Substation abnormity monitoring method, device, equipment and medium

Abstract

The invention discloses a transformer substation abnormity monitoring method, device and equipment and a medium, and the method comprises the steps: obtaining a historical alarm data set of a transformer substation; based on the historical alarm data set, a K-means clustering algorithm is combined with a machine learning algorithm to construct an equipment abnormity monitoring model; and obtaining real-time equipment operation data of each piece of power transformation equipment in the transformer substation, and inputting the real-time equipment operation data into the equipment anomaly monitoring model for anomaly monitoring to obtain an anomaly monitoring result of each piece of power transformation equipment in the transformer substation. The equipment abnormity monitoring model of the transformer substation is constructed by combining the K-means clustering algorithm and the machine learning algorithm, a large amount of equipment operation data can be processed in real time, the abnormal state can be quickly recognized, automatic, real-time and precise monitoring of the operation state of the transformer substation is achieved, and then the operation and maintenance efficiency and safety of the transformer substation are remarkably improved. The equipment fault risk is reduced, and the stable operation of the transformer substation is powerfully guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of power monitoring, and particularly to a substation anomaly monitoring method, device, equipment and medium. Background Art

[0002] With the rapid development of the power grid, the safe operation of the power system has become increasingly important. As an important part of the power system, the safety and reliability of substations directly affect the operation efficiency and stability of the entire power system.

[0003] In related technologies, usually, operation and maintenance personnel monitor whether there are abnormalities in the current state of equipment by manually reading message information to avoid equipment failures or accidents. However, with the continuous development of substations and the gradual expansion of their scale, the massive monitoring data has reached a scale where it is impossible to perform efficient and accurate diagnostic analysis by relying on manual reading of message information. This will lead to false alarms, missed alarms or inaccurate diagnoses in the anomaly diagnosis of operation and maintenance personnel, affecting the stability of substation operation. Summary of the Invention

[0004] The present invention provides a substation anomaly monitoring method, device, electronic equipment and medium to solve the technical problem that manual reading of message information cannot perform efficient and accurate diagnostic analysis, resulting in false alarms, missed alarms or inaccurate diagnoses in anomaly diagnosis, affecting the stability of substation operation.

[0005] In a first aspect, a substation anomaly monitoring method is provided, including:

[0006] Obtaining a historical alarm data set of the substation;

[0007] Based on the historical alarm data set, constructing an equipment anomaly monitoring model by using the K-means clustering algorithm in combination with a machine learning algorithm;

[0008] Obtaining real-time equipment operation data of each substation equipment, and inputting the real-time equipment operation data into the equipment anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

[0009] In a second aspect, a substation anomaly monitoring device is provided, including:

[0010] A first obtaining module, configured to obtain a historical alarm data set of the substation;

[0011] A constructing module, configured to construct an equipment anomaly monitoring model by using the K-means clustering algorithm in combination with a machine learning algorithm based on the historical alarm data set;

[0012] A second acquisition module is configured to acquire real-time device operation data of each power transformation device in a substation; a generation module is configured to input the real-time device operation data into a device anomaly monitoring model for anomaly monitoring to obtain anomaly monitoring results of each power transformation device in the substation.

[0013] In a third aspect, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned substation anomaly monitoring method are implemented.

[0014] In a fourth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned substation anomaly monitoring method are implemented.

[0015] In the solutions implemented by the above-mentioned substation anomaly monitoring method, device, electronic device, and storage medium, by combining the K-means clustering algorithm and machine learning algorithms to construct a device anomaly monitoring model for the substation, a large amount of device operation data can be processed in real time, abnormal states can be quickly identified, the influence of subjective judgment on manual monitoring can be avoided, normal and abnormal states can be accurately distinguished, false alarms and missed alarms can be reduced, and automated, real-time, and precise monitoring of the substation operation state can be achieved. Furthermore, the operation and maintenance efficiency and safety of the substation can be significantly improved, the risk of equipment failure can be reduced, and a strong guarantee for the stable operation of the substation can be provided. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 is a flowchart of a substation anomaly monitoring method in an embodiment of the present invention;

