Low-voltage distribution network change risk research and judgment and intelligent early warning method and system

Abstract

The invention provides a low-voltage power distribution network change risk research and judgment and intelligent early warning method and system, and relates to the technical field of intelligent early warning of a power system, and the method comprises the steps: carrying out the fusion analysis of power grid operation data and environment data through a space-time convolutional neural network, and generating a correlation feature matrix of a power grid operation state and an environment risk; performing dynamic modeling on the power grid topological structure by using the physical information dynamic graph attention network, and extracting a dynamic risk propagation path; calculating fault probability distribution of each node of the power grid by adopting a Bayesian network in combination with the dynamic risk propagation path; the early warning strategy is optimized through a reinforcement learning algorithm, and an optimal early warning scheme is generated; and scheduling the first-aid repair resources by using a multi-objective optimization algorithm to generate a first-aid repair resource scheduling scheme. The method has the beneficial effects that the accuracy of fault early warning and the first-aid repair efficiency are remarkably improved, and reliable technical support is provided for intelligent operation and maintenance of the low-voltage power distribution network.

Description

Technical Field

[0001] The present invention relates to the field of intelligent early warning technology for power systems, and in particular to a method and system for analyzing and judging the risk of changes in low-voltage distribution networks and for intelligent early warning. Background Art

[0002] In recent years, the continuous expansion and increasing complexity of power systems have posed significant challenges to fault diagnosis and operation and maintenance management of low-voltage distribution networks. This is particularly true in high-hazard environments, such as remote mountainous areas and coastal areas. Grid equipment is susceptible to extreme weather conditions, such as typhoons, heavy rains, and salt spray corrosion, leading to frequent failures that are difficult to quickly locate and repair.

[0003] Currently, fault warnings for low-voltage distribution networks primarily rely on manual inspections and simple threshold alarm systems. These methods suffer from low efficiency, poor real-time performance, and insufficient adaptability. For example, traditional manual inspections cannot capture the impact of environmental changes on the grid in real time, while alarm systems based on fixed thresholds struggle to cope with complex and changing fault modes. Existing research has attempted to predict faults using statistical models or rule engines, but these methods lack the ability to model dynamic changes in grid topology and physical laws, resulting in low prediction accuracy and high false alarm rates, making them unable to meet the requirements of intelligent grid operation and maintenance. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for analyzing and assessing the risk of changes in low-voltage distribution networks and providing intelligent early warning to improve the above-mentioned problems. To achieve the above-mentioned purpose, the technical solutions adopted by the present invention are as follows:

[0005] In the first aspect, the present application provides a method for evaluating and intelligently warning the risk of changes in a low-voltage distribution network, including:

[0006] The grid operation data and environmental data are integrated and analyzed through a spatiotemporal convolutional neural network to generate a correlation feature matrix between the grid operation status and environmental risks.

[0007] Based on the correlation feature matrix, the physical information dynamic graph attention network is used to dynamically model the power grid topology and extract the dynamic risk propagation path;

[0008] Combined with the dynamic risk propagation path, the Bayesian network is used to calculate the failure probability distribution of each node in the power grid;

[0009] Based on the fault probability distribution, the early warning strategy is optimized through reinforcement learning algorithm to generate the optimal early warning plan;

[0010] According to the optimal early warning plan, the multi-objective optimization algorithm is used to schedule the emergency repair resources and generate the emergency repair resource scheduling plan.

[0011] Preferably, the grid operation data and environmental data are fused and analyzed by a spatiotemporal convolutional neural network to generate a correlation feature matrix between the grid operation status and environmental risks, which includes:

[0012] The first data of voltage, current and power factor are collected in real time through smart meters and sensors at a frequency of 1 minute. The second data of wind speed, rainfall, temperature and humidity are obtained from the weather station at a frequency of 10 minutes. The third data of terrain elevation and vegetation distribution are obtained from the geographic information system. The fourth data of salt spray concentration is obtained from the corrosion monitoring equipment at a sampling frequency of 1 hour.

[0013] The first data, second data, third data and fourth data are aligned in time and space, and a spatiotemporal data cube is constructed. The spatiotemporal data cube is processed using a spatiotemporal convolutional neural network to obtain a correlation feature matrix between the power grid operation status and environmental risks.

[0014] Preferably, the method uses a physical information dynamic graph attention network based on the correlation feature matrix to dynamically model the power grid topology and extract the dynamic risk propagation path, which includes:

[0015] Obtain the topological structure information of the power grid in the power grid management system, including node data and edge data; assign the associated feature matrix as a node to the topological structure information of the power grid to obtain the power grid topology structure with node characteristics;

[0016] constructing a dynamic graph structure according to the power grid topology structure with node characteristics, the second data, and the fourth data;

[0017] Based on the dynamic graph structure and power grid operation data, including current, voltage, and resistance, and taking circuit laws as constraints, a physical information loss function is constructed. This is then integrated into the neural network training process to obtain node features embedded with physical information.

