Distributed photovoltaic cluster clustering method and device based on multi-dimensional dynamic indexes

Abstract

The invention discloses a distributed photovoltaic cluster clustering method and device based on a multi-dimensional dynamic index, and belongs to the technical field of distributed photovoltaic, and the method comprises the steps: obtaining the related electrical information of a power distribution network node, calculating the feature data according to the related electrical information, and carrying out the clustering of a distributed photovoltaic cluster; the characteristic data comprises a reactive power-voltage sensitivity matrix, photovoltaic active power output prediction and photovoltaic reactive power output prediction; establishing evaluation indexes according to the feature data, wherein the evaluation indexes comprise cluster modularity, a power generation capability factor, an equivalent electrical distance and resource redundancy; performing normalization processing on the evaluation indexes, and distributing weights through a composite weighting method to generate comprehensive evaluation indexes; and carrying out photovoltaic cluster division through a Fast-Unfolding algorithm, and carrying out iterative optimization in combination with the comprehensive evaluation index to obtain an optimal photovoltaic cluster result. According to the method, the reasonability and the high efficiency of a cluster division result can be improved, and a solid foundation is provided for intra-cluster autonomous and inter-cluster coordinated control scheduling tasks of subsequent clusters.

Description

Technical Field

[0001] This invention relates to the field of distributed photovoltaic technology, and in particular to a method and apparatus for clustering distributed photovoltaic clusters based on multidimensional dynamic indicators. Background Technology

[0002] With the rapid development of renewable energy, distributed photovoltaic (PV) systems have been widely integrated into power grids. Unlike traditional centralized power generation, PV resources are scattered and have diverse connection points. Their output is highly dependent on weather factors, exhibiting strong volatility and uncertainty, which poses a severe challenge to the reliability and economic operation of the power grid. Therefore, it is necessary to implement scientific aggregate management of PV units and adopt a multi-dimensional dynamic evaluation mechanism to achieve adaptive cluster partitioning, thereby improving the efficiency of overall scheduling and optimization.

[0003] Existing distribution network partitioning schemes are mostly based on fixed topology models and classical optimization frameworks, such as genetic algorithms, nonlinear programming, or distance-based clustering techniques. These methods often assume that system parameters are relatively constant or only vary within a small range. However, under the reality of high penetration rates of distributed photovoltaic (PV) power, the power grid faces significant dynamic uncertainties. Both PV output and load demand exhibit high spatiotemporal variability, posing complex challenges to voltage regulation and power balance. In this context, traditional methods show obvious shortcomings in addressing such problems, specifically including: 1) Overemphasizing electrical distance between nodes while ignoring insufficient power complementarity caused by environmental factors; 2) A relatively simplistic evaluation index system that fails to fully integrate the coordination potential of resource utilization and regulation capabilities; and 3) The lack of time-varying adaptation mechanisms in clustering algorithms leads to poor applicability of partitioning results in real-world dynamic scenarios, making it unable to effectively support production applications. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a distributed photovoltaic clustering method and device based on multi-dimensional dynamic indicators, which can improve the rationality and efficiency of cluster partitioning results and provide a solid foundation for subsequent cluster control and scheduling tasks such as intra-cluster autonomy and inter-cluster coordination.

[0005] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0006] In a first aspect, the present invention provides a distributed photovoltaic clustering method based on multidimensional dynamic indicators, comprising:

[0007] Obtain relevant electrical information of distribution network nodes, and calculate feature data based on the relevant electrical information. The feature data includes reactive power-voltage sensitivity matrix, photovoltaic active power output prediction, and photovoltaic reactive power output prediction.

[0008] Based on the aforementioned feature data, evaluation indicators are constructed, including cluster modularity, power generation capacity factor, equivalent electrical distance, and resource redundancy.

[0009] The evaluation indicators are normalized and weights are assigned using a composite weighting method to generate comprehensive evaluation indicators.

[0010] The photovoltaic cluster is partitioned using an incremental Fast-Unfolding algorithm, and then iteratively optimized using the comprehensive evaluation index to obtain the optimal photovoltaic cluster result.

[0011] Optionally, the calculation of feature data based on the relevant electrical information includes:

[0012] The reactive power-voltage sensitivity matrix of distribution network nodes is calculated using the Jacobian matrix analysis method. :

[0013]

[0014] In the formula, These are the changes in voltage and reactive power.

