Generation method and device of power grid operation scene, terminal equipment and storage medium

Abstract

The invention discloses a power grid operation scene generation method and apparatus, a terminal device and a storage medium, and can input data for representing power grid operation characteristics into an operation scene generation model, so that the model automatically extracts and learns current structural characteristics and node operation characteristics of a power grid based on the input data. And finally, according to various extracted features, a corresponding target power grid operation scene is generated, so that human intervention is reduced. Compared with the prior art, the power grid topological structure characteristics and the node operation characteristics can be extracted through the operation scene generation model, and the mapping relation between various characteristics and the actual power grid operation scene can be learned, so that the accuracy of scene generation is improved. Therefore, key features can be rapidly extracted from a large amount of data and the operation scene can be generated, and especially when the scale of the power grid is large or a large amount of different scenes need to be generated, the generation demand of the power grid operation scene can be rapidly and accurately responded.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology for power grid operation, and in particular to a method, device, terminal equipment and storage medium for generating a power grid operation scenario. Background Art

[0002] When generating power grid operation scenarios, it is generally necessary to sample a set of scenarios in the probability model of random variables, so as to obtain a simulation of the operation status of the power grid under different conditions, including load level, new energy output, etc. Based on the generated operation scenarios, the power grid operation strategy can be optimized, such as adjusting the power generation plan, load distribution, etc., to reduce operating costs and improve power grid efficiency.

[0003] However, in the traditional process of generating power grid operation scenarios, operation and maintenance personnel or experts are often required to predict and set possible operation scenarios based on years of power grid operation experience and historical data. This method relies on the professional knowledge and experience of the operation and maintenance personnel and is therefore highly subjective. This not only consumes a lot of time and effort, but also causes the generated operation scenarios to be not objective and accurate enough. Therefore, when the power grid is large or a large number of different scenarios need to be generated, it cannot respond quickly and accurately to the generation of power grid operation scenarios. Summary of the invention

[0004] The embodiments of the present invention provide a method, apparatus, terminal device and storage medium for generating power grid operation scenarios. By operating the scenario generation model, it is possible to learn the mapping relationship between the characteristics of various types of data and the actual power grid operation scenarios, and it is possible to quickly extract key features from a large amount of data and generate operation scenarios. This can effectively solve the problem in the prior art that it relies on the professional knowledge and experience of operation and maintenance personnel, which not only consumes a lot of time and energy, but also causes the generated operation scenarios to be not objective and accurate enough.

[0005] An embodiment of the present invention provides a method for generating a power grid operation scenario, comprising:

[0006] Obtain the topological structure data of the power grid and the node operation data of each node in the power grid;

[0007] Generate source data according to the topological structure data and the operation data of each node, input the source data into a preset operation scenario generation model, so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes;

[0008] The generation of the operation scenario generation model includes:

[0009] Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples;

[0010] Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

[0011] Preferably, the source data is a graph structure used to characterize the operation characteristics of the power grid;

[0012] The generating source data according to the topological structure data and the operation data of each node includes:

[0013] Generating a topological adjacency matrix for characterizing the topological structure characteristics of a power grid according to the topological structure data;

[0014] Generate a feature matrix for characterizing node operation characteristics according to node operation data of each node;

[0015] A graph structure for characterizing power grid operation characteristics is generated according to the topological adjacency matrix and the characteristic matrix.

[0016] Preferably, the operation scenario generation model comprises: an encoder and a decoder;

[0017] The operation scenario generation model extracts structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features, including:

[0018] The encoder extracts structural features corresponding to the graph structure and node operation features of each node in the graph structure, and then generates global features and local features of the graph structure according to the structural features and the node operation features;

[0019] The decoder generates a target power grid operation scenario corresponding to the graph structure according to the global features and the local features.

