Power internet-of-things transformer substation auxiliary equipment monitoring system and method

Abstract

The invention relates to the technical field of monitoring systems, in particular to an electric power internet of things transformer substation auxiliary equipment monitoring system and method, and the method comprises the steps: collecting the operation data of transformer substation equipment in real time; preprocessing the collected operation data by utilizing edge calculation; analyzing the preprocessed operation data based on a deep learning algorithm, and generating a preliminary fault diagnosis and prediction result; constructing a digital twinborn model of the substation equipment, and performing equipment health assessment and fault prediction by utilizing a preliminary fault diagnosis and prediction result and combining a physical model and historical data of the equipment; and according to the equipment health assessment and fault prediction result, data visualization display is carried out through the monitoring platform, and hierarchical alarm is carried out to inform operation and maintenance personnel. Through the deep learning algorithm and the digital twinborn model, deeper fault analysis and prediction can be provided, and the problem that the equipment is not maintained in time or is excessively maintained is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of monitoring systems, and in particular to a monitoring system and method for auxiliary equipment in a power Internet of Things substation. Background Art

[0002] There are many types of auxiliary systems in a substation, but each system only realizes some basic functions such as partial data collection, equipment monitoring, and auxiliary inspection. The various systems are relatively independent, and there is still much room for improvement in advanced applications such as data sharing and information interaction.

[0003] The existing publication number CN215956107U provides a monitoring system for auxiliary equipment in a power Internet of Things substation, including a central processor. A substation equipment monitoring unit, a substation environment monitoring unit, a substation access control system, and an image monitoring system are connected to the central processor. A fire automatic alarm system and an alarm are connected to the output end of the central processor. The output end of the central processor is wirelessly connected to a monitoring platform, and the output end of the monitoring platform is wirelessly connected to a mobile terminal through a wireless signal. Through the setting of the substation equipment monitoring unit, it is used to monitor the working state signals of the equipment in the substation. Through the setting of the substation environment monitoring unit, the environment in the substation can be monitored. The substation access control system can judge the identity information of the personnel entering the substation. The structure is simple, and it can automatically monitor the equipment and environment in the substation in real time, ensuring the safety of the substation and having high practicability.

[0004] However, the above-mentioned monitoring system for auxiliary equipment in a power Internet of Things substation can only judge whether a fault occurs, but cannot provide more in-depth analysis and prediction, which may lead to untimely or excessive equipment maintenance and increase the operation and maintenance costs. Summary of the Invention

[0005] The purpose of the present invention is to provide a monitoring system and method for auxiliary equipment in a power Internet of Things substation, which solves the problem that the existing monitoring system for auxiliary equipment in a power Internet of Things substation can only judge whether a fault occurs, but cannot provide more in-depth analysis and prediction, which may lead to untimely or excessive equipment maintenance and increase the operation and maintenance costs.

[0006] To achieve the above purpose, the present invention provides a monitoring method for auxiliary equipment in a power Internet of Things substation, including the following steps:

[0007] Real-time collect the operation data of substation equipment;

[0008] Use edge computing to preprocess the collected operation data;

[0009] Analyze the preprocessed operation data based on a deep learning algorithm to generate preliminary fault diagnosis and prediction results;

[0010] Build a digital twin model of substation equipment, utilize the preliminary fault diagnosis and prediction results, combine the physical model and historical data of the equipment, and conduct equipment health assessment and fault prediction;

[0011] According to the equipment health assessment and fault prediction results, conduct data visualization display through the monitoring platform, and issue graded alarms to notify the operation and maintenance personnel;

[0012] Link the alarm system with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response.

[0013] Among them, the operation data of substation equipment is collected in real time, and the steps also include:

[0014] Collect the current, voltage, temperature, humidity, smoke concentration, liquid level height of substation equipment and the infrared thermal imaging data on the surface of the equipment in real time through the sensor network. Among them, the sensor network adopts redundant design and at least includes two groups of the same type of sensors. When a group of sensors fails or the data is abnormal, it automatically switches to another group of sensors;

[0015] Utilize multi-sensor data fusion technology, combine the data of the sensors, and generate a comprehensive description of the equipment operation status through the Bayesian fusion algorithm or the Kalman filtering algorithm;

[0016] Transmit the comprehensive equipment operation data to the edge computing device in real time.

[0017] Among them, use edge computing to preprocess the collected operation data, and the steps also include:

[0018] Receive the operation data, conduct data cleaning on the operation data, and remove the noise, outliers and duplicate data in the data;

[0019] Extract features from the cleaned data, including current peak value, voltage volatility, temperature change rate, number of times of humidity threshold exceeding, and sudden change value of smoke concentration;

[0020] Based on the preset rules, conduct preliminary analysis on the extracted feature parameters, judge whether the equipment operation is in a normal state, and generate a preliminary equipment status report;

[0021] Synchronize the cleaned data and the extracted feature parameters to the blockchain in real time;

[0022] Classify the data according to the equipment operation status and the preset priority rules, mark the high-priority data as needing to be processed first, and transmit the preprocessed data for further analysis.