[0018] Figure 2 is Figure 1 a flowchart of a specific implementation manner of step S20 in;

[0019] Figure 3 is a flowchart of data preprocessing in a specific embodiment of the present invention;

[0020] Figure 4 is a flowchart of model training in a specific embodiment of the present invention;

[0021] Figure 5 is a flowchart of substation anomaly monitoring in a specific embodiment of the present invention;

[0022] Figure 6 It is a schematic structural diagram of a substation anomaly monitoring device in an embodiment of the present invention. Detailed implementation manners

[0023] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. It should be understood that the accompanying drawings in the present invention only serve the purposes of illustration and description, and are not used to limit the protection scope of the present invention.

[0024] In addition, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in the present invention illustrate operations implemented according to some embodiments of the present invention. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without a logical context relationship may be reversed or implemented simultaneously. In addition, those skilled in the art may add one or more other operations to the flowchart or remove one or more operations from the flowchart under the guidance of the content of the present invention.

[0025] In addition, the embodiments described in the present invention are only some embodiments of the present invention, rather than all embodiments. The components of the embodiments of the present invention described and illustrated in the drawings here usually can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed present invention, but only represents selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.

[0026] It should be noted that the term "including" will be used in the embodiments of the present invention to indicate the existence of the features stated thereafter, but does not exclude the addition of other features. It should also be noted that similar reference numerals and letters denote similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.

[0027] The following will describe this case in detail with reference to the relevant drawings in the specification.

[0028] In the embodiments of this specification, with the rapid development of China's power grid and the increasing demand of the people, the requirements for substations are also getting higher and higher. The reliable and stable operation of substations is the foundation of the entire national power grid economy. The equipment condition monitoring technology of substations has received extensive attention in recent years, and good condition monitoring is the cornerstone of the stable operation of substations. The real-time monitoring of equipment conditions is an important breakthrough in substation condition monitoring in recent years. Researchers are committed to making the real-time condition monitoring of operating equipment reach a level that can meet the actual needs of substations. Considering that the physical signals in the traditional circuits in the secondary state of intelligent substations have been replaced by digital signals, the massive monitoring information has reached a scale where it is impossible to perform efficient strategy analysis simply by relying on manual reading of message information.

[0029] Based on the above problems, this application proposes a substation anomaly monitoring method. By combining the K-means clustering algorithm and machine learning algorithms to construct an equipment anomaly monitoring model for substations, it can process a large amount of equipment operation data in real time, quickly identify abnormal states, avoid the influence of subjective judgment in manual monitoring, accurately distinguish normal and abnormal states, reduce false alarms and missed alarms, realize automatic, real-time and precise monitoring of the operation state of substations, and then significantly improve the operation and maintenance efficiency and safety of substations, reduce the risk of equipment failures, and provide strong guarantee for the stable operation of substations.

[0030] Please refer to Figure 1 , this specification embodiment provides a substation anomaly monitoring method, and the method specifically includes the following steps:

[0031] S10: Obtain the historical alarm data set of the substation.

[0032] It can be understood that the execution subject of the present invention can be a substation anomaly monitoring device, or a terminal or a server, and specific limitations are not made here. This embodiment of the present invention takes the server as the execution subject for illustration.

[0033] Among them, the historical alarm data set specifically includes the relevant data of all historical alarm events of each substation equipment within the acquisition time in the substation. Specifically, the relevant data of historical alarm events include the time when the historical alarm event occurred, the type of alarm equipment, the alarm reason, and the historical equipment operation parameters corresponding to the alarm event. Since the alarm data contains early signals of equipment anomalies, by collecting a large amount of historical alarm data of each equipment in the substation and training the anomaly monitoring model, potential problems in the operation of substation equipment can be identified timely and accurately, and early warnings can be given in advance before the substation equipment actually fails.

