A substation control system based on cloud control and local area network

Abstract

The present invention relates to the field of substation control technology, and particularly refers to a substation control system based on cloud control and local area network. It includes a substation data acquisition module, a data analysis module, a cloud collaborative control module, an edge computing response module, and a security guarantee module. The substation data acquisition module collects multi-dimensional data through the local area network. The data analysis module uses the dynamic multi-variate autoregressive tree algorithm to identify abnormal and fault patterns. The cloud collaborative control module uses the sparse distributed optimization algorithm for global load scheduling. The edge computing response module realizes fast local response control based on the Markov random field model. The security guarantee module combines deep learning and multi-factor identity authentication technologies to ensure the security of data transmission. The present invention improves the automation control level and overall operation efficiency of the substation, and realizes efficient, stable, and secure intelligent power management.

Description

Technical Field

[0001] The present invention relates to the technical field of substation control, and particularly to a substation control system based on cloud control and local area network. Background Art

[0002] In modern power systems, substations are the core nodes for realizing power conversion, dispatching, and distribution. The control systems of traditional substations usually adopt a distributed architecture, mainly relying on on-site equipment control and operation. However, with the increase in power demand and the improvement of grid complexity, traditional control systems have certain limitations in terms of response speed, data processing capabilities, and information sharing. Therefore, how to improve the automation level and intelligent management ability of substations has become the focus of current research and development. The following problems still exist: The existing systems lack multi-dimensional real-time data analysis means, making it difficult to accurately identify the non-linear dependencies and abnormal patterns between multiple variables, affecting the real-time performance and accuracy of fault diagnosis; The load dispatching of traditional substations is mostly local control, lacking global optimization across stations, resulting in resource waste and unreasonable energy consumption; The existing systems mainly rely on the control of the central system, with high latency and difficulty in meeting the rapid response requirements of local equipment. Summary of the Invention

[0003] To solve the above problems, the present invention provides a substation control system based on cloud control and local area network, which solves the problem of how to achieve multi-dimensional real-time data analysis, improve the accuracy and real-time performance of fault diagnosis, and achieve global optimization and rapid response of local equipment across stations while improving the automation and intelligent management levels of substations, thereby enhancing the automation control level and overall operation efficiency of substations.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] A substation control system based on cloud control and local area network includes a substation data acquisition module, a data analysis module, a cloud collaborative control module, an edge computing response module, and a security guarantee module that are communicatively connected in sequence;

[0006] The substation data acquisition module is used to collect multi-dimensional data of the substation through the local area network;

[0007] The data analysis module is used to perform regression analysis based on the multi-dimensional data using a dynamic multi-variate autoregressive tree algorithm, capture the non-linear dependence relationship between multiple variables, identify potential anomalies and fault patterns, and generate an analysis report;

[0008] The cloud collaborative control module is used to perform global load scheduling and control on multiple substations based on the analysis report. Through sparse communication collaboration among distributed nodes, it generates control instructions and dynamically adjusts the control parameters of each device in the substation by using the sparse distributed optimization algorithm;

[0009] The edge computing response module is used to locally process the key data of the analysis report through the local area network. It uses the Markov random field model to model the spatio-temporal dependence relationship between edge nodes as a probabilistic graph structure. At the same time, after receiving the control instruction, it performs rapid response control on the local devices in the substation and realizes dynamic real-time policy adjustment;

[0010] The security guarantee module is used to monitor the overall operation status of the system, and realizes all-round security guarantee for data transmission and instruction execution based on the intrusion detection algorithm of deep learning, multi-factor identity authentication and encryption transmission technology, and pushes relevant information to the terminals of operation and maintenance personnel.

[0011] Furthermore, the multi-dimensional data includes the environmental parameters of the substation, the operation parameters and status parameters of power equipment.

[0012] Furthermore, the operation process of the data analysis module includes the following steps:

[0013] Receive the multi-dimensional data and perform preprocessing operations, including removing noise and using time series smoothing to correct anomalies;

[0014] Based on the preprocessed multi-dimensional data, use the dynamic multi-variate autoregressive tree algorithm to identify the non-linear relationship between variables and generate a key feature mapping table;

[0015] Based on the key feature mapping table, construct a multi-dimensional causal relationship graph and identify the risk path between equipment and environmental parameters;

[0016] Compare the feature pattern with the preset fault pattern library in real time, identify potential fault trends, and generate a risk index report;

[0017] Generate a comprehensive analysis report according to the multi-dimensional causal relationship graph and the fault pattern matching result, and transmit the report to the cloud collaborative control module; the comprehensive analysis report includes the equipment health index in the substation, the analysis of abnormal reasons, the fault risk level, and trend prediction.

