An intelligent auxiliary control system for a substation
By combining environmental perception, intelligent image recognition, and equipment status monitoring modules, comprehensive monitoring and multi-source data fusion of substations are achieved, solving the monitoring blind spots and data isolation problems of traditional substation monitoring systems, and improving the safety and reliability of substations.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ENERGIEDATEN TECH (SHANGHAI) CO LTD
- Filing Date
- 2026-01-30
- Publication Date
- 2026-06-12
AI Technical Summary
Traditional substation monitoring systems suffer from problems such as limited monitoring methods, isolated data, low level of intelligence, and insufficient early warning capabilities. They are unable to achieve comprehensive monitoring of the environment and equipment status, leading to missed and false judgments and making it difficult to detect potential safety hazards in advance.
The system employs an environmental perception module, an intelligent image recognition and analysis module, and an intelligent equipment status monitoring module for comprehensive monitoring. It combines data fusion and intelligent analysis modules to perform multi-source data fusion and utilizes convolutional neural networks for image recognition and time series prediction to achieve intelligent analysis and hierarchical early warning.
It enables comprehensive monitoring of substations, improves the comprehensiveness and accuracy of monitoring, reduces manual intervention, increases response speed and judgment accuracy, and ensures the safe and stable operation of substations.
Smart Images
Figure CN122203554A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent monitoring technology for substations, specifically to an intelligent auxiliary control system for substations. Background Technology
[0002] Substations are an important component of the power system, undertaking the key tasks of voltage transformation, power distribution and control. The safe and stable operation of substations is directly related to the reliability and power quality of the power system. With the continuous expansion of the power system and the continuous growth of electricity load, the operating environment of substations is becoming increasingly complex and the number of equipment is constantly increasing, which puts forward higher requirements for the monitoring and management of substations. Traditional substation monitoring mainly relies on manual inspections and simple sensor monitoring, which has the following shortcomings: First, the monitoring methods are limited and can only monitor some key parameters, making it impossible to achieve comprehensive monitoring of the environment and equipment status, which can easily lead to monitoring blind spots. Second, data is isolated, and there is a lack of effective data sharing and correlation analysis between various monitoring systems, making it impossible to comprehensively and accurately assess the overall operating status of the substation. Third, the level of intelligence is low, relying mainly on manual judgment and experience analysis, resulting in slow response speed and susceptibility to human factors, leading to missed judgments and misjudgments. Fourth, insufficient early warning capabilities make it difficult to detect potential safety hazards in advance; problems are often only discovered after equipment failure or safety accidents occur. Therefore, there is an urgent need for an intelligent auxiliary control system for substations that can achieve comprehensive monitoring, multi-source data fusion, intelligent analysis and judgment, and timely and accurate early warning, in order to improve the safety and reliability of substation operation. Summary of the Invention
[0003] The purpose of this invention is to provide an intelligent auxiliary control system for substations.
[0004] To achieve the above objectives, the present invention provides the following technical solution: an intelligent auxiliary control system for substations, comprising: An environmental sensing module is used to collect environmental parameters within the substation, including a temperature sensor, a humidity sensor, an SF6 gas concentration sensor, a smoke sensor, and a water immersion sensor. The temperature sensor and humidity sensor collect ambient temperature and ambient humidity, respectively. The SF6 gas concentration sensor collects SF6 gas concentration values. The smoke sensor detects smoke signals, and the water immersion sensor detects water immersion signals. The intelligent image recognition and analysis module includes multiple cameras and an image processing unit. The cameras are distributed at key locations in the substation to collect image data. The image processing unit analyzes the image data to identify switch status information and personnel intrusion information. The intelligent equipment status monitoring module is used to monitor the operating status of substation equipment. It includes an electrical parameter monitoring unit, a mechanical performance monitoring unit, and an insulation performance monitoring unit. The electrical parameter monitoring unit collects voltage, current, and power parameters; the mechanical performance monitoring unit collects equipment vibration and displacement parameters; and the insulation performance monitoring unit collects insulation resistance and partial discharge parameters. The data fusion and intelligent analysis module, connected to the environmental perception module, intelligent image recognition and analysis module, and intelligent equipment status monitoring module, includes a data preprocessing unit, a data fusion unit, and an intelligent analysis unit. The data preprocessing unit cleans, normalizes, and converts the format of the raw data from each module. The data fusion unit uses a weighted fusion method to fuse the preprocessed data, where the weighting coefficients for the environmental perception data are... The weighting coefficient for image recognition data is 0.3. The weighting coefficient for device status data is 0.4. The value is 0.3, and the formula for calculating the fusion value is: ,in , , These are the integrated scores of normalized environmental perception data, image recognition data, and device status data, respectively. The intelligent analysis unit fuses these scores. The value is compared with a preset threshold of 0.8 to determine if there are any security risks. A value less than 0.8 indicates a potential safety hazard. The early warning module is connected to the data fusion and intelligent analysis module. When the intelligent analysis unit determines the fusion value... A warning signal will be issued when the value is less than 0.8, indicating a potential safety hazard.
