Active Distribution Network State Estimation Method and System Based on the Fusion of GCN and Transformer

Abstract

The present invention discloses an active distribution network state estimation method and system based on the fusion of GCN and Transformer. First, the present invention establishes a distribution network graph model, abstracting the nodes in the distribution network and the lines connecting these nodes into a graph structure; secondly, constructs a physical model of a photovoltaic power station, and uses a five-parameter model to describe the relationship between the terminal voltage and current; then constructs a state prediction equation based on Transformer, and uses historical measurement data to predict the system state at a future moment; and constructs a measurement equation based on GCN, and realizes the relationship modeling between the node state and the states of its neighbor nodes through graph convolution operations; finally, uses a deep neural network to perform feature fusion on the state prediction and measurement results, and outputs a more accurate distribution network state estimation. By using GCN for the transformation of measurement data, the present invention fully exploits the spatial dependence between nodes; while Transformer is responsible for the state prediction of long-sequence time dependence, effectively improving the prediction accuracy and robustness of the model.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution networks, and particularly relates to an active distribution network state estimation method and system based on the fusion of GCN and Transformer. Background Art

[0002] The large-scale access of distributed energy, the wide application of flexible loads, and the introduction of emerging technologies such as edge computing have enabled the distribution network to have a higher level of intelligence, while also bringing new challenges. Compared with traditional distribution networks, new active distribution networks exhibit significant characteristics such as high multi-energy penetration, complex dynamic loads, and highly controllable equipment. These characteristics require the power grid to have stronger adaptability and real-time response capabilities.

[0003] In this context, the operating constraints faced by the distribution network are more complex. Problems such as equipment capacity limitations, overload, aging, reverse power flow, voltage / frequency instability, and dual uncertainties of sources and loads all pose threats to the safe operation of the distribution network. At the same time, factors such as sparse data acquisition points, equipment failures, data noise, and communication delays further weaken the accuracy of distribution network state monitoring and limit its precise perception of future operating states. With the increase in the number of controllable devices and the development of the smart grid model, distribution system state estimation (DSSE) plays an increasingly crucial role in the operation, protection, and control of distribution networks.

[0004] DSSE uses power system measurements, such as line flows, node voltages (magnitude and phase angle), and node injections (obtained from the supervisory control and data acquisition SCADA system), to estimate power system states such as the voltage magnitude and phase angle of system nodes. As a core component of the distribution network energy management system, DSSE has attracted much attention. It obtains real-time data through remote metering acquisition terminals and conducts necessary monitoring, simulation, and verification of the real-time operating state of the distribution network.

[0005] In the field of DSSE, scholars have actively explored and made some important progress. In the actual operation of the power grid, the existence of non-Gaussian noise and gross errors may lead to excessive iteration times or even non-convergence, and the estimation accuracy of the traditional weighted least square (WLS) method is relatively low. To solve this problem, some scholars have proposed non-quadratic estimation criteria, such as weighted least absolute value state estimation, Huber-M estimation, and exponential objective function estimation. However, traditional state estimation methods are difficult to fully cope with the dynamic changes of complex loads and data quality problems, and have high computational complexity and long solution times, which can no longer meet the requirements of new active distribution networks.

[0006] Many scholars have introduced artificial intelligence algorithms to improve the efficiency and accuracy of power system state estimation. Compared with traditional methods, the deep neural network (DNN) shows high time efficiency and robustness in power system state estimation. For the state estimation of the transmission network, an online estimation method based on the deep neural network can utilize the high-frequency input of phasor measurement unit data to estimate the observable state of the system.

