Grid Emergency Resource Scheduling Optimization Method Based on Multi-Dimensional Data Fusion of Source, Grid, Load and Storage

Abstract

The present invention relates to the technical field of emergency resource scheduling. Specifically, it relates to an optimization method for power grid emergency resource scheduling based on the multi-dimensional data fusion of source, network, load, and storage. The method includes the following steps: Deploy edge computing nodes at key nodes of the power grid to collect real-time power grid operation data, meteorological data, and geographical information data, use a graph neural network for spatio-temporal feature fusion, and adopt a median aggregation method to optimize the process of spatio-temporal feature fusion by the graph neural network; Based on the fused spatio-temporal features, use a long short-term memory network model for load forecasting; Based on the prediction results, construct an emergency resource scheduling optimization model through dynamic programming method to generate the optimal resource scheduling path. The design of the present invention can more accurately capture the complex structure and dynamic changes of the power grid system by integrating power grid operation data, meteorological data, geographical information, and equipment status data, and using a graph neural network for spatio-temporal feature fusion.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency resource scheduling, and specifically, to an optimization method for power grid emergency resource scheduling based on multi-dimensional data fusion of source-network-load-storage. Background Art

[0002] The power grid system consists of different types of nodes (such as substations, distribution terminals, sensor networks, etc.) and connection relationships. These nodes and edges have different attributes and functions. Directly performing graph convolution operations may ignore the essential differences between nodes and fail to accurately capture the complex structure and dynamic changes of the power grid. During the operation of the power grid, a large amount of real-time data is generated, including power grid operation data such as current, voltage, and power, meteorological data such as wind speed, temperature, and precipitation, and geographical information data such as the topological structure of the power grid and the geographical location of equipment. How to effectively integrate these multi-source heterogeneous data and extract useful information for decision support is a challenge. When a fault or emergency occurs in the power grid, quickly locating the fault point, assessing the impact range, and formulating an effective resource scheduling plan are crucial. However, existing methods often lack sufficient flexibility and adaptability. When a power outage fault occurs, it leads to an extended power outage time and affects power supply reliability. In addition, with the transformation of the energy structure, the coordinated operation of source-network-load-storage has become an important direction for the development of the power grid. The volatility of the source side (such as new energy power generation), the diversity of the load side (user load), and the dynamic regulation characteristics of energy storage have brought new challenges to power grid emergency resource scheduling. Existing methods mostly focus on the optimization of a single link and lack consideration of the dynamic coupling of multiple links of source-network-load-storage, making it difficult to achieve efficient collaborative scheduling of global resources. Therefore, an optimization method for power grid emergency resource scheduling based on multi-dimensional data fusion of source-network-load-storage is provided. Summary of the Invention

[0003] The purpose of the present invention is to provide an optimization method for power grid emergency resource scheduling based on multi-dimensional data fusion of source-network-load-storage to solve the problems of dynamic changes and heterogeneity of complex power grid systems, the effective integration of multi-source heterogeneous data, and the current situation of slow emergency response speed and low resource scheduling efficiency proposed in the above background art.

[0004] To achieve the above purpose, the present invention aims to provide an optimization method for power grid emergency resource scheduling based on multi-dimensional data fusion of source-network-load-storage, including the following steps:

[0005] S1. Deploy edge computing nodes at key nodes of the power grid to collect real-time power grid operation data, meteorological data, and geographical information data;

[0006] S2. Integrate power grid operation data, meteorological data, geographical information, and equipment status data, perform spatio-temporal feature fusion using a graph neural network, optimize the spatio-temporal feature fusion process of the graph neural network by combining node type embedding vectors with real-time state factors and adopting a median aggregation method;

[0007] S3. Based on the fused spatio-temporal features, use a long short-term memory network model for load forecasting;

[0008] S4. Based on the prediction results, construct an emergency resource scheduling optimization model through dynamic programming method to generate the optimal resource scheduling path.

[0009] As a further improvement of this technical solution, in S2, integrating power grid operation data, meteorological data, geographical information, and equipment status data, and performing spatio-temporal feature fusion using a graph neural network includes the following steps:

[0010] S2.1. Regard the power grid as a graph ;

[0011] Among them, represents the nodes in the power grid, represents the connection relationship between nodes;

[0012] S2.2. For each node Merge into a comprehensive feature vector :

[0013] ;

[0014] Among them, represents power grid operation data; represents meteorological data; represents geographical information; represents equipment status data; [[ID=4...]]

[0015] S2.3. Initialize the feature of each node , integrate all features from different sources into the initial feature , as the input of the graph neural network;

[0016] S2.4. Perform multi-layer graph convolution operations to form the feature of each node, including the spatial domain features, meteorological features, and geographical information features of the nodes;

[0017] S2.5. For the operation data of the power grid, use the sliding window technique to extract the time series features in the operation data of the power grid;

[0018] S2.6. Combine the time series features obtained from time series analysis with the spatial domain features obtained through the graph neural network to form spatio-temporal features.

