Visual early warning analysis method and device for electrical equipment damage, equipment and medium

Abstract

The invention relates to the technical field of information, and provides a visual early warning analysis method and device for electrical equipment damage, equipment and a medium, and the method comprises the steps: obtaining maintenance order data of electrical equipment, and obtaining keywords and semantic information based on the maintenance order data; classifying the keywords and the semantic information to obtain structured maintenance reason data, and generating corresponding labels; integrating the structured maintenance reason data according to time and places according to regional management division of the power system to obtain a space-time dimension maintenance data set; acquiring meteorological data of a time period corresponding to the space-time dimension maintenance data set, and processing the meteorological data to obtain meteorological characteristic data; and obtaining an output result of the damage prediction model, judging the probability of failure of the equipment in each region, performing visual display, and generating early warning information. According to the invention, comprehensive visual analysis and intelligent early warning of the maintenance data can be realized, and the probability of secondary faults of the equipment is reduced.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method, device, equipment and medium for visual early warning analysis of power equipment damage. Background Art

[0002] In modern industrial and infrastructure management, the stable operation of power equipment is essential to ensure the continuity of production and life. However, due to the lack of accurate data analysis methods and technical support, many power companies face the problem of difficulty in detecting potential equipment failures and damages in advance, resulting in sudden equipment downtime and production interruptions. Traditional maintenance methods mostly rely on regular inspections and empirical judgments, which are not only inefficient, but also difficult to accurately predict when equipment will fail. In addition, equipment damage is sometimes related to specific time periods or weather conditions. The existing management system may not fully consider the impact of these external factors, making it difficult to fully assess the risk status of the equipment. Summary of the invention

[0003] The present invention provides a visual early warning analysis method for power equipment damage, which mainly includes: Acquire power equipment maintenance order data, process text content in the maintenance order data, and obtain keywords and semantic information; Constructing a maintenance reason classification system, and classifying keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data and generate corresponding labels; According to the regional management division of the power system, the structured maintenance reason data is integrated according to time and place to obtain a maintenance data set in time and space dimensions; Acquire meteorological data of a time period corresponding to the spatiotemporal maintenance data set, and perform standardization processing on the meteorological data to obtain meteorological characteristic data; Associating the spatiotemporal maintenance data set with the meteorological characteristic data to construct a feature vector; Using machine learning algorithms, a damage prediction model is established based on the constructed feature vectors; Performing failure probability prediction based on the damage prediction model and outputting the prediction result; Obtaining the output results of the damage prediction model, determining the probability of equipment failure in each area and displaying it visually; According to the preset fault probability threshold, the warning level of each area is determined and warning information is generated.

[0004] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: The present invention discloses a method for predicting damage of power equipment maintenance based on maintenance order data and meteorological data. Aiming at the problem that the lack of unified and structured data analysis and early warning after the maintenance of power equipment leads to an increased probability of secondary failure of the equipment, the present invention combines the text content in the maintenance order and the meteorological characteristic data to construct a damage prediction model, and realizes the spatiotemporal dimension analysis and intelligent early warning of the maintenance data of power equipment. By constructing a maintenance reason classification system, the unstructured maintenance order data is converted into structured maintenance reason data, and the feature vector is constructed by associating the meteorological characteristic data, and a prediction model is established using a machine learning algorithm. By visually displaying the prediction results and generating early warning information, comprehensive visual analysis and intelligent early warning of the maintenance data are realized, thereby reducing the probability of secondary failure of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 The present invention is a flowchart of a method for visualizing early warning analysis of power equipment damage.

[0006] Figure 2 The present invention is a schematic structural diagram of a visual early warning analysis device for power equipment damage.

[0007] Figure 3 The figure is a schematic diagram of the structure of a computer device of the present invention. DETAILED DESCRIPTION

[0008] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with the accompanying drawings and embodiments. The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant inventions, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0009] like Figure 1 In this embodiment, a visual early warning analysis method for power equipment damage may specifically include: Step S101, obtaining power equipment maintenance order data, processing text content in the maintenance order data, and obtaining keywords and semantic information.

[0010] Firstly, the data of power equipment maintenance orders are collected and preprocessed, and the maintenance order text is segmented and POS tagged using natural language processing technology. For example, the jieba segmentation tool is used to segment the text, and NLTK and other tools are used for POS tagging. Then, the TF-IDF algorithm is used to extract keywords from the preprocessed maintenance order text. By calculating the TF-IDF value of each word, the top 20 words with high TF-IDF values ​​are selected as keywords. These keywords are of great significance to maintenance work. Then, the Word2Vec word embedding model is used to perform semantic analysis on the maintenance order text. The word vector model is constructed by training the maintenance order corpus, and the cosine similarity between words is calculated to mine the implicit information and potential associations in the maintenance work. For example, the similarity between "fault" and "circuit breaker" reaches 85, indicating that they have a strong association in maintenance work. According to the extracted keywords and semantic information, the K-means clustering algorithm is used to classify and cluster the maintenance order data. The number of clusters is set to 5 and the number of iterations is set to 100. Different types of maintenance problems and common failure modes, such as "cable failure" and "transformer failure", are identified to provide a basis for formulating targeted maintenance strategies. Combined with the operating status and historical maintenance records of power equipment, the Apriori association rule mining algorithm is used to set the minimum support to 0.5 and the minimum confidence to 0.8 to find the association pattern between equipment failure and maintenance work. For example, the confidence of the association rule between "transformer overheating" and "insulation aging" reaches 0.9, predicting the potential failure risk of the equipment. Based on the results of keyword and semantic analysis, the Neo4j graph database is used to construct a knowledge graph for power equipment maintenance. Keywords are used as nodes and semantic relationships are used as edges to form a structured maintenance knowledge base, providing intelligent fault diagnosis and maintenance plan recommendations for maintenance personnel. Finally, the results of keyword extraction and semantic analysis are applied to the optimization and management of maintenance work. Through in-depth mining and analysis of maintenance order data, the maintenance process is continuously optimized to improve maintenance efficiency and equipment reliability. For example, preventive maintenance strategies are formulated for equipment with high frequency failures to reduce unplanned equipment downtime and improve the comprehensive utilization rate of equipment.