[0007] The prior art has proposed a Bayesian network method for the unobservable state of the distribution system, where the Bayesian network is a neural network trained offline and used to predict the probability of the unobservable state. The prior art also proposes a method of generating virtual measurements using an offline-trained neural network to estimate the unobservable state through the mixed data of real measurements and virtual measurements. However, the neural networks in these methods are all trained offline, and their accuracy cannot be guaranteed without feedback. The prior art has also proposed a hybrid learning mechanism that incorporates the system model into the backpropagation process of the neural network, but the unobservable state has not been considered yet. The above data-driven methods still do not consider the new type of active distribution network scenario, do not model distributed new energy units, and it is difficult to consider the spatio-temporal characteristics of the distribution network topology and the impact of bidirectional power flow on traditional state estimation methods. Therefore, it is urgent to study a more robust, real-time and accurate distribution network state estimation method to effectively process the measurement data while accurately identifying the physical characteristics and dynamic behaviors of source-load objects. Summary of the Invention

[0008] Aiming at the deficiencies of the prior art, the present invention provides an active distribution network state estimation method and system based on the fusion of Graph Convolutional Neural Networks (GCN) and Transformer.

[0009] In the first aspect, the present invention provides an active distribution network state estimation method based on the fusion of GCN and Transformer, including:

[0010] Step 1: Establish a distribution network graph model, where each node and the lines connecting these nodes in the distribution network are abstracted into a graph structure;

[0011] Step 2: Construct a physical model of a photovoltaic power station, where the physical model of the photovoltaic power station includes a photovoltaic array, a DC / DC converter, and a DC / AC converter, and a five-parameter model is used to describe the relationship between the terminal voltage and current;

[0012] Step 3: Construct a state prediction equation based on Transformer to predict the system state at a future moment using historical measurement data;

[0013] Step 4: Construct a measurement equation based on GCN. Input the topological structure and measurement data of the distribution network into GCN, and model the relationship between the node state and the states of its neighbor nodes through graph convolution operations.

[0014] Step 5: Use a deep neural network to perform feature fusion on the state prediction in Step 3 and the measurement results in Step 4, integrate the spatial features extracted by GCN and the temporal features captured by Transformer, and output a more accurate state estimation of the distribution network.

[0015] In a second aspect, the present invention provides an active distribution network state estimation system based on the fusion of GCN and Transformer, including:

[0016] A graph model establishment unit for establishing a distribution network graph model, where each node and the lines connecting these nodes in the distribution network are abstracted into a graph structure.

[0017] A photovoltaic power station physical model construction unit for constructing a photovoltaic power station physical model, where the photovoltaic power station physical model includes a photovoltaic array, a DC / DC converter, and a DC / AC converter, and a five-parameter model is used to describe the relationship between the terminal voltage and current.

[0018] A Transformer model unit for constructing a state prediction equation based on Transformer and predicting the system state at future moments using historical measurement data.

[0019] A GCN model unit for constructing a measurement equation based on GCN. Input the topological structure and measurement data of the distribution network into GCN, and model the relationship between the node state and the states of its neighbor nodes through graph convolution operations.

[0020] A feature fusion and state estimation unit for using a deep neural network to perform feature fusion on the state prediction and measurement results, integrating the spatial features extracted by GCN and the temporal features captured by Transformer, and outputting a more accurate state estimation of the distribution network.

[0021] In a third aspect, the present invention provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements an active distribution network state estimation method based on the fusion of GCN and Transformer.

[0022] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements an active distribution network state estimation method based on the fusion of GCN and Transformer.

[0023] Advantages of the present invention: While effectively processing complex measurement data and state prediction of the distribution network, the present invention takes into account the spatio-temporal connection between the states of each node in the distribution network. In addition, a physical model of the photovoltaic power station, a distributed energy source, is established, and its dynamic physical characteristics are incorporated into the scope of state estimation. Compared with the traditional weighted WLS state estimation method, the present invention can improve the prediction accuracy, anti-noise ability and dynamic adaptability. Description of the Drawings

[0024] Figure 1 Schematic diagram of electrical quantities for state estimation of active distribution network;

[0025] Figure 2 Schematic diagram of two-stage structure of photovoltaic power generation station;

[0026] Figure 3 GCN-Transformer algorithm framework diagram of the embodiment of the present application;

[0027] Figure 4 Transformer model unit architecture diagram;

[0028] Figure 5 GCN model unit architecture diagram;

[0029] Figure 6 Feature fusion and state estimation unit architecture diagram. Detailed Embodiment

[0030] The following details the specific embodiments of the embodiments of the present invention with reference to the drawings. It should be understood that the specific embodiments described herein are only for illustrating and explaining the embodiments of the present invention, and are not used to limit the embodiments of the present invention.