[0020] Preferably, the encoder comprises: a graph convolution layer and a fully connected layer;

[0021] The encoder extracts the structural features corresponding to the graph structure and the node operation features of each node in the graph structure, and then generates the global features and local features of the graph structure according to the structural features and the node operation features, including:

[0022] Extracting a structural feature vector corresponding to the graph structure through the graph convolution layer;

[0023] For each node, the node operation feature of the node and the node operation features of the adjacent nodes adjacent to the node are aggregated through the graph convolution layer to generate a node feature vector corresponding to the node;

[0024] The global features and local features of the graph structure are extracted from the structural feature vector and each node feature vector through the fully connected layer.

[0025] Preferably, during each iterative training, a training sample is input into the operation scenario generation model, so that the operation scenario generation model generates a power grid operation scenario prediction result corresponding to the training sample according to the structural features and node operation features in the training sample;

[0026] The network parameters of the operation scenario generation model are adjusted according to the power grid operation scenario prediction results and the actual results of the power grid operation scenarios corresponding to the training samples.

[0027] Preferably, the power grid operation scenario prediction results include: active power prediction values ​​of load nodes, reactive power prediction values ​​of load nodes, active power prediction values ​​of power supply nodes, reactive power prediction values ​​of power supply nodes, voltage amplitude prediction values ​​of balancing nodes, phase angle prediction values ​​of balancing nodes, and section power prediction values ​​of several different sections; wherein the power supply node is used to control the voltage of the node;

[0028] The adjusting of the network parameters of the operation scenario generation model according to the power grid operation scenario prediction results and the actual results of the power grid operation scenarios corresponding to the training samples includes:

[0029] The active power of the load node, the reactive power of the load node, the active power of the power supply node, the active power of the power supply node, the voltage amplitude of the balance node and the phase angle of the balance node in the prediction result of the power grid operation scenario are input into a power flow calculation model with the goal of minimizing the power grid operation cost; wherein the constraints corresponding to the power flow calculation model are: node power constraint, node voltage constraint and section power constraint;

[0030] Under the constraints of node power constraints, node voltage constraints and section power constraints, the power flow calculation model is solved to generate target section powers of several different sections corresponding to the prediction results of the grid operation scenario when the grid operation cost is minimized;

[0031] Generate the loss value currently corresponding to the operation scenario generation model according to the power prediction value of each section, the power of each target section, the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample;

[0032] Adjusting network parameters of the running scenario generation model according to the loss value;

[0033] The construction of the power flow calculation model includes:

[0034] A power flow calculation model is constructed based on the node power data, node voltage data, power supply operation cost data and section power data.

[0035] Preferably, generating the loss value currently corresponding to the operation scenario generation model according to each section power prediction value, each target section power, the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample includes:

[0036] According to the following loss function, the loss value corresponding to the current running scenario generation model is generated:

[0037] L total =L+λL constraint ;

[0038]

[0039] Among them, L total The loss value currently corresponding to the model for the operation scenario is generated, L is the deviation value between the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample, λ is the weight parameter corresponding to the section power, L constraint is the power loss value corresponding to the cross section, N u is the number of cross sections, J key,i is the predicted power value of the i-th section, P key,i is the target section power of the i-th section.

[0040] Based on the above method embodiments, the present invention provides corresponding device embodiments.

[0041] An embodiment of the present invention provides a device for generating a power grid operation scenario, comprising: a data acquisition module and an operation scenario generation module;

[0042] The data acquisition module is used to acquire the topological structure data of the power grid and the node operation data of each node in the power grid;

[0043] The operation scenario generation module is used to generate source data according to the topological structure data and the operation data of each node, and input the source data into a preset operation scenario generation model so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes;

[0044] The generation of the operation scenario generation model includes:

[0045] Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples;

[0046] Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

[0047] Based on the above method embodiments, the present invention provides corresponding terminal device embodiments.

[0048] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, it implements a method for generating a power grid operation scenario described in the above-mentioned embodiment of the invention.

[0049] Based on the above method embodiments, the present invention provides corresponding storage medium item embodiments.

