Power distribution network reactive voltage control method based on cooperative game

Abstract

The invention provides a cooperative game-based reactive voltage control method for a power distribution network. A continuous period cluster is formed by dynamically aggregating operation periods in a day-ahead time scale, an initial node partition is generated in each cluster by combining the influence of photovoltaic reactive output on node voltage, a partition strategy is further optimized by using a cooperative game model, and finally, a dominant control node is selected in a real-time scale to realize accurate regulation and control of voltage fluctuation. Through dynamic time period cluster division, the system can adapt to the time-varying characteristics of photovoltaic output, and the flexibility of the time dimension is improved; a cooperative game model optimizes a partition strategy, so that mutual interference of reactive power regulation between partitions is effectively reduced, and the global voltage control cost is reduced; according to the real-time dominant node control mechanism, key adjusting nodes are accurately positioned through sensitivity analysis, voltage fluctuation and out-of-limit problems are quickly responded in combination with reactive tracking and a cooperative control strategy, and the stability, response efficiency and robustness of voltage control of the power distribution network are improved.

Description

Technical Field

[0001] The present application relates to the field of electrical control, and in particular to a reactive power voltage control method for a distribution network based on cooperative game. Background Art

[0002] The proportion of photovoltaic power generation in modern distribution networks continues to increase annually, and traditional distribution network voltage control methods have shown significant limitations in these scenarios. Centralized control models rely on global optimization at the master station. As the number of photovoltaic nodes increases, they face significant communication delays and heavy computational burdens. Furthermore, centralized architectures present a single point of failure risk; a master station failure can lead to system control failure.

[0003] Therefore, a distributed control method can be used to solve the corresponding problems, but it also has limitations. In some feasible implementations, the partitioning strategy based on static voltage sensitivity is difficult to adapt to the time-varying characteristics of photovoltaic output. Although the dynamic partitioning method improves adaptability, it does not fully consider the coupling effect of voltage control between partitions, resulting in a strong control interaction. In addition, existing distributed algorithms mostly use fixed control parameters, which can easily cause control oscillations in high penetration scenarios and affect system stability. If cooperative games are introduced for control, traditional game solving methods (such as Nash equilibrium calculations) are highly complex. Although simplified algorithms improve computational efficiency, they may sacrifice the partitioning synergy effect, making it impossible to balance efficiency and effectiveness.

[0004] Therefore, a distribution network partition control method based on cooperative game is needed to take into account control effectiveness, response efficiency and robustness. Summary of the Invention

[0005] The purpose of this application is to solve at least one of the above-mentioned technical deficiencies, especially the technical deficiencies in the prior art that the effectiveness, response efficiency and robustness of distribution network control cannot be taken into account at the same time.

[0006] In a first aspect, the present application provides a distribution network reactive power and voltage control method based on cooperative game, the method comprising:

[0007] Based on the day-ahead time scale, the distribution network operation periods are dynamically aggregated to form multiple continuous period clusters;

[0008] In each of the time period clusters, initial node partitions are generated based on how the node voltage corresponding to each node in the distribution network is affected by photovoltaic reactive output, each of the initial node partitions is optimized using a cooperative game model, and multiple target node partitions are determined based on voltage control costs;

[0009] Based on the real-time time scale, the corresponding dominant control node in each target node partition is selected, and the voltage fluctuation of each target node partition is controlled through reactive power tracking control or collaborative control.