[0034] In actual application scenarios, the substation can be a conventional substation or an intelligent substation. The substation equipment includes primary and secondary equipment of the substation, specifically including transformers, circuit breakers, disconnectors, instrument transformers, protection devices, etc. Obtain the network message logs of the substation equipment status. A large amount of abnormal data of each substation equipment (including primary equipment and secondary equipment) monitored or reported is included in the message logs, including fault data and alarm data. By analyzing the network message logs, each alarm event and its corresponding historical operation parameters are extracted, and a historical alarm data set is obtained by summarization.

[0035] Optionally, the alarm events specifically include voltage fluctuations, current anomalies, power changes, temperature above, humidity changes, etc. Exemplarily, the alarm events can be too high transformer oil temperature, data fluctuations or loss of current transformers, data delay of merging units, transformer overheating, etc.

[0036] S20: Based on the historical alarm data set, use the K-means clustering algorithm combined with machine learning algorithms to construct an equipment anomaly monitoring model.

[0037] In this step, clustering analysis is performed on the data in the historical alarm data set through K-means clustering to identify the characteristics of different equipment in the alarm situation. Combining machine learning algorithms to train the clustering results, an equipment anomaly monitoring model is constructed, enabling the model to identify potential problems in the operation of the equipment and timely monitor equipment anomalies.

[0038] In an embodiment of the present application, as Figure 2 shown, a construction scheme for an equipment anomaly monitoring model is provided. In S20, that is, based on the historical alarm data set, use the K-means clustering algorithm combined with machine learning algorithms to construct an equipment anomaly monitoring model, which specifically includes the following steps S21 - S23:

[0039] S21: Preprocess the historical alarm data set to generate a sample data set.

[0040] In this step, the historical alarm data set is huge in quantity and inconsistent in format. If the original alarm data is directly used for model learning, it will lead to low accuracy of the model training results and slow learning rate. To improve the accuracy and efficiency of subsequent model training, a preprocessing operation is performed on a large amount of historical alarm data in the data set to obtain a sample data set, so as to improve the data quality and usability.

[0041] In an embodiment of the present application, a data preprocessing scheme is provided. In S21, that is, preprocess the historical alarm data set to generate a sample data set, which specifically includes the following steps S211 - S213:

[0042] S211: Clean the historical operation parameters in the historical alarm data set.

[0043] In this step, the original data usually contains noise, missing values, duplicate data, or error values. These low-quality data will interfere with model training, causing the subsequent model to learn incorrect patterns and reducing the accuracy and reliability of model training. Therefore, data cleaning is performed on the historical operation parameters in the dataset to remove invalid, redundant, or incorrect information in the data and ensure the accuracy and availability of the data.

[0044] Optionally, the substation data contains noise caused by environmental factors, equipment vibration, etc. The noise of the data is removed through a noise algorithm to improve the signal-to-noise ratio of the data and ensure the accuracy of the data.

[0045] S212: Normalize the cleaned historical operation parameters.

[0046] In this step, the historical operation parameters are the operation status data of each device in the substation when an alarm event occurs, such as current, voltage, temperature, power, etc. These are usually continuous values and have different dimensions and ranges. Inconsistent dimensions will affect the convergence speed and performance of the model. Therefore, it is necessary to normalize the cleaned historical operation parameters to scale parameters with different dimensions and ranges to a unified scale, thereby improving the training efficiency and performance of the model. Specifically, extract the historical operation parameters from the historical alarm dataset and normalize the extracted operation parameters (such as Min-Max normalization) to convert each operation parameter into a normalized value.

[0047] S213: Use the normalized historical operation parameters as features and their corresponding historical alarm events as labels to process the format of the historical alarm dataset to obtain a list of two-dimensional arrays, which is used as the sample dataset.

[0048] In this step, to improve the learning efficiency of the subsequent model, the historical alarm events are encoded to convert them into numerical forms, and the normalized historical operation parameters are used as features, and the encoded historical alarm events are used as labels for sample combination. Finally, all sample combinations are aggregated into a list of two-dimensional arrays as the sample dataset for model training.

[0049] Through the above methods, data cleaning, normalization processing, and sample combination are performed on the original data, reducing the influence of noise and outliers, ensuring that the features are in the same dimension, and ensuring that different features have the same influence on the clustering result, thereby improving the accuracy of the subsequent model training result.