[0023] Among them, analyze the preprocessed operation data based on the deep learning algorithm to generate preliminary fault diagnosis and prediction results, and the steps also include:

[0024] Input the preprocessed operation data into a pre-trained deep learning model, which includes a convolutional neural network and / or a recurrent neural network and its variants;

[0025] Classify and identify the extracted feature information through the deep learning model to determine whether there is a fault mode in the operation data, and determine the fault type, location and severity;

[0026] Use a generative adversarial network to generate fault data samples;

[0027] Dynamically adjust the parameters of the deep learning model based on the reinforcement learning algorithm;

[0028] Combine the historical operation data and real-time operation data of the device, and generate preliminary fault diagnosis and prediction results through the deep learning model;

[0029] Output the preliminary fault diagnosis and prediction results to the digital twin model.

[0030] Among them, to build a digital twin model of substation equipment, use the preliminary fault diagnosis and prediction results, combine the physical model and historical data of the equipment, and conduct equipment health assessment and fault prediction. The steps also include:

[0031] Build a digital twin model of substation equipment, which includes a virtual representation of the equipment's geometric structure, physical characteristics, electrical parameters, and operating status;

[0032] Synchronize the digital twin model with the physical model of the equipment in real time;

[0033] Import the generated preliminary fault diagnosis and prediction results into the digital twin model, and combine the historical operation data and real-time operation data of the equipment to comprehensively evaluate the health status of the equipment;

[0034] Based on the digital twin model, use machine learning algorithms or physical simulation technologies to predict the future operating status of the equipment, and generate a detailed fault prediction report, including the probability of fault occurrence, expected time, impact range, and possible fault modes;

[0035] Evaluate the improvement effect of different maintenance plans on the equipment health status through virtual maintenance simulation;

[0036] Dynamically adjust the parameters of the digital twin model according to the equipment health assessment and fault prediction results.

[0037] Among them, according to the equipment health assessment and fault prediction results, conduct data visualization display through the monitoring platform and send hierarchical alarm notifications to the operation and maintenance personnel. The steps also include:

[0038] Transmit the device health assessment and fault prediction results to the monitoring platform in real time. The monitoring platform includes a data visualization interface and an alarm management system;

[0039] On the data visualization interface of the monitoring platform, display the real-time operating status, health indicators, fault probability, predicted fault time, and impact range of the device in a graphical and tabular manner;

[0040] Classify the alarm information into multiple levels according to the severity and urgency of the fault prediction results, including level 1 alarm, level 2 alarm, and level 3 alarm.

[0041] Among them, according to the device health assessment and fault prediction results, perform data visualization display through the monitoring platform, and notify the operation and maintenance personnel by grading alarms. The steps also include:

[0042] For a level 1 alarm, the monitoring platform immediately triggers an audible and visual alarm, and notify the operation and maintenance personnel and relevant management personnel of the alarm information in real time by means of SMS, email, or push on the mobile terminal;

[0043] For a level 2 alarm, the monitoring platform triggers an audible alarm, and notify the operation and maintenance personnel of the alarm information by means of push on the mobile terminal or email;

[0044] For a level 3 alarm, the monitoring platform only notifies the operation and maintenance personnel of the alarm information by means of push on the mobile terminal to prompt them to pay attention to the device status;

[0045] The monitoring platform records the sending status and receiving confirmation of the alarm information to form an alarm log.

[0046] A monitoring system for auxiliary equipment in a power Internet of Things substation includes a collection module, a preprocessing module, a preliminary analysis module, an evaluation module, a monitoring platform, and an alarm module. The preprocessing module is connected to the collection module, the preliminary analysis module is connected to the preprocessing module, the evaluation module is connected to the preliminary analysis module, the monitoring platform is connected to the evaluation module, and the alarm module is connected to the monitoring platform;

[0047] The collection module is used to collect the operation data of substation equipment in real time;

[0048] The preprocessing module is used to preprocess the collected operation data by using edge computing;

[0049] The preliminary analysis module is used to analyze the preprocessed operation data based on deep learning algorithms to generate preliminary fault diagnosis and prediction results;

[0050] An evaluation module for building a digital twin model of substation equipment, using preliminary fault diagnosis and prediction results, combining the physical model and historical data of the equipment to conduct equipment health assessment and fault prediction;

[0051] A monitoring platform for visualizing data based on equipment health assessment and fault prediction results, and notifying operation and maintenance personnel by hierarchical alarms;

[0052] An alarm module for linking with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response.

[0053] A monitoring system and method for auxiliary equipment of a power Internet of Things substation according to the present invention. It collects the operation data of substation equipment in real time through a sensor network, preprocesses the collected operation data using edge computing, analyzes the preprocessed data based on deep learning algorithms such as convolutional neural networks, recurrent neural networks and their variants to generate preliminary fault diagnosis and prediction results, constructs a digital twin model of substation equipment, including a virtual representation of the geometric structure, physical characteristics, electrical parameters and operation status of the equipment, imports the preliminary fault diagnosis and prediction results into the digital twin model, combines historical and real-time operation data for equipment health assessment, uses machine learning algorithms or physical simulation techniques to predict the future operation status of the equipment, generates a detailed fault prediction report, visualizes the data through a monitoring platform, presents information such as the real-time operation status, health indicators, and fault probability of the equipment in a graphical and tabular manner, classifies the alarm information into multiple levels according to the severity and urgency of the fault prediction results, and notifies operation and maintenance personnel through different notification methods. Link the alarm system with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response. Through deep learning algorithms and digital twin models, it can provide more in-depth fault analysis and prediction, avoid problems such as untimely or excessive equipment maintenance. Significantly improves the intelligence level and reliability of the monitoring system for auxiliary equipment of a power Internet of Things substation, solves the deficiencies existing in the prior art, and has broad application prospects. Description of the Drawings

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art.

