A two-ticket illegal information identification method, system, device and medium

Abstract

The application discloses a two-ticket illegal information identification method, system, device and medium, and relates to the technical field of power operation management, and the method specifically comprises the following steps: in response to a received illegal information identification request, determining an operation ticket and a work ticket corresponding to the illegal information identification request, creating a ticket database, extracting key information of the operation ticket and the work ticket according to the ticket database, and checking whether the format expression of the key information is compliant; creating an operation step checking model, obtaining operation steps recorded in the operation ticket and the work ticket, and checking whether the operation steps are compliant according to the operation step checking model; and creating a power knowledge graph, and checking whether the execution task and the associated requirements of the operation ticket and the work ticket are compliant according to the power knowledge graph. The compliance check is performed through an automatic process, the auditing efficiency of the operation ticket and the work ticket is greatly improved, and the risk of illegal problems is reduced.

Description

Technical Field

[0001] This invention relates to the field of power operation management technology, and in particular to a method, system, device and medium for identifying violations using two types of tickets. Background Technology

[0002] In power systems, ensuring the implementation of the "two-ticket" system is a fundamental means of guaranteeing safety. The operation ticket, also known as a switching operation ticket, is the process of switching electrical equipment from one state to another. The sequence of operations can be written on the operation ticket, and the operation of the equipment must be strictly performed according to the order on the operation ticket. This is also a rigid system regulation to make switching operations more standardized, regulated, and procedural.

[0003] To avoid various safety accidents while working on electrical equipment, and also to clarify responsibilities after an accident occurs, facilitate accident handling, and prevent similar accidents from happening again, a work permit system must be implemented when working on high-voltage equipment.

[0004] However, during the execution of operation tickets or work tickets, problems often arise due to incomplete information or misunderstandings. In particular, the execution of work tasks often involves exceeding the scope of work, i.e., "operating with additional tickets," which leads to violations. Generally, the supervision of "two tickets" is usually done by manually logging into the power grid management platform to check the data. However, the data volume is too large, and it is impossible for humans to see all the data. Furthermore, it is difficult to analyze related data. Summary of the Invention

[0005] This invention provides a method, system, device, and medium for identifying violations using two tickets, which solves the technical problems of existing methods that rely on manual login to the power grid management platform for data verification, but the data volume is too large, making it impossible for humans to view all the data and making it difficult to analyze related data.

[0006] The first aspect of this invention provides a method for identifying violation information of two tickets, comprising:

[0007] In response to a received violation information identification request, determine the operation ticket and work ticket corresponding to the violation information identification request;

[0008] Create a ticket database, extract key information from the operation ticket and the work ticket based on the ticket database, and verify whether the format of the key information is compliant;

[0009] Create an operation step verification model, obtain the operation steps recorded in the operation ticket and the work ticket, and verify whether the operation steps are compliant based on the operation step verification model;

[0010] Create a power knowledge graph, and verify the compliance of the execution tasks and associated requirements of the operation ticket and work ticket based on the power knowledge graph.

[0011] Optionally, the step of extracting key information from the operation ticket and the work ticket based on the ticket database, and verifying whether the format of the key information is compliant, includes:

[0012] Input the content information of the operation ticket and work ticket into the ticket database, scan and obtain the key information of the ticket content information;

[0013] The key information includes equipment information, time information, operation information, personnel information, and safety measures;

[0014] Verify whether there are any missing or incorrect characters in the equipment information, time information, operation information, personnel information, and safety measure information;

[0015] Verify whether there are any errors or omissions in the identification symbols of the equipment information, time information, operation information, personnel information, and safety measures information;

[0016] Verify whether the semantic order of the equipment information, time information, operation information, personnel information, and safety measure information is correct.

[0017] Optionally, the step of creating the operation step verification model includes:

[0018] Collect the historical operation step data recorded in the operation ticket and the work ticket, classify and label the historical operation step data according to the preset operation step order, and obtain the operation step dataset.

[0019] The preset operating sequence is: normal operation steps, change operation steps, emergency operation steps, maintenance operation steps, power outage operation steps, and start-up operation steps.

[0020] The operation step dataset includes an operation step training set and an operation step verification set;

[0021] The training set of the operation steps is segmented into words and converted into word sequences. The word sequences are then mapped into word embedding vectors using a word embedding model.

[0022] The word embedding vectors are input into the LSTM model for training, and the training results are output.

[0023] The gap between the training results and the operation step validation set is evaluated using the binary cross-entropy loss function, and the model is optimized to obtain the operation step verification model.

[0024] Optionally, the step of segmenting the training set of the operation steps into words and converting it into a word sequence, and mapping the word sequence into word embedding vectors through a word embedding model, includes:

[0025] The training set for the above operation steps was divided into several lexical units using a word segmentation tool;

[0026] Iterate through each vocabulary unit and assign a unique integer identifier to each vocabulary unit to generate a vocabulary list in which vocabulary units correspond to integer identifiers;

[0027] The training set of the operation steps is converted into an integer sequence using the vocabulary;

[0028] Load the pre-trained Word2Vec word vector model and construct an embedding layer based on the Word2Vec word vector model;

[0029] The integer sequence is input into the embedding layer and word embedding vectors are output.