[0018] Even further, the formula of the multi-variate autoregressive tree algorithm is as follows:

[0019]

[0020] Where, denotes the predicted output variable at the current moment t, which is used to analyze the results of key data in the substation; m represents the number of multi-dimensional input features, that is, different types of data monitored in the substation; k represents the multi-level branch structure set on each input feature; denotes the influence weight of the input feature of the substation on the current output prediction; p and q represent the orders of the lag terms; denotes the l-th order lag coefficient of the i-th input feature in the j-th branch; denotes the non-linear mapping function of the i-th input feature at the lag time t - l; denotes the weight coefficient of the n-th order lag term of the i-th feature in the j-th branch; denotes the non-linear conversion function at the lag time t - n; denotes the bias term of the feature branch.

[0021] Furthermore, the operation process of the cloud collaborative control module includes the following steps:

[0022] Receive and parse the comprehensive analysis report transmitted by the data analysis module, and extract key data, where the key data includes equipment health status, fault risk level, and trend prediction;

[0023] Based on the comprehensive analysis results and the real-time load demands of multiple substations, adopt the sparse distributed optimization algorithm to determine the load distribution and control priorities of each substation, and generate control instructions, including equipment start / stop and parameter adjustment;

[0024] Allocate the control instructions to each substation node, use the distributed collaborative mechanism to optimize the instructions in real time, and perform dynamic adjustment in combination with the node feedback data;

[0025] Transmit the finally optimized control instructions to the edge computing response module through the local area network, initiate control execution, and monitor the execution feedback in real time.

[0026] Even further, the formula of the sparse distributed optimization algorithm is as follows:

[0027]

[0028] Among them, denotes minimizing the global load scheduling cost function F; N represents the number of substations participating in load scheduling and control; t represents the execution status of the control and optimization algorithm in a specific time period; denotes the control parameter of the i-th substation node at the moment t; denotes the load scheduling cost function of the i-th node; denotes other substation nodes that sparsely communicate with the i-th substation node through the local area network in the distributed system; represents the sparse regularization coefficient; represents the sum of squares of the control parameter differences between node i and node j.

[0029] Furthermore, the operation process of the edge computing response module includes the following steps:

[0030] Receive the control instructions transmitted by the cloud collaborative control module, parse the instruction content, and extract the control parameters of relevant devices;

[0031] Perform rapid localization processing on the key data in the analysis report generated by the data analysis module, construct the spatio-temporal dependence relationship between edge nodes using the Markov random field model, and map the key data into a probability graph structure;

[0032] Combine the real-time environmental data and control instructions of the local area network to perform response control on each local device in the substation, and give priority to processing the operation instructions of key devices;

[0033] Based on the real-time feedback mechanism, monitor the device response situation, and compare the immediate operating state of the device with the expected control effect. If there is a deviation, dynamically adjust the control strategy;

[0034] Summarize the operating status and control adjustment situation of the edge nodes, generate a feedback report, and transmit the key operating data and adjustment information to the cloud collaborative control module.

[0035] Furthermore, the Markov random field model specifically includes constructing a dependence graph structure between the edge nodes of the substation, where the nodes represent device state variables and the edges represent the spatio-temporal associations between devices; the node relationships are constrained through the conditional probability distribution function to form a joint probability model of the state variables.

[0036] Furthermore, the security guarantee module realizes the dynamic management of user permissions based on the multi-factor identity authentication module, and encrypts the data transmission content using symmetric and asymmetric encryption technologies.

[0037] The beneficial effects of the present invention are as follows:

[0038] Through the substation data acquisition module, the present invention realizes the real-time acquisition of multi-dimensional data of the substation. The data analysis module uses the dynamic multi-variate autoregressive tree algorithm for in-depth regression analysis, which can effectively identify the non-linear dependence relationship between multi-variables, thereby accurately identifying potential anomalies and fault patterns, and improving the accuracy and real-time performance of fault diagnosis. The cloud collaborative control module is based on the sparse distributed optimization algorithm, and through the sparse communication collaboration between distributed nodes, it can perform global load scheduling for multiple substations, realizing efficient resource allocation and energy consumption optimization. At the same time, the system can dynamically generate control instructions according to the analysis results, automatically adjust the control parameters of each device in the substation, thereby improving the intelligence level and control accuracy of the system. The edge computing response module processes key data within the local area network, models the spatio-temporal dependence relationship using the Markov random field model, ensures the rapid response control of local devices in the substation, and realizes the combination of local processing and real-time policy adjustment. This not only reduces the computing load on the cloud but also improves the response speed and reliability of the system. The security guarantee module constructs a perfect security system through deep learning intrusion detection, multi-factor identity authentication, and encrypted transmission technology, ensuring the security of data transmission and control instruction execution. It monitors the running state of the system in real time, effectively prevents data leakage and illegal access, and pushes relevant information to the operation and maintenance personnel, enhancing the overall security protection level of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a schematic diagram of the modules of a substation control system based on cloud control and local area network according to the present invention.