[0018] Combining the node features embedded with physical information and the dynamic graph structure, the dynamic graph attention network algorithm is used to calculate the attention coefficient between nodes and dynamically adjust the edge weight according to the attention coefficient;

[0019] Based on edge weights and node characteristics, the node status is iteratively updated through multi-layer PI-DGAT to simulate the dynamic propagation process of risk in the power grid and obtain the dynamic risk propagation path.

[0020] Preferably, the dynamic risk propagation path is combined to calculate the failure probability distribution of each node in the power grid using a Bayesian network, which includes:

[0021] A Bayesian network structure is constructed using a dynamic risk propagation path matrix, historical fault data, and real-time device status data. Network nodes correspond to power grid devices, network edges represent risk propagation relationships between devices, and edge weights are determined by the propagation probabilities in the dynamic risk propagation path matrix. This generates a Bayesian network topology that reflects the dynamic risk propagation logic between power grid devices.

[0022] Combining the Bayesian network topology and historical fault data, the basic fault probability of the root node and the conditional fault probability of non-root nodes are calculated by training the conditional probability table. The conditional probability of the non-root node is derived from the state of its parent node and the dynamic propagation probability, thus obtaining the conditional probability distribution of the Bayesian network.

[0023] The real-time device status data is input into the trained Bayesian network as evidence, and a probabilistic inference algorithm is used to calculate the posterior probability of each node under different fault types, and the fault probability distribution of the power grid nodes is output.

[0024] Preferably, the optimization of the early warning strategy by a reinforcement learning algorithm to generate an optimal early warning plan includes:

[0025] Obtain the fault probability distribution and user feedback data, integrate and process them to obtain the state space, and use the state space as the input of the reinforcement learning algorithm;

[0026] Using the state space and action space, after training with the Q-learning algorithm, a Q-value table is obtained, where the Q-value table is used to generate the optimal early warning strategy;

[0027] Based on the Q-value table and the current state, the optimal action is selected according to the Q-value table, and then the optimal early warning plan is obtained, where the optimal early warning plan includes the warning level and response measures.

[0028] Secondly, this application also provides a low-voltage distribution network change risk assessment and intelligent early warning system, including:

[0029] Generation module: used to integrate and analyze power grid operation data and environmental data through spatiotemporal convolutional neural networks to generate a correlation feature matrix between power grid operation status and environmental risks;

[0030] Extraction module: used to dynamically model the power grid topology based on the correlation feature matrix and use the physical information dynamic graph attention network to extract the dynamic risk propagation path;

[0031] Calculation module: used to combine dynamic risk propagation paths and use Bayesian networks to calculate the failure probability distribution of each node in the power grid;

[0032] Optimization module: used to optimize the early warning strategy based on the fault probability distribution through reinforcement learning algorithm to generate the optimal early warning plan;

[0033] Dispatching and early warning module: used to dispatch emergency repair resources based on the optimal early warning plan using a multi-objective optimization algorithm.

[0034] Thirdly, this application also provides a low-voltage distribution network change risk assessment and intelligent early warning device, including:

[0035] memory for storing computer programs;

[0036] A processor is used to implement the steps of the low-voltage distribution network change risk assessment and intelligent early warning method when executing the computer program.

[0037] Fourthly, the present application also provides a readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned early warning method based on low-voltage distribution network change risk analysis and intelligentization are implemented.

[0038] The beneficial effects of the present invention are:

[0039] The present invention uses a spatiotemporal convolutional neural network to extract the correlation characteristics between the operating status of the power grid and environmental risks, combines it with a physical information dynamic graph attention network to model the dynamic changes of the power grid topology, captures the impact of environmental factors (such as typhoons, heavy rains, and salt spray corrosion) on the power grid in real time, and calculates the fault probability distribution of each node through a Bayesian network. It can accurately predict the location and type of fault, solving the problem that traditional methods cannot dynamically adapt to environmental changes and lack the constraints of physical laws.

[0040] The present invention incorporates circuit laws (such as Ohm's law and Kirchhoff's law) as constraints into the neural network training process, constructs a physical information loss function, and ensures that the model output conforms to the basic operating laws of the power grid. This not only improves the model's prediction accuracy, but also enhances the interpretability of the results, enabling operation and maintenance personnel to intuitively understand the logic and root causes of fault propagation.

[0041] This invention optimizes the early warning strategy through a reinforcement learning algorithm, designs a reward function based on user feedback data (such as the impact range of power outages and the importance of key users), generates the optimal early warning plan and emergency repair resource scheduling plan, and the emergency repair scheduling plan based on the multi-objective optimization algorithm can maximize resource utilization and risk coverage while ensuring the shortest emergency repair time, significantly improving the power supply reliability and operation and maintenance efficiency of the power grid.

