repeated key fields, and delete unreasonable associations, and finally obtain the associations between the key fields.

[0089] Optionally, if the server generates implicit relationships between measure texts through a preset association rule mining strategy, the measure texts may be converted into corresponding key fields, thereby obtaining implicit relationships between key fields.

[0090] In this embodiment, the implicit relationship between key fields can be obtained through the association rule mining strategy, which can improve the correlation between the fields and provide support for the subsequent construction of the power knowledge graph.

[0091] In an exemplary embodiment, the method for generating a power emergency plan text further includes:

[0092] Based on the association between each key field, relationship extraction is performed on the emergency response measures library to obtain the initial triples generated by the emergency response measures library; entity disambiguation and coreference resolution are performed on the entities in the initial triples to obtain the target triples, and multiple target triples are merged to obtain the power knowledge graph.

[0093] Specifically, the server can perform triple extraction on each field in the emergency response measures library and the measure text of each field based on the association relationship between each key field and the knowledge extraction strategy, and obtain multiple initial triples. The initial triples can be of various types such as entity-attribute-attribute value, entity-relationship-entity, for example, key field-attribute-measure text, key field-relationship-key field.

[0094] The server can perform entity disambiguation through different types of entity linking methods, such as by training a classifier to link entities to the correct candidate entities, or using neural network structures such as convolutional neural networks, recurrent neural networks, attention mechanisms, etc. to learn the representation and context information of entities to complete entity disambiguation.

[0095] The server can train machine learning models with labeled data to perform coreference resolution, thereby linking pronouns or noun phrases to their referent entities, usually using classifiers or neural network models. The server can also use clustering, similarity calculation and other techniques to aggregate co-referential entities together without the need for labeled training data.

[0096] After performing entity disambiguation and public knowledge resolution on multiple initial triples, the server obtains the target triples, and links multiple identical entities through the association relationships between the multiple target triples to obtain a graph network consisting of multiple target triples, which is used as the power knowledge graph.

[0097] In this embodiment, relations are extracted from the emergency response measures library through the association between key fields to obtain multiple initial triples, and entity disambiguation and coreference resolution are performed on the initial triples to obtain target triples, thereby constructing an electric power knowledge graph based on the target triples, which can improve the accuracy of the electric power knowledge graph.

[0098] In an exemplary embodiment, as shown in FIG3 , the specific implementation process of the step “generating an emergency plan corresponding to each keyword based on the triples and keywords contained in the power knowledge graph” includes steps S302 to S306. Among them:

[0099] Step S302: Search the electric power knowledge graph according to the keywords to obtain key triples corresponding to the keywords.

[0100] Among them, there is at least one entity in the key triple that is the entity corresponding to the keyword.

[0101] Specifically, the server traverses multiple keywords, searches the electric power knowledge graph for the current keyword, obtains triples containing the keyword, and determines them as key triples.

[0102] Step S304: based on the association relationship corresponding to each entity in the key triple, obtain the associated triples adjacent to the key triple.

[0103] Among them, the entity or attribute value in the associated triplet has an associated relationship with the key triplet. The adjacent distance between the key triplet and the associated triplet can be represented by the number of associated relationships. The longer the distance, the lower the correlation between the associated triplet and the key triplet.

[0104] Specifically, the server can determine multiple associated triples corresponding to each key triple in the power knowledge graph according to the size of the adjacent distance.

[0105] Step S306, obtaining the first emergency plan included in the key triplet, and obtaining the second emergency plan included in the associated triplet, and determining the first emergency plan and the second emergency plan as emergency plans corresponding to the keywords.

[0106] Specifically, for a keyword, the server obtains the emergency plan corresponding to each entity in each key triple corresponding to the keyword as the first emergency plan, and obtains the emergency plan corresponding to each entity in each associated triple as the second emergency plan. The first emergency plan and the second emergency plan are both emergency plans corresponding to the keyword.

