A power distribution network carbon emission factor calculation method based on graph convolution network learning

Abstract

This invention discloses a method for calculating the carbon emission factor of a distribution network based on graph convolutional network learning, comprising: constructing an input matrix representing the electrical carbon information of nodes and an adjacency matrix representing the topological relationship based on the extracted topological connection relationship of the distribution system and the measurement data of each node; constructing a carbon emission factor calculation model of the distribution network based on a spectral domain graph convolutional network, and training the carbon emission factor calculation model of the distribution network based on the input matrix and the adjacency matrix; establishing a mapping relationship between the carbon emission factor and the input electrical carbon information based on the carbon emission factor calculation model; and calculating the carbon emission factor values ​​of each node of the distribution network at the measurement time based on the mapping relationship.

Description

Technical Field

[0001] This invention relates to the fields of electrical technology and information technology, and more specifically, to a method for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning. Background Technology

[0002] As a public infrastructure connecting energy production and consumption, the power grid plays a crucial role in providing services. The distribution network, a vital component of the power grid, not only has inherent requirements for carbon reduction in production but also, as a provider of high-quality electricity services, directly faces users and plays an irreplaceable role in promoting user-side carbon asset management and demand-side response. Currently, carbon flow theory has been proposed for carbon emission accounting in distribution networks. Based on power flow state estimation information from the transmission network, it can calculate indirect carbon emissions in real time and accurately quantify the differences in carbon emission factors among different nodes within the smallest spatial scale region of the average carbon emission factor.

[0003] Unlike transmission networks, which have comprehensive electrical measurement systems, distribution networks typically have measurement systems distributed at the source and load sides. Considering engineering practicality and economy, the goal of installing measurement equipment at all nodes is difficult to achieve and cannot cover the full observability of the system. Furthermore, due to inherent limitations of the devices themselves and data transmission issues, SCADA equipment always has some degree of error. Therefore, it is urgent to establish a carbon emission calculation model based on the existing measurement information of the distribution network. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for calculating the carbon emission factor of power distribution networks based on graph convolutional network learning.

[0005] According to one aspect of the present invention, a method for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning is provided, comprising: Based on the extracted power distribution system topology connections and the measurement data of each node, an input matrix representing the node's electrical carbon information and an adjacency matrix representing the topology connections are constructed. A carbon emission factor calculation model for distribution networks is constructed based on spectral domain graph convolutional networks, and the model is trained according to the input matrix and adjacency matrix. Based on the carbon emission factor calculation model, a mapping relationship between the carbon emission factor and the input electrical carbon information is established; Based on the mapping relationship, the carbon emission factor values ​​of each node in the distribution network at the measurement time are calculated.

[0006] Optionally, based on the extracted distribution system topology connections and the measurement data of each node, an input matrix representing the node electrocarbon information and an adjacency matrix representing the topological relationships are constructed, including: Extract the topological connections of the power distribution system to obtain its graph-structured network. The input matrix and adjacency matrix of node carbon information are constructed based on the measurement data of each node in the graph structure network. The measurement data of each node includes the active power of the unit, the active power of the load, the carbon emission data of the unit, and the power generation data of the unit.

[0007] Optionally, the expression for the input matrix is: In the formula, ; X The matrix formed by the carbon information vectors of each node; W i Let i be the carbon information vector of node i (1≤i≤N); The matrix formed by the carbon information of the nodes at time t. For m time points The nodal carbon information input matrix is ​​formed by (1≤k≤m); for The transpose of .

[0008] Optionally, a model for calculating the carbon emission factor of a power distribution network is constructed based on graph convolutional network learning, including: Based on the defined Laplace matrix of the spectral domain graph convolution network, construct the graph convolution operation formula; A calculation model for carbon emission factors in power distribution networks is constructed based on the graph convolution operation formula.

[0009] Optionally, the graph convolution operation formula is: In the formula, D It is a diagonal matrix; A It is an adjacency matrix; θ 0 represents the initial Chebyshev coefficient.

[0010] Optionally, the carbon emission factor calculation model for the power distribution network includes: an input layer, a feature extraction layer, and an output layer, wherein... The input layer consists of an input matrix and an adjacency matrix, where the node injection power represents the node electrical information, the unit carbon emission intensity represents the node carbon information, and the dynamic carbon emission factor represents the non-source side node carbon information. The feature extraction layer consists of two layers of graph convolutional networks. Each graph convolutional network takes the output of the previous layer as input and extracts features through graph convolution operation formula. The output layer consists of a fully connected layer that maps the data output from the feature extraction layer to the output space.

