Power grid hidden danger data extraction method and device, computer equipment and computer readable storage medium

By extracting the feature vectors of power grid data and matching them with the benchmark words in the pre-constructed hidden danger classification set, combining power business and natural disaster correlation data, the problem of low accuracy in power grid data extraction in the existing technology is solved, and more accurate is achieved Identification and prevention of hidden dangers.

CN119066448APending Publication Date: 2024-12-03GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411001666.9
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

The existing grid hidden danger data extraction methods are based on the characteristics of power business as the classification benchmark, resulting in low extraction accuracy, ignoring other related factors such as natural disasters, resulting in insufficient accuracy in identifying and preventing grid hidden dangers.

Method used

By obtaining the preprocessed power grid data, extracting the feature vectors, and traversing them with the benchmark words in the pre-constructed hidden danger classification set, determining the matching target benchmark words, and considering the multi-dimensional data associated with power business and natural disasters. , update the hidden danger classification collection and determine the response strategy.

Benefits of technology

It improves the accuracy of the identification of grid hidden danger data, can more comprehensively analyze grid data, warn and prevent hidden dangers, and reduce accident risks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119066448A_ABST
    Figure CN119066448A_ABST
Patent Text Reader

Abstract

The invention relates to a hidden danger data extraction method and device of a power grid, computer equipment and a computer readable storage medium, and relates to the technical field of big data. The method comprises the following steps: acquiring preprocessed power grid data, and extracting a feature vector of the power grid data; traversing and comparing the feature vector with a plurality of reference words in a pre-constructed hidden danger classification set, and determining a matched target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters; determining hidden danger data in the power grid data according to the target reference word; the hidden danger data is used for updating the pre-constructed hidden danger classification set and determining a coping strategy of the hidden danger data. The method can improve the accuracy of hidden danger data extraction.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of big data technology, and in particular to a method, device, computer equipment, computer-readable storage medium, and computer program product for extracting hidden danger data of a power grid. Background Art

[0002] In the process of troubleshooting hidden dangers in the power grid, it is necessary to extract hidden danger data from a large amount of power grid data, so as to take corresponding fault prevention measures based on the extracted hidden danger data. In the current power grid information system, hidden danger data extraction for the power grid is extracted after classification based on the characteristics of power services that are prone to hidden dangers, focusing on the root data and characterization data of hidden dangers. Therefore, the existing hidden danger data extraction method for the power grid has the problem of low accuracy in hidden danger data extraction because it uses the characteristics of power services as the classification basis for hidden danger data. Summary of the invention

[0003] Based on this, it is necessary to provide a method, device, computer equipment and computer-readable storage medium for extracting hidden danger data of a power grid, which can improve the accuracy of hidden danger data extraction, in order to solve the above technical problems.

[0004] In a first aspect, the present application provides a method for extracting hidden danger data of a power grid, comprising:

[0005] Acquire preprocessed power grid data and extract feature vectors of the power grid data;

[0006] The feature vector is compared with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters;

[0007] According to the target reference word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

[0008] In one embodiment, obtaining preprocessed power grid data and extracting a feature vector of the power grid data includes:

[0009] Performing text preprocessing on the power grid data to obtain preprocessed power grid data;

[0010] Performing word segmentation processing on the preprocessed power grid data to obtain feature information;

[0011] The characteristic information is subjected to vector conversion processing to obtain a characteristic vector of the power grid data.

[0012] In one embodiment, the feature vector is compared with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word, including:

[0013] respectively calculating the similarity between the feature vector and each of the reference words in the pre-constructed hidden danger classification set;

[0014] According to the similarity between the feature vector and each of the reference words, a reference word with the highest similarity is determined as a matching target reference word.

[0015] In one embodiment, the pre-constructed hidden danger classification set is constructed by the following method:

[0016] Obtaining preprocessed sample power grid hidden danger data;

[0017] The preprocessed sample power grid hidden danger data and the text vectors in the preset power word library and natural disaster related word library are respectively subjected to similarity calculation to obtain a similarity calculation result;

[0018] Performing clustering processing on the preprocessed sample power grid hidden danger data according to the similarity calculation result to obtain a hidden danger clustering result;

[0019] According to the hidden danger clustering results, a hidden danger classification set is constructed.