[0005] As a further aspect of the present invention: the image processing unit uses a convolutional neural network model to analyze image data. The convolutional neural network model includes a feature extraction layer and a classification layer. The feature extraction layer extracts image features, and the classification layer outputs recognition results.
[0006] As a further aspect of the present invention: the data fusion and intelligent analysis module includes: The data preprocessing unit cleans, normalizes, and converts the format of the raw data from each module; The data fusion unit uses a weighted fusion method to fuse the preprocessed data. In the weighted fusion method, the weight coefficient of environmental perception data is 0.3, the weight coefficient of image recognition data is 0.4, and the weight coefficient of device status data is 0.3. The intelligent analysis unit analyzes the fused data and determines whether there are any anomalies by comparing the fused data with the preset normal operating data range.
[0007] As a further aspect of the present invention, the intelligent analysis unit also includes a trend prediction function, which establishes a time series prediction model based on historical data to predict the parameter change trend in future periods.
[0008] As a further aspect of the present invention: the early warning module includes: The graded early warning unit divides the early warning into three levels according to the severity of the safety hazard. When the monitored parameters exceed the normal range by less than 30%, it is a level 1 early warning; when they exceed the normal range by 30% to 60%, it is a level 2 early warning; and when they exceed the normal range by more than 60%, it is a level 3 early warning. The multi-channel notification unit issues warning signals through three methods: audible and visual alarms, SMS notifications, and application push notifications.
[0009] As a further aspect of the present invention: in the intelligent image recognition and analysis module, when the image processing unit recognizes the switch status information, it determines the switch's closed and open status by extracting the color and position features of the switch's closing indicator and opening indicator.
[0010] As a further aspect of the present invention: in the intelligent image recognition and analysis module, when the image processing unit identifies personnel intrusion information, it detects human targets in the image through a human target detection algorithm and determines whether the human target is located in a preset restricted area.
[0011] As a further aspect of the present invention: in the environmental sensing module, the detection range of the SF6 gas concentration sensor is 0 to 5000 ppm, and the response time is less than 10 seconds.
[0012] As a further aspect of the present invention: the auxiliary control system further includes a data storage module for storing the raw data collected by each module, the fusion analysis results and the early warning records. The data storage module uses a time-series database to store time-series data.
[0013] Compared with the prior art, the beneficial effects of the present invention by adopting the above technical solution are as follows: 1. Comprehensive monitoring coverage: Through the collaborative work of the environmental perception module, intelligent image recognition and analysis module, and intelligent equipment status monitoring module, comprehensive monitoring of substation environmental parameters, security conditions, and equipment operating status is achieved, eliminating monitoring blind spots and improving the comprehensiveness of monitoring.
[0014] 2. Multi-source data fusion: The data fusion and intelligent analysis module efficiently integrates massive amounts of data from different monitoring modules and uses a weighted fusion method to perform data fusion, which can comprehensively understand the substation's operating status from multiple dimensions and greatly improve the accuracy of judging equipment failures and safety hazards.