[0011] Step S102: construct a maintenance reason classification system, and classify keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data and generate corresponding tags.

[0012] Collect historical operating status data related to different equipment names in the power system, and form an initial data set after integrating equipment model and voltage level data information. Extract fault location information for each data in the initial data set, and obtain the fault level corresponding to the fault cause information according to the pre-established fault level and fault cause correspondence table. Obtain the equipment name operating status data and equipment model information corresponding to the fault phenomenon from the inspection record database. When there are multiple data records for this equipment name, the equipment model is also fused to obtain a specific data object with an equipment model label and operating status information. Traverse the specific data objects, perform frequency analysis based on the fault level in each object and combine it with the service life, and determine the fault level with the highest frequency and exceeding the preset frequency threshold as the target fault. Use the random forest algorithm to build a fault cause analysis model, collect the target fault related data in the historical database, including the fault cause information, service life and voltage level of the target fault, input all the collected target fault data into the fault cause analysis model, and obtain the output result after model training. Construct a classifier based on support vector machine. The training input variables of the classifier come from the results of model training. When the classifier has at least two different operating status data outputs, the target fault data distribution corresponding to each operating status data is determined by constructing a clustering algorithm. The one with the largest distribution density is determined as the association table between the operating status and the cause of the fault. Obtain the voltage level data of the specific equipment name corresponding to a specific time point and the historical inspection records of this specific equipment, analyze the data and this record, determine the predicted fault cause of the equipment according to the association table between the operating status and the cause of the fault, and generate and provide relevant treatment measures for the predicted fault cause.

[0013] Exemplarily, historical operating status data of various equipment names in the power system are collected, for example, 1,000 transformer data, including "overload", "normal", "slightly hot" and other states, 800 circuit breaker data, including "normal opening position", "abnormal closing position" and other states. At the same time, the equipment model, such as the transformer model "S-11-1000 / 10" and the voltage level "10kV" and other information are integrated to form an initial data set. For each data in the data set, the fault location information is extracted, such as "low oil level of the transformer body", "fault of the circuit breaker opening coil", etc., combined with the corresponding table of fault level and fault cause, such as "low oil level" corresponding to the cause "poor sealing" belongs to the "serious" fault level. The equipment operating status and model corresponding to the fault phenomenon are extracted from the inspection record database. For example, if the fault phenomenon is "low oil level of the transformer", 10 related records are extracted, including the equipment model, and the status information may be "normal", "overload operation", and "high oil temperature". For multiple records with the same device name, the model information is fused, such as "transformer S-11-1000 / 10, overload, high oil temperature" as a specific data object. After the analysis, it is determined that the "serious" fault level has the highest frequency and exceeds the threshold of 60%, and the proportion of serious faults in the data with a service life of more than 5 years reaches 75%, so the "serious" fault level is defined as the target fault. Construct a random forest model, collect 500 data related to the "serious" fault level in the historical data set, including causes such as "poor sealing" and "insulation aging", as well as information such as age "6 years" and voltage level "35kV", and input them into the model training. For example, the number of decision trees of the random forest model is set to 100, and the maximum depth is set to 10. After obtaining the model output, further construct a classifier based on the support vector machine, and use the model training results, such as the predicted fault causes, as the input of the support vector machine model, such as "poor sealing" corresponding to "operating state 1" and "insulation aging" corresponding to "operating state 2", for training. The classifier training uses the radial basis kernel function, the penalty parameter is set to 1, and the kernel function parameter is set to 1. When the classifier has at least two outputs, a judgment based on the K-means clustering algorithm is constructed. For example, the number of clusters is set to 3, and the operating status data and the corresponding target fault data are input into the clustering model. The clustering result may cluster the "overload" status data together, and find that the corresponding fault cause is "poor sealing" The data accounts for 90%, which exceeds the preset threshold of 80%. According to the clustering results, the association table between the operating status and the cause of the fault is finally determined. For example, overload corresponds to poor sealing, and abnormal heating corresponds to insulation aging. For specific equipment such as "transformer S-11-1000 / 10", the voltage level "10kV" and historical inspection records at the current time point are obtained. For example, the record shows that the state was "overload" during the last inspection. Through the association table, the cause of the fault is predicted to be "poor sealing", and then the treatment measures are generated: "It is recommended to inspect and replace the seals."Get the text description data of the maintenance work order, perform word segmentation, remove stop words and other preprocessing on the text to obtain a keyword list. Use word embedding models such as Word2Vec to convert keywords into semantic vectors and obtain the semantic representation of keywords. According to the semantic vectors of keywords, calculate the similarity matrix between keywords. Based on the similarity matrix, use clustering algorithms such as K-Means to cluster keywords and obtain several maintenance reason categories. For each maintenance reason category, count the word frequency of keywords in the category, and select the keyword with the highest word frequency as the label of the category. Map the maintenance reason category and the corresponding label to obtain structured maintenance reason data. For a new maintenance work order, extract its keywords, determine its category based on the similarity between the keywords and the existing maintenance reason categories, and label it accordingly. First, use the jieba library to perform word segmentation on the text description data of the maintenance work order, and use the stop word list to remove meaningless words, and finally obtain a keyword list. Then, use the Word2Vec model to train the keywords, set the word vector dimension to 200, the window size to 5, and the number of iterations to 10, and finally obtain the semantic vector corresponding to each keyword. Next, the similarity matrix between keywords is calculated based on cosine similarity, and the K-Means clustering algorithm is used to cluster the keywords into 10 maintenance reason categories. The cluster center is continuously optimized during the clustering process until the clustering results converge. For each maintenance reason category, the word frequency of the keywords in the category is counted, and the three keywords with the highest word frequency are selected as the labels of the category to form a mapping relationship between the maintenance reason category and the label. Finally, for the new maintenance work order, its keywords are extracted, and the cosine similarity between the keywords and the center points of the existing maintenance reason categories is calculated. The category with the highest similarity is selected as the maintenance reason of the work order, and the corresponding label is added to realize the automatic classification of the maintenance reason. Through this method, the unstructured maintenance work order text can be effectively converted into structured maintenance reason data, providing support for subsequent data analysis and decision-making.

