Substation secondary equipment operation monitoring system and method

Abstract

The invention relates to the technical field of substation equipment, and particularly discloses a substation secondary equipment operation monitoring system and method, which adopts an artificial intelligence technology based on deep learning to carry out real-time monitoring and data analysis on current, voltage and temperature data in the operation process of substation secondary equipment. According to the method, time sequence correlation response characteristics among current, voltage and temperature of the secondary equipment of the transformer substation are captured, and potential abnormal correlation interaction modes among multi-modal parameters are mined, so that intelligent identification and abnormal early warning of the operation state of the secondary equipment of the transformer substation are realized. Through the mode, the accuracy and the response speed of equipment fault detection can be remarkably improved, the conditions of false alarm and missing alarm are reduced, and stable operation of the secondary equipment of the transformer substation is ensured.

Description

Technical Field

[0001] The present application relates to the technical field of substation equipment, and more specifically, to a substation secondary equipment operation monitoring system and method. Background Art

[0002] With the rapid development of power systems and technological progress, substations are an indispensable and important part of power systems. Their safe and stable operation is crucial to ensuring the overall performance of the power grid. Substation secondary equipment refers to electrical equipment used for auxiliary functions such as measurement, control, protection and communication. While ensuring the normal operation of primary equipment, it also undertakes the tasks of monitoring the state of the power grid, improving the quality of power and ensuring the reliability of power supply.

[0003] However, in actual operation, due to environmental factors, equipment aging, improper operation and other reasons, secondary equipment may have various faults, such as sensor failure, line short circuit, overheating and other problems. These problems may not only cause local malfunctions, but in serious cases may even cause power outages in the entire substation or even a larger area, bringing huge impacts on social production and people's lives. Therefore, how to effectively monitor and promptly detect potential faults in secondary equipment has become a major challenge facing the power industry.

[0004] Traditional substation secondary equipment monitoring methods mainly rely on regular inspections and simple threshold alarm mechanisms, which have the disadvantages of slow response, high false alarm rate, and inability to warn of complex faults in advance. In recent years, with the development of emerging technologies such as big data, cloud computing, and artificial intelligence, it has become possible to use advanced information processing technology to intelligently analyze the operating data of substation secondary equipment.

[0005] Therefore, an intelligent substation secondary equipment operation monitoring system is expected. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a substation secondary equipment operation monitoring system and method, which uses artificial intelligence technology based on deep learning to perform real-time monitoring and data analysis on the current, voltage and temperature data of the substation secondary equipment during operation, so as to capture the timing correlation response characteristics between the current, voltage and temperature of the substation secondary equipment, and mine the potential abnormal correlation interaction patterns between multi-modal parameters, so as to realize the intelligent identification of the operating status of the substation secondary equipment and abnormal warning. In this way, the accuracy and response speed of equipment fault detection can be significantly improved, the false alarms and missed alarms can be reduced, and the stable operation of the substation secondary equipment can be ensured.

[0007] Accordingly, according to one aspect of the present application, a substation secondary equipment operation monitoring system is provided, comprising:

[0008] Data collection layer, data processing layer and operation monitoring layer;

[0009] The data collection layer is used to collect the time queue of the operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value;

[0010] The data processing layer is used to requeue the time queue of the operating status data to obtain a time queue of real-time current value, a time queue of real-time voltage value and a time queue of real-time temperature value;

[0011] The operation monitoring layer is used to extract real-time current timing implicit features, real-time voltage timing implicit features and real-time temperature timing implicit features from the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value, and determine whether there is any abnormality in the operation status of the monitored substation secondary equipment based on the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features.

[0012] According to another aspect of the present application, a method for monitoring operation of secondary equipment in a substation is provided, comprising:

[0013] A time queue for collecting operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value;

[0014] Requeue the time queue of the operating status data to obtain a time queue of real-time current values, a time queue of real-time voltage values, and a time queue of real-time temperature values;

[0015] Real-time current timing implicit features, real-time voltage timing implicit features and real-time temperature timing implicit features are extracted from the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value, and based on the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features, it is determined whether there is any abnormality in the operating status of the monitored substation secondary equipment.

[0016] Compared with the prior art, the substation secondary equipment operation monitoring system and method provided by this application uses artificial intelligence technology based on deep learning to perform real-time monitoring and data analysis on the current, voltage and temperature data of the substation secondary equipment during operation, so as to capture the time-series correlation response characteristics between the current, voltage and temperature of the substation secondary equipment, and to mine the potential abnormal correlation interaction mode between multi-modal parameters, thereby realizing intelligent identification and abnormal warning of the operation status of the substation secondary equipment. In this way, the accuracy and response speed of equipment fault detection can be significantly improved, the false alarm and missed alarm can be reduced, and the stable operation of the substation secondary equipment can be ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 The block diagram is a substation secondary equipment operation monitoring system according to an embodiment of the present application.

[0019] Figure 2 This is a data flow diagram of a substation secondary equipment operation monitoring system according to an embodiment of the present application.

[0020] Figure 3 It is a block diagram of the operation monitoring layer in the substation secondary equipment operation monitoring system according to an embodiment of the present application.

[0021] Figure 4 It is a block diagram of a timing correlation pattern analysis module in a substation secondary equipment operation monitoring system according to an embodiment of the present application.

[0022] Figure 5 It is a block diagram of a fine-grained interactive analysis module in a substation secondary equipment operation monitoring system according to an embodiment of the present application.

