Substation power flow estimation method and system based on graph technology

Abstract

The invention discloses a transformer substation power flow estimation method and system based on a graph technology, and the method comprises the steps: obtaining a topological relation between a line and transformer substation equipment, constructing graph structure data, carrying out the matrix conversion of any subset of the graph structure data, and enabling the matrix conversion to serve as the input of a graph neural network model; training the graph neural network model based on the input data of the graph neural network model in combination with the missing line power flow data; and predicting and complementing data missing generated in real time according to the trained graph neural network model, summing the power flow of the transformer substation connected with the line in the main graph, and updating the power flow data of the transformer substation. The method does not need to depend on other types of equipment for measurement, is accurate in data updating, builds each time sequence power on each line as a plurality of attributes of line nodes, and guarantees the learning of the graph structure data association relationship by capturing the association relationship of each time point of the line and combining with the graph convolutional neural network and the full-link neural network. And the model can be better trained and generalized.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid analysis, and in particular to a substation power flow estimation method and system based on graph technology. Background Art

[0002] In substations, missing measurement data caused by technical problems or data affiliation will affect the power flow calculation results of the substation area, leading to inaccurate calculation results and other problems. When the measurement data of substations and lines is missing, the power flow estimation results are prone to inaccuracy; in the graph computing application of power grids, there are few related studies, and the existing related power flow calculation methods often require multiple iterations to solve, which is time-consuming and requires a large amount of data from various types of equipment; there is currently a lack of related graph neural network applications to predict and supplement data when the topological relationship between lines and substations is already understood. Summary of the invention

[0003] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a substation power flow estimation method and system based on graph technology to solve the problems mentioned in the background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a substation power flow estimation method based on graph technology, including: obtaining a topological relationship between a line and substation equipment and constructing graph structure data, and performing matrix conversion on any subset of the graph structure data as input to a graph neural network model;

[0008] Based on the input data of the graph neural network model and in combination with the missing line flow data, the graph neural network model is trained;

[0009] According to the trained graph neural network model, the missing data generated in real time are predicted and supplemented, the substation flow connected to the line is summed in the main graph, and the substation flow data is updated.

[0010] As a preferred solution of the substation power flow estimation method based on graph technology described in the present invention, wherein: obtaining the topological relationship between the line and the substation equipment and constructing the graph structure data includes: the mathematical expression of the graph structure data is: G = (V, E); wherein V is a node set and E is an edge set;

[0011] The graph structure data subset is expressed as the input of the graph neural network model as: G′=(V′,E′), and Among them, the subset represents the association relationship of all line V′ nodes.

[0012] As a preferred solution of the substation power flow estimation method based on graph technology described in the present invention, the graph neural network model includes: an input layer, a graph convolution layer, a full connection layer and an output layer, and the output features of the graph convolution layer are used as the input of the full connection layer;

[0013] The graph convolution layer is expressed as:

[0014] X L+1 =σ(D -1 (A+I)X L W L )

[0015] Among them, the left side of the equation is the output, the right side of the equation D is the degree matrix, A is the adjacency matrix, which is used to represent the connection relationship between nodes, I is the identity matrix, X is the parameter matrix, where the length of the parameter matrix is ​​the number of lines, the width is the number of timing power, and σ is the nonlinear function sigmoid;

[0016] The W matrix is ​​obtained through graph convolutional neural network machine learning.

[0017] As a preferred solution of the graph technology-based substation power flow estimation method described in the present invention, the fully connected layer includes: an input layer, a hidden layer and an output layer, and the neurons in each layer are connected to all the neurons in the next layer to form a fully connected structure to generate predicted output.

[0018] As a preferred solution of the substation power flow estimation method based on graph technology described in the present invention, training the graph neural network model includes:

[0019] The back-propagation algorithm is used to update the iterative graph neural network model parameters. After each training cycle, the validation set is used to evaluate the model performance.

[0020] After each training, when the training loss and validation loss remain stable and less than the expected threshold, the model reaches the convergence condition and outputs the prediction result.