[0019] As a further improvement of this technical solution, in S2.4, perform multi-layer graph convolution operations, including the following steps:

[0020] S2.41. Initialize the node features , and use as the input node feature vector and input it into the first layer of graph convolution;

[0021] S2.42. For each node , determine the set of neighbor nodes of node ;

[0022] S2.43. Define the weight matrix and bias term for each layer of convolution operation;

[0023] S2.44. In the first layer of graph convolution operation, the node feature vector and the feature vectors of its neighbor nodes are aggregated through the graph convolution formula, combined with the node type embedding vector and the real-time state factor, and the median aggregation method is used to optimize the graph convolution operation;

[0024] S2.45. Based on the updated features of the previous layer, the node features continue to be aggregated with the feature vectors of its neighbor nodes until the predetermined number of layers is reached ;

[0025] S2.46. After completing the convolution of all layers, obtain the final feature representation of each node .

[0026] As a further improvement of this technical solution, in S2.44, the node feature vector and the feature vectors of its neighbor nodes are aggregated through the graph convolution formula, including the following steps:

[0027] S2.441. For the current node, traverse all its neighbor nodes, multiply the feature representation of each neighbor node in the layer by the weight matrix of the current layer, and add the bias;

[0028] S2.442. Normalize the weighted features of the neighbor nodes;

[0029] S2.443. After summing the normalized features of all neighbor nodes, generate the new feature of the current node in the layer through the non-linear activation function.

[0030] As a further improvement of this technical solution, in S2.44, combining the node type embedding vector and the real-time state factor and using the median aggregation method to optimize the graph convolution operation is as follows:

[0031] For the problem of grid node type heterogeneity, assign node type embedding vectors to different node types , encode the node type features into the graph convolution operation for optimization, and calculate the attention coefficient , dynamically quantify the importance weights of different neighbor nodes to the current node to obtain the attention coefficient ;

[0032] Filter the low-confidence neighbors to obtain the filtered neighbor set :

[0033] Use the median aggregation method to suppress the influence of outliers on feature aggregation to obtain the neighbor node features after weighted median processing :

[0034] By combining the node type embedding vector with the real-time state factor , and using the median aggregation method to optimize the graph convolution formula after the graph convolution operation.

[0035] As a further improvement of this technical solution, in S2.5, for the operation data of the power grid, use the sliding window technique to extract the time series features in the operation data of the power grid, including the following steps:

[0036] S2.51. Determine the time window size and set the sliding step;

[0037] S2.52. Starting from the starting point of the time series data, initialize the position of the first window, and move the window position sequentially through the sliding window method until the window slides to the last time point;

[0038] S2.53. Extract key features within each window;

[0039] S2.54. Construct a feature vector from the extracted features of each sliding window, and the feature vector is the time series feature.

[0040] As a further improvement of this technical solution, in S3, based on the fused spatio-temporal features, use the long short-term memory network model for load forecasting, including the following steps:

[0041] S3.1. Input the fused spatio-temporal features into the long short-term memory network model;

[0042] S3.2. Process the time series data through the long short-term memory network layer to capture the time-dependent relationship of load changes;

[0043] S3.3. Each long short - term memory network unit outputs a hidden layer state;

[0044] S3.4. The output of the long short - term memory network is converted into a load prediction value through a fully - connected layer;

[0045] S3.5. Use the mean square error as the loss function to measure the difference between the predicted value and the true value of the long short - term memory network model, and train the long short - term memory network model;

[0046] S3.6. Use the trained long short - term memory network model to predict the load.

[0047] As a further improvement of this technical solution, in S4, an emergency resource scheduling optimization model is constructed through the dynamic programming method based on the prediction result to generate an optimal resource scheduling path, including the following steps:

[0048] S4.1. Define the state variable and the set of all positions of the resources ;

[0049] S4.2. Define the decision variable and select the scheduling path in the current state t;

[0050] S4.3. Define the state transition cost and the state transition equation, and introduce the task - dependency graph and the hierarchical optimization strategy to optimize the state transition equation;

[0051] S4.4. Set the initial state and the termination condition. The initial state is the starting point where the resources are located, and the termination state is that all tasks are completed;

[0052] S4.5. Through backward recursion, obtain the cumulative cost of each path reaching the final state, and select the path with the minimum cost among the cumulative costs of the paths as the optimal path;

[0053] S4.6. Backtrack the optimal decision of each state to obtain the optimal resource scheduling path.

[0054] As a further improvement of this technical solution, in S4.3, the state transition equation is:

[0055] ;

[0056] Among them, represents the minimum cumulative cost from the current time and node to the completion of all tasks; represents the node where the resource is located at time ; represents the time when the resource is at The next node to arrive at; Indicates the set of neighbor nodes that can be moved to from the current node ; Indicates the set of neighbor nodes that can be moved to from the node To The comprehensive cost of; Indicates the minimum cumulative cost from the current time And the node To the completion of all tasks;

[0057] Regarding the problem of the order of power grid emergency resource scheduling, a task dependency graph constraint is introduced into the state transition equation to optimize the scheduling path. Only when the prerequisite tasks of the node are completed, the node is added to the set :

[0058] Define the task dependency matrix , if task Depends on task , then ;

[0059] Update the set of feasible nodes: , where Indicates all the tasks that have been completed when reaching state ;

[0060] The power grid area is divided into sub-areas, and the state transition equation is optimized through a hierarchical optimization strategy. Coarse-grained paths are generated based on the key nodes of the sub-areas, and the paths are refined within the sub-areas. The global layer state variable is the sub-area ID, and the local layer is the detailed node ID;

[0061] Introduce a task dependency graph and a hierarchical optimization strategy to optimize the state transition equation.