[0050] Another embodiment of the present invention provides a storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a method for generating a power grid operation scenario as described in the above-mentioned embodiment of the invention.

[0051] The following beneficial effects are achieved by implementing the present invention:

[0052] The embodiment of the present invention provides a method, device, terminal device and storage medium for generating a power grid operation scenario. The present invention can obtain the topological structure data of the power grid and the node operation data of each node, so that these data used to represent the power grid operation characteristics are input into the operation scenario generation model, so that the model can automatically extract and learn the current structural characteristics and node operation characteristics of the power grid based on the input data, and finally generate the corresponding target power grid operation scenario according to the various extracted characteristics, thereby reducing human intervention and improving the objectivity and accuracy of the generated scenario. Compared with the prior art, the present invention uses the operation scenario generation model obtained by training to enable it to extract the topological structure characteristics and node operation characteristics of the power grid, and learn the mapping relationship between various characteristics and the actual power grid operation scenario, thereby improving the accuracy of the generated scenario. The present invention can quickly extract key features from a large amount of data and generate an operation scenario through the operation scenario generation model, especially when the power grid is large in scale or a large number of different scenarios need to be generated, it can quickly and accurately respond to the generation needs of the power grid operation scenario. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flowchart of a method for generating a power grid operation scenario provided by an embodiment of the present invention.

[0054] Figure 2 It is a schematic diagram of the generation process of a graph structure provided by an embodiment of the present invention.

[0055] Figure 3 It is a schematic diagram of model parameter updating provided by an embodiment of the present invention.

[0056] Figure 4 It is a structural schematic diagram of a device for generating a power grid operation scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0058] like Figure 1 FIG. 1 is a flow chart of a method for generating a power grid operation scenario provided by an embodiment of the present invention. The method for generating a power grid operation scenario includes:

[0059] Step S1: obtaining topological data of the power grid and node operation data of each node in the power grid;

[0060] Step S2: Generate source data according to the topological structure data and the operation data of each node, input the source data into a preset operation scenario generation model, so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes;

[0061] The generation of the operation scenario generation model includes:

[0062] Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples;

[0063] Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

[0064] For step S1, schematically, the topological structure data includes the connection relationship between each node in the power grid, so that the structure of the power grid can be understood based on the topological structure data. The node operation data includes the specific operation information of each node in the power grid, such as load, section power, unit output curve, etc. These data reflect the operation status of the power grid at a specific point in time, which is crucial for predicting possible future operation scenarios.

[0065] For step S2, in a preferred embodiment, the source data is a graph structure used to characterize the operation characteristics of the power grid.

[0066] When the source data is input into the preset running scenario generation model, the graph structure is input into the running scenario generation model, so that the running scenario generation model can identify and extract features from the graph structure. Then:

[0067] The generating source data according to the topological structure data and the operation data of each node includes:

[0068] Generating a topological adjacency matrix for characterizing the topological structure characteristics of a power grid according to the topological structure data;

[0069] Generate a feature matrix for characterizing node operation characteristics according to node operation data of each node;

[0070] A graph structure for characterizing power grid operation characteristics is generated according to the topological adjacency matrix and the characteristic matrix.

[0071] Indicatively, Figure 2 As shown, the embodiment of the present invention can represent the topological structure data and the node operation data of each node in a graph, that is, by combining the constructed feature matrix with the adjacency matrix, a graph structure for characterizing the operation characteristics of the power grid can be obtained, that is, a visual graph representation can be obtained.

[0072] The visualization graph representation obtained by the present invention can intuitively represent the connection relationship between each node in the power grid and the operating characteristics of the nodes. Therefore, by integrating the topological structure data and node operation data into the graph structure, the model can more easily understand and process these complex data.

[0073] Inputting the topological structure data and node operation data of the power grid into the model in the form of a graph structure can also enable the model to learn the behavior patterns of the power grid under different structures and operating conditions, thereby capturing the interactions and mutual influences between nodes in the power grid, predicting the overall behavior of the power grid, and obtaining accurate power grid operation scenarios.