[0031] The embodiment of the present application provides an active distribution network state estimation method based on the fusion of GCN and Transformer, including:

[0032] For the new active distribution network scenario, the present application first gives the specific modeling of the active distribution network state estimation considering the photovoltaic power station model:

[0033] Build a distribution network graph model, where each node and the lines connecting these nodes in the distribution network are abstracted into a graph structure, providing a basic data structure for subsequent steps, specifically as follows:

[0034] For a distribution network composed of N nodes, it is modeled as a graph , where represents all nodes, represents all lines.

[0035] As Figure 1 shown, for each node Let represent the corresponding complex voltage, be the voltage amplitude, be the voltage phase angle, , and Let , represent the active and reactive power at the sending end, , represent the active and reactive power at the receiving end.

[0036] Distribution System State Estimation (DSSE) aims to calculate a more accurate system state from the usually noisy measurement values . This process involves the dynamic state equation and measurement equation of the system, as shown in the following formulas:

[0037] (1)

[0038] (2)

[0039] where is the state variable of the system at time, including the voltage amplitude and voltage phase angle; is the real-time measurement value at time, including the active and reactive power injection at nodes, the active and reactive power of branches, and the voltage amplitude of nodes; is the state transition function at time; is the state-measurement connection function; is the measurement error. The key to state estimation lies in the processing of the non-linear equation .

[0040] On the other hand, establishing an accurate photovoltaic power generation system model can improve the accuracy and calculation efficiency of the state estimation of active distribution networks.

[0041] Construct a physical model of a photovoltaic power station. The model includes a photovoltaic array, a DC / DC converter, and a DC / AC converter, and uses a five-parameter model to describe the relationship between the terminal voltage and current, providing a theoretical basis for the explicit fitting function.

[0042] Photovoltaic power stations are usually divided into two structures. One is a two-stage structure with a DC / DC converter and a DC / AC converter; the other is a single-stage structure with only a DC / AC converter. Figure 2 shows the two-stage structure, including a photovoltaic array, DC / DC converter andDC / AC Converter

[0043] Specifically, given the light intensity and temperature , the output current is defined as , and the output voltage is U PVS , with the value denoted as . For the duty cycle , it is defined that DC / DC the output voltage and current of the converter are respectively and . is DC / DC and DC / AC the line resistance between the converters. For DC / AC the converter, the modulation ratio is , the input voltage and current are respectively and , the output voltage amplitude and phase angle are respectively and , P ac and Q ac are the active power and reactive power on the AC side respectively. The equivalent impedance is expressed as , and the turns ratio of the transformer is assumed to be 1: . The active power and reactive power from the photovoltaic system to the point of common coupling ( Point of Common Coupling,PCC ) are respectively and . PCC The voltage amplitude and phase angle of are respectively , PCC The active and reactive power injections at are respectively . For the state estimation of a single photovoltaic power station, the state variable .

[0044] Furthermore, the five-parameter model is one of the most commonly used models for photovoltaic arrays. Its mathematical expression is a complex implicit transcendental equation that describes the relationship between the terminal voltage and current. In the embodiments of this application, an explicit fitting function is used to approximately express the voltage-current relationship to approach the accuracy result of the five-parameter model, reducing the computational time-consuming for obtaining the measurement equation and the Jacobian matrix. Its expression is as follows:

[0045] (3)

[0046] This model uses logarithmic terms to simulate the voltage-current relationship of the diode, and the remaining terms are used to simulate the effects of the internal resistance and the changes in light and temperature.

[0047] Furthermore, based on the given five-parameter model of the photovoltaic array, a set of light intensity, temperature, output voltage, and output current are randomly generated within the feasible region, denoted as k , , , and respectively. The corresponding entries in the vectors , , and satisfy the relationships given by the five-parameter model of the photovoltaic array. The fitting parameter vector is denoted as .

