Available Transmission Capacity Calculation Method, System, Equipment and Medium

Abstract

Available Transmission Capacity Calculation Method, System, Device and Medium, which obtain the grid node power to be calculated; input the grid node power to be calculated into a pre-trained graph convolutional network model for calculation to obtain the available transmission capacity calculation result; and output the available transmission capacity calculation result. The graph convolutional network model proposed by the present invention does not rely on complex physical models, can fully exploit historical data, use prior knowledge to quickly find the ATC result, the time cost is much lower than that of the probability sampling algorithm, and the calculation accuracy is high, which can improve the maximum transmission capacity.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and particularly relates to a method, system, device and medium for calculating available transfer capability. Background Art

[0002] To solve the problems of the gradual depletion of traditional fossil energy and the increasing environmental pollution, the penetration rate of renewable energy represented by wind turbines and photovoltaic in the power grid has gradually increased. When calculating ATC, it is necessary to consider the impact of the uncertainty of renewable energy output on the transmission capacity of the power system, and reserve a part of the necessary transmission capacity in advance within the allowable range. A large amount of renewable energy makes the available transfer capability model of the power grid tend to be complex, and thus more difficult to solve. To ensure the economic and stable operation of the power grid, it is urgent to develop a new method with high calculation accuracy, independent of physical models, strong adaptability and capable of achieving the maximum transmission capacity.

[0003] Most of the existing neural networks are improved in terms of structure design and parameter training. The more complex the structure, the more parameters are often required, and the higher the requirements for the correction algorithm. The Radial Basis Function Neural Network (RBFNN) is selected by most scholars and widely used because of its simple structure and the ability to approximate any nonlinear function. Since the expression form of RBFNN is simple and it is an explicit expression from input variables to output variables, RBFNN can be used not only to approximate some expressions in the Nataf transformation, but also to approximate the deterministic solution model of ATC.

[0004] However, the input variables of the available transfer capability calculation model belong to graph data in a non-Euclidean space, which not only includes the injection power of each node, but also the connection relationship between each node. The existing algorithms based on classifier training ignore the topological structure of the network, simplify the input variables into Euclidean space data, and lose some useful information of the input variables, resulting in difficulty in further improving the accuracy. Summary of the Invention

[0005] To overcome the problems in the prior art, the purpose of the present invention is to provide a method, system, device and medium for calculating available transfer capability, and the calculation method has high accuracy and can achieve the maximum transmission capacity.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] Obtain the power of the grid nodes to be calculated;

[0008] Input the power of the grid nodes to be calculated into a pre-trained graph convolutional network model for calculation to obtain the calculation result of the available transfer capability;

[0009] Output the calculated results of the available transmission capacity.

[0010] Furthermore, the pre-trained graph convolutional network model is obtained by inputting the normalized adjacency matrix and the node injection power into the graph convolutional network for training; the node injection power is obtained by preprocessing the collected active power and reactive power of each node in the power grid history.

[0011] Furthermore, collect the active power and reactive power of each node in the power grid history, and preprocess the collected data to obtain the node injection power, including the following steps: map the active power and reactive power of each node in the power grid history to the interval [0, 1] using the deviation normalization method to obtain the node injection power.

[0012] Furthermore, the graph convolutional network includes a first graph convolutional layer, a second graph convolutional layer, a third fully connected layer, and a fourth fully connected layer.

[0013] Furthermore, input the normalized adjacency matrix and the node injection power into the graph convolutional network for training to obtain the trained graph convolutional network model, including the following steps:

[0014] The input of the first graph convolutional layer is the normalized adjacency matrix A″ and the node injection power X. The output vector of the first graph convolutional layer is calculated by the following formula: H (1) = ReLU(A″XW1 + b1)

[0015] where, H (1) is the output vector of the first graph convolutional layer, W1 is the weight matrix of the first graph convolutional layer, and b1 is the bias vector of the first graph convolutional layer;

[0016] The input of the second graph convolutional layer is the output vector H (1) of the first graph convolutional layer. The output vector of the second graph convolutional layer is calculated by the following formula:

[0017] H (2) = ReLU(A″H (1) W2 + b2)

[0018] where: W2 and b2 are the weight matrix and bias vector of the second graph convolutional layer respectively, and H (2) is the output vector of the second graph convolutional layer;