[0074] In a preferred embodiment, the operation scenario generation model of the present invention includes: an encoder and a decoder;

[0075] Then, when the operation scenario generation model extracts the structural features and the node operation features in the source data and generates the target power grid operation scenario corresponding to the source data according to the structural features and the node operation features, it specifically includes:

[0076] The encoder extracts structural features corresponding to the graph structure and node operation features of each node in the graph structure, and then generates global features and local features of the graph structure according to the structural features and the node operation features;

[0077] The decoder generates a target power grid operation scenario corresponding to the graph structure according to the global features and the local features.

[0078] It is understandable that the encoder maps the extracted features to a latent space, and represents these features with the mean and variance of the latent space, so that the model can capture the inherent laws and patterns of the input data, providing a basis for the subsequent target scene generation. In the process of extracting features, the encoder not only pays attention to the overall structure of the power grid (global features), but also pays attention to the specific operating status of each node (local features), so that the global features and local features can together constitute a complete description of the power grid operation scenario, so that the decoder can reconstruct the target power grid operation scenario from the points in the latent space.

[0079] The decoder can map the representation of the latent space back to the original data space through a series of neural network layers, thereby generating an operating scenario corresponding to the input graph structure. During the decoding process, the decoder needs to integrate global features and local features to generate the target scenario. Since global features provide the overall structural information of the power grid, and local features provide the specific operating status information of each node. By integrating this information, the decoder can generate a scenario that conforms to the actual operating laws of the power grid. This scenario can include information such as load distribution, power flow, and voltage level of the power grid.

[0080] When the topological structure data and node operation data of the power grid are input into the model in the form of a graph structure, the operation mode data in the graph structure can be mapped to the parameters of a latent space through the encoder, and represented by the mean and variance of the latent space, thereby realizing the extraction and potential encoding of the operation mode features. The decoder maps the points of the latent space back to the original data space through another series of neural network layers, thereby realizing the reconstruction of the operation mode data and obtaining the target power grid operation scenario corresponding to the graph structure. The embodiment of the present invention can realize the feature extraction of the operation mode data and the generation of the target power grid operation scenario through encoding and decoding.

[0081] In a preferred embodiment, when the encoder extracts the global features and local features of the graph structure, it can process and extract the features based on a fused graph neural network, and the graph neural network includes: a graph convolution layer and a fully connected layer.

[0082] Then, the encoder is used to extract the structural features corresponding to the graph structure and the node operation features of each node in the graph structure, and then the global features and local features of the graph structure are generated according to the structural features and the node operation features. The specific process may be:

[0083] Extracting a structural feature vector corresponding to the graph structure through the graph convolution layer;

[0084] For each node, the node operation feature of the node and the node operation features of the adjacent nodes adjacent to the node are aggregated through the graph convolution layer to generate a node feature vector corresponding to the node;

[0085] The global features and local features of the graph structure are extracted from the structural feature vector and each node feature vector through the fully connected layer.

[0086] Specifically, the operation of the graph convolution layer may include: first, constructing an adjacency matrix with self-connection, that is, a topology-based adjacency matrix A, adding the unit matrix I N To include self-connection, we get the adjacency matrix with self-connection A′=A+I N .

[0087] Next, calculate the degree matrix D of A′;

[0088] Finally, the graph convolution operation is implemented. That is, for the graph convolution layer of the lth layer, the input feature is H (l) When , the trainable parameter matrix is ​​W (l) . The information of neighboring nodes is aggregated to update the representation of each node, thereby capturing the relationship between nodes and the structural information of the graph, and then obtaining a more powerful feature expression. The specific calculation formula is shown in the following formula:

[0089] H (l+1) =f(H (l) ,A)=σ(D -0.5 A'D -0.5 H (l) W (l) );

[0090] In the formula, A'=A+I N , A' is the adjacency matrix with self-connection, A is the topological adjacency matrix, I N is the identity matrix. ii =∑ j A ij ', D is the degree matrix of A', W (l) is the trainable parameter in the convolutional layer of the graph convolutional neural network, H (l) represents the input feature of the lth layer. f represents the graph convolution operation. After the matrix multiplication in the above formula, the activation function σ=RELU is used for forward propagation.