[0042] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the embodiments of the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 Schematic diagram of the process of the low-voltage distribution network change risk assessment and intelligent early warning method according to an embodiment of the present invention;

[0045] Figure 2 Schematic diagram of the structure of the low-voltage distribution network change risk assessment and intelligent early warning system according to an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of the structure of the low-voltage distribution network change risk assessment and intelligent early warning equipment described in an embodiment of the present invention.

[0047] In the figure: 701, generation module; 702, extraction module; 703, calculation module; 704, optimization module; 705, dispatching and warning module; 800, low-voltage distribution network change risk assessment and intelligent warning equipment; 801, processor; 802, memory; 803, multimedia component; 804, I / O interface; 805, communication component. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0049] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are used only to distinguish the description and should not be understood as indicating or implying relative importance.

[0050] Example 1:

[0051] This embodiment provides a low-voltage distribution network change risk assessment and intelligent early warning method.

[0052] See also Figure 1 , the figure shows that the method includes step S100, step S200, step S300, step S400 and step S500.

[0053] S100. Perform fusion analysis on power grid operation data and environmental data through spatiotemporal convolutional neural network to generate a correlation feature matrix between power grid operation status and environmental risks.

[0054] It can be understood that step S100 includes steps S101 and S102, wherein:

[0055] S101. Collect first data of voltage, current, and power factor in real time at a frequency of 1 minute through smart meters and sensors; obtain second data of wind speed, rainfall, temperature, and humidity at a frequency of 10 minutes from a weather station; obtain third data of terrain elevation and vegetation distribution from a geographic information system; and obtain fourth data of salt spray concentration at a sampling frequency of 1 hour from corrosion monitoring equipment.

[0056] S102. Align the first data, the second data, the third data, and the fourth data in time and space, and construct a spatiotemporal data cube. Use a spatiotemporal convolutional neural network to process the spatiotemporal data cube to obtain a correlation feature matrix between the power grid operation status and the environmental risk. The calculation formula of the spatiotemporal data cube is as follows:

[0057]

[0058] Among them, X represents the fusion result of power grid operation data and environmental data, Indicates that each element in the data cube is a real number, N is the number of grid nodes, T is the time step, which indicates the number of time points for data collection, and D is the number of features of each node at each time point.

[0059] It should be noted that the 3x3 convolution kernel is used to extract local spatial features. The formula is:

[0060] F space =Conv2D(X,K space )

[0061] Where K space is the spatial convolution kernel, F space is the spatial feature map;

[0062] Use 1D convolution kernel to capture dynamic changes and output the associated feature vector of each grid node Where D′ is the feature dimension.

[0063] Among them, the correlation characteristic matrix between the grid operation status and environmental risks is F, the rows of the matrix represent the grid nodes, and the columns represent the characteristic dimensions (such as the voltage fluctuation-wind speed correlation index and the cable aging-salt spray correlation index).

[0064] S200. Based on the correlation feature matrix, the physical information dynamic graph attention network is used to dynamically model the power grid topology and extract the dynamic risk propagation path.

[0065] It can be understood that step S200 includes S201, S202, S203, S204 and S205, wherein:

[0066] S201, obtaining topological structure information of a power grid in a power grid management system, including node data and edge data; assigning a correlation feature matrix as a node to the topological structure information of the power grid to obtain a power grid topology structure with node features;

[0067] It should be noted that node data represents devices in the power grid (such as transformers and cable connectors), with each node containing information such as device ID, type, and location. Edge data represents the connections between devices (such as cable connections), with each edge containing information such as the starting and ending nodes, connection type (such as cable type), and length. Node data V and edge data E are derived from the power grid management system through APIs or database queries. Environmental risks and grid operating status characteristics are integrated into the topology, providing multidimensional input for dynamic modeling.

[0068] S202, constructing a dynamic graph structure according to the power grid topology structure with node characteristics, the second data, and the fourth data;

[0069] It can be understood that, in this embodiment, the edge weight E(t) of the dynamic graph structure G(t)=(V, E(t)) is driven by environmental data:

[0070] E ij (t)=α·W(t)+β·R(t)+γ·S(t)

[0071] Here, α, β, and γ are weight coefficients, obtained through training using historical data. For example, high wind speeds W(t) may cause cable sway to intensify, reducing edge weights. Short-term (minute-level) edge weights reflect real-time environmental changes, while long-term (hour-level) edge weights reflect the cumulative effects of equipment aging. In this step, dynamic edge weights enable the model to respond to environmental changes in real time, improving the accuracy of risk propagation paths.