[0019] The input of the third fully connected layer is the output vector of the second graph convolutional layer. The output vector of the third fully connected layer is calculated by the following formula:

[0020] H (3) = ReLU(H (2) W3 + b3)

[0021] where: W3 and b3 are the weight matrix and bias vector of the third fully connected layer, respectively, and H (3) is the output vector of the third fully connected layer;

[0022] The input of the fourth fully connected layer is the output vector of the third fully connected layer, and the output result of the fourth fully connected layer is calculated by the following formula:

[0023] Y = Sigmoid(H (3) W4 + b4)

[0024] where: W4 and b4 are the weight matrix and bias vector of the fourth fully connected layer, respectively, and H (3) is the output vector of the third fully connected layer, and Y represents the output result.

[0025] Furthermore, the normalized adjacency matrix A″ is calculated by the following formula:

[0026]

[0027] where: D is a diagonal matrix, and A′ is a self-connection.

[0028] Furthermore, the self-connection A′ is calculated by the following formula:

[0029] A′ = A + I

[0030] where A is the adjacency matrix and I is the identity matrix.

[0031] Furthermore, input the normalized adjacency matrix and the node injection power into the graph convolutional network for training to obtain a trained graph convolutional network model, including the following steps: continuously correct the graph convolutional network according to the gap between the output result of the graph convolutional network and the actual result until the gap between the output result and the actual result meets the preset requirements, and obtain a trained graph convolutional network model.

[0032] An available transfer capability calculation system includes:

[0033] A power acquisition unit for acquiring the power of grid nodes to be calculated;

[0034] A calculation unit for inputting the power of grid nodes to be calculated into a pre-trained graph convolutional network model for calculation to obtain an available transfer capability calculation result;

[0035] An output unit for outputting the available transfer capability calculation result.

[0036] A computer device, the computer device includes a memory and a processor, and a computer program capable of running on the processor is stored on the memory. When the computer program is executed by the processor, the available transmission capacity calculation method as described above is implemented.

[0037] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is caused to execute the available transmission capacity calculation method as described above.

[0038] Compared with the prior art, the beneficial effects of the present invention are:

[0039] The graph convolutional network model proposed by the present invention does not rely on complex physical models, can fully mine historical data, and use prior knowledge to quickly find the ATC result. The time cost is far lower than that of the probability sampling algorithm. For the operation of the cross-regional power market, according to the calculated available transmission capacity, the power flow distribution of the power grid can be optimized, the transmission capacity of the power grid can be improved, and the available transmission capacity is an important technical reference index for each market participant to conduct trading activities.

[0040] Furthermore, the present invention can map the complex non-linear relationship between the available transmission capacity and the node injection power data through a deep graph convolutional architecture.

[0041] Furthermore, the present invention inputs a normalized adjacency matrix and node injection power into the graph convolutional network for training to obtain a trained graph convolutional network; the present invention can use the adjacency matrix to characterize the topological information between power grid nodes by using the graph convolutional network model, so as to effectively mine the correlation of node injection power. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flowchart of the method of the present invention.

[0043] Figure 2 It is a schematic structural diagram of the system of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0044] The present invention will be described in detail below.

[0045] The following detailed descriptions are all exemplary descriptions, aiming to provide further detailed descriptions of the present invention. Unless otherwise specified, all technical terms adopted by the present invention have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. The terms used in the present invention are only for describing specific embodiments, and are not intended to limit the exemplary embodiments according to the present invention.

[0046] The rapid development of deep learning technology has made it possible to quickly calculate the available transfer capacity (ATC) strategies that do not rely on physical models but instead mine historical data and prior knowledge. For this purpose, a method for calculating ATC based on the Graph Convolutional Network (GCN) is proposed. This invention analyzes the characteristics of the ATC calculation method based on GCN; then, a deep learning model based on GCN is constructed, and a mapping model between the node injection power and the ATC is established through training with historical data; finally, the model is continuously corrected by accumulating historical data.

[0047] The following are relevant noun explanations based on the content of this invention:

[0048] Available Transfer Capacity (ATC): The available transmission capacity of the power grid between regions.