[0055] Figure 1 It is a flowchart of the steps of the method for monitoring auxiliary equipment of a power Internet of Things substation according to the first embodiment of the present invention.

[0056] Figure 2 It is a principle block diagram of the monitoring system for auxiliary equipment of a power Internet of Things substation according to the second embodiment of the present invention.

[0057] In the figure: 201 - acquisition module, 202 - pre - processing module, 203 - preliminary analysis module, 204 - evaluation module, 205 - monitoring platform, 206 - alarm module. Specific implementation mode

[0058] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to explain the present invention, and should not be construed as a limitation to the present invention.

[0059] The first embodiment of this application is:

[0060] Please refer to Figure 1 , where Figure 1 is the step - flow chart of the monitoring method for auxiliary equipment of a power Internet of Things substation in the first embodiment of the present invention.

[0061] The present invention provides a monitoring method for auxiliary equipment of a power Internet of Things substation, including the following steps:

[0062] S101: Real - time collect the operation data of substation equipment;

[0063] Specifically, deploy a high - precision sensor network inside the substation for real - time collection of equipment operation data. The sensor network includes current sensors, voltage sensors, temperature sensors, humidity sensors, smoke concentration sensors, liquid level sensors, and infrared thermal imaging sensors. The sensor network adopts a redundant design, and at least two groups of each type of sensor are deployed. When a group of sensors fails or data is abnormal, the system automatically switches to another group of sensors to ensure the continuity and reliability of data collection. The redundant design of the sensors determines whether the sensors are working properly by real - time monitoring the output data of the sensors and combining preset thresholds. For example, when the output data of a group of sensors exceeds the normal range or the data difference from another group of sensors exceeds the preset threshold, the system determines that this group of sensors is faulty and automatically switches to the standby sensor group. Each sensor collects the operation data of substation equipment in real - time according to a preset sampling frequency, such as 10 times per second. The sampling frequency can be dynamically adjusted according to the actual operation status and environmental conditions of the equipment to optimize the data collection efficiency.

[0064] The collected data is first pre - processed on the sensor nodes, including data formatting and preliminary calibration. For example, the analog signals collected by current sensors and voltage sensors are converted into digital signals through an analog - to - digital converter (ADC) and subjected to preliminary linear calibration to ensure the accuracy of the data. To improve the accuracy and reliability of the data, multi - sensor data fusion technology is adopted, combining data from different types of sensors to generate a comprehensive description of the device operating state. The data fusion uses Bayesian fusion algorithm or Kalman filter algorithm. The specific process is as follows: Bayesian fusion algorithm: For each type of sensor data, calculate its probability density function (PDF). For example, for a temperature sensor, calculate the probability density function of temperature based on historical data and current measurement values. Use Bayes' formula to fuse the probability density functions of different sensors. Suppose sensor A and sensor B measure the same physical quantity, and their probability density functions are P(A) and P(B) respectively, then the fused probability density function is: Through iterative calculation, gradually update the fused probability density function, and finally obtain a comprehensive description of the device operating state. Kalman filter algorithm: The Kalman filter realizes data fusion through two steps: state estimation and measurement update. First, predict the estimated value of the current state according to the dynamic model of the system (such as the physical model of the device). Then, combine the measurement value of the sensor and use the Kalman gain to correct the predicted value to obtain a more accurate state estimate. The comprehensive device operation data after data fusion processing is transmitted to the edge computing device in real - time through wired or wireless communication links. An encryption protocol (such as TLS / SSL) is used during the data transmission process to ensure data security. At the same time, data compression technology is used to reduce the transmission bandwidth requirement and improve the transmission efficiency. The edge computing device further pre - processes and stores the received data, providing basic data support for subsequent deep - learning analysis and digital - twin model synchronization. The efficient acquisition, fusion, and transmission of substation device operation data are realized, providing an accurate and reliable data basis for subsequent fault diagnosis, health assessment, and emergency response.