[0050] In actual application scenarios, such as Figure 3As shown in the figure, it is a schematic flowchart of data preprocessing. Among them, a large number of network message logs in the substation are obtained, which contain fault-related data and alarm-related data of each substation equipment. The alarm-related data in the network message log is read, and it is judged whether the reading is successful. If the reading is successful, the alarm-related data is subjected to data cleaning (including removing outliers and zero values). Then, the cleaned data is normalized to obtain sample data. If the reading fails (such as the message log is empty or garbled and cannot be read), the system returns to read other network message logs, and at the same time, the error log prompt information is sent to the terminal of the operation and maintenance personnel.

[0051] S22: Use the K-means clustering algorithm to perform clustering analysis on the sample data set to obtain clustering clusters.

[0052] In this step, the K-means clustering algorithm is used to cluster the sample data in the sample data set. Finally, clustering clusters are obtained, making the sample data more compact, reducing redundant information, making the subsequent model training data cleaner. When the model is training, it only needs to focus on the representative features of each cluster, thereby accelerating the training speed, reducing the computational complexity, and improving the generalization ability of the model on new data.

[0053] In an embodiment of the present application, a clustering analysis scheme is provided. In S22, that is, the K-means clustering algorithm is used to perform clustering analysis on the sample data set to obtain clustering clusters, which specifically includes the following steps S221-S224:

[0054] S221: Randomly select k different centroids in the sample data set as clustering centers.

[0055] S222: Assign all sample data to the nearest clustering center, and set the sample data belonging to the same clustering center as a cluster to obtain k clusters.

[0056] S223: Calculate the average distance from the sample data to the clustering center of its own cluster in each cluster, and determine the new centroid according to the average distance.

[0057] S224: Use the new centroid as the clustering center, reassign all historical abnormal operation parameters to the nearest clustering center, and re-determine the new centroid until all the calculated new centroids no longer change. The finally obtained clusters are used as clustering clusters, where the new centroid is the centroid with the smallest average distance to all sample data in the cluster.

[0058] For steps S221 - S224, the pre - processed sample data (eigenvalues) are input into the K - means algorithm for clustering analysis. The clustering algorithm divides the samples into different clusters, such that the samples within the same cluster have a high similarity, while the samples between different clusters have a low similarity. Randomly select k centroids from n sample data as the initial clustering centers. Set the above - mentioned strategy for determining the new centroids and start the loop. Calculate the distance from each sample point to each centroid. Assign the sample to the corresponding centroid based on the closest distance. Obtain K clusters. For each cluster, calculate the average distance of all the sample points assigned to that cluster as the new centroid. If the calculated new centroid is the same as the original centroid, it means that continuing the loop will not change the clusters, and the loop ends. Otherwise, find the new centroid until all clusters no longer change.

[0059] In the actual application scenario, the hyperparameter k represents the number of classes and needs to be specified manually. The algorithm itself cannot determine how many classes to divide into. The elbow method can be used to determine the optimal value of K. Specifically, the elbow method: can be used to estimate the number of clusters; plot the cost function values for different K values and find the K value corresponding to the position where the decrease in the distortion degree is the largest during the increase of K value (i.e., the elbow); parameters: the centroid position of the class and the position of the internal observations; cost function: the sum of the distortions of each class; the distortion of each class is equal to the sum of the squares of the distances between the centroid of the class and the positions of its internal members; the optimal solution aims to minimize the cost function where u k is the centroid position of the k - th class; effect evaluation - silhouette coefficient:

[0060]

[0061] where S is the evaluation index of the density and dispersion degree of the class; a is the average distance between samples within each class; b is the average distance between samples in a class and all samples in the nearest class; the results show that under the default parameters, the algorithm can converge quickly. When users apply it specifically, they can modify the parameters appropriately according to the situation.

[0062] Through the above - mentioned method, clustering analysis of data features can divide the data set into different clusters, and each cluster represents different data subsets or categories. It helps in the management and organization of data and provides more targeted data subsets for further data analysis and modeling.

[0063] S23: Construct a BP neural network, input the clustering clusters as training data into the BP neural network for model training, obtain the final training model, and use it as the device anomaly monitoring model.