[0015] Photovoltaic active power output prediction based on irradiance and temperature. :

[0016]

[0017] In the formula, Irradiance, For temperature, The rated power of photovoltaic power under standard test conditions. For standard test irradiance, For standard test temperature, The power temperature coefficient;

[0018] Photovoltaic reactive power output prediction based on changes in the target power factor. :

[0019]

[0020] In the formula, For the target power factor, The power factor is marked with a symbol. Indicates the lag power factor / inductive reactive power output. This indicates the leading power factor / capacitive reactive power output.

[0021] Optionally, the cluster modularity for:

[0022]

[0023] In the formula, Let represent the electrical coupling strength between node i and node j in the distribution network. For time, The normalization coefficient is... For balance coefficient, ; This is a cluster indicator function, which is 1 when node i and node j belong to the same cluster, and 0 otherwise. For the expectation of environmental similarity, For time window index value, These are the weighting coefficients;

[0024] The total number of coupling strength clusters in the distribution network is expressed as:

[0025]

[0026]

[0027] In the formula, For weight parameters, Let be the scene similarity between node i and node j. The sensitivity components of nodes i and j are given by the reactive-voltage sensitivity matrix. The value obtained from the data represents the reactive-voltage interaction strength between node i and node j.

[0028] Environmental similarity expectation The expression is:

[0029]

[0030] In the formula, For time windows The number of days, For nodes i and j in the th... Radiance of the sky For nodes i and j in the th... The average daily temperature of the day; These are the threshold values ​​for irradiance and daily average temperature. When the condition inside the parentheses is satisfied, ,otherwise .

[0031] Optionally, the power generation capacity factor for:

[0032]

[0033] In the formula, For the measured active power output and the predicted active power output of node i. Let be the power generation capacity factor of node i;

[0034] The equivalent electrical distance for:

[0035]

[0036]

[0037] In the formula, For the number of clusters, For the internal electrical compactness of the x-th cluster, Let be the equivalent impedance magnitude between node i and node j in the distribution network;

[0038] resource redundancy for:

[0039]

[0040]

[0041] In the formula, Let i be the actual reactive power output and the maximum reactive power output. Let be the rated capacity of the inverter at node i. Let i be the parallel compensation capacity. Let i be the resource redundancy of node i.

[0042] Optionally, the comprehensive evaluation indicators for:

[0043]

[0044] In the formula, This is a comprehensive evaluation index between node i and node j. Components of cluster modularity The value after normalization The power generation capacity factors of nodes i and j The value after normalization Components of equivalent electrical distance The value after normalization Resource redundancy for nodes i and j The value after normalization; The weights are the indicator weights.

[0045] Optionally, the step of partitioning the photovoltaic cluster using the incremental Fast-Unfolding algorithm and iteratively optimizing it in conjunction with the comprehensive evaluation index includes:

[0046] Step S10: Initialize each node to form its own photovoltaic cluster and calculate the comprehensive evaluation index;

[0047] Step S20: Use the comprehensive evaluation index as the weight of the edges between nodes to construct a time-varying graph;

[0048] Step S30: Randomly select nodes in the time-varying graph and calculate the cluster module degree increment caused by a node moving to the photovoltaic cluster where its neighboring node is located. ;

[0049] Step S40: If the cluster module degree increment The maximum value in If the value is greater than 0, the node will be moved to the corresponding photovoltaic cluster, and the newly generated photovoltaic cluster will be treated as a node.

[0050] Step S50, repeat steps S10-S40 until the cluster module degree is reached. convergence;

[0051] If the comprehensive evaluation index in step S20 is greater than the index threshold, or the maximum value in step S40 is... If the value is less than or equal to 0, then re-clustering is triggered.

[0052] Optionally, after obtaining the optimal photovoltaic cluster result for the current time window, the cluster modularity for the current time window can be calculated.

[0053] If the increase in cluster modularity relative to the historical best cluster modularity in the current time window exceeds the increment threshold, then the photovoltaic cluster label of the node will be smoothed.

[0054]

[0055] In the formula, Let i be the smoothed photovoltaic cluster label within time window q and the current photovoltaic cluster label. For smoothing coefficients, Let i be the smoothed photovoltaic cluster label for node i within the time window q-1.