[0065] S102: Use edge computing to pre - process the collected operation data;

[0066] Specifically, the edge computing device receives the operation data transmitted from the sensor network, including current, voltage, temperature, humidity, smoke concentration, liquid level height, and infrared thermal imaging data. The received data is first temporarily stored in the local storage unit of the edge computing device for subsequent processing. The storage format uses a time series database to ensure the timeliness and integrity of the data. The collected data is filtered to remove high-frequency noise. For example, a low-pass filter is used to filter the current and voltage data to remove noise signals above a specific frequency (such as 50 Hz). Outliers are identified and removed through statistical analysis methods. For example, the mean and standard deviation of the data are calculated, and data points exceeding the mean ± 3 times the standard deviation are regarded as outliers and removed. Duplicate data records are detected and deleted. By comparing the timestamps and values of adjacent data points, if the timestamps are the same and the values are consistent, the duplicate data points are deleted. The peak value in the current data is calculated, that is, the maximum current value within a preset time window (such as 1 minute). The volatility of the voltage data is calculated, that is, the ratio of the standard deviation of the voltage change to the average value, which reflects the stability of the voltage. The change rate of the temperature data is calculated, that is, the amount of temperature change per unit time (such as the temperature change value per minute). The number of times the humidity data exceeds a preset threshold (such as 80%) is counted, which reflects the abnormal situation of the environmental humidity. Abrupt change values in the smoke concentration data are detected, that is, the change amount between adjacent data points exceeds a preset threshold (such as 10%). Preliminary analysis is performed on the extracted feature parameters based on preset rules. For example, if the current peak exceeds 120% of the rated current of the device, it is judged that the device may be overloaded; if the temperature change rate exceeds a preset threshold (such as 5 °C per minute), it is judged that the device may have an overheating risk. According to the preliminary analysis results, a device status report is generated, and the report content includes whether the device is operating normally, whether there is a potential failure risk, and recommended maintenance measures. The cleaned data and feature parameters are packaged into data blocks, and each data block contains a timestamp, data content, and hash value. The hash value of each data block is calculated to ensure the uniqueness and immutability of the data. The data blocks are broadcast to other nodes through the blockchain network to complete the distributed storage of the data. The consensus mechanism of the blockchain (such as PoW or PoS) ensures the authenticity and integrity of the data. The data is classified and processed according to the device operation status and preset priority rules. For example, high-priority data (such as abnormal status data such as device overload and overheating) is marked as requiring priority processing. High-priority data will be preferentially transmitted to the central processor for further analysis to ensure that key issues can be processed in a timely manner. Low-priority data (such as normal operation status data) will be regularly transmitted to the central processor for long-term storage and analysis. Edge computing technology is used to efficiently preprocess the collected operation data to ensure the quality and reliability of the data, providing basic data support for subsequent deep learning analysis and digital twin model synchronization.

[0067] S103: Analyze the preprocessed operation data based on deep learning algorithms to generate preliminary fault diagnosis and prediction results;

[0068] Specifically, select suitable deep learning models, such as Convolutional Neural Network (CNN), Recurrent Neural Network (RNN) and its variants (such as LSTM, GRU). The training of the model uses historical operation data and known fault data. Optimize the model parameters through optimization algorithms such as Backpropagation and Gradient Descent to ensure that the model can accurately identify fault patterns.

[0069] Input the preprocessed operation data into the pre-trained deep learning model. The model automatically extracts feature information and conducts classification and recognition. For example, the CNN model extracts local features of the data through convolutional layers and pooling layers, and then conducts classification through fully connected layers; the RNN model extracts temporal features of the data through recurrent layers, and then conducts classification through output layers. The model determines whether there are fault patterns in the operation data according to the extracted feature information, and determines the type, location and severity of the fault. The fault types include overload, overheat, short circuit, etc.; the fault location includes specific components or areas of the device; the fault severity is evaluated according to the intensity and duration of the fault pattern. Use Generative Adversarial Network (GAN) to generate fault data samples, enhance the training data set of the model, and optimize the model's recognition ability for rare fault patterns. GAN consists of a Generator and a Discriminator. The Generator generates new fault data samples, and the Discriminator judges the authenticity of the samples. Improve the quality of the generated samples through adversarial training. Dynamically adjust the parameters of the deep learning model based on reinforcement learning algorithms (such as Q-Learning, DeepQ-Network). According to the device operation status and historical data, optimize the accuracy and timeliness of fault prediction. The reinforcement learning algorithm learns the optimal parameter adjustment strategy through interaction with the environment, improving the adaptability and robustness of the model. Combine the historical operation data and real-time operation data of the device, and generate preliminary fault diagnosis and prediction results through the deep learning model. The fault diagnosis results include the probability of fault occurrence, the predicted occurrence time and the possible influence range; the prediction results include the future operation status of the device and potential fault risks. Output the preliminary fault diagnosis and prediction results to the digital twin model to provide data support for device health assessment and further fault prediction. The digital twin model conducts virtual simulation and analysis of the device status according to the output results, providing more comprehensive device health information. Use deep learning algorithms to efficiently analyze the preprocessed operation data, generate accurate fault diagnosis and prediction results, and provide strong support for the health assessment and maintenance of substation equipment.