[0023] Figure 6 The present invention is a flowchart of a method for monitoring the operation of secondary equipment in a substation according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0025] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0026] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0027] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0028] In response to the technical problems described in the above background technology, this application proposes an intelligent substation secondary equipment operation monitoring system, which uses artificial intelligence technology based on deep learning to perform real-time monitoring and data analysis on the current, voltage and temperature data of the substation secondary equipment during operation, so as to capture the time-series correlation response characteristics between the current, voltage and temperature of the substation secondary equipment, and dig out the potential abnormal correlation interaction mode between multi-modal parameters, so as to realize the intelligent identification and abnormal warning of the operation status of the substation secondary equipment. In this way, the accuracy and response speed of equipment fault detection can be significantly improved, the false alarm and missed alarm can be reduced, and the stable operation of the substation secondary equipment can be ensured.

[0029] Figure 1 The block diagram is a substation secondary equipment operation monitoring system according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the substation secondary equipment operation monitoring system according to an embodiment of the present application. Figure 1 and Figure 2As shown, the substation secondary equipment operation monitoring system 100 includes: a data collection layer 110, which is used to collect the time queue of the operation status data of the monitored substation secondary equipment, wherein the operation status data includes real-time current value, real-time voltage value and real-time temperature value; a data processing layer 120, which is used to queue the time queue of the operation status data to obtain the time queue of real-time current value, the time queue of real-time voltage value and the time queue of real-time temperature value; an operation monitoring layer 130, which is used to extract the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features from the time queue of real-time current value, the time queue of real-time voltage value and the time queue of real-time temperature value, and determine whether the operation status of the monitored substation secondary equipment is abnormal based on the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features.

[0030] In the above-mentioned substation secondary equipment operation monitoring system, the data collection layer 110 is used to collect the time queue of the operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value. It should be understood that current, voltage and temperature are key indicators for evaluating the operating status of electrical equipment. The current reflects the load condition of the equipment, the voltage shows the stability of power transmission, and the temperature is directly related to the thermal state of the equipment and the possible risk of overheating. Therefore, in the specific implementation of the present application, the actual operation of the equipment is fully understood by real-time monitoring of the current, voltage and temperature data during the operation of the substation secondary equipment.

[0031] Specifically, in the substation secondary equipment operation monitoring system, the data collection layer is the basis of the entire system and is responsible for obtaining the time queue of the operation status data from the substation secondary equipment in real time. These operation status data include real-time current value, real-time voltage value and real-time temperature value, which are key indicators for evaluating the operation status of electrical equipment. The design and implementation of the data collection layer needs to comprehensively consider multiple aspects such as the deployment of sensors, data transmission and storage, to ensure the accuracy and real-time nature of the data.

[0032] Sensors are the core components of the data collection layer and are responsible for real-time monitoring of the operating status of substation secondary equipment. In order to obtain current, voltage and temperature data comprehensively and accurately, appropriate sensors need to be installed at key locations of the equipment. The following is a detailed implementation process of sensor deployment:

[0033] Current transformer: The current transformer is installed on the main circuit of the equipment to collect real-time current values. The current transformer converts the current signal into an electrical signal that is easy to measure by sensing the current changes in the main circuit. Selecting a high-precision and high-stability current transformer can ensure the accuracy of the current data. For example, you can choose a current transformer with an accuracy level of 0.2 or higher.

[0034] Voltage transformer: Voltage transformers are installed on the power supply side and load side of the equipment to collect real-time voltage values. Voltage transformers make voltage measurement safer and more convenient by converting high-voltage signals into low-voltage signals. Similarly, choosing a high-precision, high-stability voltage transformer can ensure the accuracy of voltage data. For example, you can choose a voltage transformer with an accuracy level of 0.2 or higher.

[0035] Temperature sensor: The temperature sensor is installed on key components of the equipment, such as contacts, coils, etc., to collect real-time temperature values. The temperature sensor senses the temperature changes on the surface or inside the equipment and converts the temperature signal into an electrical signal. Selecting a high-precision, high-stability temperature sensor can ensure the accuracy of the temperature data. For example, you can choose a temperature sensor with an accuracy level of ±0.5℃ or higher.

[0036] Data transmission is the key to ensure data real-time and reliability. The data collected by the sensor is transmitted to the data acquisition system through wired or wireless communication technology. Common communication technologies include RS-485, Modbus, Ethernet and Wi-Fi.

[0037] Wired communication: Wired communication technologies such as RS-485 and Ethernet have high transmission rates and reliability and are suitable for long-distance data transmission. For example, current transformers, voltage transformers, and temperature sensors can be connected to data acquisition modules via RS-485 communication lines, and then the data can be transmitted to the central monitoring system via Ethernet. When selecting a communication protocol, it is necessary to consider the real-time and reliability of data transmission. Commonly used protocols include Modbus RTU, Modbus TCP / IP, DL / T 645, etc.

[0038] Wireless communication: Wireless communication technologies such as Wi-Fi and LoRa are highly flexible and convenient and are suitable for short-distance data transmission. For example, data collected by sensors can be transmitted to a data acquisition system via Wi-Fi. To ensure the reliability of data transmission, redundant communication paths and error checking mechanisms can be used to prevent data loss or damage during transmission.

[0039] The design of the data collection layer needs to ensure the real-time and reliability of the data so that abnormal conditions of the equipment can be discovered and handled in a timely manner. The following are the design requirements for the data collection layer: Real-time: Ensure real-time transmission and processing of data through high-speed communication technology and efficient preprocessing algorithms. For example, high-performance embedded processors and real-time operating systems can be used to improve the speed and efficiency of data processing. Reliability: Ensure the reliability of data transmission through redundant communication paths and error checking mechanisms. For example, dual-channel communication technology can be used to automatically switch to the backup channel when the main channel fails to ensure continuous data transmission.