[0048] After setting its initial value, , and are substituted into Equation (3) to obtain the estimated voltage vector . Then, the optimal fitting parameter vector is found by solving the following least-squares problem:

[0049] (4)

[0050] Due to the large number of model parameters, an improved differential algorithm that does not depend on the initial value selection is used for fitting, and the algorithm is terminated when all the updated fitting parameters are less than . Therefore, the proposed explicit fitting function greatly simplifies the state estimation calculation while maintaining a high approximation to the widely used five-parameter model.

[0051] As described above, the state variable vector of the active distribution network is set as , where the vector is the voltage magnitude of each node, and the vector is the voltage phase angle of each node. The measurement vector of the distribution network is expressed as , where is the measurement vector of the node voltage magnitude, and are the measurement vectors of the active and reactive powers of the branch respectively, and are the measurement vectors of the active and reactive power injections of the node respectively. Assuming that the i -th node of the distribution network is connected to the k -th photovoltaic power station, the measurement equations of the power injection are as follows:

[0052] (5)

[0053] (6)

[0054] Among them, G ij is the node conductance, B ij is the node susceptance;

[0055] Assume that the active distribution network includes m PV power plants, and its state variable vector should be corrected to , and at the same time the measurement vector should be corrected to .

[0056] In this embodiment, based on Figure 2 , the derivation process of the measurement equation in the PV power plant is as follows:

[0057] (7)

[0058] (8)

[0059] (9)

[0060] (10)

[0061] (11)

[0062] (12)

[0063] (13)

[0064] (14)

[0065] (15)

[0066] (16)

[0067] (17)

[0068] (18)

[0069] (19)

[0070] Among them, the quantity with superscript is the measured value, and the quantity with superscript is the measurement error, is the fitting function of the PV array, R A 、R A-IGBT 、R A-dioThey are the three types of equivalent resistances of the photovoltaic power station respectively. V dc2 、V A-IGBT1 、V A-dio1 、V A-dio2 、V ac It is the equivalent voltage inside the photovoltaic power station model. k A-IGBT2 、k A-dio2 、M pv It is the voltage coefficient. In addition, according to the law of conservation of power, the following five virtual measurement equations can also be obtained:

[0071] (20)

[0072] (21)

[0073] (22)

[0074] (23)

[0075] (24)

[0076] Among them, They are the virtual measurement values in the photovoltaic system respectively, and their values are all 0; R D 、 R D-IGBT 、R D-dio They are the three types of equivalent resistances of the photovoltaic power station respectively. V D-IGBT1 、V D-dio1 、V D-dio2 、V pvs It is the equivalent voltage inside the photovoltaic power station model. k D-IGBT2 、k D-dio2 It is the voltage coefficient.

[0077] Therefore, the available measurement data of the photovoltaic power generation system are:

[0078] (25)

[0079] Among them, They are the virtual measurement values in the photovoltaic system respectively.

[0080] After completing the state estimation modeling of the active distribution network considering the photovoltaic power station model above, the active distribution network state estimation is carried out next:

[0081] In view of the characteristics of high-dynamic load and multi-energy penetration in the active distribution network, traditional state estimation methods have problems such as low computational efficiency, poor flexibility, and insufficient robustness. An embodiment of this application proposes a data-driven state estimation method, which fuses the Transformer prediction model (Transformer model unit 1, see specifically Figure 4 ) with the GCN (GCN model unit 2, see specifically Figure 5 ) to obtain the GCN-Transformer method, replacing the traditional state prediction equation and measurement equation, and mapping the eigenvalue extracted by the first two methods through DNN learning (feature fusion and state estimation unit 3, see specifically Figure 6 ) to the final system state value. The specific algorithm framework is as Figure 3 shown.

[0082] Construct a state prediction equation based on Transformer, and use historical measurement data to predict the system state at future moments. The optimal fitting parameter vector obtained from the least squares problem is used to improve the accuracy of the prediction.

[0083] In a certain example, in the state estimation of the distribution network, the state prediction equation usually jointly infers the system state at future moments based on the power mechanism model and the load curve. However, traditional mechanism model-based methods are difficult to consider the real-time dynamic changes of multiple loads and cannot accurately capture the temporal correlation and long-term dependence of the system state. Therefore, this embodiment adopts the Transformer model to replace the traditional state prediction equation.