[0020] In one embodiment, before obtaining the pre-processed sample power grid hidden danger data, the method further includes:

[0021] Conduct data mining on fault data in power grid data to identify fault data related to power business risks or natural disasters;

[0022] The fault data related to the hidden dangers of electric power business or natural disasters and the hidden danger data in the power grid data are determined as sample power grid hidden danger data.

[0023] In one embodiment, the preprocessed sample power grid hidden danger data and the text vectors in the preset vocabulary of multiple dimensions are respectively calculated to obtain similarity calculation results, including:

[0024] Converting the preprocessed sample power grid hidden danger data into a hidden danger vector;

[0025] The hidden danger vector and the text vector are calculated using a cosine similarity algorithm to obtain a similarity calculation result of each hidden danger vector;

[0026] When the similarity calculation result satisfies a preset similarity condition, clustering the hidden danger vector and the corresponding text vector to obtain a hidden danger clustering result;

[0027] When the similarity calculation result does not meet the preset condition, the hidden danger vector is regarded as an unclustered vector.

[0028] In one embodiment, constructing a hidden danger classification set according to the hidden danger clustering result includes:

[0029] Extracting high-frequency words of the unclustered vector;

[0030] The high-frequency words are subjected to a hidden danger relevance analysis by using crawler technology. When the high-frequency words meet a preset relevance condition, the non-clustered vectors are merged with the hidden danger clustering results to construct a hidden danger classification set.

[0031] In a second aspect, the present application also provides a device for extracting hidden danger data of a power grid, comprising:

[0032] A feature extraction module, used to obtain preprocessed power grid data and extract feature vectors of the power grid data;

[0033] A target matching module is used to traverse and compare the feature vector with multiple reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters;

[0034] The hidden danger determination module is used to determine the hidden danger data in the power grid data according to the target reference word; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

[0035] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0036] Acquire preprocessed power grid data and extract feature vectors of the power grid data;

[0037] The feature vector is compared with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters;

[0038] According to the target reference word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

[0039] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0040] Acquire preprocessed power grid data and extract feature vectors of the power grid data;

[0041] The feature vector is compared with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters;

[0042] According to the target reference word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

[0043] The above-mentioned method, device, computer equipment and computer-readable storage medium for extracting hidden danger data of the power grid, the method obtains pre-processed power grid data, extracts the feature vector of the power grid data, and traverses and compares the feature vector with multiple benchmark words in a pre-constructed hidden danger classification set to determine a matching target benchmark word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters, and considers hidden danger data associated with multiple dimensions, which can improve the accuracy of identification. Further, according to the target benchmark word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set, and to determine the response strategy for the hidden danger data. It can warn and prevent hidden dangers in the power grid and reduce the risk of accidents. By more comprehensively identifying and analyzing the power grid data, the accuracy of extracting hidden danger data of the power grid is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 It is a flowchart of a method for extracting hidden danger data of a power grid in one embodiment;

[0046] Figure 2 A schematic diagram of a process for constructing a hidden danger classification set in one embodiment;

[0047] Figure 3A schematic flow chart of a method for extracting hidden danger data of a power grid in another embodiment;

[0048] Figure 4 is a structural block diagram of a device for extracting hidden danger data of a power grid in one embodiment;

[0049] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0050] 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.

[0051] As described in the background technology, the hidden danger data extraction method of the power grid in the prior art has the problem of low accuracy in extracting hidden danger data. The inventor has found that the reason for this problem is that the relevant technology usually generates an accident time sequence diagram of the accident based on the accident text data, establishes a hidden danger database, establishes a hidden danger knowledge map based on various hidden danger text data in the hidden danger database, or trains the hidden danger classification model of the power grid through a sample data set, thereby extracting the hidden danger data in the power grid data. However, in the process of constructing the data map or classification model, it is mainly based on the characteristics, performance, root cause and equipment type of the power business as the classification benchmark, focusing on the root cause and characterization data of the hidden danger, ignoring other data related to the hidden danger and the analysis of factors that may cause the hidden danger to escalate, and the classification is weak in pertinence. Especially in the power grid operation environment, there are hidden dangers caused by natural disasters. Based on the above analysis, the hidden danger data extraction method of the power grid in the prior art has the problem of low accuracy in extracting hidden danger data of the power grid due to the weak classification pertinence and narrow adaptability.

[0052] Based on the above reasons, the present invention provides a solution for extracting hidden danger data of a power grid, which can improve the accuracy of extracting hidden danger data of the power grid.