[0040] In order to facilitate operation and maintenance personnel to monitor the operating status of the equipment in real time, the data collection layer can also provide data visualization and monitoring functions. Real-time data display: Real-time data of current, voltage and temperature are displayed through a graphical interface to help operation and maintenance personnel intuitively understand the operating status of the equipment. For example, charts can be used to show the changing trends of current, voltage and temperature, and color coding can be used to show the normal and abnormal status of the data. Alarm mechanism: When data abnormalities are detected, the system can automatically trigger the alarm mechanism to notify the operation and maintenance personnel to deal with it in time. For example, a threshold alarm can be set. When the current, voltage or temperature exceeds the preset threshold, the system automatically sends an alarm message to the operation and maintenance personnel's mobile phone or computer.

[0041] During data collection and transmission, it is necessary to ensure data security and privacy protection to prevent data from being illegally accessed and tampered with. Data encryption: Use encryption technology to encrypt transmitted data to ensure data security. For example, the SSL / TLS protocol can be used to encrypt data transmission.

[0042] Access control: Access control mechanisms are used to limit access to data and ensure that only authorized users can access data. For example, username and password authentication can be used to ensure that only legitimate operation and maintenance personnel can access data.

[0043] In the above-mentioned substation secondary equipment operation monitoring system, the data processing layer 120 is used to perform data queuing on the time queue of the operation status data to obtain the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value. It should be understood that considering that the current, voltage and temperature data may have different time series variation characteristics, therefore, in order to improve the pertinence of data time series analysis, the present application performs data queuing on the time queue of the operation status data, so as to organize different types of operation status parameters into independent time series, thereby helping the model to better learn the unique time series pattern of each parameter.

[0044] In the substation secondary equipment operation monitoring system, data requeuing of the time queue of the operation status data is a key step, which organizes different types of operation status data (such as current, voltage and temperature) into independent time series for subsequent time series analysis and feature extraction. This process not only involves the classification and sorting of data, but also needs to consider the time series characteristics of the data to ensure that different types of data can be compared and analyzed on the same time scale.

[0045] Before data is returned to the queue, the data processing layer must first process and analyze the collected operating status data to extract useful information. Since the raw data collected by the sensor may contain noise and outliers, preliminary preprocessing is required, such as using digital filtering algorithms (such as low-pass filtering, high-pass filtering, etc.) to remove high-frequency noise, using correction algorithms (such as zero-point correction, gain correction, etc.) to correct sensor deviations, and through filtering, denoising and correction operations to ensure data accuracy and reliability, the preprocessed data is then arranged in chronological order to form a time queue.

[0046] Specifically, filtering: Use digital filtering algorithms (such as low-pass filtering, high-pass filtering, etc.) to remove high-frequency noise. For example, a low-pass filter can be used to remove high-frequency interference in current and voltage data, and a high-pass filter can be used to remove low-frequency drift in temperature data. Denoising: Remove outliers in the data through statistical methods or machine learning algorithms. For example, the Z-score method can be used to detect and remove outliers in current, voltage, and temperature data. Correction: Use correction algorithms (such as zero-point correction, gain correction, etc.) to correct the deviation of the sensor. For example, the zero-point correction algorithm can be used to correct the zero-point deviation of the current transformer and the voltage transformer, and the gain correction algorithm can be used to correct the gain error of the temperature sensor.

[0047] After data preprocessing is completed, the next step is to organize different types of operating status data into independent time series. This process usually includes the following steps: Data classification: Separate different types of operating status data (such as current, voltage, and temperature) from the original data set. This process is usually implemented through data processing libraries in programming languages. For example, using the DataFrame object in the Pandas library, different types of data can be extracted by column name or index. Time alignment: Time alignment is a key step to ensure that different types of operating status data are compared and analyzed on the same time scale. Since the sampling frequency and timestamp of different sensors may not be completely consistent, the data needs to be time-aligned to ensure that the current value, voltage value, and temperature value at each moment can correspond. Time alignment methods include interpolation, resampling, and alignment. For example, the resample method in the Pandas library can be used to resample the data to a uniform time interval, or the interpolate method can be used to interpolate and fill in missing data. Interpolation: Interpolation methods are used to fill missing values ​​in the data to ensure the continuity of the data. Common interpolation methods include linear interpolation, polynomial interpolation, and spline interpolation. For example, you can use the interpolate method in the Pandas library for linear interpolation. Resampling: The resampling method is used to resample the data to a uniform time interval to ensure that the time resolution of the data is consistent. Common resampling methods include mean resampling, maximum resampling, and minimum resampling. For example, you can use the resample method in the Pandas library to resample the data to a time interval of 1 second. Alignment: The alignment method is used to ensure that the timestamps of different time series are consistent. Common alignment methods include left alignment, right alignment, and center alignment. For example, you can use the align method in the Pandas library to align different time series.

[0048] After completing data classification and time alignment, the next step is to organize different types of operating status data into independent time series. Specifically, organize the real-time current value, real-time voltage value, and real-time temperature value into three independent time queues. This step can be achieved by creating a new data structure (such as a list, array, etc.).

[0049] After the data is returned to the queue, the generated time series needs to be verified and quality checked to ensure the integrity and accuracy of the data. This step can be achieved through statistical analysis and visualization tools. For example, a time series graph can be drawn to observe the changing trends and abnormal points of the data. If obvious abnormalities or missing data are found, it is necessary to return to the data preprocessing stage for further correction.

[0050] It should be understood that data regrouping can improve the pertinence of timing analysis by regrouping the time queues of the operating status data. Organizing different types of operating status data into independent time series can help the model better learn the unique timing pattern of each parameter, thereby improving the accuracy and effectiveness of timing analysis. Data regrouping can also simplify the data processing process. The independent time series after data regrouping can simplify the subsequent data processing and analysis process and reduce the complexity of data processing. Furthermore, data regrouping can enhance the interpretability of the model. By analyzing the time series of current, voltage, and temperature separately, the operating status and potential failure modes of the equipment can be more clearly understood, and the interpretability and explainability of the model can be enhanced.