[0040] Figure 2 is a schematic flow diagram of the operation process of the data analysis module provided by an embodiment of the present invention.

[0041] Figure 3 is a schematic flow diagram of the operation process of the cloud collaborative control module provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] Please refer to Figures 1-3 as shown, the present invention relates to a substation control system based on cloud control and local area network.

[0043] Embodiment

[0044] A substation control system based on cloud control and local area network includes a substation data acquisition module, a data analysis module, a cloud collaborative control module, an edge computing response module, and a security guarantee module that are sequentially communicatively connected;

[0045] The substation data acquisition module is used to collect multi-dimensional data of the substation through the local area network; the multi-dimensional data includes the environmental parameters of the substation, the operation parameters, and the state parameters of the power equipment.

[0046] It should be noted that in the substation data acquisition module, a set of high-precision sensors and intelligent acquisition devices are deployed to achieve comprehensive data collection of the substation. This module connects all acquisition devices to a data concentrator through a local area network, and the concentrator connects the acquisition end devices and the data analysis module through a high-speed interface. The data collected by the module includes the following categories:

[0047] Environmental parameters: Environmental variables such as temperature, humidity, and air pressure are monitored by a multi-sensor array to obtain meteorological information around the substation. This information has an important impact on the heat dissipation and working stability of the equipment in the substation.

[0048] Operating parameters of power equipment: including electrical parameters such as current, voltage, and power, and the data is collected at a sampling frequency of seconds or milliseconds to ensure sufficient real-time performance.

[0049] Status parameters: such as equipment health status, component wear condition, operating time, etc., are recorded by sensors and status monitoring devices. To ensure data quality, each acquisition node is equipped with a fault tolerance and outlier filtering algorithm to avoid interference from abnormal or incorrect data.

[0050] The data analysis module is used to perform regression analysis based on the multi-dimensional data using the dynamic multi-variate autoregressive tree algorithm, capture the non-linear dependence relationships between multi-variables, identify potential anomalies and fault patterns, and generate an analysis report;

[0051] Among them, the operation process of the data analysis module includes the following steps:

[0052] Receive the multi-dimensional data and perform preprocessing operations, including removing noise and using time series smoothing to correct anomalies;

[0053] Based on the preprocessed multi-dimensional data, use the dynamic multi-variate autoregressive tree algorithm to identify the non-linear relationships between variables and generate a key feature mapping table;

[0054] Specifically, all multi-dimensional data is used as the input of the autoregressive tree model. The algorithm initializes the operation according to the set regression tree structure and parameters, and models the data. The model adopts the method of dynamic multiple regression and can adapt to the changes in the device state and environmental conditions. The algorithm gradually analyzes the mutual dependence relationships among variables to capture complex non-linear relationships. Each branch node of the autoregressive tree model divides different branch paths according to the dependence degree and change trend of variables. Through this process, the model can identify the hidden correlations among different variables, such as the non-linear correlation between the transformer temperature and the load current. The main variables identified by the algorithm and their mutual relationships are summarized into a key feature mapping table, which records the values of each important variable and its dependence strength on other variables. For example, the table will mark the influence weight of the load current on the device temperature and give the change trend of the feature values.

[0055] Based on the key feature mapping table, a multi-dimensional causal relationship graph is constructed to identify the risk paths between the device and environmental parameters;

[0056] Specifically, using the causal inference algorithm based on the graph model, the variable relationships identified in the feature mapping table are transformed into a directed graph structure. Each node represents a parameter (such as temperature, voltage), and the directed edges between the nodes represent causal relationships. For example, if the current rises, it will cause the temperature to rise. Then, in the causal relationship graph, an edge pointing from the "current" node to the "temperature" node represents this risk path. The module will mark the high-risk paths in the causal relationship graph, and the nodes and edges in the path represent specific risk transmissions. For example, when the device load is too high and the temperature rises, the causal relationship graph will mark the high-risk path from the load to the temperature, thus warning of the potential device failure risk. A weight is assigned to each causal path, and the weight value is determined according to the dependence strength between variables and the potential risk impact. The module will prioritize the possible risk paths according to these weights, thus helping with subsequent fault mode identification.