[0091] Therefore, after multiple iterations of the graph convolutional layer, the structural features of the entire graph can be implicitly encoded in the node features of each layer, and the node feature vector corresponding to the node can be obtained by aggregating the features of all nodes (for example, taking the average or maximum pooling).

[0092] For each node, its node feature vector is obtained by aggregating the features of its neighboring nodes (including itself because self-connections are added) through the graph convolution layer. Specifically, the feature vector H of each node (l+1) Through H (l+1) =f(H (l) ,A)=σ(D -0.5 A'D -0.5 H (l) W (l) ) calculated.

[0093] Furthermore, the feature vector of each node can be extracted as a local feature in each layer (or the last layer) through the fully connected layer, and the global feature can be extracted by aggregating the node features to provide rich feature representation for the subsequent power grid operation scenario generation task.

[0094] In a preferred embodiment, Figure 3 As shown, in the decoder of the present invention, an additional layer of prediction of the model input features is added, and at the same time, based on the constructed power flow calculation model to obtain the features of interest in the operation scenario (such as the power features of the section), the deviation L between the section features obtained by the power flow calculation model and the section features of the scenario generated by the model can be obtained. constraint , the model can be guided to generate operating scenarios with key features while conforming to the data distribution, and continuously make the section power in the final operating scenario conform to the optimal power flow calculation, so that the final operating scenario conforms to the minimum grid operation cost and satisfies the node power constraints, node voltage constraints and section power constraints.

[0095] It is understandable that during the training process of the operation scenario generation model, its loss function consists of two parts: reconstruction loss and feature loss. The reconstruction loss measures the difference between the reconstructed data and the original data, that is, the deviation value L between the grid operation scenario prediction result and the actual grid operation scenario; while the feature loss measures the power loss value L corresponding to the section constraint .

[0096] Then, in each iterative training, a training sample is input into the operation scenario generation model, so that the operation scenario generation model generates a power grid operation scenario prediction result corresponding to the training sample according to the structural features and node operation features in the training sample;

[0097] The network parameters of the operation scenario generation model are adjusted according to the power grid operation scenario prediction results and the actual results of the power grid operation scenarios corresponding to the training samples.

[0098] Specifically, the power grid operation scenario prediction results include: active power prediction values ​​of load nodes, reactive power prediction values ​​of load nodes, active power prediction values ​​of power supply nodes, reactive power prediction values ​​of power supply nodes, voltage amplitude prediction values ​​of balancing nodes, phase angle prediction values ​​of balancing nodes, and section power prediction values ​​of several different sections; wherein the power supply node is used to control the voltage of the node;

[0099] Schematically, the nodes of the power grid can be divided into PQ nodes, PV nodes and balance nodes (Vθ nodes). For a power grid system with N nodes, at least 2N state variables are required, and the operation scenario can be shown as follows:

[0100]

[0101] Among them, X data is the prediction result of the power grid operation scenario, P iis the predicted active power value of the load node, Q i is the reactive power prediction value of the load node, P j is the predicted active power value of the power node, V j is the reactive power prediction value of the power node, V l is the predicted value of the voltage amplitude of the equilibrium node, θ l is the predicted value of the phase angle of the equilibrium node, N PQ 、N PV 、N Vθ They are the collection of PQ nodes, PV nodes and Vθ nodes respectively.