[0010] As an optional implementation, the method of dynamically aggregating the distribution network operation time periods based on the day-ahead time scale to form multiple continuous time period clusters includes:

[0011] Based on the day-ahead time scale, determining a preset number of operating time periods corresponding to the intraday time, determining the number of clusters and the initial cluster center corresponding to each cluster, and determining a multidimensional feature space;

[0012] The multidimensional feature space is used to indicate the operating characteristics corresponding to each of the operating time periods and the cluster centers obtained by real-time iterative optimization, including photovoltaic output characteristics, active load characteristics, and reactive load characteristics;

[0013] By iteratively optimizing the optimal segmentation boundaries of adjacent cluster centers, updating the feature vectors corresponding to the real-time cluster centers based on the multidimensional feature space until the feature vectors corresponding to two adjacent iterative optimizations meet the preset convergence conditions, and outputting the time period cluster division results with time period continuity according to the number of clusters;

[0014] The convergence condition includes that the characteristic fluctuation of the real-time cluster center corresponding to two adjacent iterative optimizations is lower than the convergence threshold, or the preset number of iterations is reached.

[0015] As an optional implementation manner, the method for determining the optimal segmentation boundary includes:

[0016] Calculating the feature differences between each of the running time periods and adjacent cluster centers according to the multidimensional feature space, and selecting the time period with the largest sum of the differences as the current optimal segmentation boundary;

[0017] The characteristic difference is calculated by weighted accumulation of the Euclidean distances of the photovoltaic output characteristic, the active load characteristic, and the reactive load characteristic.

[0018] As an optional implementation manner, generating the initial node partition according to the influence of photovoltaic reactive output on the node voltage corresponding to each node in the distribution network includes:

[0019] Taking the photovoltaic grid-connected nodes in the distribution network as partition master nodes and determining the number of partitions;

[0020] In the time period cluster, according to each node to be assigned in the distribution network and each partition master node, a voltage sensitivity coefficient is determined, and node attribution is determined based on the voltage sensitivity coefficient to generate an initial node partition;

[0021] The voltage sensitivity coefficient is used to indicate the influence of the photovoltaic reactive power output of each partition master node on each of the nodes to be allocated, and is determined according to the voltage over-limit risk probability and the voltage sensitivity coefficient.

[0022] As an optional implementation, the process of constructing the cooperative game model includes:

[0023] Determine each of the initial node partitions as a participant set, and determine the voltage control cost corresponding to each of the initial node partitions as a payment function;

[0024] Determine the constraints of photovoltaic dynamic reactive power regulation and the number of zones;

[0025] The voltage control cost includes a linear combination of the local voltage deviation, the reactive power regulation voltage fluctuation in the partition, the reactive power regulation voltage fluctuation of other partitions on the partition, and the partition deviation penalty coefficient.

[0026] As an optional implementation manner, optimizing each of the initial node partitions through a cooperative game model and determining multiple target node partitions based on voltage control costs includes:

[0027] Verify the existence of Nash equilibrium solutions in a non-cooperative game framework;

[0028] Under the condition that there is a Nash equilibrium solution, the optimization goal is to minimize the payment function of each partition. For each adjusted node partition, the partition adjustment plan is determined under the condition that the total payment function of the adjusted node partition does not exceed the payment function of the node partition before adjustment. The partition strategy and topology structure are alternately optimized to reduce the global voltage control cost and determine the final target node partition.

[0029] As an optional implementation manner, the reactive power tracking control or the coordinated control is determined based on whether a voltage exceeding a limit occurs in each target node partition, and the coordinated control includes voltage sensitivity control and reactive power tracking control executed in sequence;

[0030] The selecting of the corresponding dominant control node in each target node partition and controlling the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control specifically includes:

[0031] Selecting the corresponding dominant control node in each target node partition;

[0032] For each target node partition, when the voltage fluctuation exceeds the set fluctuation threshold, reactive power tracking control is triggered, and the photovoltaic reactive power output of each node is dynamically adjusted according to the fluctuation of the leading observation node;

[0033] Furthermore, when voltage over-limit is detected, voltage sensitivity control is triggered to perform reactive power regulation, and based on the sensitivity of the reactive power regulation of the leading control node to the voltage, reactive power regulation is allocated until the voltage over-limit phenomenon in the corresponding target node partition is resolved;