[0107] In this embodiment, key triples are determined by keywords, and associated triples are determined by key triples and associated relationships, thereby obtaining multiple triples, and emergency plans of multiple triples are obtained as emergency plans corresponding to the keyword, which can improve the completeness of the emergency plan for the keyword.

[0108] In an exemplary embodiment, the specific implementation process of the step of "merging the texts of multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information" includes:

[0109] If the association relationship is a superior-subordinate relationship, the text of the emergency plan is merged according to the superior-subordinate order between the keywords to obtain the emergency plan text corresponding to the target accident information. Alternatively, if the association relationship between the keywords is no relationship, the emergency plan corresponding to the keyword with no relationship is deleted. Alternatively, if the association relationship is a type duplication relationship, the emergency plan corresponding to the low-weight keyword is deleted.

[0110] Among them, the low-weight keyword is determined based on the weight values ​​of two keywords of the same type in the target accident information.

[0111] Specifically, the server can determine the association relationship between keywords through the power plan dictionary, and traverse the association relationship between each pair of keywords. If the association relationship between the two keywords is a superior-subordinate relationship, the emergency plans of the two keywords are merged in the order of superior and subordinate, and the superior emergency plan is used as the upper plan, and the subordinate emergency plan is used as the lower plan. If the association relationship between the two keywords is no relationship, it is re-determined whether there is no association relationship between one of the keywords and other keywords. If there is no association relationship between the keyword and other keywords, the emergency plan corresponding to the keyword is deleted. If the association relationship between the two keywords is a type of repeated key, that is, the types of the two keywords are the same, it is necessary to determine which of the two keywords is a redundant keyword, then it can be determined which of the two keywords is a low-weight keyword and used as a redundant keyword. The server deletes the emergency plan corresponding to the low-weight keyword.

[0112] In this embodiment, by merging or deleting the emergency plans corresponding to the keywords through the association relationship between the keywords, it is possible to eliminate redundant texts in the emergency plans, thereby improving the accuracy of the emergency plans.

[0113] As shown in FIG4 , the following describes in detail a specific execution module of the above-mentioned method for generating a text of an electric power emergency plan in combination with a specific embodiment, including:

[0114] Module 1, emergency response measures library generation module: identify key fields in the emergency plan through natural language processing technology, read the contents of documents such as special emergency plans and on-site disposal plans, and form an emergency response measures library of different levels for different scenarios. The key fields of the measures library include area, title, event type, event level, response level, reference documents, measures description, etc.

[0115] Module 2, feature engineering module, includes: (1) obtaining text data from power field texts, and performing deduplication and filling on redundant data, large missing data, stop word related data, and abnormal data in the text data to achieve basic feature processing and obtain preprocessed text data. (2) The server can build a power plan dictionary based on text data, and the power plan dictionary can be used to recognize common words. (3) Combined with the Hidden Markov Model HMM (Hidden Markov Model) model, new words in new power field texts can be added to the power plan dictionary, thereby completing the update of the power plan dictionary.

[0116] Module 3, association rule mining module: for the data preprocessed in step 1, preliminary rules are obtained by comparing the experimental selection item set and rule mining scheme. Then, the association rules that are not helpful for building the knowledge graph are discarded by combining the text feature recognition algorithm, so as to summarize and analyze the association relationship between different categories, and use the management relationship as the ontology construction basis of the knowledge graph.

[0117] Module 4, knowledge graph construction module: (1) Integrate the data after entity, relationship, and attribute extraction to establish triples and complete preliminary knowledge representation; (2) Perform entity disambiguation and coreference resolution on the obtained triples to achieve standard knowledge representation; (3) Build an emergency plan knowledge graph based on standard knowledge representation.

[0118] Module 5, intelligent optimization module: (1) Update and iterate the models of each algorithm used in the above method based on model evaluation indicators, such as precision rate, recall rate, etc.; (2) Combine the evaluation results obtained by the knowledge graph construction module with the actual damaged result feedback to determine the differentiated data, and adjust the model parameters with the differentiated data.