[0102] When adjusting the network parameters of the operation scenario generation model according to the power grid operation scenario prediction results and the actual results of the power grid operation scenario corresponding to the training samples, it specifically includes:

[0103] The active power of the load node, the reactive power of the load node, the active power of the power supply node, the active power of the power supply node, the voltage amplitude of the balance node and the phase angle of the balance node in the prediction result of the power grid operation scenario are input into a power flow calculation model with the goal of minimizing the power grid operation cost; wherein the constraints corresponding to the power flow calculation model are: node power constraint, node voltage constraint and section power constraint;

[0104] Under the constraints of node power constraints, node voltage constraints and section power constraints, the power flow calculation model is solved to generate target section powers of several different sections corresponding to the prediction results of the grid operation scenario when the grid operation cost is minimized;

[0105] According to the power prediction values ​​of each section, the power of each target section, the prediction results of the power grid operation scenarios and the actual results of the power grid operation scenarios corresponding to the training samples, the loss value currently corresponding to the operation scenario generation model is generated; specifically, the loss value currently corresponding to the operation scenario generation model can be generated according to the following loss function formula:

[0106] L total =L+λL constraint ;

[0107]

[0108] Among them, L total The loss value currently corresponding to the model for the operation scenario is generated, L is the deviation value between the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample, λ is the weight parameter corresponding to the section power, L constraint is the power loss value corresponding to the cross section, N u is the number of cross sections, J key,i is the predicted power value of the i-th section, Pkey,i is the target section power of the i-th section.

[0109] Adjusting network parameters of the running scenario generation model according to the loss value;

[0110] The construction of the power flow calculation model includes:

[0111] A power flow calculation model is constructed based on the node power data, node voltage data, power supply operation cost data and section power data.

[0112] Specifically, P key,i The optimal solution of the power flow calculation model can be obtained by the power flow calculation program, and the objective function of the power flow calculation model is as follows:

[0113]

[0114]

[0115] P key,i =-e i e j G ij +e i f j B ij -f i f j G ij -f i e j B ij ;

[0116] U i =e i +jf i ;

[0117]

[0118]

[0119]

[0120] Among them, F is the grid operation cost, a, b, c are the unit costs corresponding to different units, P gi is the active power of node i in the power grid, P di is the active load of grid node i, P key,i is the target section power of the i-th section in the power grid, S g is the node set with power nodes in the node, n b is the set of nodes j connected to node i, e i is the real part of the voltage phasor at node i, f iis the imaginary part of the voltage phasor at node i, e j is the real part of the voltage phasor at node j, f j is the imaginary part of the voltage phasor at node j, G ij , B ij are the real and imaginary parts of the admittance matrix between node i and node j, respectively, U i is the voltage phasor of node i, j is the imaginary unit, Q gi is the reactive power of node i in the power grid, Q di is the reactive load of grid node i, P gi is the minimum active power of node i, is the maximum active power of node i, is the minimum reactive power of node i in the power grid, is the maximum reactive power of node i in the power grid.

[0121] It can be understood that the active power P of node i in the power grid is gi The active power prediction value P of the power supply node in the power grid operation scenario prediction result can be j Calculate and obtain the active load P of grid node i di It can be calculated based on the active power prediction value of the load node in the grid operation scenario prediction result, and the voltage phasor U of node i is i The voltage amplitude of the equilibrium node can be used to predict the value V l Calculate to obtain.

[0122] The present invention can input key parameters (such as active power, reactive power, voltage amplitude, phase angle, etc.) in the power grid operation scenario prediction results into the power flow calculation model, solve the power flow calculation model under constraints, and generate the target section power. Then, according to the deviation value between the power grid operation scenario prediction results and the actual results of the power grid operation scenario corresponding to the training sample, and the power loss value obtained based on the target section power and the section power prediction value, the loss value currently corresponding to the operation scenario generation model is calculated using the loss function, so as to adjust the network parameters of the operation scenario generation model according to the loss value to optimize the model performance.

[0123] By calculating the cross-sectional power loss value in the prediction scenario, the present invention can continuously guide the model to focus on the cross-sectional power during the process of optimizing the model, thereby ultimately obtaining an operating scenario with an optimal cross-sectional power.