[0051] In summary, data queuing of the time queue of the operating status data at the data processing layer is an important link in the substation secondary equipment operation monitoring system. Through steps such as data classification, time alignment, data queuing and quality inspection, different types of operating status data can be organized into independent time series, thereby providing high-quality data support for subsequent time series analysis and feature extraction. This process not only helps to improve the accuracy and response speed of equipment fault detection, but also reduces false alarms and missed alarms, ensuring the stable operation of substation secondary equipment. Specifically, the process of data queuing includes data classification, classifying the collected operating status data according to type, that is, extracting the current value, voltage value and temperature value respectively; time alignment, through interpolation, resampling and alignment methods, ensure that different types of operating status data are compared and analyzed on the same time scale; data queuing, organizing the real-time current value, real-time voltage value and real-time temperature value into three independent time queues respectively; data verification and quality inspection, through statistical analysis and visualization tools, verifying and quality checking the generated time series to ensure the integrity and accuracy of the data. Through this series of steps, the accuracy and real-time nature of the data can be ensured, providing a solid foundation for operating status monitoring of substation secondary equipment.

[0052] In the above-mentioned substation secondary equipment operation monitoring system, the operation monitoring layer 130 is used to extract the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features from the real-time current value time series, the real-time voltage value time series and the real-time temperature value time series implicit features, and determine whether the operation status of the monitored substation secondary equipment is abnormal based on the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features. Wherein, Figure 3 FIG. 1 is a block diagram of the operation monitoring layer in the substation secondary equipment operation monitoring system according to an embodiment of the present application. Figure 3As shown, the operation monitoring layer 130 includes: a data timing encoding module 131, which is used to perform timing encoding on the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value to obtain the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features; a timing correlation pattern analysis module 132, which is used to perform timing correlation pattern analysis on the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features to obtain the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features A real-time current-real-time temperature timing correlation characteristic diagram and a real-time voltage-real-time temperature timing correlation characteristic diagram; a fine-grained interactive analysis module 133, used to perform fine-grained interactive analysis on the real-time current-real-time temperature timing correlation characteristic diagram and the real-time voltage-real-time temperature timing correlation characteristic diagram based on the core joint feature guidance to obtain a multi-modal fine-grained joint coding characteristic diagram of the equipment status; an operating status abnormality identification module 134, used to determine whether there is an abnormality in the operating status of the monitored substation secondary equipment based on the multi-modal fine-grained joint coding characteristic diagram of the equipment status.

[0053] Specifically, the data timing encoding module 131 is further used to: use a sequence encoder based on a forward LSTM model to perform timing encoding on the time queue of the real-time current value, the time queue of the real-time voltage value, and the time queue of the real-time temperature value to obtain a real-time current timing-related implicit feature vector, a real-time voltage timing-related implicit feature vector, and a real-time temperature timing-related implicit feature vector as the real-time current timing implicit feature, the real-time voltage timing implicit feature, and the real-time temperature timing implicit feature. Specifically, the present application uses a forward LSTM model that is particularly good at processing sequence data with timing dependence to perform timing encoding on the time queue of the real-time current value, the time queue of the real-time voltage value, and the time queue of the real-time temperature value, respectively, and uses the long-term memory capacity of the LSTM model to capture the characteristic trend of the current, voltage, and temperature data changing over time, thereby extracting the real-time current timing implicit feature, the real-time voltage timing implicit feature, and the real-time temperature timing implicit feature, that is, the real-time current timing-related implicit feature vector, the real-time voltage timing-related implicit feature vector, and the real-time temperature timing-related implicit feature vector.

[0054] Specifically, the timing association pattern analysis module 132 is used to perform timing association pattern analysis on the real-time current timing implicit features, the real-time voltage timing implicit features, and the real-time temperature timing implicit features, respectively, to obtain a real-time current-real-time temperature timing association feature diagram and a real-time voltage-real-time temperature timing association feature diagram. It should be understood that, considering that temperature is a comprehensive indicator of the operating status of electrical equipment, it is comprehensively affected by multiple factors such as equipment current and voltage. Therefore, in order to reveal the dynamic interaction pattern between current and temperature, and voltage and temperature, so as to deeply understand the working mechanism and potential failure mode of the equipment, the present application further performs association analysis on the real-time current timing association implicit feature vector and the real-time temperature timing association implicit feature vector, as well as the real-time voltage timing association implicit feature vector and the real-time temperature timing association implicit feature vector, and constructs a real-time current-real-time temperature timing association matrix and a real-time voltage-real-time temperature timing association matrix, so as to analyze the temperature response pattern of the equipment under different working conditions and understand the operating status of the equipment from multiple dimensions. Among them, Figure 4 FIG. 1 is a block diagram of a timing correlation mode analysis module in a substation secondary equipment operation monitoring system according to an embodiment of the present application. Figure 4 As shown, the timing association pattern analysis module 132 includes: a current-temperature timing association analysis unit 1321, which is used to perform association analysis on the real-time current timing association implicit feature vector and the real-time temperature timing association implicit feature vector to obtain a real-time current-real-time temperature timing association matrix; a voltage-temperature timing association analysis unit 1322, which is used to perform association analysis on the real-time voltage timing association implicit feature vector and the real-time temperature timing association implicit feature vector to obtain a real-time voltage-real-time temperature timing association matrix; a timing association pattern feature extraction unit 1323, which is used to input the real-time current-real-time temperature timing association matrix and the real-time voltage-real-time temperature timing association matrix into a parameter-to-parameter timing association pattern feature extractor based on a void convolutional neural network model respectively to obtain the real-time current-real-time temperature timing association feature diagram and the real-time voltage-real-time temperature timing association feature diagram.