[0053] In one embodiment, Figure 1 As shown, a method for extracting hidden danger data of a power grid is provided. This embodiment takes the method applied to a server as an example for illustration. It can be understood that the method can also be applied to a terminal, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0054] Step 102: Obtain preprocessed power grid data and extract feature vectors of the power grid data.

[0055] Among them, grid data can be various data related to the power system, including power production, consumption, transmission line status, voltage load and other information. These data can help power companies to conduct operation monitoring, load forecasting, fault diagnosis, and optimized scheduling to ensure the safe and stable operation of the power grid. Among them, preprocessing can be the process of cleaning, converting and standardizing text data before using grid data for machine learning or natural language processing tasks. The main goal of preprocessing is to convert raw grid data into a form that can be understood and processed by machines to improve the effect and performance of subsequent tasks.

[0056] The feature vector may be a vector in a preset format obtained by extracting features from preprocessed power grid data and converting the extracted features.

[0057] Optionally, the server obtains preprocessed power grid data, extracts feature information of the preprocessed power grid data, and converts the feature information into a vector in a preset format to obtain a feature vector of the power grid data.

[0058] Step 104 , traverse and compare the feature vector with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word.

[0059] Among them, the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters.

[0060] The traversal may be a process of comparing each benchmark word in the pre-constructed hidden danger classification set one by one.

[0061] The target reference word may be a reference word with the highest similarity to the feature vector in a pre-constructed hidden danger classification set.

[0062] Optionally, the feature vector is compared with multiple benchmark words in the pre-constructed hidden danger classification set one by one, and the benchmark word with the highest similarity to the feature vector is determined as the matching target benchmark word. Specifically, the similarity between the feature vector and the multiple benchmark words can be calculated one by one, so as to determine the benchmark word with the highest similarity.

[0063] Step 106, determining hidden danger data in the power grid data according to the target reference word.

[0064] Among them, the hidden danger data is used to update the pre-built hidden danger classification set and determine the response strategy for the hidden danger data.

[0065] Optionally, after traversal and comparison, it is determined that the data type of the power grid data corresponding to the feature vector of the corresponding target reference word is the hidden danger data that conforms to the pre-constructed hidden danger classification set, thereby extracting the hidden danger data and the type of the hidden danger data from the power grid data. According to the extracted hidden danger data, the pre-constructed hidden danger classification set can be updated using the hidden danger data, and after determining the type of hidden danger data, the advance response strategy for the hidden danger data of this type can be determined. For example, the extracted hidden danger data is the abnormal current data detected by a certain detection station of the power grid and the meteorological data is typhoon weather. Combining multiple hidden danger data can determine the corresponding response strategy, such as the response strategy of temporarily shutting down the lines within the disaster area to protect other lines.

[0066] In the above-mentioned method for extracting hidden danger data of the power grid, the pre-processed power grid data is obtained, and the feature vector of the power grid data is extracted, and the feature vector is traversed and compared with multiple benchmark words in the pre-constructed hidden danger classification set to determine the matching target benchmark word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing the sample power grid hidden danger data associated with power business hidden dangers or natural disasters, and considers the hidden danger data associated with multiple dimensions, which can improve the accuracy of identification. Further, according to the target benchmark word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data. It can warn and prevent hidden dangers in the power grid and reduce the risk of accidents. It realizes a more comprehensive identification and analysis of the power grid data, thereby improving the accuracy of extracting hidden danger data of the power grid.

[0067] In an exemplary embodiment, step S102 of acquiring preprocessed power grid data and extracting a feature vector of the power grid data includes:

[0068] The power grid data is subjected to text preprocessing to obtain preprocessed power grid data; the preprocessed power grid data is subjected to word segmentation processing to obtain feature information; the feature information is subjected to vector conversion processing to obtain a feature vector of the power grid data.

[0069] The text preprocessing may include cleaning text, deduplication of descriptions, and removal of stop words. The word segmentation process may be to use a word segmentation algorithm to segment text data into individual word units.

[0070] Optionally, perform text preprocessing on the grid data to remove irrelevant and duplicate information in the grid data, such as special characters and redundant spaces, and remove common words in the grid data, such as "de" (的), "shi" (是), etc., to obtain the preprocessed grid data. Perform word segmentation on the preprocessed grid data, such as using a word segmentation algorithm to split the preprocessed grid data into individual word units, and perform part-of-speech tagging and word frequency statistics, etc., to extract feature information. Perform vector conversion processing on the feature information to convert it into a vector in the vector space model.