[0014] Step S103, according to the regional management division of the power system, the structured maintenance reason data is integrated according to time and location to obtain a maintenance data set of spatiotemporal dimension.

[0015] Obtain regional management division information of the power system and determine the geographical scope and division granularity of data integration. For structured maintenance reason data, extract the time attribute and location attribute to form the time dimension and location dimension. According to the time dimension and location dimension, group and aggregate the maintenance reason data to obtain a maintenance data subset indexed by time and location. Use data fusion technology to merge maintenance data subsets at different times and locations to generate a maintenance data set covering the time and space dimensions of the entire power system.

[0016] Furthermore, the generated spatiotemporal maintenance dataset can be cleaned and preprocessed to remove abnormal data, fill in missing values, and ensure data quality. Through data visualization technology, the spatiotemporal maintenance dataset can be displayed in multiple dimensions to mine the spatiotemporal characteristics and laws contained in the data. The spatiotemporal maintenance dataset is stored in the database, and an index of time and location is established to facilitate subsequent data retrieval and analysis.

[0017] For example, first, the regional management division information of the power system is obtained from the power company, and the geographical scope of data integration is determined to be a provincial power system, and the division granularity is the prefecture-level. For the structured maintenance reason data, regular expressions are used to extract the time attributes and location attributes therein to form the time dimension (accurate to the day) and location dimension (accurate to the prefecture-level city). According to the time dimension and location dimension, the MapReduce algorithm is used to group and aggregate the maintenance reason data to obtain a maintenance data subset indexed by time and location. Using data fusion technology, the maintenance data subsets at different times and locations are merged using Apache Spark's DataFrameAPI to generate a spatiotemporal maintenance data set covering the entire power system. The generated spatiotemporal maintenance data set is cleaned and preprocessed, and the data cleaning function in Python's Pandas library is used to remove abnormal data, such as removing data with incorrect time format and data with missing location attributes, and the KNN algorithm is used to fill in missing values ​​to ensure data quality. Through data visualization technology, Tableau is used to display the spatiotemporal maintenance data set in multiple dimensions, such as displaying the changes in the number of maintenance times in various cities by time trend, and displaying the maintenance heat map of various cities by geographical location, etc., to explore the spatiotemporal characteristics and laws contained in the data. Finally, the spatiotemporal maintenance data set is stored in the PostgreSQL database, and the B+ tree index is used to establish the index of time and location to facilitate subsequent data retrieval and analysis. Through the above steps, the scattered maintenance reason data can be integrated into a unified spatiotemporal maintenance data set, laying a data foundation for subsequent maintenance reason analysis, fault prediction and other applications.

[0018] Step S104, obtaining meteorological data of a time period corresponding to the spatiotemporal maintenance data set, and performing standardization processing on the meteorological data to obtain meteorological characteristic data.

[0019] According to the start and end time of the maintenance data set in the spatiotemporal dimension, the meteorological data records of the same time period are retrieved to determine the integrity of the meteorological data records. If the missing meteorological data in the time period exceeds the preset threshold of 10%, the maintenance data at the corresponding time is deleted. The original meteorological data is obtained, including weather type, wind level, precipitation, temperature value, humidity value, air pressure value, and visibility multi-dimensional data, and the original meteorological data is standardized to obtain normalized meteorological data. According to the meteorological data recorded every hour, the maximum temperature and the minimum temperature in the meteorological data are extracted to obtain the daily temperature difference data. A decision tree model is established through daily temperature difference data and weather type, precipitation, and visibility data to obtain the time distribution characteristics of the maintenance data corresponding to each region, that is, meteorological characteristic data; at this time, according to the time distribution characteristics of the maintenance data corresponding to each region, the wind level and humidity value can be used to establish a prediction model using the random forest algorithm to train a model with a prediction accuracy higher than 85%. After obtaining the trained model, the wind level and humidity values ​​in the real-time meteorological data are input to determine the possible maintenance data of each region in the next three days at the corresponding time. The possible maintenance data for each region in the next three days is obtained, and tasks are assigned according to the personnel arranged in advance. If the number of repairs exceeds 50, a task leader will be added.