[0062] As a further improvement of this technical solution, in S4.5, through reverse recursion, the cumulative cost of each path reaching the final state is obtained, and the path with the minimum cost among the cumulative costs of the paths is selected as the optimal path, including the following steps:

[0063] S4.51. Set the initial conditions for recursion;

[0064] S4.52. Solve to obtain all the next states That can be transferred from the current state Set ;

[0065] S4.53. For each state In the set Calculate the cost of transferring from To The total cost to continue to the end point;

[0066] S4.54. Among all the transfers, select the one with the minimum total cost as the value of the minimum cumulative cost, and record the corresponding optimal transfer state at the same time;

[0067] S4.55. Repeat steps S4.52 to S4.54 until the optimal costs of all states at the initial moment are calculated.

[0068] Compared with the prior art, the beneficial effects of the present invention are:

[0069] 1. In the grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source-grid-load-storage, by integrating grid operation data, meteorological data, geographical information, and equipment status data, and using a graph neural network (GNN) for spatio-temporal feature fusion, this method can more accurately capture the complex structure and dynamic changes of the grid system. Further using a long short-term memory network (LSTM) model to predict the load can effectively capture the time-dependent relationship of load changes, thereby improving the accuracy of load prediction. Based on these prediction results, the emergency resource scheduling optimization model constructed by the dynamic programming (DP) method can generate the optimal resource scheduling path, greatly improving the efficiency and response speed of resource scheduling, especially being able to quickly respond in case of emergencies, reducing the power outage time, and enhancing the power supply reliability.

[0070] 2. In the grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source-grid-load-storage, the heterogeneity of grid node types is considered. By assigning type embedding vectors to different node types and adjusting the attention coefficient in combination with real-time state factors, the model can better understand and process the characteristics of different types of nodes, enhancing the model's adaptability to the dynamic changes of the grid. In addition, aiming at the problem that scheduling all resources at the global level will lead to too high computational complexity, a two-layer optimization strategy is adopted. First, a coarse-grained path is generated based on the key nodes in the sub-region at the global level, and then the path is refined within the sub-region. This method not only alleviates the curse of dimensionality problem but also speeds up the decision-making speed, making the emergency resource scheduling in large-scale grids more efficient and feasible. This greatly improves the ability to cope with complex grid environments and also ensures the effective utilization of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 It is the overall method flow chart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0072] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0073] Embodiment: Please refer to Figure 1 As shown, this embodiment provides a power grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source, grid, load, and storage, including the following steps:

[0074] S1. Deploy edge computing nodes at key power grid nodes (including substations, distribution terminals, and sensor networks) to collect real-time power grid operation data (current, voltage, power), meteorological data (wind speed, temperature, precipitation), geographical information data (topological structure of the power network, geographical location of equipment), and new energy output data (photovoltaic, wind power) on the source side, so as to realize the real-time monitoring of all links of the source, grid, load, and storage;

[0075] S2. Integrate power grid operation data, meteorological data, geographical information, and equipment status data, and use a graph neural network (GNN) for spatio-temporal feature fusion;

[0076] In this embodiment, the graph neural network (GNN) is a neural network model specifically designed to process graph-structured data. It learns the representations of nodes, edges, or the entire graph through information propagation of the nodes and edges of the graph; the power grid system itself is a complex network composed of different types of nodes (such as substations, distribution terminals, sensor networks, etc.) and edges (power line connections, communication networks, etc.). By regarding the power grid as a graph and using the graph neural network (GNN), the complex spatial dependence relationships between nodes can be effectively captured, which is crucial for understanding the topological structure of the power grid and its dynamic changes; through the graph convolution operation and sliding window technology of the graph neural network, time series features are extracted, realizing the effective fusion of spatio-temporal features of the power grid system. This method not only considers the spatial characteristics of the power grid system but also combines its temporal variation laws, thereby improving the accuracy of the model in aspects such as future load prediction;

[0077] Integrating power grid operation data, meteorological data, geographical information, and equipment status data, and using a graph neural network (GNN) for spatio-temporal feature fusion includes the following steps:

[0078] S2.1. Regard the power grid as a graph ;

[0079] Among them, represents the nodes in the power grid (substations, distribution terminals, sensor networks, load points, etc.), Represents the connection relationship between nodes (such as power line connection, communication network, etc.);

[0080] S2.2. For each node Merge into a comprehensive feature vector :

[0081] ;

[0082] Among them, Represents the power grid operation data (including the output of the source side, the demand of the load side, and the charge and discharge power of the energy storage, to characterize the coordinated operation characteristics of the source-grid-load-storage); Represents meteorological data; Represents geographical information; Represents the equipment status data (including status parameters such as the remaining capacity and charge and discharge efficiency of the energy storage equipment);

[0083] S2.3. Initialize the features of each node of , integrate all features from different sources (such as power grid data, meteorological data, etc.) into the initial features , as the input of the graph neural network;

[0084] S2.4. Perform multi-layer graph convolution operations. The core of the graph neural network is graph convolution, which updates the features of nodes through a message passing mechanism. In the graph neural network, the features of nodes are affected by the information of their neighbor nodes in each layer of the network, forming the features of each node , including the spatial domain features, meteorological features, and geographical information features of the nodes;

[0085] Among them, performing multi-layer graph convolution operations includes the following steps:

[0086] S2.41. Initialize the node features , and as the input node feature vector, and input it into the first layer of graph convolution;

[0087] S2.42. For each node , determine the set of neighbor nodes of node ;

[0088] S2.43. Define the weight matrix and bias term for each layer of convolution operation;

[0089] S2.44. In the first layer of graph convolution operation, the node feature vector and the feature vectors of its neighbor nodes are aggregated through the graph convolution formula, combined with the node type embedding vector and the real-time state factor, and the median aggregation method is used to optimize the graph convolution operation;

[0090] Furthermore, the node feature vector and the feature vectors of its neighbor nodes are aggregated through the graph convolution formula, including the following steps:

[0091] S2.441. For the current node, traverse all its neighbor nodes, multiply the feature representation of each neighbor node at the -th layer by the weight matrix of the current layer, and add the bias;

[0092] S2.442. Normalize the weighted features of the neighbor nodes, that is, multiply by a scaling factor to balance the influence of the number of neighbor nodes of different nodes;

[0093] S2.443. After summing the normalized features of all neighbor nodes, generate the new feature of the current node at the -th layer through a non-linear activation function;

[0094] The graph convolution formula is:

[0095] ;

[0096] where, represents the set of neighbors of node ; represents the weight matrix of the -th layer; represents the bias term of the -th layer; represents the ReLU activation function; represents the index of the current node; represents node ; represents node at the -th layer after graph convolution, represents the neighbor node of node ; represents the layer index of the convolutional network;

[0097] The power grid itself is a heterogeneous network composed of different types of data nodes and connection relationships. Different types of nodes and edges may have different attributes and functions. In this case, if the differences in node types are not considered and graph convolution operations are directly performed, the model may ignore the essential differences between nodes, thus failing to accurately capture the complex structure and dynamic changes of the power grid. Encoding the node type features into the graph convolution operation can enable the model to better adapt to the heterogeneous network characteristics of the power grid and improve the modeling ability of the power grid system;

[0098] Regarding the problem of heterogeneity in power grid node types (specifically manifested as differences in the physical properties of substations and sensors. Substations and sensors in the power grid have significantly different physical properties. A substation is a power facility in the power system that transforms voltage, receives and distributes electrical energy, controls the flow of electricity, and adjusts voltage. It is large in scale and complex in function, involving a large number of power equipment and complex electrical connections. Sensors, on the other hand, are mainly used to monitor various physical quantities in the power grid, such as voltage, current, temperature, etc. Their functions are relatively single, and their physical structures are also relatively simple. Such differences in physical properties will lead to significant differences in their roles and behavior patterns in the power grid; by assigning type embedding vectors to different node types, the characteristic information brought about by these physical property differences can be encoded into the graph convolution operation, enabling the model to better understand and process the characteristics of these different types of nodes), node type embedding vectors are assigned to different node types (substations, sensors). (The node type embedding vector includes node function characteristics, node connection degree characteristics, and geographical location characteristics), encoding the node type characteristics into the graph convolution operation for optimization, and calculating the attention coefficient , dynamically quantifying the importance weights of different neighbor nodes for the current node (by capturing the characteristic differences of different types of nodes through the embedding vector, enhancing the model's understanding ability of heterogeneous networks. In the power grid, the influence degrees of different neighbor nodes on the current node are different, and this influence degree may change with the operating state and time of the power grid. Through the attention mechanism, the model can dynamically adjust the importance weights of neighbor nodes according to the node type and current state information, so as to more flexibly capture the relationships between nodes. For example, during a fault, the substation adjacent to the fault node may have a greater impact on the current node, and the real-time monitoring data provided by sensors may also become more important. The attention coefficient enables the model to dynamically adjust the weights according to these situations, improving the model's adaptability to the dynamic changes of the power grid):

[0099] ;

[0100] Among them, represents normalizing the calculation results of all neighbor nodes, represents the weight matrix of the k-th layer, and [;] represents vector concatenation;

[0101] Introduce a real-time state factor (including equipment health and data confidence) to correct the attention coefficient, and this real-time state factor will affect the attention coefficient , used to correct the weights of the influence of neighbor nodes on the target node (by type embedding and the state factor , distinguishing key nodes (such as substations in high-fault areas) from ordinary nodes), the real-time state factor The introduction, essentially, deeply combines the dynamic operation characteristics of the power grid with the attention mechanism of graph convolution, enabling the model to dynamically adjust neighbor weights to adapt to the instantaneous changes of the power grid, distinguish key nodes to ensure the feature dominance of core facilities, and integrate dynamic and static information to enhance the modeling ability for heterogeneous networks:

[0102] ;

[0103] Among them, represents the corrected attention coefficient, taking into account the device health of the equipment or the confidence of the data;

[0104] Filter low-confidence neighbors (such as sensors with missing data). To reduce the impact of low-confidence nodes on the model, it is necessary to filter out neighbor nodes with poor device health or low data confidence:

[0105] ;

[0106] Among them, represents the filtered neighbor set, only containing those neighbor nodes with a relatively high real-time status factor ; represents the filtering threshold;