[0071] In this embodiment, by performing preprocessing of data cleaning on the grid data to remove redundant text data, the subsequent analysis efficiency can be improved, and word segmentation is performed on the preprocessed grid data to extract feature information, and the feature information is converted into feature vectors, which are converted into a form that can be understood and processed by machines to improve the effect and efficiency of subsequent tasks.

[0072] In an exemplary embodiment, step S104 traverses and compares the feature vector with multiple reference words in a pre-constructed hidden danger classification set to determine the matching target reference word, including:

[0073] Calculate the similarity between the feature vector and each reference word in the pre-constructed hidden danger classification set respectively; according to the similarity between the feature vector and each reference word, determine the reference word with the highest similarity as the matching target reference word.

[0074] Among them, the similarity can be used to compare the semantic or syntactic relationship between two texts. Common text similarity calculation methods include cosine similarity based on the bag-of-words model, edit distance, Jaccard similarity coefficient, etc. By calculating the similarity, similar texts can be identified in text processing for tasks such as text classification, clustering, and information retrieval.

[0075] Optionally, calculate the similarity between the feature vector and each reference word in the pre-constructed hidden danger classification set respectively, sort the similarities between the feature vector and each reference word in descending order, and determine the reference word with the highest similarity as the matching target reference word. In one case, the similarity between the determined target reference word and the feature vector may be lower than a preset threshold, and it can be determined that the feature vector does not conform to the scope of the pre-constructed hidden danger classification set, then it can be judged that the grid data corresponding to the feature vector is not hidden danger data, otherwise, it is judged that the grid data corresponding to the feature vector is hidden danger data, and the type of the hidden danger data is the same as the type of the target reference word.

[0076] In this embodiment, by sorting the similarity between the feature vector and the benchmark word, it is possible to effectively determine whether the feature vector corresponds to the target benchmark word, and further confirm whether it belongs to the pre-constructed hidden danger classification set, thereby judging whether the corresponding power grid data is hidden danger data. This realizes automatic classification of power grid data into hidden danger data or non-hidden danger data, and determines the type of hidden danger data without manual intervention, thereby improving processing efficiency.

[0077] In an exemplary embodiment, Figure 2 As shown, the pre-constructed hidden danger classification set described in the above embodiment is constructed by the following method, including:

[0078] Step S202, obtaining pre-processed sample power grid hidden danger data.

[0079] The sample grid hidden danger data may be data records of potential safety hazards in the power system, including failure statistics of power equipment, equipment aging, equipment maintenance records, grid operation status, and other information.

[0080] Optionally, the server obtains the preprocessed sample power grid hidden danger data, and the preprocessing method can be the same as the preprocessing method of the power grid data in the above embodiment, which will not be repeated here. However, when the sample power grid hidden danger data is segmented, the special word library of the power industry will also be analyzed to achieve a more accurate segmentation result.

[0081] Step S204 , performing similarity calculations on the preprocessed sample power grid hidden danger data and the text vectors in the preset power word library and natural disaster related word library to obtain similarity calculation results.

[0082] Among them, the power word library can be a special word library for the power industry, covering power business, power plants, transmission lines, transformers, power grids, power equipment, etc. The natural disaster-related word library can contain words related to various natural disasters, such as earthquakes, hurricanes, typhoons, floods, wildfires, etc.

[0083] Optionally, the preprocessed sample power grid hidden danger data is converted into vectors and similarity is calculated with text vectors in a preset power vocabulary and a natural disaster related vocabulary. The similarity algorithm can adopt a Euclidean distance algorithm, a cosine similarity algorithm, a Pearson correlation coefficient algorithm, etc. to obtain a similarity calculation result.

[0084] Step S206 , clustering the preprocessed sample power grid hidden danger data according to the similarity calculation result to obtain hidden danger clustering results.

[0085] Among them, clustering can be an unsupervised learning method used to classify objects in a data set into groups or clusters with similar characteristics.

[0086] Optionally, based on the similarity calculation results, taking any text vector as the cluster center, the sample power grid hidden danger data whose similarity calculation results meet the preset range are clustered to obtain multiple text clusters, as well as the power grid hidden danger data whose similarity calculation results do not meet the preset similarity conditions. A clustering algorithm is used to perform additional clustering analysis on the power grid hidden danger data that does not meet the preset similarity conditions to obtain hidden danger clustering results.