[0020] For example, assume that there are 1,000 maintenance data sets in a certain area from January 1, 2023 to December 31, 2023, and each data is set to include the maintenance start and end time, such as 9:00 to 12:00 on January 5, 2023. Using these start and end time information, the meteorological records of the corresponding time period are retrieved from the meteorological database. Take each hour as a minimum time unit, such as the maintenance start time of 9:00 and the end time of 12:00, which lasts for a total of 3 hours. If there is a meteorological record once an hour in a certain time period, then there will be 3 records in 3 hours. It can be defined that there are N time units in total. If there are n records in total, if n is less than 9N, it means that there are missing records. If the meteorological data records in a certain time period are less than 90% of the number of records that should be there, for example, a 4-hour time period should have 4 meteorological records, but there are only 3 in reality, then the maintenance data corresponding to the time period will be deleted from the data set. Assuming that 900 maintenance data are left after screening, the corresponding number of meteorological data records will be obtained. Then, for the obtained meteorological data, such as weather type is divided into 1-sunny, 2-cloudy, 3-overcast, 4-rainy; wind force level is quantified as an integer from 1 to 10; precipitation is converted to millimeter units; temperature range is set to -20 to 40 degrees Celsius; humidity range is 0% to 100%; air pressure range is 800 to 1100 hPa; visibility is 0 to 30 kilometers, and a linear normalization formula is used. For example, for the temperature value, assuming that the original data is 30 degrees Celsius, the standardized value is (30-(-20)) / (40-(-20))=0.83. Similarly, the standardized values ​​of the remaining data can be obtained. Then extract the daily maximum and minimum temperature values ​​by date. For example, if the maximum temperature on a certain day is 25 degrees Celsius and the minimum temperature is 15 degrees Celsius, the daily temperature difference is 10 degrees Celsius. Based on the daily temperature difference data, combined with weather type, precipitation, and visibility data, a decision tree model is established. For example, if the temperature difference is greater than 10 degrees Celsius and the weather type is sunny, the precipitation is less than 1 mm, and the visibility is higher than 20 kilometers, the time distribution of the maintenance data for that day is marked as 0. For another example, if the temperature difference is less than 5 degrees Celsius and the weather type is rainy, the precipitation is greater than 10 mm, and the visibility is less than 5 kilometers, the time distribution of the maintenance data for that day is marked as 1. The 900 processed data can be used to train a classification model based on the Gini index calculation node to calculate the regional maintenance time distribution characteristics.

[0021] Use another 100 test data for evaluation. If the regional maintenance time distribution characteristic value obtained by the decision tree model in region A is 0, then continue to use the wind level and humidity values ​​of these 800 data to input into the random forest model containing 100 trees. After internal voting of the random forest algorithm model, the final output predicts that the number of maintenance in the region is 30. For the trained model, if the prediction accuracy of the model after testing 100 data is higher than 85%, the wind and humidity values ​​obtained in the next time are standardized, and the real-time wind level and humidity values ​​are input into the model in the same way to determine the possible number of maintenance in each region in the next three days at the corresponding time. Assuming that it is predicted that there will be 60 maintenance in a certain area in the next three days, which exceeds the preset threshold of 50, the electronic workflow system is automatically adjusted to add a maintenance task leader for the area, and the tasks are intelligently allocated according to the current load of the personnel.

[0022] Step S105, associating the spatiotemporal maintenance data set with the meteorological characteristic data to construct a characteristic vector.

[0023] According to the fault time and fault location, obtain the meteorological data information of the fault time to obtain the original meteorological data set. According to the original meteorological data set, calculate the value of each data and compare it with the preset first threshold. If it is greater than the first threshold, extract the data to obtain the filtered meteorological data set. According to the three data dimensions of meteorological temperature, meteorological humidity and meteorological wind force in the filtered meteorological data set, establish a spatial rectangular coordinate system, obtain the three-dimensional vector value composed of the three data, and obtain a data vector set. According to the data vector set, calculate the Euclidean distance between each vector and the vector. If the distance is greater than the second threshold, the vector is a noise point, delete the vector, and obtain the denoised data vector set.

[0024] Furthermore, the maintenance record database corresponding to each device can be queried according to the equipment type. If the equipment has no maintenance record, it will be marked and a list of unmaintained equipment will be obtained, otherwise the next step will be performed. According to the denoised data vector set and maintenance records, the occurrence time of each maintenance record and the acquisition time of all meteorological data are calculated, and the two time differences are calculated. If the absolute value of the time difference is less than 30 minutes, the data is associated with the maintenance record and used as a positive sample to obtain a maintenance data set containing positive samples. According to the maintenance data set containing positive samples, the attributes of the faulty equipment are extracted and associated with a list of unmaintained equipment. The equipment in the list with the same attributes as the repaired equipment is extracted and a negative sample set is constructed. The positive and negative samples are input into the isolation forest algorithm and the local anomaly factor algorithm to train the model. When the error is less than the preset value, the classification model is obtained.

[0025] For example, suppose a device fails between 9:00 and 12:00 on January 5, 2023. The geographical location information (such as longitude and latitude or prefecture-level division) within the time period is used to retrieve the meteorological data records of the same time period from the meteorological database. For example, if the device is located in Haidian District, Beijing, the weather type, wind force level, precipitation and other meteorological data of the area during the above time period are obtained to form an original meteorological data set. The temperature threshold is set to -20 to 40 degrees Celsius, the humidity threshold is set to 0% to 100%, and the air pressure threshold is set to 800 to 1100 hPa. Data points that exceed these thresholds are marked and excluded to ensure that the meteorological data used is within a reasonable range. For example, if the temperature record at a certain moment is 45 degrees Celsius, the data point will be excluded and not used for subsequent analysis. Next, a three-dimensional rectangular coordinate system is constructed based on the temperature, humidity and wind force in the filtered meteorological data as the three axes. For example, if the temperature at a certain moment is 25 degrees Celsius, the humidity is 60%, and the wind force is level 3, then the meteorological data at that moment can be represented as a three-dimensional vector (25,60,3). In this way, all the filtered meteorological data are converted into three-dimensional vectors to form a data vector set. Set a Euclidean distance threshold, such as 10 units. For each pair of three-dimensional vectors, calculate the Euclidean distance between them. If the distance between two vectors exceeds 10 units, one of the vectors is considered to be an outlier (noise point) and is deleted from the data set. After this step, the denoised feature vector set is finally obtained, ensuring the quality and reliability of the data. Finally, the denoised feature vector is fused with other attributes in the spatiotemporal maintenance data set (such as equipment type, fault location, fault cause, etc.) to form a complete feature vector. For example, a fault event occurs on a specific type of transformer, the fault location is the high voltage side, and the fault cause is overheating. These attributes are combined with the denoised meteorological feature vector to form a comprehensive feature vector, such as (transformer, high voltage side, overheating, 25, 60, 3), where the first three elements represent the equipment type, fault location and fault cause, and the last three elements represent the meteorological feature data, thereby providing high-quality input data for subsequent machine learning models.