[0034] If the reactive adjustment amount is lower than the target adjustment amount corresponding to the voltage over-limit phenomenon being resolved, active adjustment is triggered; if the reactive adjustment amount is higher than the target adjustment amount, reactive tracking control is triggered;

[0035] The selection of the dominant control node includes: determining the reactive sensitivity and active sensitivity of each node in each target node partition to other nodes in the partition, selecting a node in the partition whose comprehensive sensitivity to other nodes exceeds a preset sensitivity threshold as the dominant control node and configuring a dynamic priority weight;

[0036] The selection of the leading observation node includes: determining a node in each target node partition whose absolute value of voltage fluctuation in a corresponding time period is lower than a preset fluctuation threshold as the leading observation node.

[0037] In a second aspect, the present application provides a distribution network reactive power and voltage control device based on cooperative game, the device comprising:

[0038] A processing module is used to dynamically aggregate the distribution network operation time periods to form multiple continuous time period clusters based on the day-ahead time scale;

[0039] The processing module is further configured to generate, within each time period cluster, an initial node partition based on how the node voltage corresponding to each node in the distribution network is affected by photovoltaic reactive output, optimize each of the initial node partitions using a cooperative game model, and determine a plurality of target node partitions based on voltage control costs;

[0040] The processing module is further configured to select a corresponding dominant control node in each target node partition based on a real-time time scale, and control the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control.

[0041] In a third aspect, the present application provides a computer device comprising one or more processors and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the method described in the first aspect are performed.

[0042] In a fourth aspect, the present application provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the method described in the first aspect.

[0043] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0044] Based on any of the above embodiments, the cooperative game-based reactive power and voltage control method for distribution networks proposed in this application comprehensively covers the electrical physical quantities and operating status data during the operation of the distribution network through the collaborative collection of polymorphic measurement information and system operation logs, providing a rich data foundation for fault tracing. In the data preprocessing stage, the structure of node feature data is optimized through denoising, deduplication, and discretization processing, reducing the computational complexity. By constructing an information difference graph, the information interconnection relationship and changes between target components before and after the fault are intuitively reflected, providing a key basis for the dynamic analysis of fault propagation. Furthermore, ranking learning based on time series feature sets generates ranking scores, and combined with the information difference graph, the fault propagation path is accurately determined, realizing a complete closed loop from data collection to path tracing. This method significantly improves the accuracy and efficiency of fault location through multi-dimensional data fusion and intelligent analysis. In the subsequent process, the introduction of visual display, multi-level fault reporting, and intelligent alarm mechanism not only enhances the timeliness of fault response, but also provides scientific support for operation and maintenance decision-making. Compared with traditional methods, this solution achieves comprehensive optimization in data quality, analysis depth and response efficiency, providing reliable technical support for the stable operation and rapid fault handling of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0046] Figure 1 A flowchart of a method for controlling reactive power and voltage in a distribution network based on cooperative game theory according to an embodiment of the present application is provided;

[0047] Figure 2 A flowchart of a method for controlling reactive power and voltage in a distribution network based on cooperative game theory according to an embodiment of the present application is provided;

[0048] Figure 3 A flowchart of a method for controlling reactive power and voltage in a distribution network based on cooperative game theory according to an embodiment of the present application is provided;

[0049] Figure 4 This is a diagram of the internal structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0051] With the rapid increase in the penetration rate of photovoltaic power generation, a high proportion of photovoltaic access has become an important feature of modern distribution networks. While this trend promotes the consumption of clean energy, it also brings significant voltage control challenges to the operation of distribution networks, including voltage over-limit, increased fluctuations, and enhanced control coupling.

[0052] Traditional distribution network voltage control methods exhibit significant limitations in scenarios with a high penetration of photovoltaic (PV) power. Centralized control relies on global optimization at a master station. As the number of PV nodes increases, this approach faces significant communication latency and computational overhead. Field data indicates that for systems with more than 100 PV nodes, centralized optimization can take up to several minutes to solve, making it difficult to meet real-time control requirements. Furthermore, centralized architectures present a single point of failure risk; a master station failure can lead to system control failure.