[0087] Step S208: construct a hidden danger classification set according to the hidden danger clustering result.

[0088] Optionally, preset indicators are used to evaluate the hidden danger clustering results, such as the silhouette coefficient, Calinski-Harabasz (variance ratio standard) index and other indicators, and the clustering algorithm or clustering parameters are optimized according to the evaluation results. After obtaining the clustering results, the clustering results need to be manually evaluated to prevent misjudgment. The reasons for misjudgment are generally due to the business cognitive bias of the input person during the input process, which leads to the belief that a certain fault may be caused by a natural disaster, and the input volume reaches a certain value, and the correlation with natural disasters caused by time and space coincidence is high, and the data volume reaches a certain value.

[0089] In this embodiment, by using a clustering algorithm to perform cluster analysis on power grid hidden danger data that meets the preset range and does not meet the preset range, the data can be effectively classified and different hidden danger types can be identified. By using preset indicator evaluation and manual evaluation for secondary analysis, the occurrence of misjudgment can be reduced and the accuracy of identifying and handling power grid hidden dangers can be improved.

[0090] In an exemplary embodiment, before step S202 of obtaining the pre-processed sample power grid hidden danger data, the method further includes:

[0091] Data mining is performed on the fault data in the power grid data to determine the fault data related to power business hidden dangers or natural disasters; the fault data related to power business hidden dangers or natural disasters and the hidden danger data in the power grid data are determined as sample power grid hidden danger data.

[0092] Among them, data mining can be carried out on information such as location, time, cause of failure and manifestation of failure in the fault data to determine whether it is related to natural disasters or other hidden dangers.

[0093] Optionally, the location, time, cause of failure, and manifestation of failure in the relevant information of the fault data in the power grid data are mined to determine whether the fault data is associated with power business hidden dangers or natural disasters, thereby determining the fault data associated with power business hidden dangers or natural disasters, and determining the fault data associated with power business hidden dangers or natural disasters and the hidden danger data in the power grid data as sample power grid hidden danger data. The hidden danger data can be recorded in advance based on the data associated with the hidden dangers generated by the power grid.

[0094] In this embodiment, fault data related to power business hazards or natural disasters and hidden danger data in power grid data are determined as sample power grid hidden danger data, which helps to establish a more complete power grid hidden danger database, thereby constructing a hidden danger data set, further improving the accuracy of power grid hidden danger data extraction.

[0095] In an exemplary embodiment, step S204 performs similarity calculations on the preprocessed sample power grid hidden danger data and the text vectors in the preset vocabulary of multiple dimensions, respectively, to obtain similarity calculation results, including:

[0096] The preprocessed sample power grid hidden danger data is converted into hidden danger vectors; the cosine similarity algorithm is used to calculate the hidden danger vectors and text vectors to obtain the similarity calculation results of each hidden danger vector; when the similarity calculation results meet the preset similarity conditions, the hidden danger vectors and the corresponding text vectors are clustered to obtain the hidden danger clustering results; when the similarity calculation results do not meet the preset conditions, the hidden danger vectors are regarded as unclustered vectors.

[0097] Among them, the cosine similarity algorithm can obtain the similarity calculation result by calculating the cosine value of the angle between the vector of the sample power grid hidden danger data and the text vector.

[0098] Optionally, the preprocessed sample power grid data is converted into a hidden danger vector, and the cosine similarity algorithm is used to calculate the similarity between the hidden danger vector and the text vector, and the similarity calculation result of each hidden danger vector is obtained by calculating the cosine value of the angle between the vector of the sample power grid hidden danger data and the text vector. The similarity can be calculated by the following formula:

[0099] (\text{cos_sim}(A,B)=\frac{\vec{A}\cdot\vec{B}}{|\vec{A}|\times|\vec{B}|})

[0100] Among them, (\vec{A}) and (\vec{B}) are two text vectors, (\cdot) represents the dot product, and (|\vec{A}|) and (|\vec{B}|) are the moduli of vectors A and B respectively.

[0101] Furthermore, when the similarity calculation result meets the preset similarity condition, the hidden danger vector and the corresponding text vector are clustered to obtain the hidden danger clustering result; when the similarity calculation result does not meet the preset condition, the hidden danger vector is regarded as an unclustered vector. The unclustered vector is subjected to secondary analysis to further analyze and obtain hidden danger data.