[0124] Schematically, during the operation of the power grid, certain sections (i.e., specific parts of the power grid or connections between nodes) generally carry a large amount of power transmission tasks, so their operating status is crucial to the overall performance and stability of the power grid. By calculating the section power loss value, the present invention can identify these key sections and give them more attention during the optimization process, thereby ensuring that the model can give priority to the performance optimization of these key sections when adjusting parameters.

[0125] By continuously optimizing the model through the section power loss value, the present invention can continuously adjust the section power distribution to minimize the total power loss, so that the finally generated power grid operation scenario can improve the energy utilization efficiency.

[0126] like Figure 4 As shown, based on the above-mentioned embodiments of the method for generating various power grid operation scenarios, the present invention provides corresponding device item embodiments;

[0127] An embodiment of the present invention provides a device for generating a power grid operation scenario, comprising: a data acquisition module and an operation scenario generation module;

[0128] The data acquisition module is used to acquire the topological structure data of the power grid and the node operation data of each node in the power grid;

[0129] The operation scenario generation module is used to generate source data according to the topological structure data and the operation data of each node, and input the source data into a preset operation scenario generation model so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes;

[0130] The generation of the operation scenario generation model includes:

[0131] Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples;

[0132] Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

[0133] It should be noted that the device embodiments described above are merely schematic, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, and may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without paying creative labor.

[0134] Those skilled in the art can clearly understand that, for the sake of convenience and simplicity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0135] Based on the above-mentioned embodiments of the method for generating various power grid operation scenarios, the present invention provides corresponding embodiments of terminal equipment items.

[0136] An embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a method for generating a power grid operation scenario as described in any method embodiment of the present invention.

[0137] The terminal device may be a computing terminal device such as a desktop computer, a notebook, a palm computer, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0138] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0139] The memory can be used to store the computer program, and the processor realizes various functions of the terminal device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device or other volatile solid-state storage device.

[0140] Based on the above-mentioned embodiments of the method for generating various power grid operation scenarios, the present invention provides corresponding storage medium item embodiments.

[0141] An embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a method for generating a power grid operation scenario as described in any method embodiment of the present invention.

[0142] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each method embodiment described above can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0143] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for generating a power grid operation scenario, characterized in that: include: Obtain the topological structure data of the power grid and the node operation data of each node in the power grid; Generate source data according to the topological structure data and the operation data of each node, input the source data into a preset operation scenario generation model, so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes; The generation of the operation scenario generation model includes: Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples; Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

2. A method for generating a power grid operation scenario according to claim 1, characterized in that: The source data is a graph structure used to characterize the operation characteristics of the power grid; The generating source data according to the topological structure data and the operation data of each node includes: Generating a topological adjacency matrix for characterizing the topological structure characteristics of a power grid according to the topological structure data; Generate a feature matrix for characterizing node operation characteristics according to node operation data of each node; A graph structure for characterizing power grid operation characteristics is generated according to the topological adjacency matrix and the characteristic matrix.

3. A method for generating a power grid operation scenario according to claim 2, characterized in that: The operation scenario generation model includes: an encoder and a decoder; The operation scenario generation model extracts structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features, including: The encoder extracts structural features corresponding to the graph structure and node operation features of each node in the graph structure, and then generates global features and local features of the graph structure according to the structural features and the node operation features; The decoder generates a target power grid operation scenario corresponding to the graph structure according to the global features and the local features.

4. A method for generating a power grid operation scenario according to claim 3, characterized in that: The encoder comprises: a graph convolution layer and a fully connected layer; The encoder extracts the structural features corresponding to the graph structure and the node operation features of each node in the graph structure, and then generates the global features and local features of the graph structure according to the structural features and the node operation features, including: Extracting a structural feature vector corresponding to the graph structure through the graph convolution layer; For each node, the node operation feature of the node and the node operation features of the adjacent nodes adjacent to the node are aggregated through the graph convolution layer to generate a node feature vector corresponding to the node; The global features and local features of the graph structure are extracted from the structural feature vector and each node feature vector through the fully connected layer.