[0026] Next, query the equipment maintenance record database according to the equipment type. For example, if the equipment type is "equipment A", it is found that there are 1,000 maintenance records for this type of equipment, and there are 200 equipment of the same type without maintenance records. Then mark these 200 equipment and record them in the list of unmaintained equipment. Then calculate the difference between the occurrence time of the maintenance record and the acquisition time of the meteorological data. For example, the occurrence time of a maintenance record is 15:45 on November 11, 2023, and the acquisition time of the meteorological data is 15:30 on November 11, 2023. The absolute value of the time difference between the two is 15 minutes, which is less than 30 minutes. At this time, the meteorological data such as temperature 25 degrees Celsius, humidity 60%, and wind force 3 is associated with the maintenance record and marked as a positive sample. Traverse all 1,000 maintenance records and 92 meteorological data, associate all maintenance records that meet the time difference requirements with the meteorological data, and mark them as positive samples. Assuming that there are 150 that meet the conditions, a maintenance data set containing 150 positive samples is obtained. Furthermore, the attributes of the faulty devices in the 150 maintenance records, such as manufacturer, model, etc., are extracted and matched with the list of unrepaired devices. From the 200 devices in the list of unrepaired devices, devices with the same attributes as the 150 faulty devices are extracted. For example, if there are 50 devices with the same manufacturer and model attributes as the faulty devices, these 50 devices are added to the negative sample set. Finally, the 150 positive samples and 50 negative samples are used as training data and input into the isolation forest algorithm and the local anomaly factor algorithm for training. The number of trees in the isolation forest algorithm is set to 100, and the number of neighbors in the local anomaly factor algorithm is set to 20. The parameters are continuously adjusted and trained until the model error is less than 0.5, and the final classification model is obtained.

[0027] Step S106, using a machine learning algorithm to establish a damage prediction model based on the constructed feature vector.

[0028] Obtain the historical operation data and fault records of the equipment in each region, pre-process the historical operation data and fault records of the equipment in each region, and extract the features related to the equipment failure. Based on the extracted features, the random forest algorithm is used to establish the equipment damage prediction model. Through training, the model can predict the probability of failure in the future based on the equipment operation data. Each region is divided into several sub-regions, and the equipment in each sub-region is predicted to have a fault probability distribution in the future. The fault probability distribution of each sub-region is analyzed to identify the high-incidence areas of faults and determine the areas that need to be monitored and maintained. According to the importance of the equipment and the impact of the fault on the business, the equipment whose failure probability exceeds the preset probability threshold is warned, and the equipment maintenance plan is generated. During the actual operation of the equipment, the equipment operation data is continuously collected, the fault prediction model is regularly updated, and the equipment maintenance plan is dynamically adjusted. According to the fault prediction results and the actual fault situation, the feature selection and algorithm parameters of the damage prediction model are continuously optimized to improve the accuracy of fault prediction.

[0029] For example, suppose that a power company has power grid equipment covering multiple regions, including substations, transmission lines, etc. First, the historical operation data and fault records of the equipment in each region are obtained from the database. These data contain information such as equipment type, fault location, fault cause, and maintenance time. Next, the data are preprocessed, such as removing duplicate data, filling missing values, and extracting features related to equipment failures, such as equipment service life, historical fault frequency, etc. Then, the above extracted features are combined with the previously constructed feature vector to form a complete input data set. The equipment failure prediction model is trained using the random forest algorithm. The model can predict the probability of equipment failure in the future (such as a week or a month) based on the input feature vector. In order to improve the prediction accuracy, the entire area can be divided into several sub-areas, such as by city or county, and the equipment in each sub-area is predicted separately to obtain the probability distribution of failure of the equipment in each sub-area in the future. By analyzing the failure probability distribution of each sub-area, high-incidence areas of failure can be identified. For example, if the probability of failure in a certain section of a city is significantly higher than that in other areas, it is determined as an area that needs to be monitored and maintained. For those devices whose failure probability exceeds a preset threshold (such as 5%), the system will automatically generate early warning information and make corresponding maintenance suggestions, such as arranging emergency inspections or preventive maintenance. In addition, the system will continue to collect equipment operation data and regularly update the fault prediction model to adapt to changes in equipment status. Finally, the feature selection and algorithm parameters of the model are continuously optimized according to the actual fault situation to improve the accuracy and reliability of fault prediction. For example, if it is found that certain types of equipment are more likely to fail in hot weather, the weight of temperature-related features can be increased in the model to further improve the prediction effect.

[0030] Step S107: perform failure probability prediction based on the damage prediction model and output the prediction result.

[0031] Based on the damage prediction model, the failure probability prediction is performed and the prediction results are output. Specifically, the latest operating data of the equipment in each region is obtained from the real-time monitoring system or the historical database, including the current meteorological conditions (such as temperature, humidity, wind force level) and the operating parameters of the equipment (such as load conditions, operating time, etc.). Then, these latest acquired data are combined with the previously constructed feature vector to form a new input feature vector. The trained damage prediction model is used to predict the failure probability of the newly input feature vector. The model calculates and outputs the probability value of the failure of the equipment in each region in the future according to the input feature vector, and stores these prediction results in a structured format. Finally, the output result data is parsed to extract the failure probability value of the equipment in each region for subsequent analysis and decision-making. For example, the probability of a transformer in a substation failing in the next week is 8%, while the probability of another circuit breaker in the same substation failing is 2%. These prediction results not only help management understand the health status of the equipment in the future, but also provide a scientific basis for formulating a reasonable maintenance plan.