[0102] In this embodiment, by calculating the cosine value of the angle between the hidden danger vector and the text vector, the similarity between the hidden danger data can be quantified, which helps to accurately identify relevant hidden danger data and improve the accuracy and effect of clustering.

[0103] In an exemplary embodiment, step S208 constructs a hidden danger classification set according to the hidden danger clustering result, including:

[0104] Extract high-frequency words from unclustered vectors; perform hidden danger relevance analysis on high-frequency words through crawler technology. When the high-frequency words meet the preset relevance conditions, merge the unclustered vectors with the hidden danger clustering results to construct a hidden danger classification set.

[0105] Among them, high-frequency words can refer to words or phrases that appear frequently in a given text set. High-frequency words are usually those words that appear repeatedly in the text and have a high frequency. In the fields of text mining, information retrieval and natural language processing, identifying and analyzing high-frequency words can help understand the focus and key information of the text content.

[0106] Crawler technology can be an automated program that crawls web page content on the Internet and stores or indexes it for subsequent analysis. Crawler technology is also called web crawler, web spider or web robot, and is usually implemented by scripts or software written by programmers. Crawler technology can automatically access web pages, extract content on the pages and organize, store or use this information for data analysis.

[0107] Optionally, extract the high-frequency words of the sample power grid hidden danger data corresponding to the unclustered vector, and perform hidden danger association analysis on the high-frequency words through crawler technology. For example, crawl the Internet power industry news and fault data through high-frequency words, and analyze whether there are words related to natural disasters in the crawled Internet data. If there is a correlation, crawl the Internet data, perform secondary analysis on the Internet data, and judge the correlation with natural disasters. If the correlation with natural disasters is high, this part of the power grid hidden danger data is also determined to form faults or risks due to natural disasters. Merge the hidden danger clustering results after evaluation and optimization with the hidden danger data obtained after crawler analysis to construct a hidden danger classification set. In addition, the corresponding response strategies and causes of each type of hidden danger data can be added to the hidden danger classification set, so that when the hidden danger data is extracted later, the response strategies and causes can be quickly obtained and analyzed.

[0108] In this embodiment, by extracting high-frequency words of sample power grid hidden danger data corresponding to unclustered vectors and combining crawler technology to perform hidden danger association analysis on these high-frequency words, it can help discover the potential correlation between hidden danger data and factors such as natural disasters, thereby improving the depth and breadth of hidden danger data analysis and further improving the accuracy of subsequent hidden danger data extraction.

[0109] In an exemplary embodiment, Figure 3 As shown, another method for extracting hidden danger data of a power grid is provided, comprising:

[0110] Step S302: collecting hidden danger data and fault data.

[0111] Exemplarily, hidden danger data and fault data are extracted from power grid data, and the fault data is obtained by analyzing data that may be associated with the hidden dangers through data mining.

[0112] Step S304: perform text preprocessing on the hidden danger data and fault data.

[0113] Exemplarily, the hidden danger data and fault data are preprocessed by text data cleaning, description deduplication, stop word removal and word segmentation to obtain feature text.

[0114] Step S306: extract features from the preprocessed hidden danger data and fault data to obtain feature vectors.

[0115] Exemplarily, the preprocessed text is converted into vectors in a vector space model for subsequent clustering analysis.

[0116] Step S308: performing cluster analysis based on the power industry-specific vocabulary and natural disaster-related vocabulary to obtain clustering results.

[0117] Exemplarily, a cosine similarity algorithm is used for clustering to measure the similarity between the feature vector and vectors in the power industry-specific vocabulary and natural disaster-related vocabulary.

[0118] Step S310: Use preset indicators to evaluate the quality of the clustering results and optimize the clustering algorithm or parameters.

[0119] Step S312, perform expert analysis on the clustering results and classify the hidden danger data again to obtain an optimized clustering result.

[0120] For example, after cluster analysis, we find fault clusters related to natural disasters (natural disaster vocabulary or temporal and spatial correlation). This part of the results needs to be confirmed by experts to prevent misjudgment. There are two reasons for misjudgment: During the process of entering the cause of the fault, the business cognitive bias of the inputter leads to the belief that a certain fault may be caused by a typhoon, and the input amount reaches a certain value. The correlation with natural disasters caused by time and space coincidence is high, and the amount of data reaches a certain value. After cluster analysis, we find the hidden danger descriptions related to natural disaster vocabulary, and obtain the relevant hidden danger cluster analysis results. This part of the results needs to be confirmed by experts to prevent the system from misjudging due to the business cognitive bias of the inputter during the hidden danger entry process and reaching a certain input amount.