[0070] S104: Build a digital twin model of substation equipment, and use the preliminary fault diagnosis and prediction results, combined with the physical model and historical data of the equipment, to conduct equipment health assessment and fault prediction;

[0071] Specifically, a digital twin model of substation equipment is constructed. The model includes the geometric structure, physical properties, electrical parameters, and virtual representation of the operating state of the equipment. The geometric structure is realized through 3D modeling technology (such as CAD models), accurately reflecting the physical appearance and internal structure of the equipment. Physical properties include material attributes, heat conduction coefficients, mechanical strengths, etc., obtained through experimental data or theoretical calculations. Electrical parameters include the rated voltage, current, power factor, etc., of the equipment, obtained from equipment nameplate data or factory technical documents. The virtual representation of the operating state is achieved through real-time data synchronization, reflecting the current working state of the equipment. Through Internet of Things technology, the digital twin model is synchronized with the physical model of the equipment in real time. The specific process is as follows: Deploy sensors on the physical equipment to collect real-time operating data of the equipment (such as current, voltage, temperature, etc.). Transmit the collected data to the digital twin model through wireless or wired communication. The digital twin model updates its virtual state according to the real-time data to ensure the consistency between the model and the physical equipment. Import the preliminary fault diagnosis and prediction results generated based on deep learning algorithms into the digital twin model. Combine the historical operating data and real-time operating data of the equipment to comprehensively evaluate the health status of the equipment. For example, evaluate the health index of the equipment by analyzing historical fault data and current operating data. Use the digital twin model to comprehensively evaluate the health status of the equipment. The evaluation indicators include the health index, failure probability, remaining service life, etc., of the equipment. The health index is calculated by comparing the key parameters (such as temperature, current, voltage) of the equipment with the normal operating range. The failure probability is calculated by analyzing historical fault data and current operating status, combined with the prediction results of the deep learning model. The remaining service life is estimated based on the cumulative operating time, historical fault frequency, and current health status of the equipment. Based on the digital twin model, use machine learning algorithms (such as time series analysis, random forest) or physical simulation technologies (such as finite element analysis, thermodynamic simulation) to predict the future operating state of the equipment. Generate a detailed fault prediction report, including the probability of failure occurrence, estimated time, scope of influence, and possible failure modes. The probability of failure occurrence is obtained through model prediction, and the estimated time is calculated by analyzing the fault development trend. The scope of influence is obtained by analyzing the connection relationship of the equipment and the system topology structure, and the possible failure modes are obtained from the classification results of the deep learning model. Conduct virtual maintenance simulation through the digital twin model to evaluate the improvement effect of different maintenance plans on the health status of the equipment. Virtual maintenance simulation includes operations such as replacing components and adjusting parameters, and evaluates the health index and failure probability of the equipment after maintenance through simulation calculations. According to the simulation results, select the optimal maintenance plan to provide decision support for actual maintenance. According to the equipment health assessment and fault prediction results, dynamically adjust the parameters of the digital twin model. The adjusted parameters include the physical properties, electrical parameters, and virtual representation of the operating state of the equipment.Dynamic adjustment is achieved through a feedback mechanism. Based on actual operation data and simulation results, the model parameters are updated in real time to ensure the accuracy and adaptability of the model. The digital twin model is used to comprehensively evaluate the health status of substation equipment and predict its future operation status, providing a scientific basis for equipment maintenance and management.

[0072] S105: According to the results of equipment health assessment and fault prediction, data visualization is performed through the monitoring platform 205, and classified alarms are sent to notify the operation and maintenance personnel.

[0073] Specifically, the results of equipment health assessment and fault prediction are transmitted to the monitoring platform 205 in real time. The monitoring platform 205 includes a data visualization interface and an alarm management system. Data transmission is achieved through wired or wireless communication links to ensure the real-time and accuracy of data. An encryption protocol (such as TLS / SSL) is used to protect data security during the transmission process. On the data visualization interface of the monitoring platform 205, the real-time operation status, health indicators, fault probability, predicted fault time, and influence range of the equipment are displayed in a graphical and tabular manner. According to the severity and urgency of the fault prediction results, the alarm information is divided into multiple levels, including level 1 alarm, level 2 alarm, and level 3 alarm. Level 1 alarm (severe fault): The monitoring platform 205 immediately triggers an audible and visual alarm, emitting sound and light signals through the alarm device to alert the on-site personnel. The alarm information is notified to the operation and maintenance personnel and relevant management personnel in real time by means of SMS, email, or mobile terminal push. The alarm information includes the fault type, predicted occurrence time, influence range, and recommended emergency measures. Level 2 alarm (medium fault): The monitoring platform 205 triggers an audible alarm, emitting a sound signal through the alarm device. The alarm information is notified to the operation and maintenance personnel by means of mobile terminal push or email. The alarm information includes the fault type, predicted occurrence time, influence range, and recommended maintenance measures. Level 3 alarm (minor fault or early warning): The monitoring platform 205 notifies the operation and maintenance personnel of the alarm information only by means of mobile terminal push, prompting them to pay attention to the equipment status. The alarm information includes the fault type, predicted occurrence time, influence range, and recommended observation measures. The monitoring platform 205 records the sending status and receiving confirmation of the alarm information to form an alarm log. The alarm log includes information such as the alarm time, alarm level, alarm content, sending method, recipient, and confirmation time. The alarm log is stored in the database of the monitoring platform 205, supporting query and statistical functions, which is convenient for operation and maintenance personnel to analyze and trace alarm events. Through the monitoring platform 205, the results of equipment health assessment and fault prediction are intuitively displayed, and classified alarm notifications are sent according to the severity of the fault, ensuring that the operation and maintenance personnel can respond in a timely manner and take corresponding measures to improve the operation safety and reliability of substation equipment.

[0074] S106: Link the alarm system with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response.