[0055] More specifically, the current-temperature timing association analysis unit 1321 is further used to: calculate the Kronecker product between the real-time current timing association implicit feature vector and the transposed vector of the real-time temperature timing association implicit feature vector to obtain the real-time current-real-time temperature timing association matrix. That is, the real-time current-real-time temperature timing association matrix is ​​constructed by calculating the Kronecker product between the real-time current timing association implicit feature vector and the transposed vector of the real-time temperature timing association implicit feature vector. Similarly, the real-time voltage-real-time temperature timing association matrix is ​​constructed by calculating the Kronecker product between the real-time voltage timing association implicit feature vector and the transposed vector of the real-time temperature timing association implicit feature vector.

[0056] More specifically, the timing association pattern feature extraction unit 1323 is used to input the real-time current-real-time temperature timing association matrix and the real-time voltage-real-time temperature timing association matrix into the parameter timing association pattern feature extractor based on the hole convolution neural network model to obtain the real-time current-real-time temperature timing association feature diagram and the real-time voltage-real-time temperature timing association feature diagram. That is, in order to further abstract the deep-level timing association features between parameters, the present application uses a hole convolution neural network model to extract features from the real-time current-real-time temperature timing association matrix and the real-time voltage-real-time temperature timing association matrix. Compared with traditional pooling operations, hole convolution can expand the receptive field while maintaining time resolution, so that the model can capture a wider range of contextual information without losing time details, thereby helping to capture subtle changes and detailed abnormal patterns in the operating state of the equipment, and more comprehensively understand the complex relationship between current, voltage and temperature.

[0057] Specifically, the fine-grained interactive analysis module 133 is used to perform fine-grained interactive analysis based on the core joint feature guidance on the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram to obtain a multi-modal fine-grained joint encoding feature diagram of the equipment status. It should be understood that the present application takes into account that the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram capture the timing correlation patterns between current and temperature, and voltage and temperature, respectively. However, the operating status of the secondary equipment of the substation is determined by multiple parameters. In order to more comprehensively understand the overall status of the equipment, the present application further performs fine-grained interactive encoding on the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram to capture the comprehensive timing correlation between current, voltage and temperature, thereby realizing a multi-dimensional and in-depth analysis of the operating status of the secondary equipment of the substation. Among them, Figure 5FIG. 1 is a block diagram of a fine-grained interactive analysis module in a substation secondary equipment operation monitoring system according to an embodiment of the present application. Figure 5 As shown, the fine-grained interaction analysis module 133 includes: a fine-grained common feature extraction unit 1331, which is used to perform fine-grained common feature extraction on the real-time current-real-time temperature timing association feature diagram and the real-time voltage-real-time temperature timing association feature diagram to obtain a set of multimodal joint fine-grained core feature vectors of device status; a joint feature semantic clustering unit 1332, which is used to input the set of multimodal joint fine-grained core feature vectors of device status into a clustering network to obtain a multimodal joint feature semantic clustering center vector of device status; a global semantic interaction unit 1333, which is used to guide the set of multimodal joint fine-grained core feature vectors of device status to perform cross-domain context semantic interaction based on the multimodal joint feature semantic clustering center vector of the device status to obtain the multimodal fine-grained joint encoding feature diagram of the device status.

[0058] More specifically, the fine-grained common feature extraction unit 1331 is further used to: perform feature fine-grained decoupling on the real-time current-real-time temperature timing correlation feature graph and the real-time voltage-real-time temperature timing correlation feature graph to obtain a set of real-time current-real-time temperature local timing correlation feature vectors and a set of real-time voltage-real-time temperature local timing correlation feature vectors; and convert each corresponding set of real-time current-real-time temperature local timing correlation feature vectors in the set of real-time current-real-time temperature local timing correlation feature vectors and the set of real-time voltage-real-time temperature local timing correlation feature vectors into a real-time current-real-time temperature local timing correlation feature vector. The real-time current and real-time voltage-real-time temperature local timing correlation feature vectors are input into a fine-grained common feature extraction network to obtain a set of multi-modal joint fine-grained core feature vectors of the device state, wherein the fine-grained common feature extraction network calculates the point-by-point subtraction vector, point-by-point multiplication vector and point-by-point addition vector between the real-time current-real-temperature local timing correlation feature vector and the real-time voltage-real-temperature local timing correlation feature vector, and cascades and fuses the three into a neural network layer based on the tanh function to obtain the multi-modal joint fine-grained core feature vector of the device state.

[0059] The above-mentioned fine-grained common feature extraction unit 1331 can be expressed as follows:

[0060] Decouple(F1)={x1,x2,...,x i ,...,x n}

[0061] Decouple(F2)={y1,y2,...,y i ,...,y n}

[0062]

[0063] Among them, F1 represents the real-time current-real-time temperature time series correlation characteristic diagram, F2 represents the real-time voltage-real-time temperature time series correlation characteristic diagram, Decouple(·) represents characteristic decoupling, x1, x2, x i and x n They represent the first, second, i-th and n-th real-time current-real-time temperature local time series correlation feature vectors in the set of real-time current-real-time temperature local time series correlation feature vectors, ⊙ represents point multiplication, ⊕ represents point addition, represents point subtraction, Concat(·;·;·) represents cascade, W VT and b VT Represent the weight matrix and bias term respectively, tanh represents the hyperbolic tangent function, v i Represents the i-th device state multimodal joint fine-grained core feature vector in the set of device state multimodal joint fine-grained core feature vectors.

[0064] Specifically, in order to refine the feature granularity of interactive coding, the present application first performs feature fine-grained decoupling on the real-time current-real-time temperature timing correlation feature graph and the real-time voltage-real-time temperature timing correlation feature graph, decomposing the complex interaction pattern between current, voltage and temperature into more subtle feature components, so as to better identify and distinguish the local detail information in the feature graph, thereby generating a set of real-time current-real-time temperature local timing correlation feature vectors and a set of real-time voltage-real-time temperature local timing correlation feature vectors. Then, a fine-grained common feature extraction network is used to identify and extract the common feature patterns between the two feature sets, extract the intrinsic connection between data in different domains, and reduce the impact of domain differences on subsequent feature context interaction analysis, thereby generating a set of multi-modal joint fine-grained core feature vectors of device status.