[0030] Optionally, the step of inputting the word embedding vector into the LSTM model for training and outputting the training result includes:

[0031] The training parameters of the LSTM model are set according to the text parameters of the word embedding vector;

[0032] Set the Sigmoid function as the activation function of the LSTM model to make the output of the LSTM model either compliant or non-compliant;

[0033] The word embedding vectors are input into the LSTM model for training to obtain the initial operation step verification model;

[0034] The operation step verification set is input into the initial operation step verification model, and the training results are output.

[0035] Optionally, the step of creating the power knowledge graph includes:

[0036] Obtain the type of task to be executed for the operation ticket and the work ticket;

[0037] Create a named entity recognition model;

[0038] Based on the named entity recognition model, obtain the respective power language entities for the equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements in the task being executed;

[0039] Each type of the task being executed is treated as a node, and an association graph is constructed with the relevant electrical language entities.

[0040] All the related graphs are combined to generate an electricity knowledge graph.

[0041] Optionally, the step of creating a named entity recognition model includes:

[0042] Obtain the electrical language text of the task being performed;

[0043] The power language texts are classified and labeled according to a preset order to obtain a power language dataset.

[0044] The preset requirements are in the following order: equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements.

[0045] The power language dataset includes a power language training set and a power language validation set;

[0046] Load the pre-trained BERT model, input the power language training set into the BERT model for training, and obtain the initial named entity recognition model;

[0047] The power language validation set is input into the initial named entity recognition model, and the power language entity classification result is output.

[0048] The gap between the classification results of the power language entity and the power language validation set is evaluated using a multi-class cross-entropy loss function, and the model is optimized to obtain a named entity recognition model.

[0049] A second aspect of the present invention provides a two-ticket violation information identification system, comprising:

[0050] The response module is used to respond to a received violation information identification request and determine the operation ticket and work ticket corresponding to the violation information identification request;

[0051] The first verification module is used to create a ticket database, extract key information of the operation ticket and the work ticket from the ticket database, and verify whether the format of the key information is compliant.

[0052] The second verification module is used to create an operation step verification model, obtain the operation steps recorded in the operation ticket and the work ticket, and verify whether the operation steps are compliant according to the operation step verification model.

[0053] The third verification module is used to create a power knowledge graph and verify whether the execution tasks and associated requirements of the operation ticket and the work ticket are compliant based on the power knowledge graph.

[0054] A third aspect of the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the two-ticket violation information identification method as described in any of the preceding claims.

[0055] The fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, it implements the two-ticket violation information identification method as described in any of the preceding claims.

[0056] As can be seen from the above technical solutions, the present invention has the following advantages:

[0057] In response to a received violation information identification request, the system identifies the corresponding operation ticket and work ticket, creates a ticket database, extracts key information from the operation ticket and work ticket based on the database, verifies the compliance of the format of the key information, creates an operation step verification model, obtains the operation steps recorded on the operation ticket and work ticket, verifies the compliance of the operation steps based on the operation step verification model, and creates a power knowledge graph to verify the compliance of the execution tasks and related requirements of the operation ticket and work ticket. This system solves the technical problems of existing methods that rely on manual login to the power grid management platform for data verification, which suffers from excessive data volume, making it impossible for humans to view all data and analyze related data. By automating the compliance check process, the system significantly improves the efficiency of operation ticket and work ticket review and reduces the risk of violations.

[0058] Specifically, the technical effects include the following:

[0059] 1. An automated compliance check process has been introduced, which can greatly reduce the workload of manual review and improve the processing efficiency of operation tickets and work tickets.

[0060] 2. Perform formatting checks on the extracted key information, including checks for missing words, typos, incorrect or missing symbols, and semantic order, to ensure that the content of the operation tickets and work tickets is accurate and conforms to the specifications.

[0061] 3. By checking the compliance of the operation steps, we ensure that the steps in executing the task are error-free, thereby further improving the accuracy of automated auditing.

[0062] 4. Associating the execution task with the electrical language entity to verify the compliance of the execution task helps ensure that the execution tasks of operation tickets and work tickets are consistent with the relevant requirements. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 A flowchart illustrating the steps of a method for identifying traffic violations using two tickets, as provided in an embodiment of the present invention.

[0065] Figure 2 A flowchart illustrating the steps of a method for identifying traffic violations using two tickets, as provided in an embodiment of the present invention.

[0066] Figure 3 This is a structural block diagram of a two-ticket violation information identification system provided in an embodiment of the present invention. Detailed Implementation

[0067] This invention provides a method, system, device, and medium for identifying violations using two tickets, which solves the technical problem that existing methods rely on manual login to the power grid management platform for data verification, but the data volume is too large, making it impossible for humans to view all the data and making it difficult to analyze related data.