[0056] Secondly, the present invention provides a distributed photovoltaic clustering device based on multidimensional dynamic indicators, comprising:

[0057] The data processing module is configured to acquire relevant electrical information of the distribution network nodes and calculate feature data based on the relevant electrical information. The feature data includes a reactive power-voltage sensitivity matrix, photovoltaic active power output prediction, and photovoltaic reactive power output prediction.

[0058] The indicator calculation module is configured to construct evaluation indicators based on the feature data, the evaluation indicators including cluster modularity, power generation capacity factor, equivalent electrical distance and resource redundancy.

[0059] The indicator integration module is configured to normalize the evaluation indicators and allocate weights through a composite weighting method to generate comprehensive evaluation indicators.

[0060] The clustering module is configured to divide the photovoltaic clusters using an incremental Fast-Unfolding algorithm, and then iteratively optimize the results by combining the comprehensive evaluation index to obtain the optimal photovoltaic cluster result.

[0061] Thirdly, the present invention provides an electronic device, including a processor and a storage medium;

[0062] The storage medium is used to store instructions;

[0063] The processor is configured to operate according to the instructions to perform the steps according to the method described above.

[0064] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0065] Fifthly, the present invention provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0066] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0067] This invention provides a distributed photovoltaic (PV) clustering method and apparatus based on multi-dimensional dynamic indicators. By introducing a composite weighting method to comprehensively consider multi-dimensional dynamic characteristics such as time-varying modularity, power generation capacity, electrical distance, and resource redundancy, it overcomes the insufficient adaptability of traditional static partitioning methods to environmental changes and operational fluctuations, achieving accurate characterization of the dynamic similarity of distributed PV units. Through time smoothing and threshold triggering mechanisms, it effectively balances the stability and sensitivity of the partitioning results, suppressing frequent oscillations in the cluster structure caused by random weather fluctuations, and ensuring the engineering practicality of the partitioning scheme. It employs an incremental Fast-Unfolding algorithm for iterative optimization, achieving rapid convergence based on inheriting historical partitioning results. This significantly improves the partitioning efficiency and real-time computation of large-scale distributed PV clusters in dynamic scenarios, providing reliable support for subsequent accurate intra-cluster autonomy and inter-cluster coordinated control. Attached Figure Description

[0068] Figure 1 This is a flowchart illustrating the distributed photovoltaic clustering method based on multidimensional dynamic indicators provided in this embodiment of the invention. Detailed Implementation

[0069] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0070] Example 1

[0071] like Figure 1 As shown, this embodiment of the invention provides a distributed photovoltaic clustering method based on multi-dimensional dynamic indicators, including the following steps:

[0072] Step S1: Obtain relevant electrical information of the distribution network nodes, and calculate characteristic data based on the relevant electrical information. The characteristic data includes reactive power-voltage sensitivity matrix, photovoltaic active power output prediction, and photovoltaic reactive power output prediction.

[0073] (1) The characteristic data calculated based on relevant electrical information includes:

[0074] The reactive power-voltage sensitivity matrix of distribution network nodes is calculated using the Jacobian matrix analysis method. :

[0075]

[0076] In the formula, These are the changes in voltage and reactive power.

[0077] The specific acquisition steps are as follows: Obtain real-time data on voltage amplitude, phase angle, and active and reactive power for each node from the SCADA system, and establish the power flow Jacobian matrix accordingly. This matrix is ​​divided into blocks representing the partial derivatives of active power with respect to phase angle, active power with respect to voltage amplitude, reactive power with respect to phase angle, and reactive power with respect to voltage amplitude, denoted as matrices H, N, M, and L, respectively. Ignoring the impact of active power disturbances on voltage, the sensitivity matrix... It can be calculated using the following formula:

[0078]

[0079] This matrix reflects the degree to which changes in reactive power at nodes affect voltage amplitude.

[0080] (2) Calculate photovoltaic active power output prediction based on irradiance and temperature :

[0081]

[0082] In the formula, Irradiance, For temperature, The rated power of photovoltaic power under standard test conditions. For standard test irradiance, For standard test temperature, This is the power temperature coefficient.