5. The method for generating a power grid operation scenario according to claim 1, characterized in that: In each iterative training, a training sample is input into the operation scenario generation model, so that the operation scenario generation model generates a power grid operation scenario prediction result corresponding to the training sample according to the structural characteristics and node operation characteristics in the training sample; The network parameters of the operation scenario generation model are adjusted according to the power grid operation scenario prediction results and the actual results of the power grid operation scenarios corresponding to the training samples.

6. A method for generating a power grid operation scenario according to claim 5, characterized in that: The grid operation scenario prediction results include: active power prediction values ​​of load nodes, reactive power prediction values ​​of load nodes, active power prediction values ​​of power supply nodes, reactive power prediction values ​​of power supply nodes, voltage amplitude prediction values ​​of balancing nodes, phase angle prediction values ​​of balancing nodes, and section power prediction values ​​of several different sections; wherein the power supply node is used to control the voltage of the node; The adjusting of the network parameters of the operation scenario generation model according to the power grid operation scenario prediction results and the actual results of the power grid operation scenarios corresponding to the training samples includes: The active power of the load node, the reactive power of the load node, the active power of the power supply node, the active power of the power supply node, the voltage amplitude of the balance node and the phase angle of the balance node in the prediction result of the power grid operation scenario are input into a power flow calculation model with the goal of minimizing the power grid operation cost; wherein the constraints corresponding to the power flow calculation model are: node power constraint, node voltage constraint and section power constraint; Under the constraints of node power constraints, node voltage constraints and section power constraints, the power flow calculation model is solved to generate target section powers of several different sections corresponding to the prediction results of the grid operation scenario when the grid operation cost is minimized; Generate the loss value currently corresponding to the operation scenario generation model according to the power prediction value of each section, the power of each target section, the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample; Adjusting network parameters of the running scenario generation model according to the loss value; The construction of the power flow calculation model includes: A power flow calculation model is constructed based on the node power data, node voltage data, power supply operation cost data and section power data.

7. A method for generating a power grid operation scenario according to claim 6, characterized in that: The generating of the loss value currently corresponding to the operation scenario generation model according to each section power prediction value, each target section power, the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample includes: According to the following loss function, the loss value corresponding to the current running scenario generation model is generated: THE total =L+λL constraint ; Among them, L total The loss value currently corresponding to the model for the operation scenario is generated, L is the deviation value between the power grid operation scenario prediction result and the actual result of the power grid operation scenario corresponding to the training sample, λ is the weight parameter corresponding to the section power, L constraint is the power loss value corresponding to the cross section, N u is the number of cross sections, J key,i is the predicted power value of the i-th section, P key,i is the target section power of the i-th section.

8. A device for generating a power grid operation scenario, characterized in that: include: Data acquisition module and running scenario generation module; The data acquisition module is used to acquire the topological structure data of the power grid and the node operation data of each node in the power grid; The operation scenario generation module is used to generate source data according to the topological structure data and the operation data of each node, and input the source data into a preset operation scenario generation model so that the operation scenario generation model extracts the structural features and node operation features in the source data, and generates a target power grid operation scenario corresponding to the source data according to the structural features and node operation features; wherein the structural features are used to characterize the topological structure characteristics of the power grid; and the node operation features are used to characterize the node operation characteristics of the power grid nodes; The generation of the operation scenario generation model includes: Using the topological structure data samples of the power grid and the node operation data samples of each node in the power grid as training samples; Taking each training sample and the actual power grid operation scenario corresponding to each training sample as input, and the predicted power grid operation scenario of each training sample as output, the operation scenario generation model to be trained is iteratively trained until the model converges to generate a preset operation scenario generation model.

9. A terminal device, characterized in that: It comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a method for generating a power grid operation scenario as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute a method for generating a power grid operation scenario as described in any one of claims 1 to 7.

Images