[0015] 3. High level of intelligence: It adopts a convolutional neural network model for image recognition and a time series prediction model for trend prediction, realizing intelligent analysis and intelligent early warning, reducing manual intervention, and improving response speed and judgment accuracy.
[0016] 4. Tiered early warning mechanism: Early warnings are issued in tiers based on the severity of safety hazards, and warning signals are sent through multiple channels to ensure that relevant personnel can be informed in a timely manner and take corresponding measures, effectively preventing the occurrence of safety accidents.
[0017] 5. High reliability: Through cross-validation of multi-source data, misjudgments and omissions caused by a single data source are effectively avoided, improving the reliability of the system and providing a solid technical guarantee for the safe and stable operation of the substation. Attached Figure Description
[0018] Figure 1 This is a schematic diagram of the overall architecture of the intelligent auxiliary control system for substations of the present invention; Figure 2 This is a schematic diagram of the structure of the environmental sensing module of the present invention; Figure 3 This is a schematic diagram of the intelligent image recognition and analysis module of the present invention; Figure 4 This is a schematic diagram of the intelligent device status monitoring module of the present invention; Figure 5 This is a schematic diagram of the data fusion and intelligent analysis module of the present invention; Figure 6 This is a schematic diagram of the system workflow of the present invention. Detailed Implementation
[0019] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.
[0020] Example 1, such as Figure 1 As shown, this embodiment provides an intelligent auxiliary control system for substations, including an environmental perception module 1, an intelligent image recognition and analysis module 2, an intelligent equipment status monitoring module 3, a data fusion and intelligent analysis module 4, an early warning module 5, and a data storage module 6.
[0021] I. Environmental Perception Module like Figure 2As shown, the environmental sensing module 1 is used to collect environmental parameters within the substation, including a temperature sensor 11, a humidity sensor 12, an SF6 gas concentration sensor 13, a smoke sensor 14, and a water immersion sensor 15.
[0022] Temperature sensor 11 is a Pt100 platinum resistance temperature sensor with a measurement range of -40℃ to 85℃ and a measurement accuracy of ±0.5℃. It is installed in the substation distribution room, switch room and main control room to collect ambient temperature data with a collection cycle of 30 seconds.
[0023] The humidity sensor 12 is a capacitive humidity sensor with a measurement range of 0%RH to 100%RH and a measurement accuracy of ±3%RH. It is installed in conjunction with the temperature sensor 11 to collect ambient humidity data with a collection cycle of 30 seconds.
[0024] The SF6 gas concentration sensor 13 has a detection range of 0 to 5000 ppm and a response time of less than 10 seconds. It uses the infrared absorption principle for detection and is installed near SF6 gas-insulated switchgear to monitor SF6 gas leakage in real time. The acquisition cycle is 10 seconds. When the SF6 gas concentration exceeds 1000 ppm, it is determined that there is a risk of leakage (according to GB / T 8905-2018 standard).
[0025] The smoke sensor 14 is a photoelectric smoke sensor with a sensitivity level of 1. It is installed on the ceiling of each room in the substation to detect smoke signals. The acquisition cycle is 5 seconds, and it reports immediately when a smoke signal is detected.
[0026] The water immersion sensor 15 is a conductivity-type water immersion sensor, which is installed in the cable trench of the substation, on the ground of the power distribution room and around the equipment foundation to detect water immersion signals. The acquisition cycle is 5 seconds, and the sensor reports immediately when a water immersion signal is detected.
[0027] II. Intelligent Image Recognition and Analysis Module like Figure 3 As shown, the intelligent image recognition and analysis module 2 includes multiple cameras 21 and an image processing unit 22, which is connected to the cameras 21 via a network.
[0028] Camera 21 is a network camera with a resolution of 1920×1080 pixels, supports H.264 video encoding, and is installed in key locations such as the main transformer area, high-voltage switch room, power distribution room, main control room entrance and perimeter of the substation, with a total of 15 points, for collecting image data, with a frame rate of 25 frames per second.
[0029] The image processing unit 22 uses an edge computing server, equipped with an Intel Xeon processor and an NVIDIA GPU accelerator card, to analyze image data and identify switch status information and personnel intrusion information.