[0011] Optionally, the loss function of the distribution network carbon emission factor calculation model is: In the formula, It is a vector The second norm; m At that moment; Carbon emission factor vector for: In the formula, for t The GCN model calculates the next node at time step. i The carbon emission factor value, calculated based on GCN, requires the carbon flow calculation result as the target value, assuming... t Carbon emission factor vector under time-varying carbon flow calculation for: In the formula, for t Time-dependent carbon flow calculation at the next node i Carbon emission factor value.

[0012] Optionally, the mapping function that constructs the mapping relationship is: In the formula, For the input matrix, A It is an adjacency matrix. and b These are trainable parameters. It is a non-linear activation function. For graph convolutional network mapping functions, Y The output of this method is a carbon emission factor. .

[0013] Optionally, the evaluation indicators for the carbon emission factor calculation model of the distribution network include: the average absolute error between the calculated value and the target value, and the average relative error between the calculated value and the target value.

[0014] According to another aspect of the present invention, a device for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning is provided, comprising: The module is used to construct an input matrix representing the electrical carbon information of nodes and an adjacency matrix representing the topological relationship, based on the extracted power distribution system topology connection relationship and the measurement data of each node. The training module is used to construct a carbon emission factor calculation model for the distribution network based on the spectral domain graph convolutional network, and to train the carbon emission factor calculation model for the distribution network according to the input matrix and the adjacency matrix. A module is established to create a mapping relationship between carbon emission factors and input electrical carbon information based on a carbon emission factor calculation model. The calculation module is used to calculate the carbon emission factor values ​​of each node in the distribution network at the measurement time based on the mapping relationship.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, the storage medium storing a computer program for performing the method of any of the above aspects of the present invention.

[0016] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: a processor; a memory for storing executable instructions of the processor; the processor being configured to read the executable instructions from the memory and execute the instructions to implement the method described in any of the preceding aspects of the present invention.

[0017] Therefore, this invention constructs an input matrix representing the electrical carbon information of nodes and an adjacency matrix representing the topological relationships, based on the topological connection relationship of the power distribution system and the measurement data of each node. Secondly, using the input matrix and adjacency matrix as model input, the carbon emission factor at the measurement time is calculated based on carbon flow theory, and this is used as the target output. The GCN model is then trained to fit and learn the carbon emission factor calculation function, establishing a mapping relationship between the carbon emission factor and the input electrical carbon information, thereby generating the carbon emission factor values ​​of each node at the measurement time. Attached Figure Description

[0018] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures: Figure 1 This is a flowchart illustrating a method for calculating carbon emission factors in power distribution networks based on graph convolutional network learning, provided in an exemplary embodiment of the present invention. Figure 2 This is another flowchart illustrating the method for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning, provided by an exemplary embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of a power distribution network carbon emission factor calculation device based on graph convolutional network learning provided in an exemplary embodiment of the present invention; Figure 4 This is the structure of an electronic device provided in an exemplary embodiment of the present invention. Detailed Implementation

[0019] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0020] It should be noted that, unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of the invention.

[0021] Those skilled in the art will understand that the terms "first," "second," etc., in the embodiments of the present invention are only used to distinguish different steps, devices, or modules, and do not represent any specific technical meaning, nor do they indicate a necessary logical order between them.

[0022] It should also be understood that in the embodiments of the present invention, "multiple" can refer to two or more, and "at least one" can refer to one, two or more.

[0023] It should also be understood that any component, data or structure mentioned in the embodiments of the present invention can generally be understood as one or more unless explicitly defined or given contrary instructions in the context.

[0024] Furthermore, the term "and / or" in this invention is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this invention generally indicates that the preceding and following related objects have an "or" relationship.

[0025] It should also be understood that the description of the various embodiments in this invention emphasizes the differences between the various embodiments, and the similarities or similarities can be referred to each other. For the sake of brevity, they will not be described in detail.

[0026] At the same time, it should be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn according to actual scale.

[0027] The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the invention or its application or use.