[0084] The Transformer model is an efficient model designed for long sequence prediction. Based on the sparse self-attention mechanism, it can extract global temporal dependencies from historical data and effectively capture the long-term dependence characteristics in the operation of the distribution network. In the embodiment of the application, the state prediction process takes historical measurement data as input, where represents the measurement vector at time, such as node voltage, power injection, etc. The above measurement data is processed through the encoder-decoder architecture of the Transformer.

[0085] Specifically, given historical measurement data, the encoder part of the Transformer first converts it into a high-dimensional temporal feature representation. Through the self-attention mechanism, the model can select key features according to the importance weights in the time series. The role of the self-attention mechanism can be expressed as:

[0086] (26)

[0087] Among them, , and are the query matrix, the key matrix, and the value matrix respectively, is the dimension of the key matrix. This mechanism enables the model to assign weights to each historical time step, focusing on the time steps that are most important for predicting the current moment state. Through the multi-head attention mechanism, the Transformer can capture multi-level temporal dependencies simultaneously.

[0088] After obtaining the temporal feature representation, the decoder part of the Transformer predicts the state estimate at future moments based on these features , where is the predicted time step. The predicted state vector includes key system states such as voltage amplitude and phase angle;

[0089] (27)

[0090] Construct a measurement equation based on GCN (Graph Convolutional Network), input the topological structure and measurement data of the distribution network into GCN, and model the relationship between the node state and the states of its neighbor nodes through graph convolution operations to provide spatial features for feature fusion in step 8; specifically:

[0091] In traditional state estimation, the measurement equation uses the non-linear equation (2) to correlate the physical measurement data with the state vector , where represents the physical mapping relationship between the state vector and the measurement vector .

[0092] In the embodiment of this application, GCN is used to characterize the measurement equation, and its input includes the topological structure and measurement data of the distribution network. The injected active and reactive power and node voltage amplitude data in the measurement vector are used as the node set, and the active and reactive power transmitted by the branch and the branch impedance data are used as the edge set. The initial state matrix is expressed as:

[0093] (28)

[0094] Among them, is the number of nodes, and each represents the node set data of node .

[0095] In the GCN model, each node 's state and the state of its neighbor nodes are related through graph convolution operations, which can be expressed as:

[0096] (29)

[0097] where represents the state of node after graph convolution at the -th layer, is the set of neighbor nodes of node , is the weight matrix of the graph convolution layer, is the bias, is a non-linear activation function, such as ReLU. Through such neighborhood aggregation operations, each node can continuously update its state representation, comprehensively considering the measurement information of itself and its neighbor nodes, and finally obtain a high-dimensional representation of the system state .

[0098] Furthermore, to improve the robustness of measurement data processing, the embodiments of this application also introduce an attention mechanism in the GCN, which assigns different weights to different neighbor nodes, so as to focus on the nodes that have a greater impact on state estimation. The attention mechanism is defined as follows:

[0099] (30)

[0100] where represents the attention weight between node and node in the -th layer of graph convolution, is the attention parameter vector. Through such an attention mechanism, the GCN can automatically focus on the measurement data that is most critical for state estimation in the case of noisy or missing measurement data, thereby improving the accuracy and robustness of the estimation.

[0101] Using a deep neural network (DNN) to perform feature fusion on the state prediction and measurement results, integrating the spatial features extracted by the GCN and the temporal features captured by the Transformer, and outputting a more accurate distribution network state estimation, where the enhanced robustness in the previous step ensures the accuracy of the final state estimation.

[0102] Specifically, assuming that the output of the GCN is , representing the high-dimensional spatial features of the nodes, and the output of the Transformer is , reflecting the temporal estimation features of the nodes, the concatenated joint feature vector can be expressed as:

[0103] (31)

[0104] This combined feature vector is input into the DNN for further processing.