[0032] Assume that a power company has established the damage prediction model in the previous steps, and now needs to use the model to predict the failure risk of equipment in each region in the next week. First, obtain the latest operating data of equipment in each region from the real-time monitoring system, such as the current meteorological conditions such as temperature, humidity, wind force level, and operating parameters such as equipment load and operating time. Next, combine these latest acquired data with the previously constructed feature vector to form a new input feature vector. For example, for a transformer in a substation, its feature vector may contain information such as equipment type (transformer), fault location (high voltage side), fault cause (overheating), and current temperature (25 degrees Celsius), humidity (60%), wind force (level 3), etc. Then, input these feature vectors into the trained damage prediction model to perform fault probability prediction. The model will calculate and output the probability value of equipment failure in each region in the next week based on the input feature vector. For example, the probability of a transformer in a substation failing in the next week is 8%, while the probability of another circuit breaker in the same substation failing is 2%. These prediction results will be stored in the database in a structured table form for subsequent query and analysis. Finally, the output result data is parsed to extract the failure probability value of the equipment in each area, generate a report listing the failure probability of each device in each area, and mark high-risk areas according to the preset threshold (such as 5%). By continuously updating the equipment operation data and regularly retraining the model, the accuracy and reliability of the prediction can be continuously improved to ensure the safe and stable operation of the power system.

[0033] Step S108, obtaining the output result of the damage prediction model, determining the probability of equipment failure in each area and performing visual display.

[0034] The output result data generated by the damage prediction model is obtained, and the data contains the failure probability information of the equipment in each area. The output result data is parsed to extract the failure probability value of the equipment in each area. According to the preset failure probability threshold, the possibility of equipment failure in each area is judged, and the area above the failure probability threshold is marked as a high-risk area. Using data visualization technology, the failure probability values ​​of the equipment in each area are mapped to different colors or patterns to generate an intuitive visualization chart. In the visualization chart, high-risk areas are marked with different colors or patterns, which facilitates maintenance personnel to quickly locate the problem area. The visualization chart is superimposed on the regional map so that maintenance personnel can intuitively understand the relationship between the equipment failure probability and the geographical location. According to the visualization results, the high-risk areas that need to be focused on and repaired are determined, and maintenance resources are reasonably allocated to improve maintenance efficiency and accuracy.

[0035] Exemplarily, the output result data generated by the damage prediction model includes the failure probability information of the equipment in each area. The data is parsed by Python's pandas library to extract the failure probability value of the equipment in each area. According to the preset failure probability threshold of 7, the possibility of equipment failure in each area is judged, and the area above the threshold is marked as a high-risk area. Using data visualization libraries such as Matplotlib and Seaborn, the failure probability values ​​of the equipment in each area are mapped to different colors or patterns. For example, the area with a failure probability above 7 is represented by red, the area of ​​5-7 is represented by orange, the area of ​​3-5 is represented by yellow, and the area below 3 is represented by green, generating an intuitive visualization chart. In the visualization chart, high-risk areas are marked with different colors or patterns, which is convenient for maintenance personnel to quickly locate the problem area. The visualization chart is superimposed on the regional map, and the failure probability data is combined with the geographic location information using Python's Basemap library to generate a regional equipment failure probability distribution map, so that maintenance personnel can intuitively understand the relationship between the equipment failure probability and the geographic location. Based on the visualization results, we can identify high-risk areas that need to be focused on and repaired. By analyzing the characteristics of equipment type, usage time, environmental factors, etc. in high-risk areas, we can use machine learning algorithms such as decision trees and random forests to establish a maintenance priority model, prioritize equipment in high-risk areas, and rationally allocate maintenance resources to improve maintenance efficiency and accuracy. By monitoring equipment operation data in real time and dynamically updating the failure probability prediction results, we can achieve continuous evaluation of equipment health status and risk warning, providing data support for preventive maintenance.

[0036] Step S109, determining the warning level of each area according to the preset fault probability threshold, and generating warning information.

[0037] The warning information includes warning content, relevant maintenance methods and precautions. Specifically, first, according to the probability value of equipment failure in each area, the warning level of each area is determined by comparing with the preset warning level classification standards (such as low risk, medium risk, and high risk). For example, three warning levels are set: low risk (failure probability is less than 3%), medium risk (failure probability is between 3%-5%), and high risk (failure probability is higher than 5%). Then, for different levels of warnings, detailed warning information is automatically generated, including specific warning content (such as which type of equipment may fail in the future), recommended relevant maintenance methods (such as emergency inspections, preventive maintenance, etc.) and precautions (such as operating safety tips, environmental conditions requirements, etc.). These warning information can be conveyed to relevant departments and personnel through various channels to ensure that necessary measures are taken in time to prevent failures or mitigate their impact. For example, assuming that the probability of a transformer in a substation failing in the next week is 8%, the substation is marked as a high-risk area. The system will automatically generate a detailed warning message, the content of which is as follows: The warning content is that the probability of failure of the transformer in substation A in the next week is 8%, which belongs to the high-risk area; the relevant maintenance methods include immediately arranging technicians to conduct an emergency inspection of the transformer, focusing on checking whether there is overheating on the high-voltage side. If abnormalities are found, preventive maintenance measures will be taken immediately, such as replacing aging parts or upgrading the cooling system, and arranging regular monitoring to record temperature changes every 4 hours; the precautions remind technicians to wear protective equipment when conducting inspections, ensure personal safety, pay attention to weather conditions, and give priority to handling this issue if there is a high temperature warning in the next few days. During the maintenance, keep close communication with the dispatch center to ensure that the normal operation of the power grid is not affected. The warning information will be sent to the relevant person in charge and the technical team through the internal management system, and a paper file will also be generated for archiving. For other areas with medium and low risks, the system will also generate corresponding warning information, but the content and urgency are different. For example, medium-risk areas may recommend increasing the inspection frequency, while low-risk areas only need to maintain regular maintenance. In addition, all warning information will be updated regularly to reflect the latest equipment status and changes in meteorological conditions to ensure the real-time and effectiveness of the warning system.