[0075] Specifically, the monitoring platform 205 receives the results of device health assessment and fault prediction in real time. When a fire risk is detected (such as excessive smoke concentration or abnormal temperature rise), a fire alarm signal is immediately triggered. The fire alarm signal is generated by the alarm management system of the monitoring platform 205 and sent to the automatic fire extinguishing system and the emergency evacuation system according to preset rules. After receiving the fire alarm signal from the monitoring platform 205, the automatic fire extinguishing system immediately starts the fire extinguishing procedure. The fire extinguishing procedure includes: automatically opening the fire sprinklers or releasing fire extinguishing agents (such as carbon dioxide, dry powder, etc.) to extinguish the fire in the fire area. Cutting off the power supply of electrical equipment in the fire area to prevent the fire from spreading. Adjusting the ventilation system, closing the air intake passage in the fire area to reduce oxygen supply, and at the same time opening the smoke exhaust passage to discharge the smoke. The operating status of the fire extinguishing system is real-time fed back to the monitoring platform 205 through sensors to ensure the effectiveness and safety of the fire extinguishing operation. After receiving the fire alarm signal from the monitoring platform 205, the emergency evacuation system immediately starts the evacuation procedure. The evacuation procedure includes: emitting an alarm sound and flashing lights through the audible and visual alarm to remind people to evacuate. Playing voice prompts through the broadcast system to guide the evacuation direction and evacuation route of people. Lighting up the evacuation indicator lights to guide people to evacuate along the safe passage. Automatically opening the access control system of the evacuation passage to ensure that people can evacuate quickly. The operating status of the emergency evacuation system is real-time fed back to the monitoring platform 205 through sensors to ensure the effectiveness and safety of the evacuation operation. The monitoring platform 205 monitors the operating status of the automatic fire extinguishing system and the emergency evacuation system in real time, and obtains real-time data such as the temperature, smoke concentration, and personnel evacuation situation in the fire area through the sensor network. According to the real-time data, the monitoring platform 205 dynamically adjusts the fire extinguishing and evacuation strategies. For example: if the temperature in the fire area continues to rise, the monitoring platform 205 can increase the release amount of the fire extinguishing agent or adjust the angle of the fire sprinklers. If the evacuation passage is blocked, the monitoring platform 205 can re-guide the evacuation direction of people through the broadcast system. The monitoring platform 205 records the whole process of the emergency response, including the fire alarm time, the start time of the fire extinguishing system, the start time of the evacuation system, the execution situation of the fire extinguishing and evacuation operations, and the recovery time, to form an emergency response log. When the fire is extinguished or the danger is lifted, the monitoring platform 205 sends a stop instruction to the automatic fire extinguishing system and the emergency evacuation system. The automatic fire extinguishing system shuts down the fire extinguishing device and restores the ventilation system to the normal state. The emergency evacuation system shuts down the audible and visual alarm and the voice prompt system and restores the access control system to the normal state. The monitoring platform 205 records the end time of the emergency response and restores the system to the normal operating state. It realizes the linkage of the alarm system with the automatic fire extinguishing system and the emergency evacuation system, ensuring that it can quickly respond and take effective emergency measures in case of emergencies such as fires, and protecting the safety of personnel and equipment.

[0076] Analyzing the device operation data using convolutional neural networks (CNNs), recurrent neural networks (RNNs), and their variants can accurately identify fault patterns and predict the time and probability of faults occurring. The historical operation data and real-time data of the device are used for device health assessment through a digital twin model to detect potential faults in advance, reducing device downtime and maintenance costs. Real-time preprocessing of the collected operation data is performed through edge computing, including data cleaning, feature extraction, and preliminary analysis, reducing the amount of data transmitted and improving the system's response speed. The Bayesian fusion algorithm or Kalman filter algorithm is used to fuse multi-source data to generate a comprehensive description of the device operation state, improving the accuracy and reliability of the data. It solves the problem in the prior art that only the occurrence of faults can be judged, and provides more in-depth fault analysis and prediction through deep learning and digital twin technologies, avoiding untimely or excessive device maintenance.

[0077] The second embodiment of this application is:

[0078] Based on the first embodiment, please refer to Figure 2 , where Figure 2 is the principle block diagram of the power Internet of Things substation auxiliary equipment monitoring system according to the second embodiment of the present invention.

[0079] A power Internet of Things substation auxiliary equipment monitoring system in this embodiment includes a collection module 201, a preprocessing module 202, a preliminary analysis module 203, an evaluation module 204, a monitoring platform 205, and an alarm module 206.

[0080] For this specific embodiment, the preprocessing module 202 is connected to the collection module 201, the preliminary analysis module 203 is connected to the preprocessing module 202, the evaluation module 204 is connected to the preliminary analysis module 203, the monitoring platform 205 is connected to the evaluation module 204, and the alarm module 206 is connected to the monitoring platform 205;

[0081] The collection module 201 is used to collect the operation data of substation equipment in real time;

[0082] The preprocessing module 202 is used to preprocess the collected operation data using edge computing;

[0083] The preliminary analysis module 203 is used to analyze the preprocessed operation data based on deep learning algorithms to generate preliminary fault diagnosis and prediction results;

[0084] The evaluation module 204 is used to build a digital twin model of substation equipment, and use the preliminary fault diagnosis and prediction results, combined with the physical model and historical data of the equipment, to perform equipment health assessment and fault prediction;

[0085] The monitoring platform 205 is used to perform data visualization display based on the device health assessment and fault prediction results, and notify the operation and maintenance personnel of hierarchical alarms;

[0086] The alarm module 206 is used to be linked with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response.