[0072] S203. Based on the dynamic graph structure and power grid operation data, where the power grid operation data includes current, voltage, and resistance, and circuit laws are used as constraints, a physical information loss function is constructed. The physical information loss function is integrated into the neural network training process to obtain node features embedded with physical information. The calculation formula is as follows:

[0073]

[0074] Where, represents the physical information loss function, I i represents the current at node i, V i represents the voltage at node i, R i represents the resistance of node i, represents the neighbor set of node j;

[0075] It should be noted that when historical fault data is insufficient, physical constraints can prevent the model from overfitting and improve the prediction ability of rare faults (such as multi-point grounding caused by salt spray corrosion). Through the physical loss term, the risk propagation path output by the model can be associated with specific physical quantities (such as current anomalies), which makes it easier for operation and maintenance personnel to locate the root cause, ensure that the risk propagation model complies with the basic operating laws of the power grid, and avoid prediction results that violate the laws of physics.

[0076] S204. Combining the node features embedded with physical information and the dynamic graph structure, the dynamic graph attention network algorithm is used to calculate the attention coefficient between nodes, and the edge weights are dynamically adjusted according to the attention coefficient. The intensity of risk propagation between different nodes is distinguished through the attention mechanism to avoid errors caused by the uniform propagation assumption.

[0077] S205. Based on edge weights and node characteristics, the node status is updated through multi-layer PI-DGAT iteration to simulate the dynamic propagation process of risk in the power grid and obtain the dynamic risk propagation path.

[0078] It should be noted that risk propagation is achieved through multi-layer PI-DGAT, and the node status update formula is:

[0079]

[0080] Where, represents the state of node i at time t, represents the neighbor set of node i, α ij (t) represents the attention coefficient of node i and node j at time t, W (l) represents the weight matrix of the lth layer, represents the feature representation of node j at layer l.

[0081] Therefore, step S200 achieves a refined characterization of the risk propagation of low-voltage distribution networks through dynamic graph structure modeling, physical information embedding, attention mechanism and typhoon data fusion, organically combining environmental factors, physical laws and deep learning, which not only ensures the physical rationality of the model, but also improves the prediction accuracy in complex environments.

[0082] S300, combined with the dynamic risk propagation path, uses the Bayesian network to calculate the failure probability distribution of each node in the power grid.

[0083] It can be understood that step S300 includes S301, S302 and S303, wherein:

[0084] S301. Using the dynamic risk propagation path matrix, historical fault data, and real-time device status data, a Bayesian network structure is constructed, where network nodes correspond to power grid devices, network edges represent risk propagation relationships between devices, and edge weights are determined by the propagation probabilities in the dynamic risk propagation path matrix. This generates a Bayesian network topology that reflects the dynamic risk propagation logic between power grid devices. The dynamic propagation path matrix is:

[0085] P ij = the probability of risk propagation from node i to node j

[0086] It is generated through spatiotemporal convolutional neural network (ST-CNN) and dynamic graph attention network (DGAT) to reflect the real-time impact of environmental factors (wind speed, salt spray concentration) on risk propagation.

[0087] S302. Combine the Bayesian network topology and historical fault data, and calculate the basic fault probability of the root node and the conditional fault probability of the non-root nodes by training the conditional probability table. The conditional probability of the non-root node is derived from the state of its parent node and the dynamic propagation probability to obtain the conditional probability distribution of the Bayesian network. The formula for the conditional probability of the non-root node is:

[0088]

[0089] xi∈{0,1}: parent node Status (0 normal, 1 fault), is the parent node k i The propagation probability to the current node j.

[0090] S303: Input the real-time device status data as evidence into the trained Bayesian network, use a probabilistic inference algorithm to calculate the posterior probability of each node under different fault types, and output the fault probability distribution of the grid nodes.

[0091] It should be noted that according to the Bayesian network topology and historical fault data, the probability table (CPT) is used as the training condition, in which the failure probability of the root node is obtained by the statistics of historical fault data, and the failure probability of the non-root node is obtained by the parent node state and the dynamic propagation probability P ij The joint calculation results in a trained Bayesian network conditional probability table, which is used to describe the fault propagation relationship between nodes. Finally, the trained Bayesian network and real-time equipment status data are used to perform probability inference and output the fault probability distribution of each node in the power grid, providing a basis for fault warning and emergency repair decision-making. The fault probability distribution matrix is:

[0092] Q i，c =P(X i =1|Fault type=c)

[0093] It can be further associated with the degree of fault impact (such as the number of users experiencing power outages and economic losses) to generate a prioritized emergency repair plan.

[0094] S400, based on the fault probability distribution, optimizes the early warning strategy through reinforcement learning algorithm to generate the optimal early warning plan.

[0095] It can be understood that step S400 includes S401, S402 and S403, wherein:

[0096] S401, obtaining the fault probability distribution and user feedback data, integrating and processing them to obtain a state space, and using the state space as the input of the reinforcement learning algorithm;

[0097] S402, using the state space and action space, through Q-learning algorithm training processing, to obtain a Q value table, wherein the Q value table is used to generate an optimal early warning strategy;

[0098] S403. Based on the Q-value table and the current state, the optimal action is selected according to the Q-value table, thereby obtaining the optimal early warning plan. The optimal early warning plan includes the warning level and the response measures, and its calculation formula is as follows:

[0099]

[0100] Where, is the optimal action, Q(s t , a) is the state-action value function.