[0064] In this step, a BP neural network (Backpropagation Neural Network) structure is constructed. The clustering clusters are used as training data and input into the BP neural network. The feature vectors of each abnormal sample data are used as inputs, and the identification of the normal / abnormal state of the device is used as the target output to train the model. Finally, the trained BP neural network model is obtained as the device anomaly monitoring model, enabling it to monitor the operating state of the substation equipment and determine whether there is an anomaly in the equipment.

[0065] In an embodiment of the present application, a model training scheme is provided. In S32, that is, a BP neural network is constructed, and the clustering clusters are used as training data and input into the BP neural network for model training to obtain the final training model, which is used as the device anomaly monitoring model. It specifically includes the following steps S321 - S323:

[0066] S321: Based on business requirements, set the labels for the operating states of the substation equipment, and label each cluster.

[0067] S322: Divide the labeled clustering clusters into a training set and a test set.

[0068] S323: Input the training set into the constructed BP neural network for model training, and use the test set to test the training model to adjust the model parameters, and finally obtain the device anomaly monitoring model.

[0069] For steps S321 - S323, according to business requirements, define the labels for the operating states of each substation equipment, specifically including the normal state and the abnormal state. During clustering analysis, the clustering results of the clustering clusters already correspond to the equipment states, and each cluster can be directly labeled. Subsequently, the labeled clustering clusters are divided into a training set and a test set according to a ratio. Construct a BP neural network, input the training set into the BP neural network, calculate the output, calculate the loss function value, and use the backpropagation algorithm to update the weights and biases. Repeat the above steps until the loss function converges or reaches the maximum number of iterations. Subsequently, use the test set to evaluate the model performance and adjust the parameters to obtain the final device anomaly monitoring model.

[0070] In an actual application scenario, as Figure 4 shown, it is a schematic flowchart of model training. First, determine the initial weights and parameters and perform encoding to obtain the initial classification and threshold. Subsequently, assign the initial parameter threshold to the algorithm, and use the historical alarm data (i.e., the sample data set) in the preprocessed database for modeling training, calculate the test error, and determine whether it exceeds the error range. If so, perform parameter training until suitable parameters are found; if not (i.e., within the error range), output the trained substation anomaly monitoring model.

[0071] S30: Obtain the real-time device operation data of each substation equipment, and input the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

[0072] In this step, collect the real-time device operation data of various substation equipment in the substation, input the real-time device operation data into the trained device anomaly monitoring model, and the model analyzes the anomaly trend of the substation equipment operation according to the characteristics of the input data.

[0073] In an embodiment of the present application, a specific device operation monitoring scheme is provided. In S30, that is, obtain the real-time device operation data of each substation equipment, and input the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation, which specifically includes the following steps S31 - S33:

[0074] S31: Collect the real-time device operation data of each substation equipment in the substation.

[0075] S32: Input the real-time operation data into the device anomaly monitoring model to obtain the anomaly monitoring values corresponding to each substation equipment.

[0076] S33: Compare the anomaly monitoring values with the preset anomaly thresholds. If any anomaly monitoring value is greater than the preset anomaly threshold, determine that the operation state of the substation equipment corresponding to the anomaly monitoring value is abnormal.

[0077] For steps S31 - S33, collect the actual device operation data of each substation equipment in the substation, input the operation data into the pre-trained device anomaly monitoring model, and the model outputs an anomaly monitoring value. The larger the anomaly monitoring value, the more likely the device operation state is abnormal. Compare the anomaly monitoring value output by the model with the preset threshold. If the anomaly monitoring value is greater than the preset anomaly threshold, it is determined that the device operation state is abnormal; if the anomaly monitoring value is less than or equal to the preset anomaly value, it is determined that the device operation state is normal.

[0078] Through the above method, collect the device operation data in real time, use the anomaly monitoring model to calculate the anomaly monitoring value, realize the automatic anomaly monitoring of each substation equipment in the substation, and effectively improve the operation safety of the substation.