[0068] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0069] Please see Figure 1 , Figure 1 The flowchart illustrates the steps of a two-ticket violation information identification method provided in Embodiment 1 of the present invention.

[0070] This invention provides a method for identifying violation information from two tickets, comprising:

[0071] Step 101: In response to the received violation information identification request, determine the operation ticket and work ticket corresponding to the violation information identification request.

[0072] A violation information identification request refers to a request message for identifying violations in operation tickets and work tickets.

[0073] In this embodiment of the invention, in response to receiving a request for identifying violation information for operation tickets and work tickets, the operation ticket and work ticket corresponding to the violation information identification request are determined.

[0074] Step 102: Create a ticket database, extract key information from operation tickets and work tickets based on the ticket database, and verify whether the format of the key information is compliant.

[0075] A bill database refers to a database of rules that pre-define formatting specifications. For example, regular expressions can be used to define rules within the database to identify text with specific patterns or structures. This can be a simple string matching or a combination of multiple regular expressions.

[0076] In this embodiment of the invention, a bill database with pre-set format specifications is created. Key information of operation tickets and work tickets is extracted from the bill database, and the format expression of the key information is checked for compliance.

[0077] Step 103: Create an operation step verification model, obtain the operation steps recorded in the operation ticket and work ticket, and verify whether the operation steps are compliant based on the operation step verification model.

[0078] The operation procedure verification model refers to the verification model used to check whether the operation procedure conforms to the process specification.

[0079] It is worth mentioning that since the invoice database can only verify the text format—that is, the format specifications—it cannot verify the substantive content, such as the operation steps. Therefore, it is necessary to further verify whether the operation steps conform to the process specifications.

[0080] In this embodiment of the invention, a verification model is created to check whether the operation steps conform to the process specifications. The operation steps recorded in the operation ticket and work ticket are extracted, and the operation steps are verified to check whether they are compliant based on the operation step verification model.

[0081] Step 104: Create a power knowledge graph and verify the compliance of the execution tasks and associated requirements of the operation tickets and work tickets based on the power knowledge graph.

[0082] A power knowledge graph is a structured data representation method used to store and organize knowledge and information related to the power industry. It is a graphical data model that typically consists of entities (nodes) and the relationships between them (edges) to represent the connections between complex knowledge and concepts in the power industry.

[0083] In this embodiment of the invention, a power knowledge graph is created, and the execution tasks and associated requirements of operation tickets and work tickets are verified based on the power knowledge graph to ensure compliance.

[0084] In this invention, in response to a received violation information identification request, the corresponding operation ticket and work ticket are determined, a ticket database is created, key information of the operation ticket and work ticket is extracted from the ticket database, and the format of the key information is verified for compliance. An operation step verification model is created, the operation steps recorded on the operation ticket and work ticket are obtained, and the compliance of the operation steps is verified according to the operation step verification model. A power knowledge graph is created, and the compliance of the execution tasks and related requirements of the operation ticket and work ticket is verified according to the power knowledge graph. This invention solves the technical problem that existing methods rely on manual login to the power grid management platform for data verification, but the data volume is too large, making it impossible for humans to view all the data, and the related data is difficult to analyze. By automating the compliance check process, the efficiency of operation ticket and work ticket review is greatly improved, and the risk of violation issues is reduced.

[0085] Please see Figure 2 , Figure 2 This is a flowchart illustrating the steps of a two-ticket violation information identification method provided in Embodiment 2 of the present invention.

[0086] This invention provides a method for identifying violation information from two tickets, comprising:

[0087] Step 201: In response to the received violation information identification request, determine the operation ticket and work ticket corresponding to the violation information identification request.

[0088] In this embodiment of the invention, the specific implementation process of step 201 is similar to that of step 101, and will not be repeated here.

[0089] Step 202: Create a ticket database.

[0090] In this embodiment of the invention, a database of invoices with pre-defined formatting rules is created. The formatting rules differ for different key information. For example, equipment information includes substation name, equipment model, equipment number, switch name, transformer capacity, and transmission line name; time information includes operation date, maintenance plan date, power outage time, startup time, and planned maintenance time; operation information includes disconnecting power, operating circuit breakers, turning switches on or off, checking transformer oil levels, starting generators, and switching transmission lines; personnel information includes operators, on-duty engineers, safety inspectors, patrol inspectors, and technical maintenance personnel; and safety measures information includes electrical work permit numbers, safety warning signs, use of personal protective equipment, emergency power outage procedures, accident reporting procedures, high-voltage equipment discharge operations, and fire alarm levels.

[0091] Step 203: Input the information on the operation ticket and work ticket into the ticket database, and scan and obtain the key information on the ticket content.

[0092] Key information includes equipment information, time information, operation information, personnel information, and safety measures.

[0093] It is worth mentioning that key information includes, but is not limited to, equipment information, time information, operation information, personnel information, and safety measures, which can be set according to needs.

[0094] In this embodiment of the invention, the content information of the operation ticket and the work ticket is input into the ticket database, and the key information of the content information of the ticket is scanned and obtained. The key information includes equipment information, time information, operation information, personnel information and security measures.