[0021] As a preferred solution of the substation flow estimation method based on graph technology described in the present invention, it also includes: based on the input data of the graph neural network model, combined with the missing line flow data training, a graph neural network model is obtained based on other nodes to predict the lines or time points that are prone to data missing, and the output is obtained.

[0022] As a preferred solution of the graph-based substation power flow estimation method described in the present invention, the matrix conversion of any subset of graph structure data includes: adjacency matrix, degree matrix and attribute matrix.

[0023] In a second aspect, the present invention provides a substation power flow estimation system based on graph technology, comprising:

[0024] A data acquisition and conversion module is used to acquire the topological relationship between the line and the substation equipment and construct graph structure data, and perform matrix conversion on any subset of the graph structure data as input to the graph neural network model;

[0025] A model training module, used to train the graph neural network model based on the input data of the graph neural network model and the missing line flow data;

[0026] The model prediction module is used to predict and complete the missing data generated in real time according to the trained graph neural network model, sum the substation flow connected to the line in the main graph, and update the substation flow data.

[0027] In a third aspect, the present invention provides a computing device, comprising:

[0028] Memory and processor;

[0029] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the substation power flow estimation method based on graph technology are implemented.

[0030] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the substation power flow estimation method based on graph technology.

[0031] Compared with the prior art, the present invention has the following beneficial effects: the present invention abstracts substations and lines as nodes in a graph for graph neural network training and prediction, without relying on other types of equipment measurements, and performs weighted summation on the substations connected to the lines in the main graph after training, and updates the substation flow with more accurate data. The time series power on each line is built as multiple attributes of the line node, which can capture the correlation between each time point of the line. Combined with the graph convolutional neural network and the fully linked neural network, it ensures that the model can be better trained and generalized while learning the correlation between graph structure data. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0033] Figure 1 A schematic diagram of a method flow of a substation power flow estimation method and system based on graph technology according to an embodiment of the present invention;

[0034] Figure 2 A graph showing nodes and edges in a graph database of a graph-based power flow estimation method and system for a substation according to an embodiment of the present invention;

[0035] Figure 3 A schematic diagram of the actual application of nodes and edges in a graph-based power flow estimation method and system for a substation according to an embodiment of the present invention;

[0036] Figure 4 A schematic diagram of the connection between a graph convolutional network and a fully connected neural network of a graph technology-based substation power flow estimation method and system according to an embodiment of the present invention;

[0037] Figure 5 A schematic diagram of a single hidden layer fully-linked neural network structure of a substation power flow estimation method and system based on graph technology described in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0039] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0040] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0041] The present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0042] At the same time, in the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper, lower, inner and outer" are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first, second or third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0043] In the present invention, unless otherwise clearly specified and limited, the terms "install, connect, connect" should be understood in a broad sense, for example: it can be a fixed connection, a detachable connection or an integral connection; it can also be a mechanical connection, an electrical connection or a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0044] Example 1

[0045] Reference Figure 1 to Figure 5 , is an embodiment of the present invention, which provides a substation power flow estimation method based on graph technology, such as Figure 1 As shown, including:

[0046] S100: Obtain the topological relationship between the line and the substation equipment and construct graph structure data, and perform matrix conversion on any subset of the graph structure data as input to the graph neural network model;

[0047] S200: Based on the input data of the graph neural network model and the missing line flow data, the graph neural network model is trained;

[0048] S300: Predict and complete the missing data generated in real time based on the trained graph neural network model, sum the substation flow connected to the line in the main graph, and update the substation flow data.

[0049] It should be noted that, since the power flow of a substation at the same voltage level can be calculated by the sum of its input lines and output lines, when the measurement data of the substation or line equipment is missing, the substation power flow will be difficult to calculate. The present invention combines a method for supplementing the missing measurement data of lines and substations with a graph neural network, and builds the line as a subgraph node in a graph model for graph neural network training and prediction, without relying on other types of equipment measurements. After training, a weighted summation is performed on the substations connected to the line in the main graph to update the data for a more accurate substation power flow. Each time series power on each line is built as multiple attributes of the line node, which can capture the correlation between each time point of the line. Combined with the graph convolutional neural network and the fully linked neural network, it ensures that the model can be better trained and generalized while learning the correlation between graph structure data.