[0087] Using a monitoring system for auxiliary equipment of a power Internet of Things substation according to this embodiment, the acquisition module 201 collects the operation data of substation equipment in real time, including current, voltage, temperature, humidity, smoke concentration, liquid level height, and infrared thermal imaging data on the surface of the equipment. The sensor network adopts a redundant design to ensure the reliability and continuity of data acquisition. The collected operation data is transmitted to the preprocessing module 202, and edge computing technology is used for data cleaning, feature extraction, and preliminary analysis. The preprocessed data is synchronized to the blockchain to ensure the authenticity and integrity of the data. The preprocessed data is input into the preliminary analysis module 203, and the data is analyzed based on deep learning algorithms (such as CNN, RNN, and their variants) to generate preliminary fault diagnosis and prediction results. At the same time, a generative adversarial network (GAN) is used to generate fault data samples to optimize the model's recognition ability for rare faults. The preliminary analysis results are transmitted to the evaluation module 204, and a digital twin model of the substation equipment is constructed by combining the physical model and historical data of the equipment. The health status of the equipment is comprehensively evaluated through the digital twin model, and machine learning algorithms or physical simulation technologies are used to predict the future operation status of the equipment to generate a detailed fault prediction report. The output result of the evaluation module 204 is transmitted to the monitoring platform 205, and information such as the real-time operation status, health indicators, and fault probability of the equipment is displayed through a data visualization interface. According to the severity and urgency of the fault, the monitoring platform 205 issues graded alarm notifications to the operation and maintenance personnel: Level 1 alarm (severe fault): Trigger an audible and visual alarm, and push notifications to the operation and maintenance personnel and management personnel via text message, email, or mobile terminal. Level 2 alarm (medium fault): Trigger an audible alarm, and push notifications to the operation and maintenance personnel via mobile terminal or email. Level 3 alarm (minor fault or early warning): Push notifications to the operation and maintenance personnel via mobile terminal to prompt attention to the equipment status. The monitoring platform 205 transmits the alarm information to the alarm module 206, and the alarm module 206 is linked with the automatic fire extinguishing system and the emergency evacuation system: in case of emergencies such as fire, the fire extinguishing system is automatically activated, the power supply is cut off, the ventilation system is adjusted, and personnel are guided to evacuate through audible and visual alarms, voice prompts, and evacuation indicator lights. The monitoring platform 205 monitors the operation status of the emergency response system in real time and dynamically adjusts the fire extinguishing and evacuation strategies according to the real-time data. When the fire is extinguished or the danger is lifted, the alarm module 206 receives the stop instruction from the monitoring platform 205 and restores the system to the normal operation state. The monitoring platform 205 records the whole process of the emergency response, including the alarm time and the execution of fire extinguishing and evacuation operations, to form an alarm log for subsequent analysis and traceability. Fault diagnosis, health assessment, and prediction are realized through deep learning and digital twin technologies. Linkage with the automatic fire extinguishing system and the emergency evacuation system realizes automated emergency response. Redundant sensor networks and blockchain technologies are adopted to ensure the reliability of data acquisition and transmission. Through data visualization and graded alarm mechanisms, the work efficiency of operation and maintenance personnel is optimized.

[0088] The above-disclosed are only one or more preferred embodiments of the present application, and the scope of rights of the present application cannot be limited thereby. Those of ordinary skill in the art can understand all or part of the processes of implementing the above embodiments, and the equivalent changes made according to the claims of the present application still fall within the scope covered by the present application.

Claims

1. A method for monitoring auxiliary equipment of a power Internet of Things substation, characterized in that: The following steps are involved: Collect operation data of substation equipment in real time; Use edge computing to pre-process the collected operation data; Analyze pre-processed operating data based on deep learning algorithms to generate preliminary fault diagnosis and prediction results; Build a digital twin model of substation equipment, and use preliminary fault diagnosis and prediction results, combined with the physical model and historical data of the equipment, to conduct equipment health assessment and fault prediction; Based on the equipment health assessment and fault prediction results, the monitoring platform can visualize the data and notify the operation and maintenance personnel through graded alarms; Link the alarm system with the automatic fire extinguishing system and emergency evacuation system to achieve automated emergency response.

2. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 1, characterized in that: Collecting the operation data of the substation equipment in real time, the steps also include: The sensor network collects the current, voltage, temperature, humidity, smoke concentration, liquid level and infrared thermal imaging data of the substation equipment in real time. The sensor network adopts a redundant design and includes at least two groups of sensors of the same type. When one group of sensors fails or the data is abnormal, it automatically switches to the other group of sensors. Utilize multi-sensor data fusion technology, combine sensor data, and generate a comprehensive description of equipment operation status through Bayesian fusion algorithm or Kalman filter algorithm; Transmit comprehensive equipment operation data to edge computing devices in real time.

3. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 2, characterized in that: The collected operation data is preprocessed by edge computing, and the steps further include: Receive operation data, clean the operation data, and remove noise, outliers, and duplicate data; Extract features from the cleaned data, including current peak value, voltage fluctuation rate, temperature change rate, humidity threshold exceeding times, and smoke concentration mutation value; Based on the preset rules, the extracted characteristic parameters are preliminarily analyzed to determine whether the equipment is operating normally and generate a preliminary equipment status report; Synchronize the cleaned data and extracted feature parameters to the blockchain in real time; According to the equipment operation status and preset priority rules, the data is classified and processed, high-priority data is marked as requiring priority processing, and the pre-processed data is transmitted for further analysis.

4. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 3, characterized in that: Analyzing the preprocessed operating data based on a deep learning algorithm to generate preliminary fault diagnosis and prediction results, the steps further include: Inputting the preprocessed running data into a pre-trained deep learning model, wherein the deep learning model includes a convolutional neural network and / or a recurrent neural network and variants thereof; The extracted feature information is classified and identified through deep learning models to determine whether there is a fault mode in the operating data and determine the fault type, location and severity; Generate fault data samples using generative adversarial networks; Dynamically adjust the parameters of deep learning models based on reinforcement learning algorithms; Combine the historical and real-time operation data of the equipment to generate preliminary fault diagnosis and prediction results through deep learning models; Output preliminary fault diagnosis and prediction results to the digital twin model.

5. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 4, characterized in that: Construct a digital twin model of the substation equipment, use the preliminary fault diagnosis and prediction results, combine the physical model and historical data of the equipment, and perform equipment health assessment and fault prediction. The steps also include: Build a digital twin model of substation equipment, which includes a virtual representation of the equipment’s geometry, physical properties, electrical parameters, and operating status; Synchronize the digital twin model with the physical model of the equipment in real time; Import the generated preliminary fault diagnosis and prediction results into the digital twin model, and conduct a comprehensive assessment of the health status of the equipment by combining the historical and real-time operation data of the equipment; Based on the digital twin model, use machine learning algorithms or physical simulation technology to predict the future operating status of the equipment and generate a detailed fault prediction report, including the probability of fault occurrence, estimated time, impact range, and possible failure mode; Evaluate the effect of different maintenance solutions on improving equipment health status through virtual maintenance simulation; Dynamically adjust the parameters of the digital twin model based on equipment health assessment and fault prediction results.

6. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 5, characterized in that: According to the equipment health assessment and fault prediction results, data visualization is performed through the monitoring platform, and graded alarms are issued to notify the operation and maintenance personnel. The steps also include: Transmit equipment health assessment and fault prediction results to the monitoring platform in real time, which includes a data visualization interface and an alarm management system; The data visualization interface of the monitoring platform displays the real-time operating status, health indicators, failure probability, expected failure time and impact range of the equipment in graphical and tabular form; According to the severity and urgency of the fault prediction results, the alarm information is divided into multiple levels, including level one alarm, level two alarm and level three alarm.

7. The method for monitoring auxiliary equipment of power Internet of Things substation according to claim 6, characterized in that: According to the equipment health assessment and fault prediction results, data visualization is performed through the monitoring platform, and graded alarms are issued to notify the operation and maintenance personnel. The steps also include: For a level 1 alarm, the monitoring platform immediately triggers an audible and visual alarm, and notifies the operation and maintenance personnel and relevant management personnel of the alarm information in real time via SMS, email or mobile terminal push; For level 2 alarms, the monitoring platform triggers an audible alarm and notifies the operation and maintenance personnel of the alarm information via mobile terminal push or email; For level 3 alarms, the monitoring platform only notifies the operation and maintenance personnel of the alarm information through mobile terminal push notifications, prompting them to pay attention to the device status; The monitoring platform records the sending status and receiving confirmation of the alarm information to form an alarm log.

8. A monitoring system for auxiliary equipment of a power Internet of Things substation, applicable to the monitoring method for auxiliary equipment of a power Internet of Things substation as claimed in any one of claims 1 to 7, characterized in that: It includes a collection module, a preprocessing module, a preliminary analysis module, an evaluation module, a monitoring platform and an alarm module, wherein the preprocessing module is connected to the collection module, the preliminary analysis module is connected to the preprocessing module, the evaluation module is connected to the preliminary analysis module, the monitoring platform is connected to the evaluation module, and the alarm module is connected to the monitoring platform; Collection module, used to collect the operation data of substation equipment in real time; A preprocessing module is used to preprocess the collected operation data using edge computing; The preliminary analysis module is used to analyze the pre-processed operation data based on the deep learning algorithm to generate preliminary fault diagnosis and prediction results; The evaluation module is used to build a digital twin model of substation equipment, using preliminary fault diagnosis and prediction results combined with the physical model and historical data of the equipment to conduct equipment health assessment and fault prediction; The monitoring platform is used to visualize data based on equipment health assessment and fault prediction results, and to notify operation and maintenance personnel through graded alarms; The alarm module is used to cooperate with the automatic fire extinguishing system and the emergency evacuation system to achieve automated emergency response.

Images