[0057] Compare the feature pattern with the preset fault mode library in real time, identify the potential fault trend, and generate a risk index report;

[0058] Specifically, the module contains a preset fault mode library that stores common historical fault modes and their characteristic values. For example, the transformer overheating fault mode may include the characteristic that both current and temperature are higher than the normal threshold. By comparing the current key feature mapping table with the fault mode library in real time one by one, the module can identify the similarity between the current features and the existing fault modes. If a high degree of match is detected between the current features and a known fault mode, the module will record this situation as a potential fault trend. A risk index is generated based on the matching result, and the higher the index value, the closer the current system state is to the fault mode. For example, when the voltage, current, and temperature all exceed the threshold, the risk index will increase significantly. The module generates a risk index report based on the risk index and marks high-risk devices and potential fault types.

[0059] Generate a comprehensive analysis report according to the multi-dimensional causal relationship diagram and the fault mode matching result, and transmit the report to the cloud collaborative control module; the comprehensive analysis report includes the equipment health index, abnormal cause analysis, fault risk level, and trend prediction in the substation.

[0060] Furthermore, the formula of the multivariate autoregressive tree algorithm is as follows:

[0061]

[0062] Among them, represents the predicted output variable at the current time t, which is used to analyze the results of key data in the substation; m represents the number of multi-dimensional input features, that is, different types of data monitored in the substation; k represents the multi-level branch structure set on each input feature; represents the influence weight of the input feature of the substation on the current output prediction; p and q represent the orders of the lag terms; represents the l-th order lag coefficient of the i-th input feature in the j-th branch; represents the non-linear mapping function of the i-th input feature at the lag time t-l, which converts the lag effect of the environmental parameters or the state of power equipment in the substation to identify non-linear dependence relationships; represents the weight coefficient of the n-th order lag term of the i-th feature in the j-th branch, which converts the lag effect of the environmental parameters or the state of power equipment in the substation to identify non-linear dependence relationships; represents the non-linear conversion function at the lag time t-n; represents the bias term of the feature branch.

[0063] The cloud collaborative control module is used to perform global load scheduling and control on multiple substations based on the analysis report. Through sparse communication collaboration between distributed nodes, control instructions are generated to dynamically adjust the control parameters of each device in the substation.

[0064] Among them, the operation process of the cloud collaborative control module includes the following steps:

[0065] Receive and parse the comprehensive analysis report transmitted by the data analysis module, and extract key data, where the key data includes the device health status, failure risk level, and trend prediction;

[0066] Based on the comprehensive analysis results and the real-time load demands of multiple substations, adopt a sparse distributed optimization algorithm to determine the load distribution and control priorities of each substation, and generate control instructions, including device start / stop and parameter adjustment;

[0067] Specifically, obtain the current load demand data from each substation node, including real-time current, voltage, power demand and other data of each site. This demand data is updated regularly through a local area network connection to ensure the real-time nature of the scheduling decision-making basis. Based on the collected load demands and the key data in the comprehensive analysis report, the module initializes the sparse distributed optimization algorithm. This algorithm divides the entire scheduling task into several subtasks and distributes them to different substation nodes to achieve distributed computing. The algorithm optimizes the load distribution according to data such as the device health status, risk level, and trend prediction of each substation. For example, substations with a lower health index and a high failure risk are given priority to reduce the load or temporarily shut down some devices to reduce the failure risk. The algorithm dynamically allocates the load ratio and generates control priorities according to the health status, load demand, and environmental data of each substation node. After the optimization calculation is completed, the module generates specific control instructions, and the instruction content includes device start / stop operations (such as shutting down or starting power generation equipment) and key parameter adjustments (such as current and voltage settings) to meet the scheduling requirements of the global load.

[0068] Distribute the control instructions to each substation node, use the distributed collaborative mechanism to optimize the instructions in real time, and perform dynamic adjustment in combination with the node feedback data;

[0069] Specifically, the generated control instructions are sent to the control nodes of each substation through the local area network. The instructions include specific requirements for equipment operation and parameter adjustment. The module uses an instruction distributor to divide the control instructions into node instruction packets, and each substation node receives its own instruction content. After each substation node receives the instructions, it starts to perform preliminary control operations. At this time, the cloud collaborative control module monitors the execution status and feedback data of each node in real time through a distributed collaborative mechanism. For example, if an abnormality occurs when a certain node performs start-stop operations, the module will make dynamic adjustments based on the feedback data, reallocate tasks or adjust the control strategy. The module receives and analyzes the feedback data of each node, including the current load status, equipment response situation, and environmental parameters, etc., to ensure the effectiveness of instruction execution. If the feedback data indicates that the load of a certain node is too high or the equipment status is abnormal, the module will dynamically adjust the load distribution strategy to avoid overloading or failures. For example, if a substation node feedbacks that the temperature exceeds the expectation, the module will issue an instruction to reduce the load of that node.