[0107] Adopt the median aggregation method (weighted median instead of mean) to suppress the impact of outliers on feature aggregation (in the power grid scenario, sensor data may generate outliers (such as mutated current / voltage values) due to equipment failures or communication interferences. If the mean aggregation is used, these outliers will significantly distort the feature representation, leading to misjudgments by the model. By weighting with the attention coefficient combined with the real-time status factor (such as data confidence), the median aggregation can not only suppress outliers but also retain the contributions of high-confidence neighbors, such as accidental outlier data still existing after filtering; the weighted median can more flexibly reflect the priorities of key nodes (such as substations with high health), while reducing the interference of low-confidence nodes):

[0108] ;

[0109] Among them, represents the weighted median operation; represents the neighbor node features after weighted median processing;

[0110] To sum up, by combining the node type embedding vector with the real-time status factor , and adopting the median aggregation method to optimize the graph convolution formula after graph convolution operation is:

[0111] ;

[0112] Among them, represents the characteristics of the optimized nodes; represents a multi-layer perceptron for further processing the characteristics of nodes ; represents the weight coefficient for controlling the contribution ratio of the node's own characteristics (the result after being processed by the MLP) in the overall feature update.

[0113] S2.45. Based on the updated characteristics of the previous layer, the node characteristics continue to be aggregated with the feature vectors of its neighbor nodes. At each layer, the node characteristics are updated according to the characteristics of the neighbor nodes, continuously transmitting information until the predetermined number of layers is reached. For each additional layer, the node will be able to obtain information from more distant neighbors, thereby capturing a larger range of spatial dependence relationships;

[0114] S2.46. After completing the convolution of all layers, the final feature representation of each node is obtained .

[0115] S2.5. For the operation data of the power grid, the sliding window technique is used to extract the time series characteristics in the operation data of the power grid;

[0116] Among them, for the operation data of the power grid, using the sliding window technique to extract the time series characteristics in the operation data of the power grid includes the following steps:

[0117] S2.51. Determine the time window size (i.e., the length of the time period processed each time) and set the sliding step (i.e., the distance the window moves each time);

[0118] S2.52. Starting from the starting point of the time series data, initialize the position of the first window, and move the window position sequentially through the sliding window method until the window slides to the last available time point;

[0119] S2.53. In each window, extract key characteristics, including statistical characteristics and frequency domain characteristics;

[0120] S2.54. Construct a feature vector from the extracted features of each sliding window, that is, the time series characteristics;

[0121] S2.6. Combine (concatenate) the time series characteristics obtained from time series analysis with the spatial domain characteristics obtained through the graph neural network to form spatio-temporal characteristics. The graph convolution operation aggregates the time series characteristics of neighbor nodes in the spatio-temporal domain to achieve the integration of spatio-temporal characteristics.

[0122] S3. Based on the fused spatio-temporal characteristics, use the long short-term memory network (LSTM) model for load forecasting;

[0123] In this embodiment, LSTM is a special type of Recurrent Neural Network (RNN) that is particularly good at processing and predicting time series data. It can effectively capture long-term dependencies in time series, which is especially important for load forecasting because electrical loads are usually affected by historical load patterns, seasonal variations, and periodic factors such as weekdays / weekends. By integrating the spatio-temporal features of the power grid (including geographical location information, the status of neighboring nodes, etc.), LSTM can utilize the spatial features extracted by the Graph Neural Network (GNN) to enhance the performance of the prediction model. This method not only considers the changing trends in the time dimension but also integrates the mutual influences between different geographical locations, making the prediction more accurate and comprehensive.

[0124] Based on the fused spatio-temporal features, a Long Short-Term Memory Network (LSTM) model is used for load forecasting, which includes the following steps:

[0125] S3.1: Input the fused spatio-temporal features into the Long Short-Term Memory Network model. These spatio-temporal features are a combination of power grid operation data, meteorological data, geographical information, and equipment status information.

[0126] S3.2: Process the time series data through the Long Short-Term Memory Network layer to capture the time dependencies of load changes. In this layer, LSTM has two main gating mechanisms: the input gate and the forget gate, which are used to control the storage and forgetting of information and can effectively capture long-term dependencies.

[0127] S3.3: Each Long Short-Term Memory Network unit outputs a hidden layer state that captures the information of the historical input sequence.

[0128] S3.4: The output of the Long Short-Term Memory Network is converted into a load prediction value through a fully connected layer (i.e., a multi-layer perceptron). Here, the output layer will predict the value of the power grid load.

[0129] S3.5: Use the Mean Squared Error (MSE) as the loss function to measure the difference between the predicted value and the true value of the Long Short-Term Memory Network model, and train the Long Short-Term Memory Network model.

[0130] The Mean Squared Error is:

[0131] ;

[0132] Where, represents the mean squared error; represents the number of data points for prediction; represents the th true load value of the data point; represents the th load value predicted by the Long Short-Term Memory Network model for the data point.

[0133] S3.6. Use the trained long short-term memory network model to predict the load.

[0134] S4. Based on the prediction results, construct an emergency resource scheduling optimization model through the dynamic programming (DP) method to generate the optimal resource scheduling path.

[0135] In this embodiment, dynamic programming decomposes complex problems into a series of interrelated sub-problems and uses the method of backward recursion to solve them, and can obtain the global optimal solution from the initial state to the target state. This is particularly important for emergency resource scheduling because it involves multiple constraints and variables, and directly solving may lead to local optima; emergency resource scheduling is usually a complex process involving multiple decision points, and each decision will affect subsequent choices. Dynamic programming can effectively handle this type of multi-stage decision-making problem, describe the relationship between different decisions through the state transition equation, and thus achieve step-by-step optimization; by backtracking the optimal decisions of each state to determine the overall optimal path, this process can not only obtain the final solution, but also allow checking and verification of the intermediate steps to ensure that each decision is made based on the current best information.