[0079] In an embodiment of the present application, after confirming that the substation equipment is abnormal, there are also the following steps:

[0080] Based on the anomaly event mapping set, determine the anomaly events and predicted faults of the abnormal substation equipment corresponding to the anomaly monitoring value.

[0081] Generate warning messages based on abnormal events, abnormal monitoring values, predicted faults, and abnormal substation equipment, and send the warning messages to the terminals of the operation and maintenance personnel.

[0082] In the specific implementation of this embodiment, the abnormal event mapping set is a pre-constructed knowledge base used to associate abnormal monitoring values with specific abnormal events and predicted faults. Exemplarily, the abnormal monitoring value range is [0.7, 0.9), the corresponding abnormal event is current overload, and the corresponding predicted fault is circuit breaker failure. According to the abnormal monitoring value, find the corresponding abnormal event and predicted fault in the abnormal event mapping set. Then, use the device ID, abnormal monitoring value, abnormal event, predicted fault, and timestamp as the content of the warning message to generate a warning message, and finally send the warning message to the terminals of the operation and maintenance personnel to help the operation and maintenance personnel quickly locate and handle equipment abnormalities.

[0083] Through the above method, the operation and maintenance efficiency of the substation can be significantly improved, and the losses caused by equipment failures can be reduced.

[0084] In a practical application scenario, as Figure 5 shown, it is a schematic flowchart of substation anomaly monitoring in this application. First, collect the historical alarm data set of the substation, preprocess the data, and establish a data model of the database based on the preprocessed alarm data set. Further, establish a substation anomaly monitoring model through the K-means algorithm combined with the BP neural network algorithm. Monitor the real-time device operation data in the substation through the substation anomaly monitoring model to obtain abnormal monitoring values, and compare the abnormal monitoring values with the preset abnormal thresholds. If the abnormal monitoring value is greater than the preset abnormal threshold, it is determined that the substation equipment is abnormal; if the abnormal monitoring value is less than or equal to the preset abnormal threshold, it is confirmed that the substation equipment is normal.

[0085] It can be seen that in the above solution, by combining the K-means clustering algorithm and the machine learning algorithm to construct an equipment anomaly monitoring model for the substation, a large amount of device operation data can be processed in real time, abnormal states can be quickly identified, the influence of subjective judgment on manual monitoring can be avoided, normal and abnormal states can be accurately distinguished, false alarms and missed alarms can be reduced, and the automatic, real-time, and precise monitoring of the substation operation state can be realized. Furthermore, the operation and maintenance efficiency and safety of the substation can be significantly improved, the equipment failure risk can be reduced, and a strong guarantee for the stable operation of the substation can be provided.

[0086] In an embodiment, a substation anomaly monitoring device is provided, and the substation anomaly monitoring device corresponds one-to-one to the substation anomaly monitoring method in the above embodiment. As Figure 6 shown, the substation anomaly monitoring device 100 includes: a first acquisition module 101, a construction module 102, a second acquisition module 103, and a generation module 104. The detailed descriptions of each functional module are as follows:

[0087] The first acquisition module 101 is configured to acquire the historical alarm data set of the substation;

[0088] The construction module 102 is configured to construct an equipment anomaly monitoring model based on the historical alarm data set by using the K-means clustering algorithm in combination with a machine learning algorithm;

[0089] The second acquisition module 103 is configured to acquire the real-time equipment operation data of each substation equipment; the generation module 104 is configured to input the real-time equipment operation data into the equipment anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

[0090] In one embodiment, the construction module 102 specifically includes:

[0091] The first generation unit is configured to preprocess the historical alarm data set to generate a sample data set;

[0092] The second generation unit is configured to perform clustering analysis on the sample data set by using the K-means clustering algorithm to obtain clustering clusters;

[0093] The construction unit is configured to construct a BP neural network, input the clustering clusters as training data into the BP neural network for model training to obtain a final training model, and use it as the equipment anomaly monitoring model.