[0053] Existing distributed voltage control methods still have key shortcomings. Partitioning strategies based on static voltage sensitivity (such as modularity optimization) struggle to adapt to the time-varying nature of PV output. When PV output fluctuates by more than 20%, the rationality of partitioning decreases significantly. While dynamic partitioning methods improve adaptability, they fail to fully consider the coupling effects of voltage control between partitions, resulting in significant control interactions. Furthermore, existing distributed algorithms often use fixed control parameters, which can easily lead to control oscillations in high-penetration scenarios, compromising system stability.

[0054] The application of cooperative game theory to distribution network zoning optimization also faces challenges. Traditional game-solving methods (such as Nash equilibrium calculations) are highly complex and difficult to meet the time requirements of real-time control. Existing simplified algorithms improve computational efficiency but may sacrifice zoning coordination, leading to increased voltage regulation deviations. Furthermore, information asymmetry between photovoltaic nodes has not yet been effectively addressed, reducing the robustness of control strategies.

[0055] In this context, a new voltage control method is urgently needed that balances dynamic partitioning adaptability, control accuracy, and computational efficiency. This method must address three key issues: first, how to construct a dynamic partitioning mechanism to maintain partitioning rationality and control effectiveness under PV output fluctuations; second, how to design a coordinated control strategy to achieve inter-partition voltage decoupling and rapid intra-partition voltage regulation under limited communication conditions; and third, how to establish an efficient optimization framework to ensure voltage control accuracy while reducing computational complexity. Addressing these issues will provide important technical support for the safe and stable operation of high-PV distribution networks.

[0056] Therefore, this application proposes a distribution network reactive voltage control method based on cooperative game, which is used to solve the voltage over-limit and fluctuation problems caused by a high proportion of photovoltaic access to the distribution network, and overcome the shortcomings of existing centralized control and traditional distributed control in terms of partition rationality, control accuracy and real-time performance. Specifically, this application comprehensively covers the electrical physical quantities and operating status data during the operation of the distribution network through the collaborative collection of polymorphic measurement information and system operation logs, providing a rich data foundation for fault tracing. In the data preprocessing stage, the structure of the node feature data is optimized and the computational complexity is reduced through denoising, deduplication and discretization. By constructing an information difference graph, the information interconnection relationship and changes between the target components before and after the fault are intuitively reflected, providing a key basis for the dynamic analysis of fault propagation. Furthermore, the ranking score is generated based on the ranking learning of the time series feature set, and the fault propagation path is accurately determined in combination with the information difference graph, realizing a complete closed loop from data collection to path tracing. This method significantly improves the accuracy and efficiency of fault location through multi-dimensional data fusion and intelligent analysis. In subsequent processes, the introduction of visual display, multi-level fault reporting, and intelligent alerting mechanisms not only enhances the timeliness of fault response but also provides scientific support for operation and maintenance decision-making. Compared with traditional methods, this solution achieves comprehensive optimization in data quality, analysis depth, and response efficiency, providing reliable technical support for the stable operation of the distribution network and rapid fault resolution.

[0057] The method provided in this application is described in detail below based on corresponding implementation methods in some actual application scenarios.

[0058] See also Figure 1 , Figure 1 A flow chart of a method for controlling reactive power and voltage in a distribution network based on cooperative game is provided as a flowchart of an embodiment of the present application. Figure 1 As shown, the method includes:

[0059] S101. Based on the day-ahead time scale, dynamically aggregate the distribution network operation time periods to form multiple continuous time period clusters;

[0060] S102. Within each of the time period clusters, generating initial node partitions based on how node voltages corresponding to nodes in the distribution network are affected by photovoltaic reactive output, optimizing each of the initial node partitions using a cooperative game model, and determining multiple target node partitions based on voltage control costs;

[0061] S103 . Based on the real-time time scale, select the corresponding dominant control node in each target node partition, and control the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control.