[0136] Constructing an emergency resource scheduling optimization model through the dynamic programming (DP) method based on the prediction results to generate the optimal resource scheduling path includes the following steps:

[0137] S4.1. Define the state variable , representing the location where a certain resource (such as a repair person or equipment) is located, the tasks that have been completed, and the set of all locations of the resources ;

[0138] S4.2. Define the decision variable (the decision variable refers to the specific path selected to transfer from the current state to the next state, that is, the decision to move from the current node to the next node). Select the scheduling path at the current state t. The decision at each moment is to select the next scheduling path, that is, from the current node to the next node.

[0139] S4.3. Define the state transition cost and the state transition equation. The core of dynamic programming is the state transition equation, which describes the cost of transferring from one state to another. The state transition cost refers to the cost or resources consumed for transferring from one state to another. The state transition equation is used to describe how to calculate the best way to reach the next state based on the current state and determine the corresponding transfer cost accordingly.

[0140] Further, the state transition equation is:

[0141] ;

[0142] ;

[0143] Among them, represents the minimum cumulative cost from the current time and node to the completion of all tasks; represents the node (substation, maintenance center, etc.) where the resource is located at time ; represents the next node that the resource reaches at time ; represents the set of neighbor nodes that can be moved from the current node (i.e., the optional next positions); represents the comprehensive cost (including time, resource consumption, traffic impact, etc.) of moving from node to ; represents the minimum cumulative cost (recursive calculation) from the current time and node to the completion of all tasks; represents the time cost (distance divided by speed) of moving from node to ; represents the resource consumption (manpower, equipment loss); represents the road congestion level; represents the weight coefficient of the time cost; represents the weight coefficient of the resource consumption; represents the weight coefficient of the road congestion level;

[0144] There is often an inherent logical sequence among tasks in power grid emergency handling. For example, when repairing a transmission line fault, it may be necessary to first detect and locate the fault point, and then arrange maintenance personnel and equipment for repair work; after the repair is completed, testing and acceptance are also required to ensure that the line returns to normal operation. The task dependency graph (DAG) can accurately reflect these dependencies among tasks. By introducing DAG constraints for the scheduling path, it can ensure that resource scheduling proceeds in a reasonable task order, avoiding unreasonable scheduling that violates the task logic; if the task sequence is not considered, it may lead to resources being assigned to a task at an inappropriate time, resulting in resource conflicts and waste. For example, dispatching maintenance personnel and equipment to the site without completing the fault detection may cause the repair work to not proceed smoothly due to the inability to accurately understand the fault situation, wasting human and material resources. After introducing DAG constraints, it can ensure that the resources required for subsequent tasks are only scheduled after the completion of the preceding tasks, improving the utilization efficiency of resources;

[0145] Regarding the problem of the order of power grid emergency resource scheduling, a task dependency graph (DAG) is introduced into the state transition equation to optimize the scheduling path. Only when the prerequisite tasks of a node are completed, the node is added to the set. This can make the scheduling process more targeted, avoid unnecessary calculations and attempts, and speed up the decision-making speed of emergency resource scheduling.

[0146] Define the task dependency matrix , if task depends on task , then ;

[0147] Among them, the task dependency matrix is used to clarify the dependency relationships between various tasks in the power grid emergency resource scheduling, so as to optimize the scheduling path; first, all tasks involved in the power grid emergency resource scheduling process need to be determined. These tasks can be different repair works, equipment detections, power supply restorations, etc. operations. Number these tasks. If there are a total of tasks, they are respectively marked as ; conduct a detailed analysis of each task to determine the sequence and dependency relationships between them. For example, when performing power equipment maintenance, it may be necessary to conduct equipment detection first before carrying out specific maintenance work. This means that the maintenance task depends on the detection task; construct a matrix , the rows and columns of the matrix respectively correspond to each task; the element in the matrix represents the dependency relationship between task and task , represents the task number corresponding to the row, represents the task number corresponding to the column; among them, ; according to the analysis results of the task dependency relationships, assign values to the elements in the matrix : if task depends on task , then let , which means that task must be completed before task starts; if task does not depend on task , then let ;

[0148] Update the set of feasible nodes: , among them, represents all the tasks that have been completed when reaching state ;

[0149] When all resources are dispatched at a global level, the state space expands dramatically as the scale of the power grid increases. Power grids contain a large number of nodes (such as substations and maintenance centers) and resources (such as repair personnel and equipment), each of which has multiple possible states and combinations. This makes the dimensionality of the state variables very high. If the entire power grid is dispatched uniformly, the computational effort will increase exponentially, resulting in extremely high computational complexity, and even becoming infeasible in practical applications. By dividing the power grid into subregions, the state space of each subregion is relatively small, and the computational complexity is reduced accordingly, thereby alleviating the curse of dimensionality. Global scheduling requires considering all possible dispatch paths for all resources in the entire power grid, resulting in an extremely large search space. This means that when solving for the optimal dispatch path, a large number of possible solutions must be traversed, resulting in a huge computational cost and excessive time consumption. Through a two-layer optimization approach, a coarse-grained path is first generated at the global level based on the key nodes in the subregion, narrowing the search scope and reducing unnecessary computation. Paths are then refined within the subregions, further improving dispatch accuracy while avoiding the high complexity associated with a comprehensive search of the entire power grid.