[0028] Techniques, methods, and equipment known to those skilled in the art may not be discussed in detail, but where appropriate, they should be considered part of the specification.

[0029] It should be noted that similar labels and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be discussed further in subsequent figures.

[0030] The embodiments of this invention can be applied to electronic devices such as terminal devices, computer systems, and servers, and can operate together with a wide range of other general-purpose or special-purpose computing system environments or configurations. Well-known examples of terminal devices, computing systems, environments, and / or configurations suitable for use with electronic devices such as terminal devices, computer systems, and servers include, but are not limited to: personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and distributed cloud computing environments including any of the above systems, etc.

[0031] Electronic devices such as terminal devices, computer systems, and servers can be described in the general context of computer system executable instructions (such as program modules) executed by a computer system. Typically, program modules can include routines, programs, object programs, components, logic, data structures, etc., which perform specific tasks or implement specific abstract data types. Computer systems / servers can be implemented in distributed cloud computing environments, where tasks are executed by remote processing devices linked through communication networks. In distributed cloud computing environments, program modules can reside on local or remote computing system storage media, including storage devices.

[0032] Exemplary methods Figure 1 This is a schematic flowchart illustrating a method for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning, provided in an exemplary embodiment of the present invention. This embodiment can be applied to electronic devices, such as… Figure 1 As shown, the method 100 for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning includes the following steps: Step 101: Based on the extracted power distribution system topology connection relationship and the measurement data of each node, construct an input matrix representing the node electrical carbon information and an adjacency matrix representing the topology relationship. Step 102: Based on the spectral domain graph convolutional network, construct a carbon emission factor calculation model for the distribution network, and train the carbon emission factor calculation model for the distribution network according to the input matrix and the adjacency matrix. Step 103: Based on the carbon emission factor calculation model, establish the mapping relationship between the carbon emission factor and the input electrical carbon information; Step 104: Based on the mapping relationship, calculate the carbon emission factor values ​​of each node in the distribution network at the measurement time.

[0033] Specifically, in view of the fact that existing technologies cannot establish a mapping relationship between carbon emissions and measurement information based on the node feature information measured in the distribution network to determine the carbon emission factor value of each node, this invention provides a method for calculating the carbon emission factor of the distribution network based on graph convolutional network learning.

[0034] Step S1: Extract the topological connection relationship of the power distribution system and the measurement data of each node, and construct the input matrix representing the electrical carbon information of the nodes and the adjacency matrix representing the topological relationship; Specifically, the topological connections of a power distribution system include topological information. During the extraction of distribution network topological information, the connection relationships between nodes and lines allow the distribution network to be viewed as a graph-structured network. Its topological structure determines the transmission paths and distribution of electrical energy. Given an ordered binary system... V , E () represents the graphical structure of the distribution network G . V Represents a set of nodes. E This represents an edge set, where the connection between any two points is considered as an edge between those two points. Represents a node. Represents a node i , j The edges between them.

[0035] The measurement data at each node includes generator active power, load active power, generator carbon emissions data, and generator power generation data. The raw electrical information primarily comes from source-side SCADA and load-side µPMU measurement data. Source-side SCADA can measure active and reactive power, as well as node voltage and current amplitudes, but cannot directly measure phase angle; load-side µPMU can measure node voltage and current amplitudes and phase angles, but cannot directly measure power; however, it can indirectly calculate power. Measurement data is used to obtain generator active power and calculate load active power. Raw carbon information includes real-time measured direct carbon emissions and generator power generation data.

[0036] Adjacency matrix represents the connection relationships between nodes, defined as follows: A For the image G If the adjacency matrix is... ,but Otherwise, it is 0. Input matrix X in Reflecting node carbon information, nodes in the distribution network have N If there are 1 node, then the node i (1≤ i ≤ N The carbon information vector of ) is The matrix X formed by the carbon information vectors of each node is: (1) Assumption for t The matrix formed by the time node's electrical carbon information, if studied m At that moment, (1≤ k ≤ m The nodal electrocarbon information input matrix formed for: (2) In the formula, yes The transpose of .

[0037] Step S2: Construct a calculation model for the carbon emission factor of the power distribution network based on graph convolutional network learning and train the model.

[0038] A spectral domain graph convolutional network is used to construct a carbon emission factor calculation model. The Laplace matrix on the graph structure G is defined as follows: diagonal matrix D For the image G The degree matrix of nodes on, .