[0095] Step 204: Check whether there are any missing or incorrect characters in the equipment information, time information, operation information, personnel information, and safety measures information.

[0096] In this embodiment of the invention, the equipment information, time information, operation information, personnel information, and safety measures information are checked for missing or incorrect characters.

[0097] Step 205: Verify whether there are any errors or omissions in the equipment information, time information, operation information, personnel information, and safety measures information.

[0098] In this embodiment of the invention, the system verifies whether there are any errors or omissions in the identification symbols of the equipment information, time information, operation information, personnel information, and safety measures information.

[0099] Step 206: Verify whether the semantic order of equipment information, time information, operation information, personnel information, and safety measures information is correct.

[0100] In this embodiment of the invention, the semantic order of equipment information, time information, operation information, personnel information, and safety measures information is checked to ensure that it is correct.

[0101] Step 207: Create an operation step verification model, obtain the operation steps recorded in the operation ticket and work ticket, and verify whether the operation steps are compliant based on the operation step verification model.

[0102] Furthermore, step 207 may include the following sub-steps:

[0103] S11. Collect historical operation step data recorded in operation tickets and work tickets, classify and label the historical operation step data according to the preset operation step order, and obtain the operation step dataset.

[0104] The preset operating sequence is: normal operation steps, change operation steps, emergency operation steps, maintenance operation steps, power outage operation steps, and start-up operation steps.

[0105] The operation step dataset includes an operation step training set and an operation step validation set.

[0106] The preset operation component sequence refers to the sequence of operation steps used for classification and labeling.

[0107] In this embodiment of the invention, historical operation step data recorded in the operation ticket and the work ticket are collected, and the historical operation step data is classified and labeled according to normal operation steps, modified operation steps, emergency operation steps, maintenance operation steps, power outage operation steps and start-up operation steps to obtain an operation step dataset. The operation step dataset includes an operation step training set and an operation step verification set.

[0108] It should be noted that, firstly, historical operation step data from operation tickets and work tickets must be collected. This data includes different types of operation steps, such as normal operation steps, change operation steps, emergency operation steps, maintenance operation steps, power outage operation steps, and startup operation steps. The data must be categorized and labeled to ensure that each operation step is correctly identified as its type. In the above operation step categories, normal operation steps refer to operation steps executed according to a predetermined procedure without any abnormalities; change operation steps refer to operation steps requiring manual intervention, such as manually switching equipment tool modes or setting parameters; emergency operation steps refer to operation steps performed in emergency situations; maintenance operation steps refer to operation steps for regular maintenance and inspection of equipment to ensure equipment performance and safety; power outage operation steps refer to planned power outage operations that are notified in advance and follow relevant safety procedures; and startup operation steps refer to operation steps that are executed in a specific sequence to start the system or equipment and ensure normal operation.

[0109] It should be noted that, secondly, based on the classification label, the data is divided into an operation step training set and an operation step validation set, which are used to train and evaluate the operation step verification model.

[0110] S12. Segment the training set of operation steps into words and convert it into a word sequence. Then, map the word sequence into a word embedding vector using a word embedding model.

[0111] It should be noted that, then, the operation step text is segmented into words, each operation step is converted into a word sequence, a vocabulary is constructed, words are mapped to unique integer identifiers, and a pre-trained word embedding model (such as Word2Vec or GloVe) is used to map the word sequence into word embedding vectors, which preserve the semantic information of the operation step text.

[0112] Furthermore, S12 may include the following sub-steps:

[0113] S121. Use a word segmentation tool to divide the training set of operation steps into several word units.

[0114] In this embodiment of the invention, when using a word segmentation tool to perform word segmentation on the training set of operation steps, the differences in language need to be considered. Different word segmentation tools are suitable for different languages. For example, the NLTK (Natural Language Toolkit) library is suitable for English text, while the jieba library is suitable for Chinese text. Appropriate word segmentation tools can be used under different languages. No limitations are made on language and word segmentation tools here.

[0115] S122. Traverse each vocabulary unit and assign a unique integer identifier to each vocabulary unit to generate a vocabulary list in which vocabulary units and integer identifiers correspond to each other.

[0116] In this embodiment of the invention, an empty vocabulary is then created, and each segmented vocabulary unit is entered sequentially. Each vocabulary unit is then assigned a unique integer identifier, for example, the first vocabulary unit is assigned an integer identifier of 1, and so on, until the vocabulary is filled.

[0117] S123. Convert the training set of operation steps into an integer sequence using a vocabulary list.

[0118] In this embodiment of the invention, the vocabulary units of the operation step training set are mapped to integer identifiers according to the vocabulary list to generate an integer sequence.

[0119] S124. Load the pre-trained Word2Vec word vector model and construct the embedding layer based on the Word2Vec word vector model.