[0150] To address the problem of excessive computational complexity caused by scheduling all resources at the global level, the power grid area is divided into sub-areas. A hierarchical optimization strategy is used to optimize the state transition equation. Coarse-grained paths are generated based on the key nodes in the sub-areas, and the paths are refined within the sub-areas. The global layer state variable is the sub-area ID, and the local layer is the detailed node ID, solving the dimensionality curse problem.

[0151] In summary, the state transition equation optimized by introducing the task dependency graph and hierarchical optimization strategy is:

[0152] ;

[0153] in, represents the minimum cumulative cost after optimization; Represents the state variables at the sub-region level, used for hierarchical optimization; Represents the adjustment factor, which controls the weights of the global layer and the local layer to optimize the computational efficiency. Its value range is [0,1] and is determined by grid search. Indicates that from the next time point and sub-region level state variables The minimum cumulative cost from the start to the completion of all tasks.

[0154] S4.4. Set the initial state and termination conditions. The initial state is the starting point where the resources are located, and the termination state is when all tasks are completed.

[0155] S4.5. Obtain the cumulative cost of each path to the final state through reverse recursion, and select the path with the smallest cumulative cost as the optimal path.

[0156] Among them, through reverse recursion, the cumulative cost of each path reaching the final state is obtained, and the path with the smallest cost among the cumulative costs of the paths is selected as the optimal path, including the following steps:

[0157] S4.51. Set the initial conditions for recursion, that is, determine the optimal cost of all possible states at the final moment;

[0158] S4.52. Solve to obtain all the next states to which the current state can be transferred ;

[0159] S4.53. For each state in the set calculate the total cost of transferring from to and continuing to the end point according to the state transition equation, including the direct transfer cost plus the minimum cost from to the end point in the future;

[0160] S4.54. Among all the transfers, select the one with the smallest total cost as the value of the minimum cumulative cost, and record the corresponding optimal transfer state at the same time;

[0161] S4.55. Repeat steps S4.52 to S4.54 until the optimal cost of all states at the initial moment is calculated. At this time, the optimal cost of the initial state is the minimum scheduling cost from the starting point to the end point. In this way, the optimal path to reach the target state and its corresponding minimum cost can be found from any starting point within the entire time range. This is actually constructing a global optimal strategy to ensure that resources reach the predetermined target position at the lowest cost during the entire scheduling process;

[0162] S4.6. Backtrack the optimal decision of each state to obtain the optimal resource scheduling path.

[0163] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art of this industry should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification are only preferred examples of the present invention and are not used to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed.

Claims

1. A power grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source, grid, load and storage, characterized in that It includes the following steps: S1. Deploy edge computing nodes at key nodes of the power grid to collect real-time power grid operation data, meteorological data, and geographical information data; S2. Integrate power grid operation data, meteorological data, geographical information, and equipment status data, perform spatio-temporal feature fusion using a graph neural network, optimize the process of spatio-temporal feature fusion by combining node type embedding vectors with real-time status factors, and adopting the median aggregation method; Among them, for the problem of grid node type heterogeneity, node type embedding vectors are assigned to different node types , and the node type features are encoded into the graph convolution operation for optimization, and the attention coefficients are calculated , and the importance weights of different neighbor nodes to the current node are dynamically quantified to obtain the attention coefficients ; Filter the low-confidence neighbors to obtain the filtered neighbor set : Using the median aggregation method to suppress the influence of outliers on feature aggregation, the neighbor node features after weighted median processing are obtained : By combining node type embedding vectors with real-time state factors , and using the median aggregation method to optimize the graph convolution formula after graph convolution operations; S3. Based on the fused spatio-temporal features, use a long short-term memory network model for load forecasting; S4. Based on the prediction results, construct an emergency resource scheduling optimization model through dynamic programming method to generate the optimal resource scheduling path.

2. The grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of the source, network, load, and storage according to claim 1, wherein: In S2, integrating power grid operation data, meteorological data, geographical information, and equipment status data, and performing spatio-temporal feature fusion using a graph neural network includes the following steps: S2.

1. Consider the power grid as a graph ; Among them, represents a node in the power grid, represents the connection relationship between nodes; S2.

2. For each node Merge into a comprehensive feature vector ; S2.

3. Initialize each node features , integrate all features from different sources into initial features , which serve as the input to the graph neural network; S2.

4. Perform multi-layer graph convolution operations to form the features of each node , including the spatial domain features, meteorological features, and geographical information features of the nodes; S2.

5. For the operation data of the power grid, use the sliding window technique to extract time series features in the operation data of the power grid; S2.

6. Combine the time series features obtained from time series analysis with the spatial domain features obtained through the graph neural network to form spatio-temporal features.

3. The power grid emergency resource scheduling optimization method based on source-network-load-storage multi-dimensional data fusion according to claim 2, characterized in that: In S2.4, performing multi-layer graph convolution operations includes the following steps: S2.

41. Initialize node features , take as the input node feature vector and feed it into the first-layer graph convolution; S2.