[0039] Normalized Laplace matrix: (3) Further, we can obtain: (4) In the formula, It is an identity matrix. A real symmetric matrix that is positive semi-definite. The eigenvalue decomposition is performed as follows: (5) In the formula, for A diagonal matrix composed of eigenvalues; U To be according to The eigenvector matrix sorted by eigenvalues ​​defines the signal. x , y In the figure G The convolution on is: (6) In the formula, Represents convolution on graph G. Represents the dot product of vectors. Definition For the convolution kernel to be learned, we can obtain (7) In the formula, To project the signal x from the node domain into the frequency domain.

[0040] To avoid gradient vanishing or exploding, the convolution kernel is fitted using Chebyshev polynomials as follows: (8) In the formula, Represents a Chebyshev polynomial of order k. The matrix represents the corresponding Chebyshev coefficients, K represents the number of terms in the Chebyshev polynomial, and the matrix represents the number of terms in the polynomial. The eigenvalues ​​are between ±1, which meets the range requirements of Chebyshev polynomials.

[0041] Ignoring higher-order terms and keeping only k=0,1, according to the properties of Chebyshev polynomials, we can obtain: (9) Substitute formula (4) into formula (9) and let achievable (10) make Define a diagonal matrix satisfy The formula for graph convolution can be obtained as follows: (11) The distribution network carbon emission factor model comprises an input data layer, a feature extraction layer, and an output layer. First, the raw information of the input data is transformed into intermediate information, ultimately forming the input data layer by constructing a data input matrix and an adjacency matrix. The raw information includes raw electrical carbon information and topology information. Raw electrical information includes active power data from generating units and loads, as well as measurement data from source-side SCADA and load-side µPMU. Source-side SCADA can measure active and reactive power, node voltage, and current amplitude, but cannot directly measure phase angle; load-side µPMU can measure node voltage, current amplitude, and phase angle, but cannot directly measure power although it can indirectly calculate power. Raw carbon information includes real-time measured direct carbon emissions and power generation data from generating units. Based on the raw information, intermediate information can be calculated: source-side carbon emission intensity is used, and non-source-side average carbon emission factor is used. The electrical carbon information forms the input matrix, and the topology information forms the adjacency matrix, together constituting the data input matrix.

[0042] In the input data layer, In the input data layer, node injected power represents node electrical information, unit carbon emission intensity represents node carbon information, and unit carbon emission intensity is calculated from the raw carbon information. In the formula: This represents the carbon emissions measured by the unit during time period t. The generating capacity of the unit during time period t. The time interval measured in units.

[0043] Most of the non-source-side nodes are load nodes. Based on the relationship between load carbon flow rate and carbon emission factor in carbon flow theory, the dynamic carbon emission factor is calculated to characterize the carbon information of non-source-side nodes, specifically including: Neglecting network losses, the sum of the carbon flow rate injected into the system and the carbon flow rate of the system outflow load is equal, specifically: (12) In the above formula, and The inflow and outflow carbon flow rates.

[0044] The formula for calculating carbon flow rate is: (13) In the above formula, It is a meritorious trend. It is a carbon emission factor.

[0045] Assuming there is in the distribution network A load, defining the average outflow carbon flow rate at the load. for; (14) Then the dynamic inflow average carbon emission factor The calculation formula is: (15) Numerically, it is equal to the arithmetic mean of the carbon emission factors at each node under carbon flow calculation. For the unit's active power, For the unit's carbon emission intensity, The active power of the load is derived from known intermediate information from the graph convolutional network model, when the unit outputs... and load active power When changes occur (default unit carbon emission intensity) constant), It changes dynamically accordingly.

[0046] The input matrix and the adjacency matrix together constitute the input data layer of the model.

[0047] The feature extraction layer consists of a two-layer graph convolutional network. Each graph convolutional network takes the output of the previous layer as input and extracts features by performing the convolution operation of equation (11). n Graph convolutional networks of layers propagate in the following way: (16) In the above formula, Represents the input information of layer n. Representative process n The output information after integration by the layer graph convolutional network. For the first n The parameter matrix in a layered graph convolutional network, The activation function is non-linear, and the input information of the first layer of the graph convolutional network is the information input matrix. .