[0101] S500: Based on the optimal early warning plan, a multi-objective optimization algorithm is used to schedule emergency repair resources and generate an emergency repair resource scheduling plan.

[0102] It can be understood that in this step, emergency repair resource data and geographic information data are obtained, and after integration processing, the input data of the emergency repair resource scheduling problem is obtained; the input data of the emergency repair resource scheduling problem and the optimal early warning plan are used, and after processing by a multi-objective optimization algorithm (such as NSGA-II), the Pareto optimal solution set is obtained; based on the Pareto optimal solution set and user feedback data, after scheme selection processing, the optimal emergency repair resource scheduling plan is obtained, including the scheduling plan and path planning of emergency repair personnel, equipment, and vehicles.

[0103] Example 2:

[0104] like Figure 2 As shown, this embodiment provides a low-voltage distribution network change risk assessment and intelligent early warning system, see Figure 2 The system comprises:

[0105] Generation module 701: used to perform fusion analysis on power grid operation data and environmental data through spatiotemporal convolutional neural network to generate a correlation feature matrix between power grid operation status and environmental risks;

[0106] Extraction module 702: for dynamically modeling the power grid topology structure based on the correlation feature matrix and extracting the dynamic risk propagation path using the physical information dynamic graph attention network;

[0107] Calculation module 703: used to calculate the failure probability distribution of each node in the power grid using a Bayesian network in combination with the dynamic risk propagation path;

[0108] Optimization module 704: used to optimize the early warning strategy based on the fault probability distribution through reinforcement learning algorithm to generate the optimal early warning plan;

[0109] The dispatching and warning module 705 is used to dispatch emergency repair resources according to the optimal warning plan using a multi-objective optimization algorithm.

[0110] Specifically, the generating module 701 includes:

[0111] Acquisition unit: used to collect first data of voltage, current and power factor in real time at a frequency of 1 minute through smart meters and sensors, obtain second data of wind speed, rainfall, temperature and humidity at a frequency of 10 minutes from the weather station, obtain third data of terrain elevation and vegetation distribution from the geographic information system, and obtain fourth data of salt spray concentration at a sampling frequency of 1 hour from corrosion monitoring equipment;

[0112] The first construction unit is used to align the first data, the second data, the third data, and the fourth data in time and space, and construct a spatiotemporal data cube. The spatiotemporal data cube is processed using a spatiotemporal convolutional neural network to obtain a correlation feature matrix between the power grid operation status and environmental risks. The calculation formula of the spatiotemporal data cube is as follows:

[0113]

[0114] Among them, X represents the fusion result of power grid operation data and environmental data, Indicates that each element in the data cube is a real number, N is the number of grid nodes, T is the time step, which indicates the number of time points for data collection, and D is the number of features of each node at each time point.

[0115] Specifically, the extraction module 702 includes:

[0116] The first acquisition unit is used to acquire the topological structure information of the power grid in the power grid management system, including node data and edge data; assign the associated feature matrix as a node to the topological structure information of the power grid to obtain the power grid topology structure with node features;

[0117] A second construction unit is configured to construct a dynamic graph structure according to the power grid topology structure with node characteristics, the second data, and the fourth data;

[0118] The third construction unit is used to construct a physical information loss function based on the dynamic graph structure and power grid operation data, where the power grid operation data includes current, voltage, and resistance, and uses circuit laws as constraints. The physical information loss function is integrated into the neural network training process to obtain node features embedded with physical information. The calculation formula is as follows:

[0119]

[0120] Where, represents the physical information loss function, I i represents the current at node i, V i represents the voltage at node i, R i represents the resistance of node i, represents the neighbor set of node j;

[0121] The first computing unit is used to combine the node features embedded with physical information and the dynamic graph structure, use the dynamic graph attention network algorithm to calculate the attention coefficient between nodes, and dynamically adjust the edge weight according to the attention coefficient;

[0122] Simulation unit: It is used to update the node status through multi-layer PI-DGAT iteration based on edge weights and node characteristics, simulate the dynamic propagation process of risks in the power grid, and obtain the dynamic risk propagation path.

[0123] Specifically, the calculation module 703 includes:

[0124] The fourth construction unit is used to construct a Bayesian network structure using the dynamic risk propagation path matrix, historical fault data, and real-time device status data. The network nodes correspond to power grid devices, the network edges represent the risk propagation relationships between devices, and the edge weights are determined by the propagation probabilities in the dynamic risk propagation path matrix. This generates a Bayesian network topology that reflects the dynamic risk propagation logic between power grid devices.