[0120] In this embodiment of the invention, Word2Vec is a word embedding technique used to map words in text data to a continuous vector space. It can capture semantic relationships between words and calculate similarity scores, which is extremely useful for verifying the compliance of vocabulary in operational steps. If a word is highly similar to compliant words in the vector space, it is likely correct. Conversely, if it is similar to non-compliant words, further review may be necessary. Simultaneously, Word2Vec maps a high-dimensional vocabulary space to a low-dimensional vector space, helping to reduce the number of parameters in the operational step verification model, lowering computational complexity, and reducing the possibility of overfitting.

[0121] S125. Input the integer sequence into the embedding layer and output the word embedding vector.

[0122] In this embodiment of the invention, when constructing the embedding layer, the parameter settings of the embedding layer include the input dimension (size of the vocabulary), the output dimension (dimension of word embeddings), the maximum length of the input sequence, and the training weights.

[0123] S13. Input the word embedding vectors into the LSTM model for training and output the training results.

[0124] It should be noted that an LSTM model is then constructed for operation step verification. The word embedding vectors of the operation step training set are input into the LSTM model for training. The binary cross-entropy loss function is used to evaluate the gap between the training results and the operation step validation set, and the model is optimized. Through multiple iterations and adjustments to the model parameters, the operation step verification model is obtained.

[0125] It's worth noting that the LSTM (Long Short-Term Memory) model is a special type of recurrent neural network (RNN) designed to process and predict time-series data, as well as model sequential data such as text. The operation steps are typically ordered text sequences, with each step depending on previous steps. Therefore, LSTM can capture this sequence information, maintaining a memory of long sequences, which helps in analyzing the relationships between operation steps. Within these operation steps, some steps may depend on steps from much earlier times. Ordinary RNNs struggle to learn long-term dependencies due to the vanishing gradient problem, while LSTM, through its special structure, can effectively capture dependencies over long time periods, thus better modeling complex operation steps.

[0126] Furthermore, S13 may include the following sub-steps:

[0127] S131. Set the training parameters of the LSTM model according to the text parameters of the word embedding vector.

[0128] S132. Set the Sigmoid function as the activation function of the LSTM model to make the output of the LSTM model either compliant or non-compliant.

[0129] S133. Input the word embedding vector into the LSTM model for training to obtain the initial operation step verification model.

[0130] S134. Input the operation step verification set into the initial operation step verification model and output the training results.

[0131] In this embodiment of the invention, the hidden state dimension of the LSTM model can be set according to the dimension of the word embedding vectors, and the time step of the LSTM model can be set according to the length of the text sequence of the word embedding vectors. The hidden state dimension determines the representation dimension of each word unit in the embedding space. A larger dimension can capture more semantic information, but it will increase the computational cost of the model. The time step refers to the number of words processed in each operation step in the model. If the time step is set too short, important information may be lost, but if it is set too long, the training and inference of the model will become more costly. In addition, there are parameters such as batch size, learning rate, dropout rate, and activation function of the LSTM model. The batch size determines the number of samples processed in each training step. Dropout is a regularization technique that helps prevent the model from overfitting. The sigmoid function, as an activation function, maps the input to the range (0, 1). It is used for binary classification problems and can interpret the output as a probability. In this model, it can output compliant or non-compliant results.

[0132] S14. The gap between the training results and the operation step validation set is evaluated by using the binary cross-entropy loss function, and the model is optimized to obtain the operation step verification model.

[0133] In this embodiment of the invention, the gap between the training results and the operation step validation set is evaluated by the binary cross-entropy loss function, and the model is optimized to obtain the operation step verification model.

[0134] Step 208: Create a power knowledge graph and verify the compliance of the execution tasks and associated requirements of the operation tickets and work tickets based on the power knowledge graph.

[0135] Furthermore, step 208 may include the following sub-steps:

[0136] S21. Obtain the type of task to be executed for the operation ticket and work ticket.

[0137] S22. Create a named entity recognition model.

[0138] S23. Obtain the respective power language entities for equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements in the task execution based on the named entity recognition model.

[0139] S24. Construct an association graph by treating each type of task as a node and connecting it with the relevant power language entities.

[0140] S25. Combine all the related graphs to generate an electricity knowledge graph.

[0141] In this embodiment of the invention, the power knowledge graph is a structured data representation method used to store and organize knowledge and information related to the power field. It is a graphical data model, typically composed of entities (nodes) and the relationships between them (edges), used to represent the associations between complex knowledge and concepts in the power field.

[0142] Entities in a power knowledge graph represent various objects, concepts, and data within the power sector. Relationships in a power knowledge graph represent the connections or dependencies between entities. For example, a power knowledge graph can include relationships between equipment and equipment status, between equipment and safety regulations, and between equipment and maintenance plans. Entities also contain attributes, which are additional information related to the entity and used to describe its characteristics or attributes. For example, an equipment entity can have attributes such as equipment model, manufacturing date, and rated capacity.