[0048] The propagation process consists of three steps: feature propagation, feature transformation, and nonlinear transition. Neighbor node feature propagation is achieved by left-multiplying the updated adjacency matrix. Each node in the distribution network obtains neighbor node information through feature propagation at each layer of the GCN, and each feature propagation also smooths the feature values ​​between neighbor nodes. After local smoothing, the feature is then right-multiplied by the learnable parameter matrix. A linear transformation is then performed on the smoothed node features. Finally, a non-linear activation function is applied before the output features. For example, by correcting the linear unit, we can obtain the input data for the next layer of GCN. .

[0049] The first layer of the GCN captures the local relationships between features of first-order neighbor nodes. The second layer of the GCN further updates the information of each node through second-order neighbor nodes, based on the first layer, to improve the model's learning ability and computational accuracy. As the number of layers increases, the parameter matrix... The gradient may gradually decrease and eventually vanish, making it difficult for the model to converge. Therefore, using two layers of GCN can limit the depth of gradient propagation and avoid the potential gradient vanishing problem.

[0050] The output layer consists of a fully connected layer, whose function is to process the output of the feature extraction layer. Mapping to the output space. The mapping function for graph convolutional networks is: In the formula, For the input matrix, A It is an adjacency matrix. and b These are trainable parameters. It is a non-linear activation function. For graph convolutional network mapping functions, Y The output of this method is a carbon emission factor. .

[0051] Assumption t Carbon emission factor vector calculated by the GCN model at time step for: (17) In the formula, for t The GCN model calculates the next node at time step. i The carbon emission factor value, calculated based on GCN, requires the carbon flow calculation result as the target value, assuming... t Carbon emission factor vector under time-varying carbon flow calculation for: (18) In the formula, Fort Time-dependent carbon flow calculation at the next node i Carbon emission factor value.

[0052] If research m At time n, the loss function is defined as: (19) In the formula, It is a vector The L2 norm is used to define the loss function, which can prevent overfitting and improve the generalization ability of the model.

[0053] The samples are randomly divided into training, validation, and test sets. The training and validation sets are used to train the GCN model, and the test set is used to evaluate the GCN model. The parameter matrix is ​​updated... Implement the model loss function loss To minimize the loss function, the direction and magnitude of parameter updates depend on the gradient of the loss function. The gradient of the loss function with respect to the model parameters is calculated using backpropagation, and the model parameters are updated along the negative gradient direction. After multiple parameter updates, model training is complete.

[0054] Step 3: Establish the mapping relationship between carbon emission factors and input electrical carbon information, and calculate the carbon emission factor values ​​of each node in the distribution network at the measurement time.

[0055] Specifically, the GCN model can learn the mapping relationship between the carbon emission factor to be determined and the input electrical carbon information, thereby calculating the carbon emission factor values ​​of each node using the input electrical carbon information.

[0056] To evaluate model performance, MAE and MAPE are used as evaluation metrics. MAE is the average absolute error between the calculated value and the target value, and MAPE is the average relative error between the calculated value and the target value. The specific formulas are as follows: (20) (twenty one) In the formula, and These are the GCN model and carbon flow theory at the 1st x Nodes in each sample y The carbon emission factor calculated above, n 1 is the number of nodes in each sample. n 2 represents the number of samples in the test set. To avoid the denominator becoming meaningless when calculating MAPE due to a carbon emission factor target value of 0, this technique adds a very small positive number to the denominator. , The value is 10 -6 .

[0057] To reflect the error of the GCN model at different nodes, the mean absolute error of node y on each sample of the test set is defined. as follows: (twenty two) In one specific embodiment of the present invention, an improved IEEE-123 node system was used for testing. Excluding the interconnection node, there are a total of 114 nodes. Distributed wind turbines are connected to nodes 5, 13, 26, 57, 67, 84, and 89, distributed gas turbines are connected to nodes 21, 45, and 63, and distributed photovoltaics are connected to nodes 74 and 109.

[0058] In this embodiment, the dataset consists of node injected power, source-side unit carbon emission intensity, non-source-side node dynamic carbon emission factor values, and target values ​​of carbon emission factors calculated by carbon flow. When acquiring power flow data, due to the volatility of distributed unit output, it is assumed that each distributed unit fluctuates between its lower output limit and rated power, and any load fluctuates within 50% to 150% of the standard example load. All power flow samples obtained from the load converge. The unit carbon emission intensity is assumed to remain constant. A total of 5000 valid samples are obtained, of which 4000 are randomly selected for constructing the training set, 500 are used as the validation set, and the remaining 500 are used as the test set.