[0065] More specifically, the joint feature semantic clustering unit 1332 is used to input the set of the multimodal joint fine-grained core feature vectors of the device state into the clustering network to obtain the semantic clustering center vector of the multimodal joint feature of the device state. That is, the set of the multimodal joint fine-grained core feature vectors of the device state is input into the clustering network to realize the global sparse joint representation of the fine-grained joint features, generate the semantic clustering center vector of the multimodal joint feature of the device state, and use it as the global context information to guide the subsequent global context interaction of the joint features, so as to enhance the consistency and semantic richness of the feature representation.

[0066] The above-mentioned joint feature semantic clustering unit 1332 can be expressed as follows:

[0067]

[0068] Among them, v c represents the semantic clustering center vector of the multimodal joint feature of the device state, n is the number of feature vectors in the set of the real-time current-real-time temperature local time series correlation feature vectors, v i Represents the i-th device state multimodal joint fine-grained core feature vector in the set of device state multimodal joint fine-grained core feature vectors.

[0069] More specifically, the global semantic interaction unit 1333 is further used to: construct a query vector and a value vector based on each device state multimodal joint fine-grained core feature vector in the set of the device state multimodal joint fine-grained core feature vectors, and construct a key vector based on the device state multimodal joint feature semantic clustering center vector, input the query vector, the value vector and the key vector into a global semantic interaction encoder based on a converter structure to obtain a set of context device state multimodal joint fine-grained core feature vectors; and perform feature reshaping on each context device state multimodal joint fine-grained core feature vector in the set of context device state multimodal joint fine-grained core feature vectors to obtain the device state multimodal fine-grained joint encoding feature map.

[0070] The global semantic interaction unit 1333 can be expressed as follows:

[0071] v qi =W q c i

[0072] v k =W k v c

[0073] v vi =W b v i

[0074]

[0075] V s = {v s1 ,v s2 ,...,v si ,...,v sn}

[0076] F r = reshape(V s )

[0077] Among them, v irepresents the i-th device state multimodal joint fine-grained core feature vector in the set of device state multimodal joint fine-grained core feature vectors, v c Represents the semantic clustering center vector of multimodal joint features of device status, W q , W k and W b denote the query embedding matrix, key embedding matrix and value embedding matrix respectively, and v qi and v vi They represent the query vector and value vector corresponding to the multimodal joint fine-grained core feature vector of the i-th device state, respectively, and v k represents a key vector, L represents the length of the key vector, represents the matrix multiplication operation, (·) T represents the transpose of the vector, Softmax represents the normalized exponential function, V s Represents the set of multimodal joint fine-grained core feature vectors of contextual device states, v s1 、v s2 、v si and v sn denote the first, second, i-th and n-th contextual device state multimodal joint fine-grained core feature vectors in the set of the contextual device state multimodal joint fine-grained core feature vectors, respectively, reshape(·) denotes feature shape reshaping, and F r A multimodal fine-grained joint encoding feature map representing the device state.

[0078] Specifically, the present application constructs a query vector and a value vector based on the multimodal joint fine-grained core feature vector of the device state, and uses the semantic clustering center vector of the multimodal joint feature of the device state to construct a key vector. The global context interaction encoding of the multimodal joint feature is performed using a global semantic interaction encoder based on a converter structure. Through the self-attention mechanism, information exchange is achieved on a global scale, and the self-representation is adjusted according to the global context information to obtain a richer feature representation. Then, a feature reshape operation is performed on the generated contextual device state multimodal joint fine-grained core feature vector to restore the original feature structure and generate a device state multimodal fine-grained joint encoding feature map. In this way, the correlation features between different domain data can be extracted at a fine-grained level, thereby providing a more accurate feature representation for the operating status monitoring of substation secondary equipment.

[0079] Specifically, the abnormal operation state identification module 134 is further used to: input the multimodal fine-grained joint encoding feature map of the equipment state into the classifier-based operation state identifier to obtain an identification result, and the identification result is used to indicate whether the operation state of the monitored substation secondary equipment is abnormal. It should be known to those skilled in the art that the classifier establishes a model by learning a large amount of normal and abnormal state data, and can judge whether the equipment is in a normal operation state according to the input feature data, so as to issue an alarm in time when there is an abnormality in the operation state, and remind the operation and maintenance personnel to take corresponding measures.

[0080] In a preferred example of the present application, inputting the device state multimodal fine-grained joint coding feature map into a classifier-based operating state identifier to obtain the identification result includes: determining the number of zero eigenvalues ​​in the device state multimodal fine-grained joint coding feature map, and respectively calculating the reciprocal of the logarithm of the number of zero eigenvalues ​​with base two and the exponential value of the reciprocal of the number of zero eigenvalues ​​with base natural constants to obtain a first device state multimodal fine-grained joint coding representation value and a second device state multimodal fine-grained joint coding representation value:

[0081] α=[log2(n)] -1

[0082] β=exp[(n) -1 ]

[0083] Wherein, α is the first device state multimodal fine-grained joint coding representation value, β is the second device state multimodal fine-grained joint coding representation value, n is the number of zero eigenvalues ​​in the device state multimodal fine-grained joint coding feature map, and exp(·) represents a natural exponential function;