[0105] The DNN consists of several fully connected layers. By learning the relationships between different features, these high-dimensional features are gradually mapped to the final state estimation. Each layer of the DNN calculates a linear combination through a weight matrix and a bias vector and performs a non-linear transformation using a non-linear activation function. Through this layer-by-layer mapping, the DNN can comprehensively process the spatial information from the GCN and the temporal information from the Transformer, and gradually learn the complex non-linear relationship between the two.

[0106] The calculation form of the output layer is:

[0107] (32)

[0108] where is the final estimation result, is the output result of the th layer, is the weight matrix of the th layer, is the bias vector of the th layer.

[0109] Through the fusion of deep features and the model, the DNN can automatically learn and extract the dynamic temporal features of the Transformer and the spatio-temporal features output by the GCN, and has stronger robustness when dealing with incomplete or noisy data.

[0110] The following is a system embodiment of the present invention, which can be used to execute the method embodiment. For the details not disclosed in this embodiment, please refer to the method embodiment.

[0111] In this embodiment, an active distribution network state estimation system based on the fusion of GCN and Transformer is provided, including:

[0112] A graph model building unit for building a distribution network graph model, where each node and the lines connecting these nodes in the distribution network are abstracted into a graph structure;

[0113] A photovoltaic power station physical model construction unit for constructing a photovoltaic power station physical model, the photovoltaic power station physical model including a photovoltaic array, a DC / DC converter, and a DC / AC converter, and using a five-parameter model to describe the relationship between the terminal voltage and the current;

[0114] The Transformer model unit is used to construct a Transformer-based state prediction equation and predict the system state at future moments using historical measurement data;

[0115] The GCN model unit is used to construct a GCN-based measurement equation. Input the topological structure and measurement data of the distribution network into the GCN, and model the relationship between the node state and the states of its neighbor nodes through graph convolution operations;

[0116] The feature fusion and state estimation unit is used to fuse the state prediction and measurement results using a deep neural network, integrate the spatial features extracted by the GCN and the temporal features captured by the Transformer, and output a more accurate state estimation of the distribution network.

[0117] Similarly, based on the concept of the method embodiment, the embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the active distribution network state estimation method based on the fusion of GCN and Transformer.

[0118] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding function.

[0119] Similarly, based on the concept of the method embodiment, the embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, it implements the active distribution network state estimation method based on the fusion of GCN and Transformer.

[0120] It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and, of course, the extended storage medium supported by the computer device. The computer-readable storage medium provides storage space, and this storage space stores the operating system of the terminal. Moreover, one or more instructions suitable for being loaded and executed by the processor are stored in this storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the active distribution network state estimation method based on the fusion of GCN and Transformer in the above embodiments.

[0121] In summary, this application proposes a new state estimation method based on the combination of GCN and Transformer models to solve the problems of dynamic load, complex topology, and data quality in the new active distribution network. By using GCN for the transformation of measurement data, the spatial dependence between nodes and neighbors is fully exploited; while Transformer is responsible for the state prediction of long-sequence time dependence, effectively improving the prediction accuracy and robustness of the model. It is significantly superior to the traditional state estimation method in terms of accuracy, anti-interference ability, and computational efficiency, and this method can be applied to new power system applications with more voltage levels.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not 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: the specific implementation manners of the present invention can still be modified or equivalently replaced, and any modification or equivalent replacement without departing from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. Active distribution network state estimation method based on GCN and Transformer fusion, characterized by: Step 1: Establish a distribution network graph model, in which each node in the distribution network and the lines connecting these nodes are abstracted into a graph structure; Step 2: construct a photovoltaic power station physical model, which includes a photovoltaic array, a DC / DC converter and a DC / AC converter, and uses a five-parameter model to describe the relationship between terminal voltage and current; specifically: Given the light intensity G and temperature T, the output current is defined as I pvs , the output voltage is U PVS , the value is recorded as V pvs ; For duty cycle D pv , define the output voltage and current of the DC / DC converter as V dc1 and I dc1 , R dc is the line resistance between the DC / DC and DC / AC converters; for the DC / AC converter, the modulation ratio is M pv , the input voltage and current are V dc2 and I dc2 , the output voltage amplitude and phase angle are V ac and θ ac , P ac and Q ac are the active power and reactive power on the AC side respectively; the equivalent impedance is expressed as X pv , the transformer ratio is assumed to be 1:K pv ; The active power and reactive power from the photovoltaic system to the common coupling point are P g and Q g ; The voltage amplitude and phase angle of PCC are V pvac and θ pvac , the active and reactive power injection at PCC are P pvi and Q pvi ; For the state estimation of a single PV power station, the state variable x pv =(G,T,I pvs ,V dc1 ,V dc2 ,V ac ,θ ac ) T ; Step 3: Construct a state prediction equation based on Transformer and use historical measurement data to predict the system state at future times; Step 4: Construct a measurement equation based on GCN, input the topological structure and measurement data of the distribution network into GCN, and model the relationship between the node state and the state of its neighboring nodes through graph convolution operation; Step 5: Use a deep neural network to fuse the state prediction in step 3 and the measurement results in step 4, integrate the spatial features extracted by GCN and the temporal features captured by Transformer, and output a more accurate distribution network state estimate.