[0121] Step S314, performing secondary analysis on the unclustered results to obtain secondary analysis hidden danger data.

[0122] Step S316, merging the optimized clustering results and the hidden danger data of the secondary analysis to form a hidden danger classification set, and extracting the hidden danger data in the power grid according to the hidden danger classification set.

[0123] For example, the hidden danger cluster analysis results obtained above are merged to form a typhoon-related hidden danger classification set, which includes hidden danger classification categories and related hot words. When extracting hidden dangers, the hidden danger description is segmented, and after segmentation, the hidden danger classification set is matched to determine whether it is related to the hidden danger.

[0124] In this embodiment, big data and artificial intelligence algorithms are used to quickly screen out hidden danger data that may cause risks and failures due to natural disasters from a large amount of hidden danger data, thereby improving emergency efficiency, reducing manual workload, and effectively supporting fault analysis and processing during emergencies.

[0125] 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.

[0126] Based on the same inventive concept, the embodiment of the present application also provides a hidden danger data extraction device for a power grid for implementing the hidden danger data extraction method for a power grid 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 the embodiments of one or more hidden danger data extraction devices for power grids provided below can refer to the limitations of the hidden danger data extraction method for power grids above, and will not be repeated here.

[0127] In an exemplary embodiment, Figure 4 As shown, a device 400 for extracting hidden danger data of a power grid is provided, comprising: a feature extraction module 402, a target matching module 404 and a hidden danger determination module 406, wherein:

[0128] The feature extraction module 402 is used to obtain the pre-processed power grid data and extract the feature vector of the power grid data.

[0129] The target matching module 404 is used to traverse and compare the feature vector with multiple reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters.

[0130] The hidden danger determination module 406 is used to determine hidden danger data in the power grid data according to the target reference word; the hidden danger data is used to update the pre-built hidden danger classification set and determine the response strategy for the hidden danger data.

[0131] Furthermore, in one embodiment, the feature extraction module 402 is also used to perform text preprocessing on the power grid data to obtain preprocessed power grid data; perform word segmentation processing on the preprocessed power grid data to obtain feature information; and perform vector conversion processing on the feature information to obtain a feature vector of the power grid data.

[0132] Furthermore, in one embodiment, the target matching module 404 is also used to respectively calculate the similarity between the feature vector and each benchmark word in the pre-constructed hidden danger classification set; based on the similarity between the feature vector and each benchmark word, determine the benchmark word with the highest similarity as the matching target benchmark word.

[0133] Furthermore, in one embodiment, the power grid hidden danger data extraction device 400 also includes a set construction module, which is used to obtain preprocessed sample power grid hidden danger data; perform similarity calculation on the preprocessed sample power grid hidden danger data and the text vectors in the preset power vocabulary and natural disaster related vocabulary to obtain similarity calculation results; perform clustering processing on the preprocessed sample power grid hidden danger data according to the similarity calculation results to obtain hidden danger clustering results; and construct a hidden danger classification set according to the hidden danger clustering results.

[0134] Furthermore, in one embodiment, the set construction module is also used to perform data mining on the fault data in the power grid data to determine the fault data associated with power business hidden dangers or natural disasters; the fault data associated with power business hidden dangers or natural disasters and the hidden danger data in the power grid data are determined as sample power grid hidden danger data.

[0135] Furthermore, in one embodiment, the set construction module is also used to convert the preprocessed sample power grid hidden danger data into a hidden danger vector; use the cosine similarity algorithm to calculate the hidden danger vector and the text vector to obtain the similarity calculation result of each hidden danger vector; when the similarity calculation result meets the preset similarity condition, cluster the hidden danger vector and the corresponding text vector to obtain a hidden danger clustering result; when the similarity calculation result does not meet the preset condition, the hidden danger vector is regarded as an unclustered vector.

[0136] Furthermore, in one embodiment, the set construction module is also used to extract high-frequency words from unclustered vectors; perform hidden danger correlation analysis on the high-frequency words through crawler technology, and when the high-frequency words meet preset correlation conditions, merge the unclustered vectors with the hidden danger clustering results to construct a hidden danger classification set.