[0059] The model was trained using the Adam optimizer with 2000 iterations. Since the number of convolutional kernels in GCNs is typically a power of 2, the loss function values ​​and average time required for each iteration were statistically analyzed for different convolutional kernels (2¹~2¹⁰) after ten iterations (the first ten iterations showed the largest decrease in loss function). These results are shown in Table 1. Table 1. Effect of different numbers of convolutional kernels on the loss function after ten iterations.

[0060] In this embodiment, , The calculated MAE of the GCN model on the test set is 0.0342 kgCO2 / kWh, and the MAPE is 9.55%. MAE reflects the actual magnitude of the error; 0.0342 represents the average absolute error of the GCN model across different nodes in the test set. MAPE reflects the relative magnitude of the error; 9.55% is the average relative error of the GCN model across different nodes in the test set.

[0061] Most nodes have relatively small errors, but nodes 29, 30, 33, 51, 61, 96, and 109-113 have larger errors. These nodes with larger errors are all located at the ends of radial branches, and the errors increase the closer the node is to the end. This is because nodes near the end of branches have less information during the two-layer GCN update. The mean absolute errors of nodes 29, 30, 33, 51, 96, and 109-113 are 0.0746, 0.1061, 0.0865, 0.0742, 0.0892, 0.1037, 0.0997, 0.1009, 0.1069, 0.1498, and 0.1812, respectively.

[0062] Therefore, this invention constructs an input matrix representing the electrical carbon information of nodes and an adjacency matrix representing the topological relationships, based on the topological connection relationship of the power distribution system and the measurement data of each node. Secondly, using the input matrix and adjacency matrix as model input, the carbon emission factor at the measurement time is calculated based on carbon flow theory, and this is used as the target output. The GCN model is then trained to fit and learn the carbon emission factor calculation function, establishing a mapping relationship between the carbon emission factor and the input electrical carbon information, thereby generating the carbon emission factor values ​​of each node at the measurement time.

[0063] Exemplary device Figure 3 This is a schematic diagram of the structure of a power distribution network carbon emission factor calculation device based on graph convolutional network learning, provided in an exemplary embodiment of the present invention. Figure 3 As shown, the device 300 includes: The construction module 310 is used to construct an input matrix representing the electrical carbon information of the nodes and an adjacency matrix representing the topological relationship based on the extracted power distribution system topology connection relationship and the measurement data of each node. Training module 320 is used to construct a carbon emission factor calculation model for the distribution network based on the spectral domain graph convolutional network, and to train the carbon emission factor calculation model for the distribution network according to the input matrix and the adjacency matrix. Module 330 is established to create a mapping relationship between carbon emission factors and input electrical carbon information based on the carbon emission factor calculation model. The calculation module 340 is used to calculate the carbon emission factor values ​​of each node in the distribution network at the measurement time based on the mapping relationship.

[0064] Optionally, based on the extracted distribution system topology connections and the measurement data of each node, an input matrix representing the node electrocarbon information and an adjacency matrix representing the topological relationships are constructed, including: Extract the topological connections of the power distribution system to obtain its graph-structured network. The input matrix and adjacency matrix of node carbon information are constructed based on the measurement data of each node in the graph structure network. The measurement data of each node includes the active power of the unit, the active power of the load, the carbon emission data of the unit, and the power generation data of the unit.

[0065] Optionally, the expression for the input matrix is: In the formula, ; X The matrix formed by the carbon information vectors of each node; W i Let i be the carbon information vector of node i (1≤i≤N); The matrix formed by the carbon information of the nodes at time t. For m time points The nodal carbon information input matrix is ​​formed by (1≤k≤m); for The transpose of .

[0066] Optionally, a model for calculating the carbon emission factor of a power distribution network is constructed based on graph convolutional network learning, including: Based on the defined Laplace matrix of the spectral domain graph convolution network, construct the graph convolution operation formula; A calculation model for carbon emission factors in power distribution networks is constructed based on the graph convolution operation formula.