[0143] Creating a power knowledge graph requires collecting power-related knowledge related to equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements across different execution tasks. Based on the collected knowledge, the entities to be represented in the knowledge graph (such as equipment, time, safety measures, personnel, and materials) and their relationships (such as equipment status, time constraints, safety regulations, required personnel, and required materials) are determined. Based on the collected knowledge and the defined architecture, the power knowledge graph is then constructed.

[0144] The types of tasks to be performed include inspection tasks, maintenance tasks, power outage repair tasks, new equipment installation tasks, emergency fault repair tasks, etc. The relevant conditions of equipment requirements, time requirements, safety requirements, personnel requirements and material requirements for each task are clearly defined, thereby establishing a correlation graph, and the correlation graphs of each task are combined to form a power knowledge graph.

[0145] Furthermore, S22 may include the following sub-steps:

[0146] S221. Obtain the power language text for the task.

[0147] S222. Classify and label the power language texts according to the preset requirements to obtain the power language dataset.

[0148] The preset requirements are in the following order: equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements.

[0149] The power language dataset includes a power language training set and a power language validation set.

[0150] S223. Load the pre-trained BERT model, input the power language training set into the BERT model for training, and obtain the initial named entity recognition model.

[0151] S224. Input the power language validation set into the initial named entity recognition model and output the power language entity classification result.

[0152] S224. Evaluate the gap between the power language entity classification results and the power language validation set based on the multi-class cross-entropy loss function, and optimize the model to obtain the named entity recognition model.

[0153] In this embodiment of the invention, BERT (Bidirectional Encoder Representations from Transformers) is a deep learning model in Natural Language Processing (NLP) that can be used for various NLP tasks, such as text classification, named entity recognition, and question answering systems. The BERT model can be loaded by importing the Hugging Face Transformers library and using the BertModel and BertTokenizer classes from the Hugging Face Transformers library. One or more fully connected layers need to be added after the BERT model to receive the output of the BERT model and map it to the scores of named entity recognition labels. During training, the true values ​​of the named entity recognition labels are compared with the model's predicted values, and the loss is calculated. Then, the weights of the fully connected layers are updated through backpropagation, and a multi-class cross-entropy loss function is used to measure the loss of the named entity recognition task. During training, the performance metrics of the named entity recognition model, such as precision, recall, and F1 score, are monitored, and the model is adjusted based on the performance on the validation set.

[0154] In this invention, in response to a received violation information identification request, the corresponding operation ticket and work ticket are determined, a ticket database is created, key information of the operation ticket and work ticket is extracted from the ticket database, and the format of the key information is verified for compliance. An operation step verification model is created, the operation steps recorded on the operation ticket and work ticket are obtained, and the compliance of the operation steps is verified according to the operation step verification model. A power knowledge graph is created, and the compliance of the execution tasks and related requirements of the operation ticket and work ticket is verified according to the power knowledge graph. This invention solves the technical problem that existing methods rely on manual login to the power grid management platform for data verification, but the data volume is too large, making it impossible for humans to view all the data, and the related data is difficult to analyze. By automating the compliance check process, the efficiency of operation ticket and work ticket review is greatly improved, and the risk of violation issues is reduced.

[0155] Please see Figure 3 , Figure 3 This is a structural block diagram of a two-ticket violation information identification system provided in Embodiment 3 of the present invention.

[0156] This invention provides a two-ticket violation information identification system, comprising:

[0157] The response module 301 is used to respond to the received violation information identification request and determine the operation ticket and work ticket corresponding to the violation information identification request;

[0158] The first verification module 302 is used to create a ticket database, extract key information from operation tickets and work tickets based on the ticket database, and verify whether the format of the key information is compliant.

[0159] The second verification module 303 is used to create an operation step verification model, obtain the operation steps recorded in the operation ticket and work ticket, and verify whether the operation steps are compliant based on the operation step verification model.

[0160] The third verification module 304 is used to create a power knowledge graph and verify whether the execution tasks and associated requirements of operation tickets and work tickets are compliant based on the power knowledge graph.

[0161] Furthermore, the first verification module 302 includes:

[0162] The key information submodule is used to input the content information of operation tickets and work tickets into the ticket database, scan and obtain key information of the ticket content information;

[0163] Key information includes equipment information, time information, operational information, personnel information, and safety measures;

[0164] The missing and misspelled characters submodule is used to check whether there are missing or misspelled characters in equipment information, time information, operation information, personnel information, and safety measures information;

[0165] The identifier error and omission submodule is used to check whether there are identifier errors or omissions in equipment information, time information, operation information, personnel information, and safety measures information;

[0166] The semantic order verification submodule is used to verify whether the semantic order of equipment information, time information, operation information, personnel information, and safety measures information is correct.

[0167] Furthermore, the second verification module 303 includes:

[0168] The operation step dataset submodule is used to collect historical operation step data recorded in operation tickets and work tickets, classify and label the historical operation step data according to the preset operation step order, and obtain the operation step dataset.

[0169] The preset operating sequence is: normal operation steps, change operation steps, emergency operation steps, maintenance operation steps, power outage operation steps, and start-up operation steps.