[0050] In the embodiment of the present application, obtaining the topological relationship between the line and the substation equipment and constructing the graph structure data includes: the mathematical expression of the graph structure data is: G = (V, E); where V is a node set and E is an edge set;

[0051] The graph structure data subset is represented as the input of the graph neural network model as: G′=(V′,E′), and Among them, the subset represents the association relationship of all line V′ nodes.

[0052] It should be noted that the association relationship of the lines is one of the factors for predicting flow data; in the embodiment of the present application, the nodes are substations and lines at the same voltage level, and may also include corresponding transformers, units, loads and other containers or equipment, and the edges express the association relationship between these nodes.

[0053] In an optional embodiment, the graph structure data is as follows: Figure 2 , Figure 3 As shown in the figure, line 1 and line 3 are both outgoing lines of substation 2, so the flow data of line 1 and line 3 are highly correlated. For example, if line 1 sends n kWh of electricity to substation 2, then line 3 needs to take n kWh of electricity from the substation to balance the power, and the same is true for other lines.

[0054] It should also be noted that the number of attributes in a node expresses the number of time points. In the embodiments of the present application, it is only necessary to reflect that the node contains time series measurement data as the input and output of the graph neural network.

[0055] For example, if you want to understand the changing trend of the line flow in a day, you can create 96 attribute points for the line node, which represent the line active power flow data updated every 15 minutes in a day. It should be noted that the number of time series attribute points of the nodes in the sub-graph and the main graph must be consistent, as shown in Table 1 and Table 2:

[0056] Table 1 Line attributes

[0057] Property name Property Description Data Format p_1 Timing Power 1 double p_2 Timing Power 2 double p_3 Timing Power 3 double p_4 Timing Power 4 double p_5 Timing Power 5 double

[0058] Table 2 Substation attributes

[0059] Property name Property Description Data Format p_1 Timing Power 1 double p_2 Timing Power 2 double p_3 Timing Power 3 double p_4 Timing Power 4 double p_5 Timing Power 5 double

[0060] It should be noted that, by inputting multi-time series flow data in the node attributes, the present application not only learns the flow relationship between line equipment, but also captures the line relationship at different time points to better predict the missing data results.

[0061] In an embodiment of the present application, matrix conversion of any subset of graph structure data includes: adjacency matrix, degree matrix and attribute matrix.

[0062] In the embodiment of the present application, the graph neural network model includes: an input layer, a graph convolution layer, a full connection layer and an output layer, and the output features of the graph convolution layer serve as the input of the full connection layer;

[0063] The graph convolutional layer is represented as:

[0064] X L+1 =σ(D -1 (A+I)X L W L )

[0065] Among them, the left side of the equation is the output, the right side of the equation D is the degree matrix, A is the adjacency matrix, which is used to represent the connection relationship between nodes, I is the identity matrix, X is the parameter matrix, where the length of the parameter matrix is ​​the number of lines, the width is the number of timing power, and σ is the nonlinear function sigmoid;

[0066] The W matrix is ​​obtained through graph convolutional neural network machine learning.

[0067] For example, the adjacency matrix is ​​expressed as:

[0068]

[0069] The parameter matrix is ​​4 (number of nodes) * 5 (number of parameters of each node) in size, showing the parameters of each node. In the graph convolutional neural network, the use of the convolution kernel matrix W and the input of the adjacency matrix can enable the model to learn the hidden associations between each line well.

[0070] It should be noted that the graph convolutional neural network exists as a layer in the neural network architecture and is passed to the fully connected layer after flattening. Since the graph convolutional neural network can effectively capture the graph structure and connection information, and compared with other graph neural networks, the graph convolutional neural network can better capture the potential correlation between nodes, and the fully connected layer can capture the complex correlation information in the data and better generalize the model.