[0070] The finally optimized control instructions are transmitted to the edge computing response module through the local area network to initiate control execution and monitor the execution feedback in real time.

[0071] Furthermore, the formula of the sparse distributed optimization algorithm is as follows:

[0072]

[0073] Where, means minimizing the global load scheduling cost function F to find the optimal control parameter allocation and achieve global load balance and cost minimization for multiple substations; N represents the number of substations participating in load scheduling and control; t represents the execution status of the control and optimization algorithm in a specific time period; represents the control parameter of the i-th substation node at time t, that is, the control variable to be adjusted, such as the start-stop status of the equipment or the load distribution weight, etc., to meet the real-time load demand; represents the load scheduling cost function of the i-th node, which is used to measure the scheduling cost of the control parameter at this node, including cost factors such as the health status of substation equipment, operating parameters, and status parameters of power equipment; represents other substation nodes that sparsely communicate with the i-th substation node through the local area network in the distributed system; represents the sparse regularization coefficient, which controls the communication intensity between different substation nodes; represents the sum of the squares of the differences between the control parameters of node i and node j, which is used to measure the load scheduling coordination between two substation nodes, and promotes the load control parameters of adjacent nodes to be close, so as to achieve stable load balance.

[0074] The edge computing response module is used to locally process the key data of the analysis report through a local area network. By adopting a Markov random field model, the spatio-temporal dependence relationship between edge nodes is modeled as a probabilistic graph structure. Meanwhile, after receiving the control instruction, it quickly responds to control the local equipment of the substation and realizes dynamic real-time policy adjustment;

[0075] Among them, the operation process of the edge computing response module includes the following steps:

[0076] Receive the control instruction transmitted by the cloud collaborative control module, parse the instruction content and extract the control parameters of relevant equipment;

[0077] Quickly locally process the key data in the analysis report generated by the data analysis module. Adopt a Markov random field model to construct the spatio-temporal dependence relationship between edge nodes and map the key data into a probabilistic graph structure; The Markov random field model specifically includes constructing a dependence graph structure between the edge nodes of the substation, where the nodes represent equipment state variables and the edges represent the spatio-temporal associations between the equipment; The node relationship is constrained through a conditional probability distribution function to form a joint probability model of the state variables;

[0078] It should be noted that the module uses a Markov random field (MRF) model to establish a dependence graph structure between substation equipment according to the spatial layout and data dependence relationship of edge nodes.

[0079] Node definition: Each node in the graph structure represents a state variable of a device (such as temperature, load, voltage). The node state is updated through real-time data.

[0080] Edge definition: The edge represents the spatial or temporal association relationship between devices. For example, the edge connection between adjacent devices represents that their temperatures or loads will affect each other.

[0081] Probability constraint definition: Use a conditional probability distribution function to define the relationships of each node. Through these probability distributions, the dependence relationships between nodes are constrained to form an overall joint probability model. This model allows speculation on the possibility of device state changes and helps identify potential anomalies of key devices.

[0082] Combine the real-time environmental data and control instructions of the local area network to perform response control on each local device in the substation, and give priority to processing the operation instructions of key devices;

[0083] Specifically, collect environmental data (such as temperature, humidity, voltage fluctuations, etc.) in real time from within the local area network to ensure that control operations are based on the latest environmental status. The real-time data is collected by sensors at the edge nodes and transmitted to the response module for decision support. Control instructions are allocated according to priorities, and operation instructions for critical devices are executed first. For example, for devices with high risks or abnormal operating states, the module preferentially adjusts their loads or start / stop states to ensure overall safety. The module adjusts the parameters and operations of the target devices according to the control instructions and real-time data. For example, when the temperature of a device exceeds the safe range, the module will preferentially reduce its load or temporarily shut it down; for other devices with lower priorities, the module will adjust them step by step in sequence.