[0170] The operation step dataset includes an operation step training set and an operation step validation set;

[0171] The word embedding vector submodule is used to segment the operation step training set into words and convert it into a word sequence. The word embedding model is then used to map the word sequence into word embedding vectors.

[0172] The training results submodule is used to input word embedding vectors into the LSTM model for training and output the training results.

[0173] The operation step verification model submodule is used to evaluate the gap between the training results and the operation step validation set through the binary cross-entropy loss function and optimize the model to obtain the operation step verification model.

[0174] Furthermore, the word embedding vector submodule includes:

[0175] Lexical unit is used to divide the training set of operation steps into several lexical units using a word segmentation tool;

[0176] The vocabulary unit is used to traverse each vocabulary unit and assign a unique integer identifier to each vocabulary unit, generating a vocabulary list in which vocabulary units and integer identifiers correspond to each other.

[0177] Integer sequence unit, used to convert the training set of operation steps into an integer sequence using a vocabulary;

[0178] The embedding layer unit is used to load the pre-trained Word2Vec word vector model and construct the embedding layer based on the Word2Vec word vector model.

[0179] The word embedding vector output unit is used to input an integer sequence into the embedding layer and output a word embedding vector.

[0180] Furthermore, the training results submodule includes:

[0181] The training parameter unit is used to set the training parameters of the LSTM model based on the text parameters of the word embedding vectors;

[0182] The activation function unit is used to set the Sigmoid function as the activation function of the LSTM model, which determines whether the output of the LSTM model is compliant or non-compliant.

[0183] The initial operation step verification model unit is used to input word embedding vectors into the LSTM model for training to obtain the initial operation step verification model.

[0184] The training result output unit is used to input the operation step verification set into the initial operation step verification model and output the training results.

[0185] Furthermore, the third verification module 304 includes:

[0186] The Task Type submodule is used to obtain the type of task to be executed from operation tickets and work tickets;

[0187] The Named Entity Recognition Model submodule is used to create named entity recognition models;

[0188] The Power Language Entity submodule is used to obtain the respective power language entities for equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements in the execution task based on the named entity recognition model;

[0189] The association graph submodule is used to construct an association graph by treating each type of task as a node and connecting it with the relevant power language entities.

[0190] The Power Knowledge Graph submodule is used to combine all the related graphs to generate a power knowledge graph.

[0191] Furthermore, the named entity recognition model submodule includes:

[0192] The power language text unit is used to obtain the power language text for the task execution.

[0193] The Electricity Language Dataset Unit is used to classify and label the electricity language texts according to a preset order to obtain an electricity language dataset.

[0194] The preset requirements are in the following order: equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements.

[0195] The power language dataset includes a power language training set and a power language validation set;

[0196] The initial named entity recognition model unit is used to load the pre-trained BERT model, input the power language training set into the BERT model for training, and obtain the initial named entity recognition model.

[0197] The power language entity classification result unit is used to input the power language validation set into the initial named entity recognition model and output the power language entity classification result.

[0198] The named entity recognition model creation unit is used to evaluate the gap between the power language entity classification results and the power language validation set based on the multi-class cross-entropy loss function and optimize the model to obtain the named entity recognition model.

[0199] In this invention, in response to a received violation information identification request, the corresponding operation ticket and work ticket are determined, a ticket database is created, key information of the operation ticket and work ticket is extracted from the ticket database, and the format of the key information is verified for compliance. An operation step verification model is created, the operation steps recorded on the operation ticket and work ticket are obtained, and the compliance of the operation steps is verified according to the operation step verification model. A power knowledge graph is created, and the compliance of the execution tasks and related requirements of the operation ticket and work ticket is verified according to the power knowledge graph. This invention solves the technical problem that existing methods rely on manual login to the power grid management platform for data verification, but the data volume is too large, making it impossible for humans to view all the data, and the related data is difficult to analyze. By automating the compliance check process, the efficiency of operation ticket and work ticket review is greatly improved, and the risk of violation issues is reduced.

[0200] An electronic device according to an embodiment of the present invention includes: a memory and a processor, wherein the memory stores a computer program; when the computer program is executed by the processor, the processor performs a two-ticket violation information identification method as described in any of the above embodiments.

[0201] The memory can be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. The memory has storage space for program code used to perform any of the method steps described above. For example, the storage space for program code may include individual program codes for implementing the various steps in the methods described above. This program code can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact discs (CDs), memory cards, or floppy disks. The program code may be compressed, for example, in a suitable form. When run by a computing processing device, this code causes the computing processing device to perform the various steps in the methods described above.

[0202] This invention provides a computer-readable storage medium storing a computer program thereon, which, when executed, implements a method for identifying violation information of two tickets as described in any embodiment of this invention.