[0125] A second calculation unit is used to combine the Bayesian network topology and historical fault data to calculate the basic fault probability of the root node and the conditional fault probability of the non-root nodes by training the conditional probability table, wherein the conditional probability of the non-root node is derived from the state of its parent node and the dynamic propagation probability to obtain the conditional probability distribution of the Bayesian network;

[0126] The third computing unit is used to input the real-time device status data as evidence into the trained Bayesian network, use the probabilistic inference algorithm to calculate the posterior probability of each node under different fault types, and output the fault probability distribution of the grid nodes.

[0127] Specifically, the optimization module 704 includes:

[0128] The second acquisition unit is used to obtain the fault probability distribution and user feedback data, integrate and process them to obtain the state space, and use the state space as the input of the reinforcement learning algorithm;

[0129] Processing unit: used to use the state space and action space, and obtain the Q value table through Q-learning algorithm training, where the Q value table is used to generate the optimal early warning strategy;

[0130] Acquisition unit: used to select the optimal action based on the Q value table and the current state, and then obtain the optimal warning plan. The optimal warning plan includes the warning level and response measures. The calculation formula is as follows:

[0131]

[0132] Where, is the optimal action, Q(s t , a) is the state-action value function.

[0133] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated on here.

[0134] Example 3:

[0135] Corresponding to the above method embodiment, this embodiment also provides a low-voltage distribution network change risk assessment and intelligent early warning device. The low-voltage distribution network change risk assessment and intelligent early warning device described below and the low-voltage distribution network change risk assessment and intelligent early warning method described above can be referenced to each other.

[0136] Figure 3 FIG. 8 is a block diagram of a low-voltage distribution network change risk assessment and intelligent early warning device 800 according to an exemplary embodiment. Figure 3 As shown, the low-voltage distribution network change risk assessment and intelligent early warning device 800 includes: a processor 801 and a memory 802. The low-voltage distribution network change risk assessment and intelligent early warning device 800 also includes one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0137] The processor 801 is used to control the overall operation of the low-voltage distribution network change risk assessment and intelligent early warning device 800 to complete all or part of the steps in the low-voltage distribution network change risk assessment and intelligent early warning method described above. The memory 802 is used to store various types of data to support the operation of the low-voltage distribution network change risk assessment and intelligent early warning device 800. This data may include, for example, instructions for any application or method operating on the low-voltage distribution network change risk assessment and intelligent early warning device 800, as well as application-related data, such as contact information, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be a keyboard, a mouse or buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the low-voltage distribution network change risk assessment and intelligent early warning device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, near field communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 may include: a Wi-Fi module, a Bluetooth module or an NFC module.

[0138] In an exemplary embodiment, the low-voltage distribution network change risk assessment and intelligent early warning device 800 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned low-voltage distribution network change risk assessment and intelligent early warning method.

[0139] In another exemplary embodiment, a computer-readable storage medium including program instructions is further provided. When executed by a processor, the program instructions implement the steps of the aforementioned low-voltage distribution network change risk assessment and intelligent early warning method. For example, the computer-readable storage medium may be the aforementioned memory 802 including the program instructions. The program instructions may be executed by the processor 801 of the low-voltage distribution network change risk assessment and intelligent early warning device 800 to implement the aforementioned low-voltage distribution network change risk assessment and intelligent early warning method.

[0140] Example 4:

[0141] Corresponding to the above method embodiment, this embodiment also provides a readable storage medium. The readable storage medium described below and the low-voltage distribution network change risk assessment and intelligent early warning method described above can refer to each other.

[0142] A computer program is stored on the readable storage medium, and when the computer program is executed by the processor, the steps of the low-voltage distribution network change risk assessment and intelligent early warning method of the above method embodiment are implemented.

[0143] The readable storage medium may specifically be any readable storage medium that can store program code, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0144] In summary, a spatiotemporal convolutional neural network is used to extract the correlation characteristics between the grid's operating status and environmental risks. This is combined with a physical information dynamic graph attention network to model the dynamic changes in the grid topology. Circuit laws are incorporated as constraints to ensure the model's physical rationality. Furthermore, a Bayesian network is used to calculate the fault probability distribution of each node. A reinforcement learning algorithm is then used to optimize the early warning strategy and generate an optimal emergency repair resource scheduling plan. This approach not only captures the impact of environmental factors on the grid in real time, but also accurately predicts the risk propagation path before a fault occurs, significantly improving the accuracy of fault warnings and the efficiency of emergency repairs. This provides reliable technical support for the intelligent operation and maintenance of low-voltage distribution networks.

[0145] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

[0146] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A low-voltage distribution network change risk assessment and intelligent early warning method, characterized by: include: The grid operation data and environmental data are integrated and analyzed through a spatiotemporal convolutional neural network to generate a correlation feature matrix between the grid operation status and environmental risks. Based on the correlation feature matrix, the physical information dynamic graph attention network is used to dynamically model the power grid topology and extract the dynamic risk propagation path; Combined with the dynamic risk propagation path, the Bayesian network is used to calculate the failure probability distribution of each node in the power grid; Based on the fault probability distribution, the early warning strategy is optimized through reinforcement learning algorithm to generate the optimal early warning plan; According to the optimal early warning plan, the multi-objective optimization algorithm is used to schedule the emergency repair resources and generate the emergency repair resource scheduling plan.