[0084] Next, the device state multimodal fine-grained joint coding feature graph is expanded into a device state multimodal fine-grained joint coding feature vector; each eigenvalue of the device state multimodal fine-grained joint coding feature vector is calculated as a power function with the inverse of the number of zero eigenvalues ​​as the exponent to obtain a device state multimodal fine-grained joint coding structure state vector: the device state multimodal fine-grained joint coding structure state vector is point-multiplied with the first device state multimodal fine-grained joint coding representation value to obtain a device state multimodal fine-grained joint coding information representation vector:

[0085]

[0086] Wherein, V1 is the state vector of the device state multimodal fine-grained joint coding structure, α is the first device state multimodal fine-grained joint coding representation value, V is the device state multimodal fine-grained joint coding feature vector, and V is a row vector, represents a power function value of each eigenvalue in the calculation vector with the reciprocal of the number of zero eigenvalues ​​as the exponent;

[0087] Secondly, the autocorrelation matrix of the state vector of the device state multimodal fine-grained joint coding structure is point-multiplied with the second device state multimodal fine-grained joint coding representation value and weight hyperparameters to obtain the device state multimodal fine-grained joint coding regression understanding matrix; the device state multimodal fine-grained joint coding information representation vector is matrix-multiplied with the device state multimodal fine-grained joint coding regression understanding matrix to obtain the optimized device state multimodal fine-grained joint coding feature vector:

[0088]

[0089]

[0090] Among them, M is the device state multimodal fine-grained joint encoding regression understanding matrix, w is the weight hyperparameter, β is the second device state multimodal fine-grained joint encoding representation value, Represents the state vector of the multimodal fine-grained joint encoding structure of the device state, (·) T represents the transpose of a vector, represents matrix multiplication, V′ is the optimized multi-modal fine-grained joint encoding feature vector of the device state;

[0091] Finally, the optimized multi-modal fine-grained joint encoding feature vector of the device state is input into a classifier-based operating state identifier to obtain the identification result.

[0092] Here, when the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram respectively represent the real-time current-real-time temperature local time domain correlation collaborative features and the real-time voltage-real-time temperature local time domain correlation collaborative features, when performing global semantic interactive encoding based on cross-domain fine-grained core joint features, the cross-domain fine-grained core joint feature representation of the device state multimodal fine-grained joint encoding feature diagram will also have a remote class regression understanding representation imbalance relative to the distribution field selective state space of the heteromodal data features, thereby deviating from the temporal semantic distribution representation class understanding and affecting the accuracy of the classification results.

[0093] Based on this, the present application realizes effective long-range modeling of the multimodal fine-grained jointly encoded feature map of the device state under the field linear complexity by modeling the vector field zero dimension of the multimodal fine-grained jointly encoded feature vector of the device state obtained after the multimodal fine-grained jointly encoded feature map of the device state is expanded as the global field redundant dependency. In this way, in the feature regression process of the multimodal fine-grained jointly encoded feature map of the device state, the structured correlation self-distillation between information representation and class understanding is performed in the selective state space based on the feature distribution, so as to maintain the overall information complexity of the feature distribution of the multimodal fine-grained jointly encoded feature map of the device state while promoting the long-range dynamic class understanding balance of its feature convergence, so as to improve the class understanding consistency between the classification regression and feature extraction processes, and improve the accuracy of the recognition result obtained by inputting the multimodal fine-grained jointly encoded feature map of the device state into the classifier-based running state identifier.

[0094] In summary, the substation secondary equipment operation monitoring system based on the embodiment of the present application is explained, which uses artificial intelligence technology based on deep learning to perform real-time monitoring and data analysis on the current, voltage and temperature data of the substation secondary equipment during operation, so as to capture the time-series correlation response characteristics between the current, voltage and temperature of the substation secondary equipment, and dig out the potential abnormal correlation interaction mode between multi-modal parameters, so as to realize the intelligent identification and abnormal warning of the operation status of the substation secondary equipment. In this way, the accuracy and response speed of equipment fault detection can be significantly improved, the false alarm and missed alarm can be reduced, and the stable operation of the substation secondary equipment can be ensured.

[0095] Furthermore, a method for monitoring the operation of secondary equipment in a substation is also provided.

[0096] Figure 6 FIG. 1 is a flow chart of a method for monitoring the operation of a substation secondary device according to an embodiment of the present application. Figure 6 As shown, the substation secondary equipment operation monitoring method includes the steps of: S1, collecting the time queue of the operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value; S2, queuing the time queue of the operating status data to obtain the time queue of real-time current value, the time queue of real-time voltage value and the time queue of real-time temperature value; S3, extracting the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features from the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value, and determining whether the operating status of the monitored substation secondary equipment is abnormal based on the real-time current time series implicit features, the real-time voltage time series implicit features and the real-time temperature time series implicit features.

[0097] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned substation secondary equipment operation monitoring method have been referred to above. Figures 1 to 5 It has been introduced in detail in the description of the substation secondary equipment operation monitoring system, and therefore, its repeated description will be omitted.

[0098] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0099] In the above embodiments, the description of each embodiment has its own emphasis. For the parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0100] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0101] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0102] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A substation secondary equipment operation monitoring system, characterized in that: include: Data collection layer, data processing layer and operation monitoring layer; The data collection layer is used to collect the time queue of the operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value; The data processing layer is used to requeue the time queue of the operating status data to obtain a time queue of real-time current value, a time queue of real-time voltage value and a time queue of real-time temperature value; The operation monitoring layer is used to extract real-time current timing implicit features, real-time voltage timing implicit features and real-time temperature timing implicit features from the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value, and determine whether there is any abnormality in the operation status of the monitored substation secondary equipment based on the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features.