[0203] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0204] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0205] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0206] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0207] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0208] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying traffic violation information using two tickets, characterized in that, include: In response to a received violation information identification request, determine the operation ticket and work ticket corresponding to the violation information identification request; Create a ticket database, extract key information from the operation ticket and the work ticket based on the ticket database, and verify whether the format of the key information is compliant; Create an operation step verification model, obtain the operation steps recorded in the operation ticket and the work ticket, and verify whether the operation steps are compliant based on the operation step verification model; The steps for creating and verifying the model include: Collect the historical operation step data recorded in the operation ticket and the work ticket, classify and label the historical operation step data according to the preset operation step order, and obtain the operation step dataset. The preset operation steps are in the following order: normal operation steps, change operation steps, emergency operation steps, maintenance operation steps, power outage operation steps, and start-up operation steps. The operation step dataset includes an operation step training set and an operation step verification set; The training set of the operation steps is segmented into words and converted into word sequences. The word sequences are then mapped into word embedding vectors using a word embedding model. The word embedding vectors are input into the LSTM model for training, and the training results are output. The gap between the training results and the operation step validation set is evaluated by the binary cross-entropy loss function, and the model is optimized to obtain the operation step verification model. The step of segmenting the training set of the operation steps into words and converting it into a word sequence, and mapping the word sequence into word embedding vectors through a word embedding model, includes: The training set for the above operation steps was divided into several lexical units using a word segmentation tool; Iterate through each vocabulary unit and assign a unique integer identifier to each vocabulary unit to generate a vocabulary list in which vocabulary units correspond to integer identifiers; The training set of the operation steps is converted into an integer sequence using the vocabulary; Load the pre-trained Word2Vec word vector model and construct an embedding layer based on the Word2Vec word vector model; The integer sequence is input into the embedding layer and word embedding vectors are output. Create a power knowledge graph, and verify the compliance of the execution tasks and associated requirements of the operation ticket and work ticket based on the power knowledge graph.

2. The method for identifying violation information of two tickets according to claim 1, characterized in that, The step of extracting key information from the operation ticket and the work ticket based on the ticket database, and verifying whether the format of the key information is compliant, includes: Input the content information of the operation ticket and work ticket into the ticket database, scan and obtain the key information of the ticket content information; The key information includes equipment information, time information, operation information, personnel information, and safety measures information; Verify whether there are any missing or incorrect characters in the equipment information, time information, operation information, personnel information, and safety measure information; Verify whether there are any errors or omissions in the identification symbols of the equipment information, time information, operation information, personnel information, and safety measures information; Verify whether the semantic order of the equipment information, time information, operation information, personnel information, and safety measure information is correct.

3. The method for identifying violation information of two tickets according to claim 1, characterized in that, The step of inputting the word embedding vector into the LSTM model for training and outputting the training result includes: The training parameters of the LSTM model are set according to the text parameters of the word embedding vector; Set the Sigmoid function as the activation function of the LSTM model to make the output of the LSTM model either compliant or non-compliant; The word embedding vectors are input into the LSTM model for training to obtain the initial operation step verification model; The operation step verification set is input into the initial operation step verification model, and the training results are output.

4. The method for identifying violation information of two tickets according to claim 1, characterized in that, The steps for creating the power knowledge graph include: Obtain the type of task to be executed for the operation ticket and the work ticket; Create a named entity recognition model; Based on the named entity recognition model, obtain the respective power language entities for the equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements in the task being executed; Each type of the task being executed is treated as a node, and an association graph is constructed with the relevant electrical language entities. All the related graphs are combined to generate an electricity knowledge graph.

5. The method for identifying violation information of two tickets according to claim 4, characterized in that, The steps for creating a named entity recognition model include: Obtain the electrical language text of the task being performed; The power language text is classified and labeled according to a preset order to obtain a power language dataset. The preset requirements are in the following order: equipment requirements, time requirements, safety requirements, personnel requirements, and material requirements. The power language dataset includes a power language training set and a power language validation set; Load the pre-trained BERT model, input the power language training set into the BERT model for training, and obtain the initial named entity recognition model; The power language validation set is input into the initial named entity recognition model, and the power language entity classification result is output. The gap between the classification results of the power language entity and the power language validation set is evaluated using a multi-class cross-entropy loss function, and the model is optimized to obtain a named entity recognition model.

6. A two-ticket violation information identification system, characterized in that, The two-ticket violation information identification system is used to implement the two-ticket violation information identification method as described in any one of claims 1-5, wherein the two-ticket violation information identification system includes: The response module is used to respond to a received violation information identification request and determine the operation ticket and work ticket corresponding to the violation information identification request; The first verification module is used to create a ticket database, extract key information of the operation ticket and the work ticket from the ticket database, and verify whether the format of the key information is compliant. The second verification module is used to create an operation step verification model, obtain the operation steps recorded in the operation ticket and the work ticket, and verify whether the operation steps are compliant according to the operation step verification model. The third verification module is used to create a power knowledge graph and verify whether the execution tasks and associated requirements of the operation ticket and the work ticket are compliant based on the power knowledge graph.

7. An electronic device, characterized in that, The system includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the two-ticket violation information identification method as described in any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed, it implements the method for identifying violation information of two tickets as described in any one of claims 1-5.

Images