42. For each node , determine the set of neighbor nodes of node ; S2.

43. Define weight matrices and bias terms for each layer of convolution operation; S2.

44. In the first layer of graph convolution operation, aggregate the node feature vectors with the feature vectors of their neighbor nodes through the graph convolution formula, combine the node type embedding vectors with the real-time status factors, and optimize the graph convolution operation using the median aggregation method; S2.

45. Based on the updated features of the previous layer, the node features continue to be aggregated with the feature vectors of their neighbor nodes until a predetermined number of layers is reached. ; S2.

46. After completing the convolution of all layers, the final feature representation of each node is obtained .

4. The grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source, network, load and storage according to claim 3, characterized in that: In S2.44, aggregating the node feature vectors with the feature vectors of their neighbor nodes through the graph convolution formula includes the following steps: S2.

441. For the current node, traverse all its neighbor nodes, multiply the feature representation of each neighbor node at the layer by the weight matrix of the current layer, and add the bias; S2.

442. Normalize the weighted features of neighbor nodes; S2.

443. After the normalized features of all neighbor nodes are summed up, a new feature of the current node at the layer is generated through a non-linear activation function.

5. The power grid emergency resource scheduling optimization method based on source-network-load-storage multi-dimensional data fusion according to claim 2, wherein: In S2.5, for the operation data of the power grid, using the sliding window technique to extract time series features in the operation data of the power grid includes the following steps: S2.

51. Determine the time window size and set the sliding step; S2.

52. Starting from the starting point of the time series data, initialize the position of the first window, and move the window position sequentially through the sliding window method until the window slides to the last time point; S2.

53. Extract key features within each window; S2.

54. Construct a feature vector from the features extracted by each sliding window, and the feature vector is the time series feature.

6. The power grid emergency resource scheduling optimization method based on multi-dimensional data fusion of source, network, load and storage according to claim 1, characterized in that: In S3, based on the fused spatio-temporal features, using a long short-term memory network model for load forecasting includes the following steps: S3.

1. Input the fused spatio-temporal features into the long short-term memory network model; S3.

2. Process the time series data through the long short-term memory network layer to capture the time-dependent relationship of load changes; S3.

3. Each long short-term memory network unit outputs a hidden layer state; S3.

4. The output of the long short-term memory network is converted into a load prediction value through a fully connected layer; S3.

5. Use the mean square error as the loss function to measure the difference between the prediction value and the true value of the long short-term memory network model, and train the long short-term memory network model; S3.

6. Use the trained long short-term memory network model to predict the load.

7. The power grid emergency resource scheduling optimization method based on multi-dimensional data fusion of source, network, load and storage according to claim 1, characterized in that: In S4, based on the prediction results, an emergency resource scheduling optimization model is constructed by the dynamic programming method to generate the optimal resource scheduling path, including the following steps: S4.

1. Define state variables and all sets of positions of resources ; S4.

2. Define decision variables and select the scheduling path at the current state t. S4.

3. Define the state transition cost and the state transition equation, and introduce the task dependency graph and the hierarchical optimization strategy to optimize the state transition equation. S4.

4. Set the initial state and the termination condition. The initial state is the starting point where the resources are located, and the termination state is that all tasks are completed. S4.

5. Through backward recursion, obtain the cumulative cost of each path reaching the final state, and select the path with the minimum cost among the cumulative costs of the paths as the optimal path. S4.

6. Backtrack the optimal decision of each state to obtain the optimal resource scheduling path.

8. The grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of source, network, load and storage according to claim 7, wherein: In S4.3, the state transition equation is: ； Among them, represents the minimum cumulative cost from the current time and the node to the completion of all tasks; represents the node where the resource is at time ; represents the next node that the resource reaches at time ; represents the set of neighbor nodes that can be moved to from the current node ; represents the comprehensive cost of moving from node to ; represents the minimum cumulative cost from the current time and the node to the completion of all tasks; Regarding the problem of the sequence of power grid emergency resource scheduling, a task dependency graph is introduced into the state transition equation to optimize the scheduling path. Only when the pre-task of a node is completed, the node is added to the set : Define the task dependency matrix If task depends on task , then ; Update the set of feasible nodes: , where represents all the tasks that have been completed when reaching state . Divide the power grid area into sub-areas, optimize the state transition equation through the hierarchical optimization strategy, generate a coarse-grained path based on the key nodes of the sub-areas, refine the path within the sub-areas, the global layer state variable is the sub-area ID, and the local layer is the detailed node ID. Introduce the task dependency graph and the hierarchical optimization strategy to optimize the state transition equation.

9. The grid emergency resource scheduling optimization method based on the multi-dimensional data fusion of the source, grid, load and storage according to claim 8, characterized in that: In S4.5, through backward recursion, obtain the cumulative cost of each path reaching the final state, and select the path with the minimum cost among the cumulative costs of the paths as the optimal path, including the following steps: S4.

51. Set the initial conditions for recursion. S4.

52. Solve to obtain all the next states to which the current state can transit ; S4.

53. For the set For each state Calculate the total cost of transferring from to and continuing to the end point according to the state transition equation; S4.

54. Among all the transitions, select the one with the minimum total cost as the value of the minimum cumulative cost, and record the corresponding optimal transition state at the same time. S4.

55. Repeat steps S4.52 to S4.54 until the optimal costs of all states at the initial time are calculated. are calculated.

Images