[0119] Module 6, emergency plan visualization module: integrates external data, monitoring data, and business data to present the generation of emergency plans in a data visualization manner, and intuitively displays data such as areas, event types, measures, etc. involved in the plans over a period of time.

[0120] In this embodiment, by analyzing emergencies and emergency plans, the emergency management process and management system are studied, and natural language processing related technologies are combined to achieve data acquisition and text processing. At the same time, association rules are used to mine the correlation between accident information and plan handling solutions, and a method for generating emergency plan text is constructed in combination with knowledge graphs, which improves the accuracy of generating emergency plan text.

[0121] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0122] Based on the same inventive concept, the embodiment of the present application also provides a device for generating a text of an electric power emergency plan for realizing the method for generating a text of an electric power emergency plan involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the device for generating a text of an electric power emergency plan provided below can refer to the limitations of the method for generating a text of an electric power emergency plan above, and will not be repeated here.

[0123] In an exemplary embodiment, as shown in FIG5 , a device 500 for generating a text of an electric power emergency plan is provided, comprising: a keyword recognition module 501 , a plan generation module 502 and a text merging module 503 , wherein:

[0124] The keyword recognition module 501 is used to recognize the target accident information based on the pre-built power emergency plan dictionary and determine the keywords contained in the target accident information;

[0125] The plan generation module 502 is used to generate emergency plans corresponding to each keyword based on the triples and keywords contained in the power knowledge graph; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the power plan dictionary;

[0126] The text merging module 503 is used to merge the texts of multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information.

[0127] Furthermore, the device also includes a dictionary generation module, which is specifically used to: perform word segmentation processing on the power field text to obtain the initial words corresponding to the power field text; pre-process the redundant data, missing data, stop word data, and abnormal data in the initial words to obtain the target words; generate feature values ​​corresponding to the target words, and obtain the power plan dictionary based on the target words and the feature values ​​of the target words.

[0128] Furthermore, the device also includes a measures library generation module, which is specifically used to: perform word segmentation processing on historical emergency texts based on the power plan dictionary to determine the characteristic values ​​of key fields contained in the historical emergency texts; classify the characteristic values ​​of key fields based on the text recognition algorithm to determine the field type corresponding to each key field; determine the measures text corresponding to the key field in the historical emergency text for the current field type; merge the measures texts corresponding to multiple key fields to obtain at least one emergency response measures library corresponding to the historical emergency text.

[0129] Furthermore, the device also includes a relationship mining module, which is specifically used to: perform rule mining on the key fields contained in the emergency response measures library and the measure texts corresponding to the key fields according to a preset association rule mining strategy, so as to obtain the association relationship between the key fields.

[0130] Furthermore, the device also includes a graph generation module, which is specifically used to: extract relationships from the emergency response measures library based on the association relationship between each key field to obtain initial triples generated by the emergency response measures library; perform entity disambiguation and co-reference resolution on the entities in the initial triples to obtain target triples, and merge multiple target triples to obtain an electric power knowledge graph.

[0131] Furthermore, the plan generation module 502 is specifically used to: search in the electric power knowledge graph according to the keywords to obtain the key triples corresponding to the keywords; based on the association relationship corresponding to each entity in the key triple, obtain the associated triples adjacent to the key triples; obtain the first emergency plan contained in the key triples, and obtain the second emergency plan contained in the associated triples, and determine that the first emergency plan and the second emergency plan are the emergency plans corresponding to the keywords.

[0132] Furthermore, the text merging module is specifically used to: if the association relationship is a superior-subordinate relationship, the text of the emergency plan is merged in the superior-subordinate order between the keywords to obtain the emergency plan text corresponding to the target accident information; if the association relationship between the keywords is no relationship, the emergency plan corresponding to the keyword with no relationship is deleted; if the association relationship is a type duplication relationship, the emergency plan corresponding to the low-weight keyword is deleted; the low-weight keyword is determined based on the weight value of two keywords of the same type in the target accident information.