[0062] The reactive power and voltage control method for distribution networks provided in this application, which is based on cooperative game theory, forms continuous time period clusters by dynamically aggregating operating time periods on a day-ahead time scale, and generates initial node partitions within each cluster based on the impact of photovoltaic reactive output on node voltage. The cooperative game model is further used to optimize the partitioning strategy, and finally the dominant control node is selected on a real-time scale to achieve precise control of voltage fluctuations. This method solves the response lag problem caused by communication delay and computational burden in traditional centralized control, as well as the control oscillation caused by the coupling effect between partitions in distributed control. Through dynamic time period clustering, the system can adapt to the time-varying characteristics of photovoltaic output and improve the flexibility of the time dimension; the cooperative game model optimizes the partitioning strategy, effectively reducing the mutual interference of reactive power regulation between partitions and reducing the global voltage control cost; the real-time dominant node control mechanism accurately locates key regulation nodes through sensitivity analysis, and combines reactive power tracking with collaborative control strategies to quickly respond to voltage fluctuations and over-limit problems. Ultimately, the stability, response efficiency and robustness of distribution network voltage control are significantly improved, providing an efficient solution for voltage management in high-proportion photovoltaic scenarios.

[0063] As an optional implementation, the method of dynamically aggregating the distribution network operation time periods based on the day-ahead time scale to form multiple continuous time period clusters includes:

[0064] Based on the day-ahead time scale, determining a preset number of operating time periods corresponding to the intraday time, determining the number of clusters and the initial cluster center corresponding to each cluster, and determining a multidimensional feature space;

[0065] The multidimensional feature space is used to indicate the operating characteristics corresponding to each of the operating time periods and the cluster centers obtained by real-time iterative optimization, including photovoltaic output characteristics, active load characteristics, and reactive load characteristics;

[0066] By iteratively optimizing the optimal segmentation boundaries of adjacent cluster centers, updating the feature vectors corresponding to the real-time cluster centers based on the multidimensional feature space until the feature vectors corresponding to two adjacent iterative optimizations meet the preset convergence conditions, and outputting the time period cluster division results with time period continuity according to the number of clusters;

[0067] The convergence condition includes that the characteristic fluctuation of the real-time cluster center corresponding to two adjacent iterative optimizations is lower than the convergence threshold, or the preset number of iterations is reached.

[0068] This implementation iteratively optimizes cluster centers in a multidimensional feature space and dynamically updates optimal segmentation boundaries based on feature differences, ultimately outputting cluster partitioning results with time-period continuity. This design ensures that the time-period clustering accurately reflects the dynamic changes in PV output, avoiding control deviations caused by traditional static partitioning that ignores temporal continuity. This improves the coordination and adaptability of the voltage control strategy within the cluster, laying a solid foundation for subsequent partitioning optimization.

[0069] As an optional implementation manner, the method for determining the optimal segmentation boundary includes:

[0070] Calculating the feature differences between each of the running time periods and adjacent cluster centers according to the multidimensional feature space, and selecting the time period with the largest sum of the differences as the current optimal segmentation boundary;

[0071] The characteristic difference is calculated by weighted accumulation of the Euclidean distances of the photovoltaic output characteristic, the active load characteristic, and the reactive load characteristic.

[0072] This implementation calculates the feature differences between the operating time period and the cluster center using weighted Euclidean distance accumulation, and determines the optimal segmentation boundary by maximizing the sum of these differences. This method enhances the discriminability of time-period clustering, accurately capturing the correlation between PV output fluctuations and load changes, and reduces clustering ambiguity caused by feature overlap, further improving the clarity of time-period cluster boundaries and the targeted nature of control strategies.