[0067] Optionally, the graph convolution operation formula is: In the formula, D It is a diagonal matrix; A It is an adjacency matrix; θ 0 represents the initial Chebyshev coefficient.

[0068] Optionally, the carbon emission factor calculation model for the power distribution network includes: an input layer, a feature extraction layer, and an output layer, wherein... The input layer consists of an input matrix and an adjacency matrix, where the node injection power represents the node electrical information, the unit carbon emission intensity represents the node carbon information, and the dynamic carbon emission factor represents the non-source side node carbon information. The feature extraction layer consists of two layers of graph convolutional networks. Each graph convolutional network takes the output of the previous layer as input and extracts features through graph convolution operation formula. The output layer consists of a fully connected layer that maps the data output from the feature extraction layer to the output space.

[0069] Optionally, the loss function of the distribution network carbon emission factor calculation model is: In the formula, It is a vector The second norm; m At that moment; Carbon emission factor vector for: In the formula, for t The GCN model calculates the next node at time step. i The carbon emission factor value, calculated based on GCN, requires the carbon flow calculation result as the target value, assuming... t Carbon emission factor vector under time-varying carbon flow calculation for: In the formula, for t Time-dependent carbon flow calculation at the next node i Carbon emission factor value.

[0070] Optionally, the mapping function that constructs the mapping relationship is: In the formula, For the input matrix, A It is an adjacency matrix. and b These are trainable parameters. It is a non-linear activation function. For graph convolutional network mapping functions, Y The output of this method is a carbon emission factor. .

[0071] Optionally, the evaluation indicators for the carbon emission factor calculation model of the distribution network include: the average absolute error between the calculated value and the target value, and the average relative error between the calculated value and the target value.

[0072] Exemplary electronic devices Figure 4 This is the structure of an electronic device provided in an exemplary embodiment of the present invention. For example... Figure 4 As shown, the electronic device 40 includes one or more processors 41 and memory 42.

[0073] The processor 41 may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions.

[0074] The memory 42 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 41 may execute the program instructions to implement the methods of the software programs of the various embodiments of the present invention described above, and / or other desired functions. In one example, the electronic device may also include an input device 43 and an output device 44, these components being interconnected via a bus system and / or other forms of connection mechanisms (not shown).

[0075] In addition, the input device 43 may also include, for example, a keyboard, a mouse, etc.

[0076] The output device 44 can output various information to the outside. The output device 44 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0077] Of course, for the sake of simplicity, Figure 4 Only some of the components of this electronic device relevant to the present invention are shown, omitting components such as buses, input / output interfaces, etc. In addition, the electronic device may include any other suitable components depending on the specific application.

[0078] Exemplary computer program products and computer-readable storage media In addition to the methods and apparatus described above, embodiments of the present invention may also be computer program products, which include computer program instructions that, when executed by a processor, cause the processor to perform the steps in the methods according to various embodiments of the present invention described in the "Exemplary Methods" section above.

[0079] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of the present invention. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0080] Furthermore, embodiments of the present invention may also be computer-readable storage media storing computer program instructions thereon, which, when executed by a processor, cause the processor to perform the steps of the methods according to various embodiments of the present invention described in the "Exemplary Methods" section above.

[0081] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0082] The basic principles of the present invention have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in the present invention are merely examples and not limitations, and should not be considered as essential features of each embodiment of the present invention. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the present invention to the necessity of employing the aforementioned specific details.

[0083] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For system embodiments, since they largely correspond to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments.

[0084] The block diagrams of devices, systems, devices, and systems involved in this invention are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, systems, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0085] The methods and systems of the present invention may be implemented in many ways. For example, they may be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above-described order of steps for the methods is for illustrative purposes only, and the steps of the methods of the present invention are not limited to the order specifically described above unless otherwise specifically stated. Furthermore, in some embodiments, the present invention may also be implemented as a program recorded on a recording medium, the program comprising machine-readable instructions for implementing the methods according to the present invention. Thus, the present invention also covers recording media storing programs for performing the methods according to the present invention.