[0133] Each module in the above-mentioned power emergency plan text generation device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0134] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG6. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data such as power knowledge graphs, power plan dictionaries, and historical emergency texts. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for generating a power emergency plan text is implemented.

[0135] Those skilled in the art will understand that the structure shown in FIG. 6 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0136] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0137] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0138] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0139] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0140] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0141] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0142] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for generating a text of an electric power emergency plan, characterized in that: The method comprises: Identify target accident information based on a pre-built power emergency plan dictionary to determine keywords included in the target accident information; Based on the triples and keywords contained in the electric power knowledge graph, an emergency plan corresponding to each keyword is generated; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the electric power plan dictionary; According to the association relationship between the keywords, the texts of the multiple emergency plans are merged to obtain the emergency plan text corresponding to the target accident information.

2. The method according to claim 1, characterized in that: The method further comprises: Perform word segmentation processing on the electric power field text to obtain initial words corresponding to the electric power field text; Preprocessing redundant data, missing data, stop word data, and abnormal data in the initial word to obtain a target word; A feature value corresponding to the target word is generated, and the power plan dictionary is obtained based on the target word and the feature value of the target word.

3. The method according to claim 1, characterized in that The method further comprises: Perform word segmentation processing on the historical emergency text based on the power plan dictionary to determine the characteristic values ​​of key fields contained in the historical emergency text; Classifying the characteristic values ​​of the key fields based on a text recognition algorithm to determine the field type corresponding to each of the key fields; For the current field type, determining the measure text corresponding to the key field in the historical emergency text; The measure texts corresponding to the multiple key fields are merged to obtain at least one emergency response measure library corresponding to the historical emergency texts.

4. The method according to claim 3, characterized in that The method further comprises: According to a preset association rule mining strategy, rule mining is performed on the key fields contained in the emergency disposal measure library and the measure texts corresponding to the key fields to obtain the association relationship between the key fields.

5. The method according to claim 4, characterized in that The method further comprises: Based on the association relationship between the key fields, performing relationship extraction on the emergency disposal measures library to obtain an initial triple generated by the emergency disposal measures library; Entity disambiguation and coreference resolution are performed on the entities in the initial triples to obtain target triples, and multiple target triples are merged to obtain the power knowledge graph.

6. The method according to claim 1, characterized in that The generating of emergency plans corresponding to the keywords based on the triples and keywords included in the electric power knowledge graph includes: Searching the electric power knowledge graph according to the keyword to obtain a key triple corresponding to the keyword; Based on the association relationship corresponding to each entity in the key triple, obtaining an associated triple adjacent to the key triple; A first emergency plan included in the key triple is obtained, and a second emergency plan included in the associated triple is obtained, and the first emergency plan and the second emergency plan are determined to be emergency plans corresponding to the keyword.

7. The method according to claim 1, characterized in that The step of merging the texts of the multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information includes: If the association relationship is a superior-subordinate relationship, the text of the emergency plan is merged according to the superior-subordinate order between the keywords to obtain the emergency plan text corresponding to the target accident information; If the correlation relationship between the keywords is no relationship, the emergency plans corresponding to the keywords with no correlation relationship are deleted; If the association relationship is a type duplication relationship, the emergency plan corresponding to the low-weight keyword is deleted; the low-weight keyword is determined based on the weight values ​​of two keywords of the same type in the target accident information.

8. A device for generating a text of an electric power emergency plan, characterized in that: The device comprises: A keyword recognition module, used to recognize target accident information based on a pre-built power emergency plan dictionary, and determine keywords contained in the target accident information; A plan generation module is used to generate emergency plans corresponding to each keyword based on the triples and keywords contained in the power knowledge graph; the triples are generated based on the entities of each field in the emergency disposal measures library and the association relationship between each field in the emergency disposal measures library; the emergency disposal measures library is obtained by extracting historical emergency texts based on the power plan dictionary; The text merging module is used to merge the texts of multiple emergency plans according to the association relationship between the keywords to obtain the emergency plan text corresponding to the target accident information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Emergency plan generation method and system based on artificial intelligence

    CN119227980A