2. The low-voltage distribution network change risk assessment and intelligent early warning method according to claim 1 is characterized in that: The grid operation data and environmental data are fused and analyzed through the spatiotemporal convolutional neural network to generate a correlation feature matrix between the grid operation status and environmental risks, which includes: The first data of voltage, current and power factor are collected in real time through smart meters and sensors at a frequency of 1 minute. The second data of wind speed, rainfall, temperature and humidity are obtained from the weather station at a frequency of 10 minutes. The third data of terrain elevation and vegetation distribution are obtained from the geographic information system. The fourth data of salt spray concentration is obtained from the corrosion monitoring equipment at a sampling frequency of 1 hour. The first, second, third, and fourth data are aligned in time and space, and a spatiotemporal data cube is constructed. The spatiotemporal data cube is processed using a spatiotemporal convolutional neural network to obtain a correlation feature matrix between the power grid operation status and environmental risks. The calculation formula for the spatiotemporal data cube is as follows: Among them, X represents the fusion result of power grid operation data and environmental data, Indicates that each element in the data cube is a real number, N is the number of grid nodes, T is the time step, which indicates the number of time points for data collection, and D is the number of features of each node at each time point.

3. The low-voltage distribution network change risk assessment and intelligent early warning method according to claim 2 is characterized in that: Based on the correlation feature matrix, the physical information dynamic graph attention network is used to dynamically model the power grid topology and extract the dynamic risk propagation path, including: Obtain the topological structure information of the power grid in the power grid management system, including node data and edge data; assign the associated feature matrix as a node to the topological structure information of the power grid to obtain the power grid topology structure with node characteristics; constructing a dynamic graph structure according to the power grid topology structure with node characteristics, the second data, and the fourth data; Based on the dynamic graph structure and power grid operation data, where the power grid operation data includes current, voltage, and resistance, and circuit laws are used as constraints, a physical information loss function is constructed. The physical information loss function is integrated into the neural network training process to obtain the node features embedded with physical information. The calculation formula is as follows: Where, represents the physical information loss function, I i represents the current at node i, V i represents the voltage at node i, R i represents the resistance of node i, represents the neighbor set of node j; Combining the node features embedded with physical information and the dynamic graph structure, the dynamic graph attention network algorithm is used to calculate the attention coefficient between nodes and dynamically adjust the edge weight according to the attention coefficient; Based on edge weights and node characteristics, the node status is iteratively updated through multi-layer PI-DGAT to simulate the dynamic propagation process of risk in the power grid and obtain the dynamic risk propagation path.

4. The low-voltage distribution network change risk assessment and intelligent early warning method according to claim 1 is characterized in that: The dynamic risk propagation path is combined with the Bayesian network to calculate the failure probability distribution of each node in the power grid, including: A Bayesian network structure is constructed using a dynamic risk propagation path matrix, historical fault data, and real-time device status data. Network nodes correspond to power grid devices, network edges represent risk propagation relationships between devices, and edge weights are determined by the propagation probabilities in the dynamic risk propagation path matrix. This generates a Bayesian network topology that reflects the dynamic risk propagation logic between power grid devices. Combining the Bayesian network topology and historical fault data, the basic fault probability of the root node and the conditional fault probability of non-root nodes are calculated by training the conditional probability table. The conditional probability of the non-root node is derived from the state of its parent node and the dynamic propagation probability, thus obtaining the conditional probability distribution of the Bayesian network. The real-time device status data is input into the trained Bayesian network as evidence, and a probabilistic inference algorithm is used to calculate the posterior probability of each node under different fault types, and the fault probability distribution of the power grid nodes is output.

5. The low-voltage distribution network change risk assessment and intelligent early warning method according to claim 1 is characterized in that: The aforementioned method optimizes the early warning strategy through reinforcement learning algorithm to generate the optimal early warning plan, which includes: Obtain the fault probability distribution and user feedback data, integrate and process them to obtain the state space, and use the state space as the input of the reinforcement learning algorithm; Using the state space and action space, after training with the Q-learning algorithm, a Q-value table is obtained, where the Q-value table is used to generate the optimal early warning strategy; Based on the Q-value table and the current state, the optimal action is selected according to the Q-value table, and then the optimal early warning plan is obtained. The optimal early warning plan includes the warning level and response measures. The calculation formula is as follows: Where, is the optimal action, Q(s t , a) is the state-action value function.