[0030] The image processing unit 22 uses a convolutional neural network model to analyze image data. The convolutional neural network model is based on the ResNet-50 architecture and includes a feature extraction layer and a classification layer. The feature extraction layer includes 5 convolutional blocks, each of which consists of multiple convolutional layers, batch normalization layers, and activation function layers, used to extract multi-level features of the image. The classification layer includes a global average pooling layer and a fully connected layer, and outputs the recognition result.
[0031] When identifying switch status information, the image processing unit 22 determines the switch's closed and open status by extracting the color and position features of the switch's closed and open indicator lights. Specifically, it first converts the RGB image to the HSV color space through color space conversion, then sets the HSV threshold range for red and green, extracts the red and green indicator light areas, determines the closed status when the green indicator light is lit, determines the open status when the red indicator light is lit, and determines the abnormal status when neither is lit. The specific switch number is determined by the area location information.
[0032] When identifying intrusion information, the image processing unit 22 detects human targets in the image using a human target detection algorithm and determines whether the human target is located in a preset restricted area. The human target detection algorithm uses the YOLO v5 target detection model, which can detect human targets in the image in real time and output the bounding box coordinates and confidence level of the target. The preset restricted area is set by marking polygonal areas in the image, including high-voltage equipment area, a 3-meter range around the main transformer, and unauthorized area. When the center point of the bounding box of the detected human target is located in the restricted area and the confidence level is greater than 0.7, it is determined to be an intrusion and an intrusion warning is immediately generated.
[0033] III. Intelligent Equipment Status Monitoring Module like Figure 4 As shown, the intelligent equipment status monitoring module 3 is used to monitor the operating status of substation equipment, including an electrical parameter monitoring unit 31, a mechanical performance monitoring unit 32, and an insulation performance monitoring unit 33.
[0034] The electrical parameter monitoring unit 31 uses a digital power acquisition device to collect voltage, current and power parameters. The voltage measurement range is 0 to 220kV with a measurement accuracy of 0.2 class. The current measurement range is 0 to 5000A with a measurement accuracy of 0.2 class. The power measurement accuracy is 0.5 class. The acquisition cycle is 1 second. When the voltage deviation exceeds 10% of the rated value or the current exceeds 120% of the rated value, it is determined that the electrical parameters are abnormal.
[0035] The mechanical performance monitoring unit 32 includes a vibration sensor and a displacement sensor. The vibration sensor is a piezoelectric accelerometer with a measurement range of 0 to 10g and a frequency response range of 1Hz to 10kHz. It is installed in key parts of the main transformer, high-voltage switchgear, and circuit breaker to collect equipment vibration data with a collection period of 0.1 seconds. The displacement sensor is a laser displacement sensor with a measurement range of 0 to 500mm and a measurement accuracy of 0.1mm. It is used to monitor the displacement changes of the equipment. When the effective value of the vibration acceleration exceeds 2g or the displacement exceeds 5mm, it is determined to be an abnormality in mechanical performance.
[0036] The insulation performance monitoring unit 33 includes an insulation resistance tester and a partial discharge detection device. The insulation resistance tester periodically measures the insulation resistance of the equipment. The measurement voltage is 2500V and the measurement range is 0 to 200GΩ. When the insulation resistance is less than 10MΩ, it is determined that the insulation performance has deteriorated. The partial discharge detection device uses a combination of ultrasonic sensors and ultra-high frequency sensors to monitor the partial discharge signal of the equipment in real time. When the local discharge exceeds 500pC, it is determined that there is a defect in the insulation.
[0037] IV. Data Fusion and Intelligent Analysis Module like Figure 5 As shown, the data fusion and intelligent analysis module 4 is connected to the environmental perception module 1, the intelligent image recognition and analysis module 2, and the equipment status intelligent monitoring module 3. It is used to receive and fuse data from each module, obtain the comprehensive operating status of the substation through data correlation analysis, and determine whether there are any safety hazards based on preset thresholds.