[0049] Graph Convolutional Network (GCN): A convolutional neural network algorithm that uses multiple graph convolutional layers to automatically extract the features of input variables and takes into account the topological structure between each node.

[0050] See Figure 1 , the method for calculating ATC based on the graph convolutional network of this invention includes the following steps:

[0051] When the graph convolutional neural network is used for ATC calculation of the power grid, the learned model will be affected by the injection power and the node connection relationship, which is different from image classification, but the algorithm process is basically the same. By inputting historical data, GCN supervised learns the potential laws of the data and obtains the mapping relationship between the node injection power, the topological structure, and the ATC result. The specific steps are as follows:

[0052] Step 1: Collect input data, where the input data includes the active power and reactive power data of each node in the power grid history, and preprocess the input data;

[0053] The specific process of Step 1 is as follows:

[0054] Considering that the data differences are too large, which is likely to cause problems such as the loss function of GCN being difficult to converge or having low accuracy, the deviation normalization method is used to map each variable to the [0,1] interval to obtain the normalized variable. The specific formula is as follows:

[0055] x′=(x - x min ) / (x max - x min )

[0056] In the formula: x and x' are the variable before normalization and the variable after normalization respectively, x max and x minThey are the maximum and minimum values of the variables respectively. In the present invention, the active power and reactive power data of each node in the power grid history are normalized through the above formula to obtain a normalized set, that is, the node injection power X is obtained.

[0057] Step 2: Normalize the adjacency matrix;

[0058] The specific process of Step 2 is as follows:

[0059] Normalize the adjacency matrix. There is no self-connection for the power grid nodes, and the values of the adjacency matrix A at the diagonal positions are all 0. In fact, the information of itself is also very important during feature extraction. To solve this problem, the adjacency matrix A is added to the identity matrix I, thereby adding a self-connection to each node. The formulaic description is:

[0060] A′ = A + I

[0061] Where A′ is the self-connection, A is the adjacency matrix, and I is the identity matrix.

[0062] Then, since the adjacency matrix A is not normalized, it is not conducive to the training of the neural network. To solve this problem, the adjacency matrix A can be normalized:

[0063]

[0064] In the formula: A″ is the normalized adjacency matrix, D is a diagonal matrix, and the values at its diagonal positions are the degrees of the corresponding nodes.

[0065] Step 3: Establish the basic structure of the graph convolutional network GCN;

[0066] The specific process of Step 3 is as follows:

[0067] GCN is composed of 2 graph convolutional layers (the first graph convolutional layer and the second graph convolutional layer) and 2 fully connected layers (the third fully connected layer and the fourth fully connected layer). The graph convolutional layer is responsible for feature extraction of the node injection power data. The fully connected layer is located at the end of the structure and is used to represent various states of the power equipment.

[0068] For the first graph convolutional layer, the input data are the normalized adjacency matrix A″ and the node injection power X, and the output vector of the first graph convolutional layer is H (1) .

[0069] The formulaic description of the first layer is:

[0070] H (1) = ReLU(A″XW1 + b1)

[0071] In the formula, H (1)is the output vector of the first graph convolutional layer, W1 is the weight matrix of the first graph convolutional layer, and b1 is the bias vector of the first graph convolutional layer.

[0072] The data change process of the second graph convolutional layer is as follows:

[0073] H (2) = ReLU(A″H (1) W2 + b2)

[0074] where: W2 and b2 are the weight matrix and bias vector of the second graph convolutional layer respectively. H (2) is the output vector of the second graph convolutional layer.

[0075] The third and fourth layers of the GCN are fully connected layers. Compared with the graph convolutional layer, the adjacency matrix does not participate in the calculation process of the fully connected layer, and its output result is obtained by multiplying the input data by the weight matrix and then adding the bias vector:

[0076] H (3) = ReLU(H (2) W3 + b3)

[0077] where: W3 and b3 are the weight matrix and bias vector of the third fully connected layer respectively. H (3) is the output vector of the third fully connected layer.

[0078] The output result of the fourth layer is:

[0079] Y = Sigmoid(H (3) W4 + b4)

[0080] where: W4 and b4 are the weight matrix and bias vector of the fourth fully connected layer respectively. H (3) is the output vector of the third fully connected layer. Y represents the output result.