6. A low-voltage distribution network change risk assessment and intelligent early warning system, based on the low-voltage distribution network change risk assessment and intelligent early warning method according to claim 1, characterized in that: include: Generation module: used to integrate and analyze power grid operation data and environmental data through spatiotemporal convolutional neural networks to generate a correlation feature matrix between power grid operation status and environmental risks; Extraction module: used to dynamically model the power grid topology based on the correlation feature matrix and use the physical information dynamic graph attention network to extract the dynamic risk propagation path; Calculation module: used to combine dynamic risk propagation paths and use Bayesian networks to calculate the failure probability distribution of each node in the power grid; Optimization module: used to optimize the early warning strategy based on the fault probability distribution through reinforcement learning algorithm to generate the optimal early warning plan; Dispatching and early warning module: used to dispatch emergency repair resources based on the optimal early warning plan using a multi-objective optimization algorithm.

7. The low-voltage distribution network change risk assessment and intelligent early warning system according to claim 6 is characterized in that: The generation module includes: Acquisition unit: used to collect first data of voltage, current and power factor in real time at a frequency of 1 minute through smart meters and sensors, obtain second data of wind speed, rainfall, temperature and humidity at a frequency of 10 minutes from the weather station, obtain third data of terrain elevation and vegetation distribution from the geographic information system, and obtain fourth data of salt spray concentration at a sampling frequency of 1 hour from corrosion monitoring equipment; The first construction unit is used to align the first data, the second data, the third data, and the fourth data in time and space, and construct a spatiotemporal data cube. The spatiotemporal data cube is processed using a spatiotemporal convolutional neural network to obtain a correlation feature matrix between the power grid operation status and environmental risks. The calculation formula of the spatiotemporal data cube is as follows: Among them, X represents the fusion result of power grid operation data and environmental data, Indicates that each element in the data cube is a real number, N is the number of grid nodes, T is the time step, which indicates the number of time points for data collection, and D is the number of features of each node at each time point.

8. The low-voltage distribution network change risk assessment and intelligent early warning system according to claim 7 is characterized in that: The extraction module includes: The first acquisition unit is used to acquire the topological structure information of the power grid in the power grid management system, including node data and edge data; assign the associated feature matrix as a node to the topological structure information of the power grid to obtain the power grid topology structure with node features; A second construction unit is configured to construct a dynamic graph structure according to the power grid topology structure with node characteristics, the second data, and the fourth data; The third construction unit is used to construct a physical information loss function based on the dynamic graph structure and power grid operation data, where the power grid operation data includes current, voltage, and resistance, and uses circuit laws as constraints. The physical information loss function is integrated into the neural network training process to obtain node features embedded with physical information. The calculation formula is as follows: Where, represents the physical information loss function, I i represents the current at node i, V i represents the voltage at node i, R i represents the resistance of node i, represents the neighbor set of node j; The first computing unit is used to combine the node features embedded with physical information and the dynamic graph structure, use the dynamic graph attention network algorithm to calculate the attention coefficient between nodes, and dynamically adjust the edge weight according to the attention coefficient; Simulation unit: It is used to update the node status through multi-layer PI-DGAT iteration based on edge weights and node characteristics, simulate the dynamic propagation process of risks in the power grid, and obtain the dynamic risk propagation path.

9. The low-voltage distribution network change risk assessment and intelligent early warning system according to claim 6 is characterized in that: The computing module includes: The fourth construction unit is used to construct a Bayesian network structure using the dynamic risk propagation path matrix, historical fault data, and real-time device status data. The network nodes correspond to power grid devices, the network edges represent the risk propagation relationships between devices, and the edge weights are determined by the propagation probabilities in the dynamic risk propagation path matrix. This generates a Bayesian network topology that reflects the dynamic risk propagation logic between power grid devices. A second calculation unit is used to combine the Bayesian network topology and historical fault data to calculate the basic fault probability of the root node and the conditional fault probability of the non-root nodes by training the conditional probability table, wherein the conditional probability of the non-root node is derived from the state of its parent node and the dynamic propagation probability to obtain the conditional probability distribution of the Bayesian network; The third computing unit is used to input the real-time device status data as evidence into the trained Bayesian network, use the probabilistic inference algorithm to calculate the posterior probability of each node under different fault types, and output the fault probability distribution of the grid nodes.

10. The low-voltage distribution network change risk assessment and intelligent early warning system according to claim 6 is characterized in that: The optimization module includes: The second acquisition unit is used to obtain the fault probability distribution and user feedback data, integrate and process them to obtain the state space, and use the state space as the input of the reinforcement learning algorithm; Processing unit: used to use the state space and action space, and obtain the Q value table through Q-learning algorithm training, where the Q value table is used to generate the optimal early warning strategy; Acquisition unit: used to select the optimal action based on the Q value table and the current state, and then obtain the optimal warning plan. The optimal warning plan includes the warning level and response measures. The calculation formula is as follows: Where, is the optimal action, Q(s t , a) is the state-action value function.

Images