[0083] (3) Calculate the photovoltaic reactive power output prediction based on the change in the target power factor. :

[0084]

[0085] In the formula, For the target power factor, The power factor is marked with a symbol. Indicates the lag power factor / inductive reactive power output. This indicates the leading power factor / capacitive reactive power output.

[0086] The specific steps are as follows: Import the network topology from the Distribution Automation System (DAS): node number, branch connection, and branch parameters. Import load data: typical daily load curves for each node, and integrate the predicted output of each photovoltaic access node.

[0087] For a system with N nodes, the power flow balance equation is: Simulate the operating state of each node under minimum load conditions, perform power flow calculations, and obtain the corresponding reactive power demand. First, establish the power flow calculation model using the node admittance matrix method:

[0088]

[0089]

[0090] in, and These are the predicted active and reactive power at node i, respectively. and These are the voltage magnitudes at nodes i and j, respectively. and These are the real and imaginary parts of the nodal admittance matrix, respectively. It is the difference in voltage phase angle between nodes i and j.

[0091] Step S2: Construct evaluation indicators based on feature data. The evaluation indicators include cluster modularity, power generation capacity factor, equivalent electrical distance, and resource redundancy.

[0092] (1) Cluster modularity for:

[0093]

[0094] In the formula, Let represent the electrical coupling strength between node i and node j in the distribution network. For time, The normalization coefficient is... For balance coefficient, ; This is a cluster indicator function, which is 1 when node i and node j belong to the same cluster, and 0 otherwise. For the expectation of environmental similarity, For time window index value, These are the weighting coefficients;

[0095] The total number of coupling strength clusters in the distribution network is expressed as:

[0096]

[0097]

[0098] In the formula, For weight parameters, Let be the scene similarity between node i and node j. The sensitivity components of nodes i and j are given by the reactive-voltage sensitivity matrix. The value obtained from the data represents the reactive-voltage interaction strength between node i and node j.

[0099] Environmental similarity expectation The expression is:

[0100]

[0101] In the formula, For time windows The number of days, For nodes i and j in the th... Radiance of the sky For nodes i and j in the th... The average daily temperature of the day; These are the threshold values ​​for irradiance and daily average temperature. When the condition inside the parentheses is satisfied, ,otherwise .

[0102] (2) Power generation capacity factor for:

[0103]

[0104] In the formula, For the measured active power output and the predicted active power output of node i. Let be the power generation capacity factor of node i.

[0105] (3) Equivalent electrical distance for:

[0106]

[0107]

[0108] In the formula, For the number of clusters, For the internal electrical compactness of the x-th cluster, Let be the equivalent impedance magnitude between node i and node j in the distribution network.

[0109] (4) Resource redundancy for:

[0110]

[0111]

[0112] In the formula, Let i be the actual reactive power output and the maximum reactive power output. Let be the rated capacity of the inverter at node i. Let i be the parallel compensation capacity. Let i be the resource redundancy of node i.

[0113] Step S3: Normalize the evaluation indicators and assign weights using a composite weighting method to generate comprehensive evaluation indicators.

[0114] Cluster modularity, power generation capacity factor, and resource redundancy are positive indicators, while equivalent electrical distance is a negative indicator. During normalization, their directions are reversed.

[0115] Comprehensive evaluation indicators for:

[0116]

[0117] In the formula, This is a comprehensive evaluation index between node i and node j. Components of cluster modularity The value after normalization The power generation capacity factors of nodes i and j The value after normalization Components of equivalent electrical distance The value after normalization Resource redundancy for nodes i and j The value after normalization; The weights are the indicator weights.

[0118] Step S4: Divide the photovoltaic clusters using the incremental Fast-Unfolding algorithm, and iteratively optimize the results by combining comprehensive evaluation indicators to obtain the optimal photovoltaic cluster results.

[0119] The specific iterative optimization process includes:

[0120] Step S10: Initialize each node to form its own photovoltaic cluster and calculate the comprehensive evaluation index;

[0121] Step S20: Use the comprehensive evaluation index as the weight of the edges between nodes to construct a time-varying graph;

[0122] Step S30: Randomly select nodes in the time-varying graph and calculate the cluster module degree increment caused by a node moving to the photovoltaic cluster where its neighboring node is located. ;

[0123] Step S40: If the cluster module degree increments The maximum value in If the value is greater than 0, the node will be moved to the corresponding photovoltaic cluster, and the newly generated photovoltaic cluster will be treated as a node.