[0038] The data fusion and intelligent analysis module 4 includes a data preprocessing unit 41, a data fusion unit 42, and an intelligent analysis unit 43.
[0039] The data preprocessing unit 41 cleans, normalizes, and converts the format of the raw data from each module. Data cleaning includes outlier removal, missing value imputation, and smoothing. Outlier detection uses the 3σ principle; data points deviating from the mean by more than three times the standard deviation are identified as outliers and removed. Missing value imputation uses linear interpolation. Normalization uses the min-max normalization method, mapping data of different dimensions to the range of 0 to 1. The normalization formula is: ; in, The value is the normalized value. These are the original values. This is the theoretical minimum value of the parameter. This is the theoretical maximum value of the parameter, determined based on historical operating data of the equipment. and And it is updated regularly.
[0040] Format conversion transforms data from different formats into standard JSON structured data.
[0041] The data fusion unit 42 uses a weighted fusion method to fuse the preprocessed data. In the weighted fusion method, the weight coefficient for environmental perception data is 0.3, the weight coefficient for image recognition data is 0.4, and the weight coefficient for device status data is 0.3. The formula for calculating the fusion value is as follows: ; in, For fusion value, This represents the weighting coefficient for the environmental perception data, with a value of 0.3. The comprehensive score is based on the normalized environmental perception data. This represents the weighting coefficient for the image recognition data, with a value of 0.4. The normalized image recognition data is used for comprehensive scoring. This is the weighting coefficient for the device status data, with a value of 0.3. This is a comprehensive score for the normalized equipment status data.
[0042] The comprehensive score for each data source is calculated based on the degree of abnormality of each monitored parameter within that data source. When a parameter is within the normal range, the score for that parameter is 1. When a parameter exceeds the normal range, the score decreases linearly according to the degree of exceedance, and the score is 0 when it exceeds the normal range by 100%.
[0043] The intelligent analysis unit 43 analyzes the fused data by comparing it with a preset normal operating data range to determine if there are any anomalies. The fused value is the normal operating value. It should be in the range of 0.8 to 1.0 when the fusion value If the value is less than 0.8, an anomaly is identified, and detailed analysis is required to pinpoint the source of the anomaly.
[0044] The intelligent analysis unit 43 also includes a trend prediction function. It establishes a time series prediction model based on historical data to predict the parameter change trend in future periods. The time series prediction model uses an LSTM long short-term memory network. The input is historical data of the past 24 hours, and the output is the predicted data for the next 2 hours. Through trend prediction, abnormal change trends of parameters can be detected in advance, enabling preventive maintenance.
[0045] V. Early Warning Module The early warning module 5 is connected to the data fusion and intelligent analysis module 4, and issues an early warning signal when it is determined that there is a potential safety hazard.
[0046] The early warning module 5 includes a tiered early warning unit 51 and a multi-channel notification unit 52.
[0047] The graded early warning unit 51 divides the early warning into three levels according to the severity of the safety hazard. When the monitored parameters exceed the normal range by less than 30%, it is a level 1 early warning, indicating a minor abnormality that requires attention. When the monitored parameters exceed the normal range by 30% to 60%, it is a level 2 early warning, indicating a moderate abnormality that requires timely handling. When the monitored parameters exceed the normal range by more than 60%, it is a level 3 early warning, indicating a serious abnormality that requires immediate handling.
[0048] The multi-channel notification unit 52 issues warning signals through three methods: audible and visual alarms, SMS notifications, and application push notifications. The audible and visual alarm device is installed in the main control room and automatically activates when an alarm occurs, emitting sound and flashing lights. The SMS notification function sends warning SMS messages to the preset mobile phone numbers of maintenance personnel through the SMS gateway. The SMS content includes the warning level, abnormal parameters, abnormal values, and occurrence time. The application push function pushes warning messages to the mobile phones of maintenance personnel through a mobile application. The message content is the same as the SMS notification, and detailed monitoring data and historical records can be viewed through the application.