[0081] Step 4: Train the GCN network to obtain a trained GCN model;

[0082] The specific process of Step 4 is as follows:

[0083] Through the active power, reactive power and ATC result data of each node in the historical power grid (i.e., the output result Y), perform GCN training. According to the gap between the output ATC result of the GCN and the actual ATC result, continuously correct the parameters of each layer of the GCN (i.e., the weight matrix and bias vector of the first graph convolutional layer, the second graph convolutional layer, the third fully connected layer and the fourth fully connected layer), and train until the gap between the output result and the actual result meets the preset requirements (the preset requirements are determined according to the actual situation) to obtain the GCN learning model corresponding to the historical data, that is, the trained GCN model.

[0084] Step 5: Input the power of the power grid nodes to be calculated into the trained GCN model, output the ATC calculation result, and improve the power transmission capacity of the power grid according to the calculation result of the transmission capacity, and guide the cross-regional power transaction.

[0085] The available transmission capacity in the calculation result of the transmission capacity is an important technical reference index for each market participant to carry out trading activities. The power grid operation and dispatching management department needs the available transmission capacity as an important technical guidance, and the market entity uses the available transmission capacity as the decision-making basis for the electric energy transaction.

[0086] The graph convolutional network (GCN) is an extension of the traditional convolutional network (CNN) in the non-Euclidean space. It can not only automatically extract the features of the input variables by using multiple graph convolutional layers, but also take into account the topological structure between each node, which is very suitable for processing complex graph data. At present, GCN has good application effects in the fields of link prediction, protein classification, drug synthesis, and cross-domain person re-identification, etc., but its application in the operation of the power system is still in the primary stage. The present invention can make full use of its adjacency matrix to characterize the topological information between the power grid nodes by using GCN, effectively mine the correlation between the nodes and the branches, and establish the mapping relationship between the available transmission capacity and the node injection power data through the deep graph convolutional architecture, so as to more quickly and accurately determine the ATC result of the transmission line.

[0087] See Figure 2 , the available transmission capacity calculation system, including:

[0088] A power acquisition unit, configured to acquire the power of the power grid nodes to be calculated;

[0089] A calculation unit, configured to input the power of the power grid nodes to be calculated into a pre-trained graph convolutional network model for calculation to obtain the calculation result of the available transmission capacity;

[0090] An output unit, configured to output the calculation result of the available transmission capacity.

[0091] A computer device, the computer device includes a memory and a processor, and a computer program capable of running on the processor is stored on the memory. When the computer program is executed by the processor, the steps of the available power transmission capacity calculation method described above are implemented. Among them, the memory may include internal memory, such as high-speed random access memory, and may also include non-volatile memory, such as at least one disk memory, etc.; the processor, network interface, and memory are interconnected through an internal bus, and this internal bus can be an Industry Standard Architecture bus, a Peripheral Component Interconnect standard bus, an Extended Industry Standard Architecture bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs. Specifically, the program may include program code, and the program code includes computer operation instructions. The memory may include internal memory and non-volatile memory, and provide instructions and data to the processor.

[0092] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor is caused to execute the steps of the available power transmission capacity calculation method described above. Specifically, the computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may include random access memory (RAM) and / or cache memory, etc. The non-volatile memory may include read-only memory (ROM), hard disk, flash memory, optical disc, magnetic disk, etc.

[0093] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, rather than all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the protection scope of the present invention.

[0094] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other.

[0095] The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0096] In the present invention, terms such as "module", "device", "system", etc. refer to related entities applied to a computer, such as hardware, a combination of hardware and software, software, or software in execution. Specifically, for example, an element may be, but is not limited to, a process running on a processor, a processor, an object, an executable element, an execution thread, a program, and / or a computer. Also, an application program or a script program running on a server, and the server can both be elements. One or more elements may be in an execution process and / or thread, and the elements may be localized on one computer and / or distributed between two or more computers, and can be run by various computer-readable media. The elements can also communicate through local and / or remote processes according to a signal having one or more data packets, for example, a signal from data that interacts with another element in a local system, a distributed system, and / or interacts with other systems through a signal on a network in the Internet.

[0097] Finally, it should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprise" and "include", not only include those elements, but also include other elements not explicitly listed, or also include elements inherent to such a process, method, article, or device. Without further limitation, an element defined by the statement "comprising..." does not exclude the existence of additional identical elements in the process, method, article, or device that includes the said element.