2. The active distribution network state estimation method based on GCN and Transformer fusion according to claim 1 is characterized by: In step 1, the electrical properties of each node are determined, including voltage amplitude, phase angle, active power and reactive power; the electrical properties of each edge are determined, including line impedance and line flow.

3. The active distribution network state estimation method based on GCN and Transformer fusion according to claim 1 is characterized by: Step 2 also includes approximating the voltage-current relationship of the photovoltaic array using an explicit fitting function.

4. The active distribution network state estimation method based on GCN and Transformer fusion according to claim 3 is characterized by: Based on the five-parameter model, multiple sets of light intensity, temperature, output voltage and output current data are randomly generated in the feasible region; the optimal fitting parameter vector is found by solving the least squares problem to improve the accuracy of state prediction.

5. The method for estimating the state of an active distribution network based on the fusion of GCN and Transformer according to claim 1 is characterized in that: In step 3, the encoder-decoder architecture of the Transformer model is used to process historical measurement data and extract global time dependencies; Multi-level temporal dependencies are captured through self-attention mechanism and multi-head attention mechanism.

6. The method for estimating the state of an active distribution network based on the fusion of GCN and Transformer according to claim 5 is characterized in that: The attention mechanism is introduced into GCN to dynamically adjust the weights according to the correlation between nodes, improve the robustness of measurement data processing, and ensure that the feature fusion in step 5 can focus on key measurement data when there is noise or missing data.

7. The method for estimating the state of an active distribution network based on the fusion of GCN and Transformer according to claim 5 is characterized in that: Step 5 is as follows: Use DNN to effectively integrate the spatial features extracted by GCN and the temporal features captured by Transformer to output a more accurate distribution network status; The fully connected layers of DNN learn the relationship between different features and gradually map these high-dimensional features to the final state estimation.

8. An active distribution network state estimation system based on the fusion of GCN and Transformer, used to implement the method described in claim 1, characterized in that: include: A graph model building unit, used to build a distribution network graph model, in which each node in the distribution network and the lines connecting these nodes are abstracted into a graph structure; A photovoltaic power station physical model building unit is used to build a photovoltaic power station physical model, wherein the photovoltaic power station physical model includes a photovoltaic array, a DC / DC converter and a DC / AC converter, and uses a five-parameter model to describe the relationship between terminal voltage and current; Transformer model unit, used to construct a Transformer-based state prediction equation and use historical measurement data to predict the system state at future times; The GCN model unit is used to construct the measurement equation based on GCN, input the topological structure and measurement data of the distribution network into GCN, and realize the relationship modeling between the node state and the state of its neighboring nodes through graph convolution operation; The feature fusion and state estimation unit is used to use deep neural networks to fuse the state prediction and measurement results, integrate the spatial features extracted by GCN and the temporal features captured by Transformer, and output a more accurate distribution network state estimate.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the active distribution network state estimation method based on GCN and Transformer fusion according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the active distribution network state estimation method based on GCN and Transformer fusion according to any one of claims 1 to 7 is implemented.

Images