[0084] Based on the real-time feedback mechanism, monitor the response of the devices, and compare the immediate operating state of the devices with the expected control effect. If there are deviations, dynamically adjust the control strategy;

[0085] Specifically, during the process of the devices executing the control instructions, collect the feedback data of the devices in real time, such as changes in parameters such as load, voltage, and temperature, and monitor the actual effects of the control operations. The module compares the actual operating state of the devices with the expected control effect to determine whether the control execution effect meets the expectations. For example, if the expected temperature drops but the actual temperature does not decrease significantly, the module identifies it as a deviation situation. When it is found that there are deviations between the actual state and the expected state of the devices, the module automatically adjusts the control strategy. For example, if the load of a device fails to decrease as expected, the module may further reduce the power of the device or start a standby device to achieve the control target. The dynamic adjustment function of the control strategy of the module ensures that the device state always meets the safe operating requirements of the system.

[0086] Summarize the operating state and control adjustment situation of the edge nodes, generate a feedback report, and transmit the key operating data and adjustment information to the cloud collaborative control module.

[0087] The security guarantee module is used to monitor the overall operating state of the system, and achieve all-round security guarantee for data transmission and instruction execution based on the deep learning-based intrusion detection algorithm, multi-factor identity authentication, and encryption transmission technology, and push the relevant information to the terminals of the operation and maintenance personnel; the security guarantee module realizes the dynamic management of user permissions based on the multi-factor identity authentication module, and encrypts the data transmission content using symmetric and asymmetric encryption technologies.

[0088] In one embodiment, to enhance the system's ability to identify potential network attacks, the security assurance module adopts a deep learning-based intrusion detection system to monitor and analyze the system's network traffic and operation behaviors in real time. The specific design and implementation methods are as follows: Extract multi-dimensional features from network traffic, including communication frequency, packet size, source IP address, destination IP address, etc. These features are combined to characterize normal and abnormal communication behaviors. A deep model combining a convolutional neural network (CNN) and a recurrent neural network (RNN) is used and trained based on substation historical data and known attack patterns. The CNN is responsible for capturing spatial features in network traffic (such as the distribution of packets), and the RNN is used to capture temporal features (such as the time series of packet transmission). The trained model analyzes network traffic in real time. When abnormal data features (such as traffic surges, connections from malicious IP addresses, etc.) are detected, the module immediately triggers an alarm and isolates suspicious communication nodes to protect the system from potential attacks. To improve the detection accuracy, the module is equipped with a self-learning function that can continuously collect new data features and update the model during the system operation to enhance the ability to identify unknown attacks.

[0089] Multi-factor authentication ensures the legitimacy of system visitors and prevents unauthorized personnel from accessing the system. The authentication adopts a multiple verification mechanism combining physical and biometric features, including the following steps: When logging in for the first time, the operation and maintenance personnel need to use two-factor authentication (2FA), such as entering a password and a mobile dynamic verification code. This ensures that only personnel with authorized devices can enter the system. Biometric technologies such as face recognition or fingerprint scanning are used to ensure that the personnel entering the system each time are registered users. The face recognition algorithm is built based on a convolutional neural network (CNN) and can accurately match the user's biometric information stored in the system database. Based on the sensitivity of operation and maintenance tasks and operations, the system dynamically adjusts the user's permissions. For example, ordinary equipment inspection operations do not require high-level permissions, while equipment debugging and control operations require high-level permission verification. The system dynamically judges the task requirements and requires additional authentication when necessary. Different roles and permissions are set for different operators to prevent unauthorized operations. For example, data analysts can only view and analyze data, while equipment operation and maintenance personnel have the permission to control the equipment.

[0090] The encryption mechanism for data transmission ensures the security of data flow inside and outside the system. The module adopts a combination of symmetric encryption and asymmetric encryption to achieve efficient and secure data encryption. The specific implementation is as follows:

[0091] Symmetric encryption (AES): The real-time data transmitted within the substation internal local area network (such as the collected device status, environmental parameters, etc.) adopts the AES encryption algorithm. Symmetric encryption has a fast calculation speed during transmission and can ensure the transmission speed while ensuring data security.

[0092] Asymmetric Encryption (RSA): For data transmission across networks (such as from a substation to the cloud), the asymmetric encryption (RSA) algorithm is used to encrypt important information and instructions. The RSA public key is used to encrypt sensitive data, and the private key is used for decryption at the receiving end, which ensures that only the legitimate cloud server can decrypt the data content.

[0093] End-to-End Encrypted Transmission: To avoid data leakage or tampering during transmission, the module uses the SSL / TLS protocol to encrypt the transmission channel, ensuring the confidentiality and integrity of data along the transmission path. The SSL certificate will be issued by a third-party security certification authority to enhance trust.