[0049] VI. Data Storage Module Data storage module 6 is used to store the raw data collected by each module, the fusion analysis results and the early warning records. Data storage module 6 uses the time series database InfluxDB to store time series data and the relational database MySQL to store configuration information and early warning records. The data retention strategy of the time series database is as follows: raw data is retained for 90 days, hourly aggregated data is retained for 1 year, and daily aggregated data is retained for 5 years.
[0050] VII. System Workflow like Figure 6 As shown, the workflow of the intelligent auxiliary control system for substations of this invention is as follows: Step S1: The environmental perception module 1, intelligent image recognition and analysis module 2, and intelligent equipment status monitoring module 3 are started simultaneously to begin collecting data; Step S2: Each monitoring module transmits the collected data to the data fusion and intelligent analysis module 4 in real time; Step S3: The data preprocessing unit 41 cleans, normalizes, and converts the format of the received raw data; Step S4: The data fusion unit 42 performs weighted fusion on the preprocessed data and calculates the fusion value; Step S5: The intelligent analysis unit 43 analyzes the fused value to determine whether there are any anomalies; Step S6: If no abnormality is found, the data is stored in the data storage module 6, and the process returns to step S1 to continue monitoring. If an abnormality is found, the process proceeds to step S7. Step S7: The graded early warning unit 51 determines the early warning level based on the severity of the anomaly; Step S8: The multi-channel notification unit 52 issues a warning signal through three methods: audible and visual alarm, SMS notification, and application push. Step S9: Store the warning information in data storage module 6, and return to step S1 to continue monitoring.
[0051] Through the above process, the present invention realizes real-time monitoring, intelligent analysis and timely early warning of substation operation status, effectively improving the safety and reliability of substation operation.
[0052] Example 2: This example, based on Example 1, provides a detailed description of the trend prediction function of the intelligent analysis unit.
[0053] The LSTM (Long Short-Term Memory) network model consists of an input layer, an LSTM layer, and an output layer. The input layer receives historical data from the past 24 hours, with one time step per hour, for a total of 24 time steps. Each time step contains 20 feature parameters, including ambient temperature, ambient humidity, SF6 gas concentration, equipment voltage, equipment current, equipment power, vibration acceleration value, etc.
[0054] The LSTM layer consists of two LSTM unit layers. The first layer has 128 neurons, and the second layer has 64 neurons. Each LSTM unit includes an input gate, a forget gate, and an output gate, which can effectively capture long-term dependencies in time series data.
[0055] The output layer is a fully connected layer that outputs the predicted data for the next 2 hours. There is one time step per hour, for a total of 2 time steps. Each time step contains the same 20 feature parameters as the input.
[0056] The model was trained using monitoring data from the past 6 months, with a training set to validation set ratio of 8:2. The loss function used was mean squared error (MSE), the optimization algorithm used was the Adam optimizer, the learning rate was set to 0.001, the batch size was set to 32, and the number of training epochs was 100.
[0057] In practical applications, a prediction is made every hour to forecast the parameter change trend for the next two hours. When the predicted value shows that a certain parameter will exceed the normal range in the future, an early warning is issued to remind maintenance personnel to pay attention and take preventive measures, thereby achieving preventive maintenance and avoiding failures.
[0058] Through trend prediction, this invention can detect abnormal changes in parameters in advance, transforming from a passive response to an active prevention, thereby further improving the safety of substation operation.