[0071] In the embodiments of the present application, Figure 4 , Figure 5 As shown, the fully connected layer includes: input layer, hidden layer and output layer. The neurons in each layer are connected to all the neurons in the next layer, forming a fully connected structure to generate predicted output.

[0072] In an embodiment of the present application, training a graph neural network model includes:

[0073] The back-propagation algorithm is used to update the iterative graph neural network model parameters. After each training cycle, the validation set is used to evaluate the model performance.

[0074] After each training, when the training loss and validation loss remain stable and less than the expected threshold, the model reaches the convergence condition and outputs the prediction result.

[0075] In an embodiment of the present application, it also includes: based on the input data of the graph neural network model, combined with the missing line flow data training, a graph neural network model is obtained based on other nodes to predict the lines or time points that are prone to data missing, and an output is obtained.

[0076] It should be noted that the embodiments of the present application can predict and complete the missing data generated in real time through the above-mentioned training and prediction methods. After completion, the substation currents connected to the line can be summed in the main graph, and the corresponding substation current data can be updated to ensure data accuracy. The present invention improves the ability to complete missing line data based on graph computing technology and graph neural network technology, and improves the efficiency and accuracy of completion; compared with existing relational databases, since the graph database model can be flexibly expanded, this technology can be flexibly applied to various types of power grid applications based on graph technology. For example, this technology uses the line subgraph of the original graph as input, and the technology can be expanded by adding more node types to the original graph.

[0077] Example 2

[0078] In an exemplary embodiment, a substation power flow estimation system based on graph technology is also provided in this embodiment, including:

[0079] The data acquisition and conversion module is used to obtain the topological relationship between the line and the substation equipment and construct the graph structure data, and perform matrix conversion on any subset of the graph structure data as the input of the graph neural network model;

[0080] Model training module, used to train the graph neural network model based on the input data of the graph neural network model and the missing line flow data;

[0081] The model prediction module is used to predict and complete the missing data generated in real time based on the trained graph neural network model, sum the substation flow connected to the line in the main graph, and update the substation flow data.

[0082] In an optional embodiment, the substation power flow estimation system based on graph technology may perform the following steps:

[0083] Obtaining the topological relationship between the line and the substation equipment and constructing the graph structure data includes: the mathematical expression of the graph structure data is: G = (V, E); where V is the node set and E is the edge set;

[0084] The graph structure data subset is represented as the input of the graph neural network model as: G′=(V′,E′), and Among them, the subset represents the association relationship of all line V′ nodes.

[0085] The graph neural network model includes: input layer, graph convolution layer, full connection layer and output layer. The output features of the graph convolution layer serve as the input of the full connection layer.

[0086] The graph convolutional layer is represented as:

[0087] X L+1 =σ(D -1 (A+I)X L W L )

[0088] Among them, the left side of the equation is the output, the right side of the equation D is the degree matrix, A is the adjacency matrix, which is used to represent the connection relationship between nodes, I is the identity matrix, X is the parameter matrix, where the length of the parameter matrix is ​​the number of lines, the width is the number of timing power, and σ is the nonlinear function sigmoid;

[0089] The W matrix is ​​obtained through graph convolutional neural network machine learning.

[0090] The fully connected layer includes: input layer, hidden layer and output layer. The neurons in each layer are connected to all the neurons in the next layer, forming a fully connected structure to generate predicted output.

[0091] Training a graph neural network model involves:

[0092] The back-propagation algorithm is used to update the iterative graph neural network model parameters. After each training cycle, the validation set is used to evaluate the model performance.

[0093] After each training, when the training loss and validation loss remain stable and less than the expected threshold, the model reaches the convergence condition and outputs the prediction result.

[0094] It also includes: based on the input data of the graph neural network model, combined with the missing line flow data training, a graph neural network model based on other nodes predicts the lines or time points that are prone to data missing, and obtains the output.