2. The substation secondary equipment operation monitoring system according to claim 1, characterized in that: The operation monitoring layer includes: A data timing encoding module, used for performing timing encoding on the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value respectively to obtain the real-time current timing implicit feature, the real-time voltage timing implicit feature and the real-time temperature timing implicit feature; A timing correlation mode analysis module, used for performing timing correlation mode analysis on the real-time current timing implicit feature, the real-time voltage timing implicit feature and the real-time temperature timing implicit feature respectively to obtain a real-time current-real-time temperature timing correlation feature diagram and a real-time voltage-real-time temperature timing correlation feature diagram; A fine-grained interactive analysis module, used to perform fine-grained interactive analysis based on core joint feature guidance on the real-time current-real-time temperature time series correlation feature diagram and the real-time voltage-real-time temperature time series correlation feature diagram to obtain a multi-modal fine-grained joint encoding feature diagram of the device state; The operating status abnormality identification module is used to determine whether there is an abnormality in the operating status of the monitored substation secondary equipment based on the multi-modal fine-grained joint coding feature map of the equipment status.

3. The substation secondary equipment operation monitoring system according to claim 2, characterized in that: The data timing encoding module is used to: A sequence encoder based on a forward LSTM model is used to perform time series encoding on the time queue of the real-time current value, the time queue of the real-time voltage value, and the time queue of the real-time temperature value to obtain a real-time current time series associated implicit feature vector, a real-time voltage time series associated implicit feature vector, and a real-time temperature time series associated implicit feature vector as the real-time current time series implicit feature, the real-time voltage time series implicit feature, and the real-time temperature time series implicit feature, respectively.

4. The substation secondary equipment operation monitoring system according to claim 3, characterized in that: The time series correlation pattern analysis module comprises: A current-temperature time series correlation analysis unit, used for performing correlation analysis on the real-time current time series correlation implicit feature vector and the real-time temperature time series correlation implicit feature vector to obtain a real-time current-real-time temperature time series correlation matrix; A voltage-temperature timing correlation analysis unit, configured to perform correlation analysis on the real-time voltage timing correlation implicit feature vector and the real-time temperature timing correlation implicit feature vector to obtain a real-time voltage-real-time temperature timing correlation matrix; The timing correlation pattern feature extraction unit is used to input the real-time current-real-time temperature timing correlation matrix and the real-time voltage-real-time temperature timing correlation matrix into the parameter timing correlation pattern feature extractor based on the void convolutional neural network model to obtain the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram.

5. The substation secondary equipment operation monitoring system according to claim 4, characterized in that: The current-temperature timing correlation analysis unit is used to: The Kronecker product between the transposed vectors of the real-time current time series correlation implicit feature vector and the real-time temperature time series correlation implicit feature vector is calculated to obtain the real-time current-real-time temperature time series correlation matrix.

6. The substation secondary equipment operation monitoring system according to claim 5, characterized in that: The fine-grained interaction analysis module includes: A fine-grained common feature extraction unit, used for performing fine-grained common feature extraction on the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram to obtain a set of multi-modal joint fine-grained core feature vectors of device states; A joint feature semantic clustering unit, used for inputting the set of the multimodal joint fine-grained core feature vectors of the device state into a clustering network to obtain a device state multimodal joint feature semantic clustering center vector; A global semantic interaction unit is used to guide the set of the device state multimodal joint fine-grained core feature vectors to perform cross-domain context semantic interaction based on the device state multimodal joint feature semantic clustering center vector to obtain the device state multimodal fine-grained joint encoding feature map.

7. The substation secondary equipment operation monitoring system according to claim 6, characterized in that: The fine-grained common feature extraction unit is used to: Performing feature fine-grained decoupling on the real-time current-real-time temperature timing correlation feature diagram and the real-time voltage-real-time temperature timing correlation feature diagram to obtain a set of real-time current-real-time temperature local timing correlation feature vectors and a set of real-time voltage-real-time temperature local timing correlation feature vectors; Each corresponding group of real-time current-real-time temperature local timing association feature vectors and real-time voltage-real-time temperature local timing association feature vectors in the set of real-time current-real-time temperature local timing association feature vectors and the set of real-time voltage-real-time temperature local timing association feature vectors are input into a fine-grained common feature extraction network to obtain a set of multi-modal joint fine-grained core feature vectors of the device state.

8. The substation secondary equipment operation monitoring system according to claim 7, characterized in that: The global semantic interaction unit is used to: Constructing a query vector and a value vector based on each device state multimodal joint fine-grained core feature vector in the set of the device state multimodal joint fine-grained core feature vectors, and constructing a key vector based on the device state multimodal joint feature semantic clustering center vector, and inputting the query vector, the value vector and the key vector into a global semantic interaction encoder based on a converter structure to obtain a set of context device state multimodal joint fine-grained core feature vectors; The feature shape of each context device state multimodal joint fine-grained core feature vector in the set of the context device state multimodal joint fine-grained core feature vectors is reshaped to obtain the device state multimodal fine-grained joint encoding feature map.

9. The substation secondary equipment operation monitoring system according to claim 8, characterized in that: The abnormal operation status identification module is used to: The equipment state multimodal fine-grained joint encoding feature map is input into a classifier-based operating state identifier to obtain an identification result, and the identification result is used to indicate whether the operating state of the monitored substation secondary equipment is abnormal.

10. A method for monitoring the operation of secondary equipment in a substation, characterized in that: include: A time queue for collecting operating status data of the monitored substation secondary equipment, wherein the operating status data includes real-time current value, real-time voltage value and real-time temperature value; Requeue the time queue of the operating status data to obtain a time queue of real-time current values, a time queue of real-time voltage values, and a time queue of real-time temperature values; Real-time current timing implicit features, real-time voltage timing implicit features and real-time temperature timing implicit features are extracted from the time queue of the real-time current value, the time queue of the real-time voltage value and the time queue of the real-time temperature value, and based on the real-time current timing implicit features, the real-time voltage timing implicit features and the real-time temperature timing implicit features, it is determined whether there is any abnormality in the operating status of the monitored substation secondary equipment.

Images