[0094] Key Management System (KMS): The module is configured with a Key Management System (KMS) that automatically generates, stores, and manages encryption keys, and regularly updates the keys to prevent risks brought by key leakage. The KMS also has a key auditing function to record all key usage and update records for security review.

[0095] The module provides a security event alarm and automated operation and maintenance notification mechanism to ensure that operation and maintenance personnel can respond to security events in a timely manner. The alarms and notifications are as follows:

[0096] Real-Time Alarm System: When the system detects potential security threats (such as intrusion behavior, unauthorized login), the module will trigger multiple alarms. These include:

[0097] Visual Alarm: Highlight the alarm information on the control center display screen, and use icons or text to prompt the specific threat type and location.

[0098] Sound Alarm: Emit an alarm sound through audio devices to attract the attention of operation and maintenance personnel.

[0099] Mobile Push: Real-time push the alarm information to the mobile phones of authorized operation and maintenance personnel via text messages, emails, or the operation and maintenance App to prompt an immediate response.

[0100] Fault Log Record: The module has a built-in log recording system that automatically records the detailed information of each security event, including the event occurrence time, threat type, response measures, etc. The logs can help operation and maintenance personnel conduct post-event analysis and optimize security policies.

[0101] Adaptive Defense Response: When detecting a major threat, the module will automatically activate preset defense strategies, such as automatically blocking specific IP addresses, reducing network access permissions, or even switching the system to the security mode, allowing only critical operations.

[0102] In summary, through the deployment of high-precision sensors and intelligent acquisition devices in the substation data acquisition module of the present invention, comprehensive and refined acquisition of environmental parameters, operation parameters, and status parameters of power equipment is achieved. This module uses the high-speed interfaces of the local area network and concentrator to achieve fast and stable data transmission, ensuring the real-time and accuracy of data collection, and providing a solid foundation for subsequent data analysis. The data analysis module adopts the dynamic multiple autoregressive tree algorithm. By constructing a key feature mapping table and a multi-dimensional causal relationship diagram, it identifies the non-linear dependence relationships between variables, effectively discovers the potential causes and risk paths of equipment failures. This process can timely identify and analyze abnormal data trends, providing real-time fault warnings and risk analysis.

[0103] In the present invention, the cloud collaborative control module is based on the sparse distributed optimization algorithm. Through the collaboration and optimization among distributed nodes, it globally optimizes and schedules the load of the substation, and dynamically adjusts the control parameters of each device. The system allocates the load according to the device health status and the priority of fault risks, reducing the load on high-risk devices, and ensuring the safety of the devices and the overall operation efficiency of the system. After receiving the cloud control instruction, the edge computing response module uses the local area network data for local processing, adopts the Markov random field model to establish the spatio-temporal dependence relationship between nodes, and performs fast response control on local devices in the substation. The module combines a real-time feedback mechanism to ensure the effectiveness of the control strategy, and automatically performs dynamic adjustment during the execution process, improving the response speed and decision-making accuracy of the system.

[0104] The security guarantee module in the present invention adopts deep learning intrusion detection algorithms, multi-factor identity authentication, and encryption technologies to ensure the security of data transmission and control instructions. The intrusion detection system in the module uses a deep learning model to monitor and analyze network traffic in real time, identify potential network attacks, and has the self-learning function to continuously improve the detection ability for unknown threats. The adaptive defense response mechanism can quickly adjust the system state when a security event occurs, protecting the operation security and stability of the system. The system generates a comprehensive analysis report based on the multi-dimensional causal relationship diagram and the matching results of the fault mode library, which includes the device health index, abnormal cause analysis, fault risk level, and trend prediction, helping the operation and maintenance personnel to more clearly understand the device state and fault risks, and making management and maintenance more accurate and efficient.

[0105] The above embodiments are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary engineering and technical personnel in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A substation control system based on cloud control and local area network, characterized in that It includes a substation data acquisition module, a data analysis module, a cloud collaborative control module, an edge computing response module, and a security guarantee module that are communicatively connected in sequence; The substation data acquisition module is used to collect multi-dimensional data of the substation through a local area network; The data analysis module is used to perform regression analysis based on the multi-dimensional data, adopt a dynamic multi-variate autoregressive tree algorithm to capture the non-linear dependence relationship between multi-variables, identify potential anomalies and fault patterns, and generate an analysis report; The cloud collaborative control module is used to perform global load scheduling and control on multiple substations based on the analysis report, adopt a sparse distributed optimization algorithm, generate control instructions through sparse communication collaboration between distributed nodes, and dynamically adjust the control parameters of each device in the substation; The edge computing response module is used to locally process the key data of the analysis report through a local area network, adopt a Markov random field model to model the spatio-temporal dependence relationship between edge nodes as a probabilistic graph structure. At the same time, after receiving the control instructions, it quickly responds to control the local devices in the substation and realizes dynamic real-time policy adjustment; The security guarantee module is used to monitor the overall operation status of the system, and realize all-round security guarantee for data transmission and instruction execution based on deep learning intrusion detection algorithms, multi-factor identity authentication, and encrypted transmission technologies, and push relevant information to the terminals of operation and maintenance personnel.