[0137] Each module in the hidden danger data extraction device 400 of the power grid can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0138] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 5As shown. 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 power grid data, sample power grid hidden danger data, power vocabulary and natural disaster related vocabulary. 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 extracting hidden danger data of a power grid is implemented.

[0139] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the 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 those shown in the figure, or combine certain components, or have a different arrangement of components.

[0140] In one 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 method embodiments when executing the computer program.

[0141] 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.

[0142] 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.

[0143] 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.

[0144] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method 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 the present application can include at least one of non-volatile memory 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), magnetic 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, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0145] The technical features of the above embodiments may be combined arbitrarily. 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 application.

[0146] 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 extracting hidden danger data of a power grid, characterized in that: The method comprises: Acquire preprocessed power grid data and extract feature vectors of the power grid data; The feature vector is compared with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters; According to the target reference word, the hidden danger data in the power grid data is determined; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

2. The method according to claim 1, characterized in that: The obtaining of preprocessed power grid data and extracting a feature vector of the power grid data includes: Performing text preprocessing on the power grid data to obtain preprocessed power grid data; Performing word segmentation processing on the preprocessed power grid data to obtain feature information; The characteristic information is subjected to vector conversion processing to obtain a characteristic vector of the power grid data.

3. The method according to claim 1, characterized in that The step of traversing and comparing the feature vector with a plurality of reference words in a pre-constructed hidden danger classification set to determine a matching target reference word includes: respectively calculating the similarity between the feature vector and each of the reference words in the pre-constructed hidden danger classification set; According to the similarity between the feature vector and each of the reference words, a reference word with the highest similarity is determined as a matching target reference word.

4. The method according to claim 1, characterized in that: The pre-constructed hidden danger classification set is constructed by the following method: Obtaining preprocessed sample power grid hidden danger data; The preprocessed sample power grid hidden danger data and the text vectors in the preset power word library and natural disaster related word library are respectively subjected to similarity calculation to obtain a similarity calculation result; Performing clustering processing on the preprocessed sample power grid hidden danger data according to the similarity calculation result to obtain a hidden danger clustering result; According to the hidden danger clustering results, a hidden danger classification set is constructed.

5. The method according to claim 4, characterized in that Before obtaining the pre-processed sample power grid hidden danger data, the method further includes: Conduct data mining on fault data in power grid data to identify fault data related to power business risks or natural disasters; The fault data related to the hidden dangers of electric power business or natural disasters and the hidden danger data in the power grid data are determined as sample power grid hidden danger data.

6. The method according to claim 4, characterized in that The preprocessed sample power grid hidden danger data and the text vectors in the preset vocabulary of multiple dimensions are respectively calculated to obtain similarity calculation results, including: Converting the preprocessed sample power grid hidden danger data into a hidden danger vector; The hidden danger vector and the text vector are calculated using a cosine similarity algorithm to obtain a similarity calculation result of each hidden danger vector; When the similarity calculation result satisfies a preset similarity condition, clustering the hidden danger vector and the corresponding text vector to obtain a hidden danger clustering result; When the similarity calculation result does not meet the preset condition, the hidden danger vector is regarded as an unclustered vector.

7. The method according to claim 6, characterized in that The step of constructing a hidden danger classification set according to the hidden danger clustering result includes: Extracting high-frequency words of the unclustered vector; The high-frequency words are subjected to a hidden danger relevance analysis by using crawler technology. When the high-frequency words meet a preset relevance condition, the non-clustered vectors are merged with the hidden danger clustering results to construct a hidden danger classification set.

8. A device for extracting hidden danger data of a power grid, characterized in that: The device comprises: A feature extraction module, used to obtain preprocessed power grid data and extract feature vectors of the power grid data; A target matching module is used to traverse and compare the feature vector with multiple reference words in a pre-constructed hidden danger classification set to determine a matching target reference word; the pre-constructed hidden danger classification set is obtained by clustering and analyzing sample power grid hidden danger data associated with power business hidden dangers or natural disasters; The hidden danger determination module is used to determine the hidden danger data in the power grid data according to the target reference word; the hidden danger data is used to update the pre-constructed hidden danger classification set and determine the response strategy for the hidden danger data.

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

  • Power data transmission method and device, electronic equipment and storage medium

    CN119788734A