[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. A substation intelligent auxiliary control system, characterized in that: include: An environmental sensing module is used to collect environmental parameters within the substation, including a temperature sensor, a humidity sensor, an SF6 gas concentration sensor, a smoke sensor, and a water immersion sensor. The temperature sensor and humidity sensor collect ambient temperature and ambient humidity, respectively. The SF6 gas concentration sensor collects SF6 gas concentration values. The smoke sensor detects smoke signals, and the water immersion sensor detects water immersion signals. The intelligent image recognition and analysis module includes multiple cameras and an image processing unit. The cameras are distributed at key locations in the substation to collect image data. The image processing unit analyzes the image data to identify switch status information and personnel intrusion information. The intelligent equipment status monitoring module is used to monitor the operating status of substation equipment. It includes an electrical parameter monitoring unit, a mechanical performance monitoring unit, and an insulation performance monitoring unit. The electrical parameter monitoring unit collects voltage, current, and power parameters; the mechanical performance monitoring unit collects equipment vibration and displacement parameters; and the insulation performance monitoring unit collects insulation resistance and partial discharge parameters. The data fusion and intelligent analysis module, connected to the environmental perception module, intelligent image recognition and analysis module, and intelligent equipment status monitoring module, includes a data preprocessing unit, a data fusion unit, and an intelligent analysis unit. The data preprocessing unit cleans, normalizes, and converts the format of the raw data from each module. The data fusion unit uses a weighted fusion method to fuse the preprocessed data, where the weighting coefficients for the environmental perception data are... The weighting coefficient for image recognition data is 0.
3. The weighting coefficient for device status data is 0.
4. The value is 0.3, and the formula for calculating the fusion value is: ,in , , These are the integrated scores of normalized environmental perception data, image recognition data, and device status data, respectively. The intelligent analysis unit fuses these scores. The value is compared with a preset threshold of 0.8 to determine if there are any security risks. A value less than 0.8 indicates a potential safety hazard. The early warning module is connected to the data fusion and intelligent analysis module. When the intelligent analysis unit determines the fusion value... A warning signal will be issued when the value is less than 0.8, indicating a potential safety hazard.
2. The intelligent auxiliary control system for substations according to claim 1, characterized in that: The image processing unit uses a convolutional neural network model to analyze image data. The convolutional neural network model includes a feature extraction layer and a classification layer. The feature extraction layer extracts image features, and the classification layer outputs recognition results.
3. The intelligent auxiliary control system for substations according to claim 1, characterized in that: The data fusion and intelligent analysis module includes: The data preprocessing unit cleans, normalizes, and converts the format of the raw data from each module; The data fusion unit uses a weighted fusion method to fuse the preprocessed data. In the weighted fusion method, the weight coefficient of environmental perception data is 0.3, the weight coefficient of image recognition data is 0.4, and the weight coefficient of device status data is 0.
3. The intelligent analysis unit analyzes the fused data and determines whether there are any anomalies by comparing the fused data with the preset normal operating data range.
4. The intelligent auxiliary control system for substations according to claim 1, characterized in that: The intelligent analysis unit also includes a trend prediction function, which establishes an LSTM (Long Short-Term Memory) network model based on historical data. The LSTM network model takes historical data from the past 24 hours as input and outputs predicted data for the next 2 hours, predicting the parameter change trend in the future period.
5. The intelligent auxiliary control system for a substation according to claim 1, characterized in that: The early warning module includes: The graded early warning unit divides the early warning into three levels according to the severity of the safety hazard. When the monitored parameters exceed the normal range by less than 30%, it is a level 1 early warning; when they exceed the normal range by 30% to 60%, it is a level 2 early warning; and when they exceed the normal range by more than 60%, it is a level 3 early warning. The multi-channel notification unit issues warning signals through three methods: audible and visual alarms, SMS notifications, and application push notifications.
6. The intelligent auxiliary control system for substations according to claim 1, characterized in that: In the intelligent image recognition and analysis module, when the image processing unit recognizes the switch status information, it determines the switch's closed and open status by extracting the color and position features of the switch's closing and opening indicator lights.
7. The intelligent auxiliary control system for substations according to claim 1, characterized in that: In the intelligent image recognition and analysis module, when the image processing unit identifies personnel intrusion information, it detects human targets in the image through a human target detection algorithm and determines whether the human target is located in a preset restricted area.
8. The intelligent auxiliary control system for a substation according to claim 1, characterized in that: In the environmental sensing module, the SF6 gas concentration sensor has a detection range of 0 to 5000 ppm and a response time of less than 10 seconds.
9. The intelligent auxiliary control system for a substation according to claim 1, characterized in that: The auxiliary control system also includes a data storage module for storing the raw data collected by each module, the fusion analysis results, and the early warning records. The data storage module uses a time-series database to store time-series data.