2. The substation control system based on cloud control and local area network according to claim 1, characterized in that The multi-dimensional data includes the environmental parameters of the substation, the operation parameters and status parameters of power equipment.

3. The substation control system based on cloud control and local area network according to claim 1, characterized in that, The operation process of the data analysis module includes the following steps: Receive the multi-dimensional data and perform preprocessing operations, including removing noise and using time series smoothing to correct anomalies; Based on the preprocessed multi-dimensional data, use the dynamic multi-variate autoregressive tree algorithm to identify the non-linear relationship between variables and generate a key feature mapping table; Based on the key feature mapping table, construct a multi-dimensional causal relationship graph to identify the risk path between equipment and environmental parameters; Compare the feature pattern with the preset fault pattern library in real time, identify potential fault trends, and generate a risk index report; Generate a comprehensive analysis report according to the multi-dimensional causal relationship graph and the fault pattern matching result, and transmit the report to the cloud collaborative control module; the comprehensive analysis report includes the equipment health index, abnormal cause analysis, fault risk level, and trend prediction in the substation.

4. A substation control system based on cloud control and local area network according to claim 3, characterized in that, The formula of the multi-variate autoregressive tree algorithm is as follows: ， Among them, represents the predicted output variable at the current moment t, which is used to analyze the results of key data in the substation; m represents the number of multi-dimensional input features, that is, different types of data monitored in the substation; k represents the multi-level branch structure set on each input feature; represents the influence weight of the input features of the substation on the current output prediction; p and q represent the orders of the lag terms; represents the l-th order lag coefficient of the i-th input feature in the j-th branch; represents the non-linear mapping function of the i-th input feature at the lag time t-l; represents the weight coefficient of the n-th order lag term of the i-th feature in the j-th branch; represents the non-linear conversion function at the lag time t-n; represents the bias term of the feature branch.

5. The substation control system based on cloud control and local area network according to claim 1, characterized in that, The operation process of the cloud collaborative control module includes the following steps: Receive and parse the comprehensive analysis report transmitted by the data analysis module, and extract key data, where the key data includes equipment health status, fault risk level, and trend prediction; Based on the comprehensive analysis result and the real-time load demand of multiple substations, adopt a sparse distributed optimization algorithm to determine the load distribution and control priority of each substation, and generate control instructions, including equipment start-stop and parameter adjustment; Allocate the control instructions to each substation node, use the distributed collaborative mechanism to optimize the instructions in real time, and perform dynamic adjustment in combination with the node feedback data; Transmit the finally optimized control instructions to the edge computing response module through a local area network, start control execution, and monitor the execution feedback in real time.

6. The substation control system based on cloud control and local area network according to claim 1, characterized in that, The operation process of the edge computing response module includes the following steps: Receive the control instructions transmitted by the cloud collaborative control module, parse the instruction content and extract the control parameters of relevant devices; Perform rapid localization processing on the key data in the analysis report generated by the data analysis module, construct the spatio-temporal dependence relationship between edge nodes using the Markov random field model, and map the key data into a probabilistic graph structure; Combine the real-time environmental data and control instructions of the local area network to perform response control on each local device in the substation, and give priority to processing the operation instructions of key devices; Based on the real-time feedback mechanism, monitor the device response situation, and compare the immediate operating state of the device with the expected control effect. If there is a deviation, dynamically adjust the control strategy; Summarize the operating status and control adjustment situation of the edge nodes, generate a feedback report, and transmit the key operating data and adjustment information to the cloud collaborative control module.

7. The substation control system based on cloud control and local area network according to claim 6, characterized in that, The Markov random field model specifically includes constructing a dependence graph structure between the edge nodes of the substation, where the nodes represent device state variables and the edges represent the spatio-temporal associations between devices; constraining the node relationships through the conditional probability distribution function to form a joint probability model of state variables.

8. A substation control system based on cloud control and local area network according to claim 1, characterized in that, The security guarantee module realizes the dynamic management of user permissions based on the multi-factor identity authentication module, and encrypts the data transmission content using symmetric and asymmetric encryption technologies.

Images