[0094] In one embodiment, the historical alarm data set includes the historical alarm events of each substation equipment in the substation and their corresponding historical operation parameters. The first generation unit is specifically configured to:

[0095] Perform data cleaning on the historical operation parameters in the historical alarm data set;

[0096] Perform normalization processing on the cleaned historical operation parameters;

[0097] Use the normalized historical operation parameters as features and the historical alarm events as labels to perform format processing on the historical alarm data set to obtain a two-dimensional array list, which is used as the sample data set.

[0098] In one embodiment, the second generation unit is specifically configured to:

[0099] Randomly select k different centroids as clustering centers in the sample data set;

[0100] Allocate all sample data to the nearest clustering center, and set the sample data belonging to the same clustering center as a cluster to obtain k clusters;

[0101] Calculate the average distance from the sample data to the clustering center of its own cluster in each cluster, and determine a new centroid according to the average distance;

[0102] Taking the new centroid as the clustering center, reassigning all historical abnormal operation parameters to the nearest clustering center, and re-determining the new centroid until all the calculated new centroids no longer change. The finally obtained clusters are used as the clustering clusters, where the new centroid is the centroid with the minimum average distance to all sample data in the cluster.

[0103] In one embodiment, the construction unit is specifically configured to:

[0104] Based on business requirements, set labels for the operation states of substation equipment, and label each cluster;

[0105] Divide the labeled clustering clusters into a training set and a test set;

[0106] Input the training set into the constructed BP neural network for model training, and use the test set to test the trained model to adjust the model parameters, and finally obtain the equipment anomaly monitoring model.

[0107] In one embodiment, the second acquisition module 103 is specifically configured to:

[0108] Collect the real-time equipment operation data of each substation equipment in the substation.

[0109] In one embodiment, the generation module 104 is specifically configured to:

[0110] Input the real-time operation data into the equipment anomaly monitoring model to obtain the anomaly monitoring values corresponding to each substation equipment;

[0111] Compare the anomaly monitoring values with the preset anomaly thresholds. If any anomaly monitoring value is greater than the preset anomaly threshold, it is determined that the operation state of the substation equipment corresponding to the anomaly monitoring value is abnormal.

[0112] The present invention provides a substation anomaly monitoring device 100. By combining the K-means clustering algorithm and the machine learning algorithm to construct an equipment anomaly monitoring model for the substation, it can process a large amount of equipment operation data in real time, quickly identify abnormal states, avoid the influence of subjective judgment in manual monitoring, accurately distinguish normal and abnormal states, reduce false alarms and missed alarms, realize automatic, real-time and precise monitoring of the substation operation state, and thus significantly improve the operation and maintenance efficiency and safety of the substation, reduce the equipment failure risk, and provide a strong guarantee for the stable operation of the substation.

[0113] For the specific limitations of the substation anomaly monitoring device, reference can be made to the limitations of the substation anomaly monitoring method in the above text, which will not be elaborated here. Each module in the above substation anomaly monitoring device can be implemented in whole or in part by software, hardware, and their combination. Each of the above modules can be embedded in the processor of the electronic device in hardware form or independent of it, or stored in the memory of the electronic device in software form, so that the processor can call and execute the operations corresponding to each of the above modules.

[0114] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented:

[0115] Obtain the historical alarm data set of the substation;

[0116] Based on the historical alarm data set, use the K-means clustering algorithm combined with the machine learning algorithm to construct a device anomaly monitoring model;

[0117] Obtain the real-time device operation data of each substation equipment in the substation, and input the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

[0118] 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 following steps are implemented:

[0119] Obtain the historical alarm data set of the substation;

[0120] Based on the historical alarm data set, use the K-means clustering algorithm combined with the machine learning algorithm to construct a device anomaly monitoring model;

[0121] Obtain the real-time device operation data of each substation equipment in the substation, and input the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

[0122] It should be noted that for the functions or steps that can be achieved by the above computer-readable storage medium or electronic device, reference can be made to the relevant descriptions on the server side and the client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0123] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. 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 methods. Among them, any reference to a memory, storage, database, or other medium used in the various embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0124] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0125] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A substation anomaly monitoring method, characterized in that, Including: Obtain the historical alarm data set of the substation; Based on the historical alarm data set, use the K-means clustering algorithm combined with the machine learning algorithm to construct an equipment anomaly monitoring model; Obtain the real-time equipment operation data of each substation equipment in the substation, and input the real-time equipment operation data into the equipment anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each substation equipment in the substation.