[0124] Step S50, repeat steps S10-S40 until the cluster module degree is reached. convergence;

[0125] If the comprehensive evaluation index in step S20 is greater than the index threshold, or the maximum value in step S40 is... If the value is less than or equal to 0, then re-clustering is triggered.

[0126] Furthermore, during the iterative optimization process, time smoothing and threshold triggering mechanisms are introduced, including:

[0127] After obtaining the optimal photovoltaic cluster result for the current time window, calculate the cluster modularity for the current time window;

[0128] If the increase in cluster modularity relative to the historical best cluster modularity in the current time window exceeds the increment threshold, then the photovoltaic cluster label of the node will be smoothed.

[0129]

[0130] In the formula, Let i be the smoothed photovoltaic cluster label within time window q and the current photovoltaic cluster label. For smoothing coefficients, Let i be the smoothed photovoltaic cluster label for node i within the time window q-1.

[0131] This smoothing mechanism ensures that nodes only perform "switching" operations when they exhibit a stable clustering tendency over multiple consecutive time windows, thereby ensuring topological stability in dynamic environments.

[0132] Example 2

[0133] This invention provides a distributed photovoltaic clustering device based on multidimensional dynamic indicators, comprising:

[0134] The data processing module is configured to acquire relevant electrical information of the distribution network nodes and calculate characteristic data based on the relevant electrical information. The characteristic data includes reactive power-voltage sensitivity matrix, photovoltaic active power output prediction and photovoltaic reactive power output prediction.

[0135] The indicator calculation module is configured to construct evaluation indicators based on feature data. The evaluation indicators include cluster modularity, power generation capacity factor, equivalent electrical distance, and resource redundancy.

[0136] The indicator integration module is configured to normalize the evaluation indicators and assign weights through a composite weighting method to generate comprehensive evaluation indicators.

[0137] The clustering module is configured to divide photovoltaic clusters using an incremental Fast-Unfolding algorithm, and then iteratively optimize the results by combining comprehensive evaluation indicators to obtain the optimal photovoltaic cluster results.

[0138] Example 3

[0139] Based on the distributed photovoltaic clustering method provided in Embodiment 1, this embodiment of the invention provides an electronic device, including a processor and a storage medium;

[0140] Storage media are used to store instructions;

[0141] The processor is used to perform operations according to instructions to execute the steps according to the method described above.

[0142] Example 4

[0143] Based on the distributed photovoltaic clustering method provided in Embodiment 1, this embodiment of the invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above method.

[0144] Example 5

[0145] Based on the distributed photovoltaic clustering method provided in Embodiment 1, this embodiment of the invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the above method.

[0146] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0148] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A distributed photovoltaic clustering method based on multidimensional dynamic indicators, characterized in that, include: Obtain relevant electrical information of distribution network nodes, and calculate feature data based on the relevant electrical information. The feature data includes reactive power-voltage sensitivity matrix, photovoltaic active power output prediction, and photovoltaic reactive power output prediction. Based on the aforementioned feature data, evaluation indicators are constructed, including cluster modularity, power generation capacity factor, equivalent electrical distance, and resource redundancy. The evaluation indicators are normalized and weights are assigned using a composite weighting method to generate comprehensive evaluation indicators. The photovoltaic cluster is partitioned using an incremental Fast-Unfolding algorithm, and then iteratively optimized using the comprehensive evaluation index to obtain the optimal photovoltaic cluster result.

2. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, The calculation of feature data based on the relevant electrical information includes: The reactive power-voltage sensitivity matrix of distribution network nodes is calculated using the Jacobian matrix analysis method. : In the formula, These are the changes in voltage and reactive power. Photovoltaic active power output prediction based on irradiance and temperature. : In the formula, Irradiance, For temperature, The rated power of photovoltaic power under standard test conditions. For standard test irradiance, For standard test temperature, The power temperature coefficient; Photovoltaic reactive power output prediction based on changes in the target power factor. : In the formula, For the target power factor, The power factor is marked with a symbol. Indicates the lag power factor / inductive reactive power output. This indicates the leading power factor / capacitive reactive power output.

3. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, The cluster modularity for: In the formula, Let represent the electrical coupling strength between node i and node j in the distribution network. For time, The normalization coefficient is... For balance coefficient, ; This is a cluster indicator function, which is 1 when node i and node j belong to the same cluster, and 0 otherwise. For environmental similarity expectations, For time window index value, These are the weighting coefficients; The total number of coupling strength clusters in the distribution network is expressed as: In the formula, For weight parameters, Let be the scene similarity between node i and node j. The sensitivity components of nodes i and j are given by the reactive-voltage sensitivity matrix. The value obtained from the data represents the reactive-voltage interaction strength between node i and node j. Environmental similarity expectation The expression is: In the formula, For time windows The number of days, For nodes i and j in the th... The day's irradiance, For nodes i and j in the th... The average daily temperature of the day; These are the threshold values ​​for irradiance and daily average temperature. When the condition inside the parentheses is satisfied, ,otherwise .

4. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, The power generation capacity factor for: In the formula, For the measured active power output and the predicted active power output of node i. Let be the power generation capacity factor of node i; The equivalent electrical distance for: In the formula, For the number of clusters, For the internal electrical compactness of the x-th cluster, Let be the equivalent impedance magnitude between node i and node j in the distribution network; resource redundancy for: In the formula, Let i be the actual reactive power output and the maximum reactive power output. Let be the rated capacity of the inverter at node i. Let i be the parallel compensation capacity. Let i be the resource redundancy of node i.

5. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, The comprehensive evaluation indicators for: In the formula, This is a comprehensive evaluation index between node i and node j. Components of cluster modularity The value after normalization The power generation capacity factors of nodes i and j The value after normalization Components of equivalent electrical distance The value after normalization Resource redundancy for nodes i and j The value after normalization; The weights are the indicator weights.

6. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, The process of partitioning photovoltaic clusters using the incremental Fast-Unfolding algorithm, combined with iterative optimization based on the comprehensive evaluation indicators, includes: Step S10: Initialize each node to form its own photovoltaic cluster and calculate the comprehensive evaluation index; Step S20: Use the comprehensive evaluation index as the weight of the edges between nodes to construct a time-varying graph; Step S30: Randomly select nodes in the time-varying graph and calculate the cluster module degree increment caused by a node moving to the photovoltaic cluster where its neighboring node is located. ; Step S40: If the cluster module degree increment The maximum value in If the value is greater than 0, the node will be moved to the corresponding photovoltaic cluster, and the newly generated photovoltaic cluster will be treated as a node. Step S50, repeat steps S10-S40 until the cluster module degree is reached. convergence; If the comprehensive evaluation index in step S20 is greater than the index threshold, or the maximum value in step S40 is... If the value is less than or equal to 0, then re-clustering is triggered.

7. The distributed photovoltaic clustering method based on multidimensional dynamic indicators according to claim 1, characterized in that, After obtaining the optimal photovoltaic cluster result for the current time window, the cluster modularity of the current time window is calculated; If the increase in cluster modularity relative to the historical best cluster modularity in the current time window exceeds the increment threshold, then the photovoltaic cluster label of the node will be smoothed. In the formula, Let i be the smoothed photovoltaic cluster label within time window q and the current photovoltaic cluster label. For smoothing coefficients, Let i be the smoothed photovoltaic cluster label for node i within the time window q-1.

8. A distributed photovoltaic clustering device based on multidimensional dynamic indicators, characterized in that, include: The data processing module is configured to acquire relevant electrical information of the distribution network nodes and calculate feature data based on the relevant electrical information. The feature data includes a reactive power-voltage sensitivity matrix, photovoltaic active power output prediction, and photovoltaic reactive power output prediction. The indicator calculation module is configured to construct evaluation indicators based on the feature data, the evaluation indicators including cluster modularity, power generation capacity factor, equivalent electrical distance and resource redundancy. The indicator integration module is configured to normalize the evaluation indicators and allocate weights through a composite weighting method to generate comprehensive evaluation indicators. The clustering module is configured to divide the photovoltaic clusters using an incremental Fast-Unfolding algorithm, and then iteratively optimize the results by combining the comprehensive evaluation index to obtain the optimal photovoltaic cluster result.

9. An electronic device, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-7.

Images