[0095] The matrix transformation of any subset of graph structure data includes: adjacency matrix, degree matrix and attribute matrix.

[0096] This embodiment further provides a computing device, which is applicable to the case of the substation power flow estimation method based on graph technology, including:

[0097] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the substation power flow estimation method based on graph technology as proposed in the above embodiment.

[0098] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for estimating power flow in a substation based on graph technology as proposed in the above embodiment is implemented.

[0099] The storage medium proposed in this embodiment and the method for implementing substation power flow estimation based on graph technology proposed in the above embodiment belong to the same inventive concept. The technical details not described in detail in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0100] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A substation power flow estimation method based on graph technology, characterized in that: include: Obtain the topological relationship between the line and the substation equipment and construct graph structure data, and perform matrix conversion on any subset of the graph structure data as input to the graph neural network model; Based on the input data of the graph neural network model and in combination with the missing line flow data, the graph neural network model is trained; According to the trained graph neural network model, the missing data generated in real time are predicted and supplemented, the substation flow connected to the line is summed in the main graph, and the substation flow data is updated.

2. The substation power flow estimation method based on graph technology according to claim 1, characterized in that: Acquiring the topological relationship between the line and the substation equipment and constructing the graph structure data includes: the mathematical expression of the graph structure data is: G = (V, E); wherein V is a node set and E is an edge set; The graph structure data subset is expressed as the input of the graph neural network model as: G′=(V′,E′), and Among them, the subset represents the association relationship of all line V′ nodes.

3. The substation power flow estimation method based on graph technology according to claim 2, characterized in that: The graph neural network model includes: an input layer, a graph convolution layer, a full connection layer and an output layer, and the output features of the graph convolution layer serve as the input of the full connection layer; The graph convolution layer is expressed as: X L+1 =σ(D -1 (A+I)X L W L ) Among them, the left side of the equation is the output, the right side of the equation D is the degree matrix, A is the adjacency matrix, which is used to represent the connection relationship between nodes, I is the identity matrix, X is the parameter matrix, where the length of the parameter matrix is ​​the number of lines, the width is the number of timing power, and σ is the nonlinear function sigmoid; The W matrix is ​​obtained through graph convolutional neural network machine learning.

4. The method for estimating power flow in a substation based on graph technology according to claim 3, characterized in that: The fully connected layer includes: an input layer, a hidden layer and an output layer. The neurons in each layer are connected to all the neurons in the next layer, forming a fully connected structure to generate predicted output.

5. The method for estimating power flow of a substation based on graph technology according to claim 4, characterized in that: Training the graph neural network model includes: The back-propagation algorithm is used to update the iterative graph neural network model parameters. After each training cycle, the validation set is used to evaluate the model performance. After each training, when the training loss and validation loss remain stable and less than the expected threshold, the model reaches the convergence condition and outputs the prediction result.

6. The method for estimating power flow in a substation based on graph technology according to claim 5, characterized in that: Also includes: Based on the input data of the graph neural network model, combined with the missing line flow data training, a graph neural network model based on other nodes predicts the lines or time points that are prone to data missing, and obtains the output.

7. The method for estimating power flow in a substation based on graph technology according to claim 6, characterized in that: The matrix transformation of any subset of graph structure data includes: adjacency matrix, degree matrix and attribute matrix.

8. A substation power flow estimation system based on graph technology, characterized in that: include: A data acquisition and conversion module is used to acquire the topological relationship between the line and the substation equipment and construct graph structure data, and perform matrix conversion on any subset of the graph structure data as input to the graph neural network model; A model training module, used to train the graph neural network model based on the input data of the graph neural network model and the missing line flow data; The model prediction module is used to predict and complete the missing data generated in real time according to the trained graph neural network model, sum the substation flow connected to the line in the main graph, and update the substation flow data.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the substation power flow estimation method based on graph technology described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the substation power flow estimation method based on graph technology as described in any one of claims 1 to 7.

Images