[0098] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0099] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementing in the process Figure 1 one process or multiple processes and / or blocksFigure 1 means for the functions specified in one or more boxes.

[0100] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce a manufacture including an instruction means that implements the functions specified in one Figure 1 process or more processes and / or boxes Figure 1 or more boxes.

[0101] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus, such that a series of operational steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in one Figure 1 process or more processes and / or boxes Figure 1 or more boxes.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent replacements can still be made to the specific embodiments of the present invention, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered by the protection scope of the claims of the present invention.

Claims

1. Calculation method for available transmission capacity, characterized in that Including the following steps: Obtain the power of the grid nodes to be calculated; Input the power of the grid nodes to be calculated into a pre-trained graph convolutional network model for calculation to obtain the available transfer capability calculation result; Output the available transfer capability calculation result; The pre-trained graph convolutional network model is obtained by inputting a normalized adjacency matrix and node injection power into the graph convolutional network for training; the node injection power is obtained by preprocessing the collected active power and reactive power of each grid node in the history; The node injection power is specifically obtained through the following steps: Map the active power and reactive power of each grid node in the history to the interval [0, 1] using the deviation normalization method to obtain the node injection power; The pre-trained graph convolutional network model is specifically obtained through the following steps: The input of the first graph convolutional layer is the normalized adjacency matrix A″ and the node injection power X. The output vector of the first graph convolutional layer is calculated by the following formula: H (1) = ReLU(A″XW1 + b1) where H (1) is the output vector of the first graph convolutional layer, W1 is the weight matrix of the first graph convolutional layer, and b1 is the bias vector of the first graph convolutional layer; The input of the second graph convolutional layer is the output vector H of the first graph convolutional layer (1) , and the output vector of the second graph convolutional layer is calculated by the following formula: H (2) = ReLU(A″H (1) W2 + b2) where: W2 and b2 are the weight matrix and bias vector of the second graph convolutional layer, respectively, and H (2) is the output vector of the second graph convolutional layer; The input of the third fully connected layer is the output vector of the second graph convolutional layer, and the output vector of the third fully connected layer is calculated by the following formula: H (3) = ReLU(H (2) W3 + b3) Where: W3 and b3 are the weight matrix and bias vector of the third fully connected layer, respectively, and H (3) is the output vector of the third fully connected layer; The input of the fourth fully connected layer is the output vector of the third fully connected layer, and the output result of the fourth fully connected layer is calculated by the following formula: Y = Sigmoid(H (3) W4 + b4) Where: W4 and b4 are the weight matrix and bias vector of the fourth fully connected layer, respectively, H (3) is the output vector of the third fully connected layer, and Y represents the output result; The normalized adjacency matrix A″ is calculated by the following formula: In the formula: D is a diagonal matrix, and A′ is a self-connection; The self-connection A′ is calculated by the following formula: A′ = A + I Where A is the adjacency matrix and I is the identity matrix.

2. The available transmission capacity calculation method according to claim 1, characterized in that Input the normalized adjacency matrix and node injection power into the graph convolutional network for training to obtain a trained graph convolutional network model, including the following steps: Continuously correct the graph convolutional network according to the gap between the output result of the graph convolutional network and the actual result until the gap between the output result and the actual result meets the preset requirements to obtain a trained graph convolutional network model.

3. Available Transmission Capacity Calculation System, which uses the available transmission capacity calculation method described in Claim 1 to calculate the transmission capacity, and is characterized in that Including: A power acquisition unit for obtaining the power of the grid nodes to be calculated; A calculation unit for inputting the power of the grid nodes to be calculated into a pre-trained graph convolutional network model for calculation to obtain the available transfer capability calculation result; An output unit for outputting the available transfer capability calculation result.

4. A computer device, characterized in that, The computer device includes a memory and a processor, and a computer program capable of running on the processor is stored on the memory. When the computer program is executed by the processor, it implements the available transfer capability calculation method according to any one of claims 1 to 2.

5. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program. When the computer program is executed by the processor, it causes the processor to execute the available transfer capability calculation method according to any one of claims 1 to 2.

Images