[0086] It should also be noted that in the systems, apparatus, and methods of the present invention, the components or steps can be disassembled and / or recombined. These disassemblies and / or recombinations should be considered equivalents of the present invention. The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the invention. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of the invention. Therefore, the invention is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0087] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the invention to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A method for calculating the carbon emission factor of a power distribution network based on graph convolutional network learning, characterized in that, include: Based on the extracted power distribution system topology connections and the measurement data of each node, an input matrix representing the node's electrical carbon information and an adjacency matrix representing the topology connections are constructed. Based on the defined Laplace matrix of the spectral domain graph convolution network, construct the graph convolution operation formula; The carbon emission factor calculation model of the distribution network is constructed according to the graph convolution operation formula, and the carbon emission factor calculation model of the distribution network is trained according to the input matrix and the adjacency matrix. Based on the carbon emission factor calculation model, a mapping relationship between the carbon emission factor and the input electrical carbon information is established; Based on the mapping relationship, the carbon emission factor values ​​of each node in the distribution network at the measurement time are calculated. The carbon emission factor calculation model for the power distribution network includes an input layer, a feature extraction layer, and an output layer, wherein... The input layer consists of an input matrix and an adjacency matrix, where the node injection power represents the node electrical information, the unit carbon emission intensity represents the node carbon information, and the dynamic carbon emission factor represents the non-source side node carbon information. The feature extraction layer consists of two-layer graph convolutional networks. Each graph convolutional network takes the output of the previous layer as input and extracts features through the graph convolution operation formula. The output layer consists of a fully connected layer that maps the data output by the feature extraction layer to the output space.

2. The method according to claim 1, characterized in that, Based on the extracted distribution system topology connections and the measurement data of each node, an input matrix representing the node electrocarbon information and an adjacency matrix representing the topological relationships are constructed, including: Extract the topological connections of the power distribution system to obtain its graph-structured network. The input matrix of node carbon information and the adjacency matrix are constructed based on the measurement data of each node in the graph structure network. The measurement data of each node includes unit active power, load active power, unit carbon emission data and unit power generation data.

3. The method according to claim 2, characterized in that, The expression for the input matrix is: In the formula, ; X The matrix formed by the carbon information vectors of each node; W i Let i be the carbon information vector of node i (1≤i≤N); The matrix formed by the carbon information of the nodes at time t. For m time points The nodal carbon information input matrix is ​​formed by (1≤k≤m); for The transpose of .

4. The method according to claim 1, characterized in that, The formula for graph convolution operation is: In the formula, D It is a diagonal matrix; A It is an adjacency matrix; θ 0 represents the initial Chebyshev coefficient.

5. The method according to claim 1, characterized in that, The loss function of the carbon emission factor calculation model for the power distribution network is: In the formula, It is a vector The second norm; m At that moment; Carbon emission factor vector for: In the formula, for t The GCN model calculates the next node at time step. i The carbon emission factor value, calculated based on GCN, requires the carbon flow calculation result as the target value, assuming... t Carbon emission factor vector under time-varying carbon flow calculation for: In the formula, for t Time-dependent carbon flow calculation at the next node i Carbon emission factor value.

6. The method according to claim 1, characterized in that, The mapping function that constructs the mapping relationship is: In the formula, For the input matrix, A It is an adjacency matrix. and b These are trainable parameters. It is a non-linear activation function. For graph convolutional network mapping functions, Y The output of this method is a carbon emission factor. .

7. The method according to claim 1, characterized in that, The evaluation indicators of the carbon emission factor calculation model for the power distribution network include: the average absolute error between the calculated value and the target value, and the average relative error between the calculated value and the target value.

8. A device for calculating carbon emission factors of a power distribution network based on graph convolutional network learning, used to implement the method described in any one of claims 1-7, characterized in that, include: The module is used to construct an input matrix representing the electrical carbon information of nodes and an adjacency matrix representing the topological relationship, based on the extracted power distribution system topology connection relationship and the measurement data of each node. The training module is used to construct a carbon emission factor calculation model for the power distribution network based on the spectral domain graph convolutional network, and to train the carbon emission factor calculation model for the power distribution network according to the input matrix and the adjacency matrix. A module is established to establish a mapping relationship between the carbon emission factor and the input electrical carbon information based on the carbon emission factor calculation model. The calculation module is used to calculate the carbon emission factor values ​​of each node in the distribution network at the measurement time based on the mapping relationship.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program for performing the method described in any one of claims 1-7.

10. An electronic device, characterized in that, The electronic device includes: processor; Memory used to store the processor's executable instructions; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method described in any one of claims 1-7.

Images