2. The method according to claim 1, characterized in that, The step of constructing an equipment anomaly monitoring model based on the historical alarm data set by using the K-means clustering algorithm combined with the machine learning algorithm specifically includes: Preprocess the historical alarm data set to generate a sample data set; Use the K-means clustering algorithm to perform clustering analysis on the sample data set to obtain clustering clusters; Construct a BP neural network, input the clustering clusters as training data into the BP neural network for model training to obtain the final training model, and use it as the equipment anomaly monitoring model.

3. The method according to claim 2, characterized in that, The historical alarm data set includes the historical alarm events of each substation equipment in the substation and their corresponding historical operation parameters. The step of preprocessing the historical alarm data set to generate a sample data set specifically includes: Clean the historical operation parameters in the historical alarm data set; Perform normalization processing on the cleaned historical operation parameters; Use the normalized historical operation parameters as features and the historical alarm events as labels to perform format processing on the historical alarm data set to obtain a two-dimensional array list, which is used as the sample data set.

4. The method according to claim 2, characterized in that, The step of using the K-means clustering algorithm to perform clustering analysis on the sample data set to obtain clustering clusters specifically includes: Randomly select k different centroids as clustering centers in the sample data set; Assign all sample data to the nearest clustering center, and set the sample data belonging to the same clustering center as a cluster to obtain k clusters; Calculate the average distance from the sample data to the clustering center of its own cluster in each cluster, and determine the new centroid according to the average distance; Use the new centroid as the clustering center, reassign all historical abnormal operation parameters to the nearest clustering center, and re-determine the new centroid until all the calculated new centroids no longer change. Take the finally obtained clusters as the clustering clusters, where the new centroid is the centroid with the smallest average distance to all sample data in the cluster.

5. The method according to claim 2, wherein The step of constructing a BP neural network, inputting the clustering clusters as training data into the BP neural network for model training to obtain the final training model, and using it as the equipment anomaly monitoring model specifically includes: Based on business requirements, set the labels of the operation states of substation equipment and label each cluster; Divide the labeled clustering clusters into a training set and a test set; Input the training set into the constructed BP neural network for model training, and use the test set to test the training model to adjust the model parameters to finally obtain the equipment anomaly monitoring model.

6. The method according to claim 1, wherein The step of obtaining the real-time device operation data of each power transformation device in the substation and inputting the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each power transformation device in the substation specifically includes: Collect the real-time device operation data of each power transformation device in the substation; Input the real-time operation data into the device anomaly monitoring model to obtain the anomaly monitoring values corresponding to each power transformation device; Compare the anomaly monitoring values with the preset anomaly thresholds. If any anomaly monitoring value is greater than the preset anomaly threshold, it is determined that the operation state of the power transformation device corresponding to the anomaly monitoring value is abnormal.

7. The method according to claim 6, wherein After determining that the operation state of the power transformation device corresponding to the anomaly monitoring value is abnormal, it further includes: Based on the anomaly event mapping set, determine the anomaly event and predicted fault of the abnormal power transformation device corresponding to the anomaly monitoring value; Generate a warning prompt message based on the anomaly event, the anomaly monitoring value, the predicted fault, and the abnormal power transformation device, and send the warning prompt message to the terminal of the operation and maintenance personnel.

8. An abnormal monitoring device for a substation, characterized in that, It includes: A first acquisition module for acquiring the historical alarm data set of the substation; A construction module for constructing a device anomaly monitoring model based on the historical alarm data set by using the K-means clustering algorithm in combination with the machine learning algorithm; A second acquisition module for acquiring the real-time device operation data of each power transformation device in the substation; a generation module for inputting the real-time device operation data into the device anomaly monitoring model for anomaly monitoring to obtain the anomaly monitoring results of each power transformation device in the substation.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the substation anomaly monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the substation anomaly monitoring method according to any one of claims 1 to 7.

Images