[0038] Reference Figure 2 The present application also proposes a visual early warning analysis device for power equipment damage, the device comprising: Data collection and processing module 100: used to obtain power equipment maintenance order data, process the text content in the maintenance order data, and obtain keywords and semantic information; Classification system construction module 200: used to construct a maintenance reason classification system, and classify keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data and generate corresponding labels; The spatiotemporal data integration module 300 is used to integrate the structured maintenance reason data by time and location according to the regional management division of the power system to obtain a spatiotemporal maintenance data set; Meteorological data processing module 400: used to obtain meteorological data of a time period corresponding to the spatiotemporal maintenance data set, and perform standardization processing on the meteorological data to obtain meteorological characteristic data; Feature vector construction module 500: used to associate the spatiotemporal maintenance data set with the meteorological feature data to construct a feature vector; Prediction model building module 600: used to use a machine learning algorithm to build a damage prediction model based on the constructed feature vector; Fault probability prediction module 700: used to predict the fault probability based on the damage prediction model and output the prediction result; Visualization display module 800: used to obtain the output results of the damage prediction model, determine the probability of equipment failure in each area and perform visual display; The warning information generating module 900 is used to determine the warning level of the area according to a preset fault probability threshold and generate warning information.

[0039] Reference Figure 3 In an embodiment of the present application, a computer device is also provided. The computer device may be a server, and its internal structure may be as follows: Figure 3As shown. The computer device includes a processor, an internal memory, a storage medium (non-volatile storage medium) and a network interface connected through a system bus. Among them, the processor designed by the computer is used to provide computing and control capabilities. The memory of the computer device includes the above-mentioned storage medium (non-volatile storage medium) and the internal memory. The storage medium (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 computer program in the storage medium (non-volatile storage medium). The database of the computer device is used to store usage data, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. Furthermore, the above-mentioned computer device can also be provided with an input device and a display screen, etc. When the above-mentioned computer program is executed by a processor, a visual early warning analysis method for power equipment damage is implemented, including the following steps: obtaining power equipment maintenance order data, processing the text content in the maintenance order data, and obtaining keywords and semantic information; constructing a maintenance reason classification system, and classifying the keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data, and generating corresponding labels; according to the regional management division of the power system, the structured maintenance reason data is integrated by time and place to obtain a spatiotemporal maintenance data set; obtaining meteorological data of a time period corresponding to the spatiotemporal maintenance data set, standardizing the meteorological data, and obtaining meteorological feature data; associating the spatiotemporal maintenance data set with the meteorological feature data to construct a feature vector; using a machine learning algorithm to establish a damage prediction model based on the constructed feature vector; predicting the probability of failure based on the damage prediction model, and outputting the prediction result; obtaining the output result of the damage prediction model, judging the probability of equipment failure in each area and displaying it visually; determining the warning level of each area according to a preset failure probability threshold, and generating warning information. Figure 3 The structure shown in is merely a block diagram of a portion of the structure related to the present application solution and does not constitute a limitation on the computer device to which the present application solution is applied.

[0040] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, a visual early warning analysis method for power equipment damage is implemented, including the following steps: constructing a maintenance reason classification system, and classifying keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data, and generating corresponding labels; integrating the structured maintenance reason data by time and place according to the regional management division of the power system to obtain a spatiotemporal maintenance data set; obtaining meteorological data of a time period corresponding to the spatiotemporal maintenance data set, and standardizing the meteorological data to obtain meteorological feature data; associating the spatiotemporal maintenance data set with the meteorological feature data to construct a feature vector; using a machine learning algorithm to establish a damage prediction model based on the constructed feature vector; predicting the probability of failure based on the damage prediction model, and outputting the prediction result; obtaining the output result of the damage prediction model, judging the probability of failure of equipment in each area and displaying it visually; determining the warning level of each area according to a preset fault probability threshold, and generating warning information. It can be understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0041] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented 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 memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0042] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, device, article or method including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, device, article or method. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, device, article or method including the element.

[0043] The above description is only a preferred embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

[0044] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. A visual early warning analysis method for power equipment damage, characterized in that: The method comprises: Acquire power equipment maintenance order data, process text content in the maintenance order data, and obtain keywords and semantic information; Constructing a maintenance reason classification system, and classifying keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data and generate corresponding labels; According to the regional management division of the power system, the structured maintenance reason data is integrated according to time and place to obtain a maintenance data set in time and space dimensions; Acquire meteorological data of a time period corresponding to the spatiotemporal maintenance data set, and perform standardization processing on the meteorological data to obtain meteorological characteristic data; Associating the spatiotemporal maintenance data set with the meteorological characteristic data to construct a characteristic vector; Using machine learning algorithms, a damage prediction model is established based on the constructed feature vectors; Performing failure probability prediction based on the damage prediction model and outputting the prediction result; Obtaining the output results of the damage prediction model, determining the probability of equipment failure in each area and displaying it visually; According to the preset fault probability threshold, the warning level of each area is determined and the warning information is generated.

2. The method according to claim 1, characterized in that: The construction of the maintenance reason classification system includes: Collect historical operating status data related to different equipment names in the power system, and form an initial data set by integrating equipment model and voltage level data information; Extract the fault location information of each data in the initial data set, and combine it with the pre-established fault level and fault cause correspondence table to obtain the fault level corresponding to the fault cause information; Obtain the device name, operating status data and device model information corresponding to the fault phenomenon in the inspection record database. If there are multiple data records for this device name, fuse the device model information to obtain a specific data object; Traversing the specific data objects, performing frequency analysis based on the fault level in each object and combining the service life, determining the fault level with the highest frequency and exceeding a preset frequency threshold as a target fault; Use the random forest algorithm to build a fault cause analysis model, collect relevant data of the target fault in the historical database, input all collected target fault data into the fault cause analysis model, and obtain the output result after training; Construct a classifier based on support vector machine. When the classifier has at least two different operating status data outputs, use clustering algorithm to determine the target fault data distribution corresponding to each operating status data, and determine the association table between operating status and fault cause with the largest density; Analyze the voltage level data and historical inspection records of the specific equipment name corresponding to the specific time point, determine the predicted fault cause of the corresponding equipment according to the association table between the operating status and the fault cause, and generate treatment measures related to the predicted fault cause.

3. The method according to claim 1, characterized in that According to the regional management division of the power system, the structured maintenance reason data is integrated according to time and location to obtain a maintenance data set in the time and space dimensions, including: Obtain regional management division information of the power system and determine the geographical scope and division granularity of data integration; Extract the time attribute and location attribute of the structured maintenance reason data to form the time dimension and location dimension; According to the time dimension and location dimension, the maintenance reason data is grouped and aggregated to obtain a maintenance data subset indexed by time and location; Data fusion technology is used to merge maintenance data subsets at different times and locations to generate a maintenance data set covering the spatiotemporal dimensions of the entire power system.

4. The method according to claim 1, characterized in that The step of acquiring meteorological data of a time period corresponding to the spatiotemporal maintenance data set and performing standardization processing on the meteorological data to obtain meteorological characteristic data includes: According to the start and end time of the maintenance data set in the spatiotemporal dimension, the meteorological data records of the same time period are retrieved. If the missing meteorological data in the time period exceeds a first preset ratio, the maintenance data at the corresponding time is deleted; Acquire meteorological data, perform standardization processing on the meteorological data, and obtain normalized meteorological data; Extract the maximum and minimum temperature values ​​from the normalized meteorological data and calculate the daily temperature difference data; The normalized meteorological data and daily temperature difference data are combined into meteorological characteristic data.

5. The method according to claim 1, characterized in that Associating the spatiotemporal maintenance data set with the meteorological characteristic data to construct a characteristic vector includes: According to the fault time and fault location, the meteorological data information of the fault time is obtained to obtain the original meteorological data set; Calculate the value of each data according to the original meteorological data set and compare it with the preset first threshold value; If it is greater than the first threshold, the data is extracted to obtain a filtered meteorological data set; According to the three data dimensions of meteorological temperature, meteorological humidity and meteorological wind force in the filtered meteorological data set, a spatial rectangular coordinate system is established to obtain the three-dimensional vector values ​​composed of the three data to obtain a data vector set; According to the data vector set, the Euclidean distance between each vector is calculated. If the distance is greater than the second threshold, the corresponding vector is a noise point. The vector is deleted to obtain the feature vector of the denoised data.

6. The method according to claim 1, characterized in that The damage prediction model is established based on the constructed feature vector using a machine learning algorithm, including: Obtain and pre-process the historical operation data and fault records of equipment in each area, and extract features related to equipment failures; Using the feature vector as input, a damage prediction model is established using a random forest algorithm, so that the model can predict the probability of failure within a future period of time based on equipment operation data; Each area is divided into several sub-areas, and fault prediction is performed on the equipment in each sub-area to obtain the probability distribution of failure of the equipment in each sub-area in the future; Analyze the fault probability distribution of each sub-area, identify the areas with high fault incidence, and determine the areas that need key monitoring and maintenance; Issue early warnings for equipment whose failure probability exceeds a preset probability threshold and generate equipment maintenance plans; Continuously collect equipment operation data, regularly update fault prediction models, and dynamically adjust equipment maintenance plans; According to the fault prediction results and actual fault conditions, the feature selection and algorithm parameters of the damage prediction model are optimized to improve the accuracy of fault prediction.

7. The method according to claim 1, characterized in that The obtaining of the output result of the damage prediction model, determining the probability of equipment failure in each area and performing visual display, includes: Obtain output result data generated by the damage prediction model, which includes failure probability information of equipment in each area; Analyze the output result data and extract the failure probability value of the equipment in each area; According to the preset failure probability threshold, the possibility of equipment failure in each area is determined, and areas with a probability higher than the failure probability threshold are marked as high-risk areas; Using data visualization technology, the failure probability values ​​of equipment in each area are mapped into different colors or patterns to generate intuitive visualization charts; Among them, in the visual chart, high-risk areas are marked with specific colors or patterns; Overlay visualization charts with regional maps to show the relationship between equipment failure probability and geographic location, identify high-risk areas that require special attention and maintenance, and allocate maintenance resources.

8. A visual early warning analysis device for power equipment damage, characterized in that: The device comprises: Data acquisition and processing module: used to obtain power equipment maintenance order data, process the text content in the maintenance order data, and obtain keywords and semantic information; Classification system construction module: used to construct a maintenance reason classification system, and classify keywords and semantic information based on the maintenance reason classification system to obtain structured maintenance reason data and generate corresponding labels; Spatiotemporal data integration module: used to integrate the structured maintenance reason data by time and location according to the regional management division of the power system to obtain a spatiotemporal maintenance data set; Meteorological data processing module: used to obtain meteorological data of a time period corresponding to the time-space dimension maintenance data set, perform standardization processing on the meteorological data, and obtain meteorological characteristic data; Feature vector construction module: used to associate the spatiotemporal maintenance data set with the meteorological feature data to construct a feature vector; Prediction model building module: used to establish a damage prediction model based on the constructed feature vector using a machine learning algorithm; Fault probability prediction module: used to predict the fault probability based on the damage prediction model and output the prediction result; Visualization display module: used to obtain the output results of the damage prediction model, determine the probability of equipment failure in each area and perform visual display; Warning information generation module: used to determine the warning level of the area according to a preset fault probability threshold and generate warning information.

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

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

Images