[0073] As an optional implementation manner, generating the initial node partition according to the influence of photovoltaic reactive output on the node voltage corresponding to each node in the distribution network includes:

[0074] Taking the photovoltaic grid-connected nodes in the distribution network as partition master nodes and determining the number of partitions;

[0075] In the time period cluster, a voltage sensitivity coefficient is determined according to each node to be assigned in the distribution network and each partition master node, and node attribution is determined based on the voltage sensitivity coefficient to generate an initial node partition.

[0076] The voltage sensitivity coefficient is used to indicate the influence of the photovoltaic reactive power output of each partition master node on each of the nodes to be allocated, and is determined according to the voltage over-limit risk probability and the voltage sensitivity coefficient.

[0077] This implementation combines the risk probability of overshooting with voltage sensitivity, determining node ownership through the voltage sensitivity coefficient and generating an initial partition centered around the PV grid-connected node. This mechanism quantifies the differential impact of PV reactive output on each node, avoiding the waste of control resources caused by subjective partitioning. It ensures the physical rationality and regulation efficiency of the initial partitioning, providing high-precision input for cooperative game optimization.

[0078] As an optional implementation, the process of constructing the cooperative game model includes:

[0079] Determine each of the initial node partitions as a participant set, and determine the voltage control cost corresponding to each of the initial node partitions as a payment function;

[0080] Determine the constraints of photovoltaic dynamic reactive power regulation and the number of zones;

[0081] The voltage control cost includes a linear combination of the local voltage deviation, the reactive power regulation voltage fluctuation in the partition, the reactive power regulation voltage fluctuation of other partitions on the partition, and the partition deviation penalty coefficient.

[0082] This implementation constructs a cooperative game model using voltage control cost as the payment function and introduces dynamic PV regulation and constraints on the number of zones. This design incorporates inter-zone coupling effects into the cost optimization objective and coordinates the regulation strategies of each zone through a game equilibrium solution, significantly reducing global control costs while avoiding voltage oscillations caused by independent zone regulation.

[0083] As an optional implementation manner, optimizing each of the initial node partitions through a cooperative game model and determining multiple target node partitions based on voltage control costs includes:

[0084] Verify the existence of Nash equilibrium solutions in a non-cooperative game framework;

[0085] Under the condition that there is a Nash equilibrium solution, the optimization goal is to minimize the payment function of each partition. For each adjusted node partition, the partition adjustment plan is determined under the condition that the total payment function of the adjusted node partition does not exceed the payment function of the node partition before adjustment. The partition strategy and topology structure are alternately optimized to reduce the global voltage control cost and determine the final target node partition.

[0086] This implementation alternately optimizes partitioning strategies and topology structures, assuming a Nash equilibrium solution exists, with the goal of minimizing the payoff function. This process ensures that total costs do not exceed pre-optimization levels through partition adjustment, achieving a dynamic balance between partition coordination and resource allocation. Ultimately, it achieves the optimal balance between cost control and stability, improving the system's economic efficiency and robustness.

[0087] As an optional implementation manner, the reactive power tracking control or the coordinated control is determined based on whether a voltage exceeding a limit occurs in each target node partition, and the coordinated control includes voltage sensitivity control and reactive power tracking control executed in sequence;

[0088] The selecting of the corresponding dominant control node in each target node partition and controlling the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control specifically includes:

[0089] Selecting the corresponding dominant control node in each target node partition;

[0090] For each target node partition, when the voltage fluctuation exceeds the set fluctuation threshold, reactive power tracking control is triggered, and the photovoltaic reactive power output of each node is dynamically adjusted according to the fluctuation of the leading observation node;

[0091] Furthermore, when voltage over-limit is detected, voltage sensitivity control is triggered to perform reactive power regulation, and based on the sensitivity of the reactive power regulation of the leading control node to the voltage, reactive power regulation is allocated until the voltage over-limit phenomenon in the corresponding target node partition is resolved;

[0092] If the reactive adjustment amount is lower than the target adjustment amount corresponding to the voltage over-limit phenomenon being resolved, active adjustment is triggered; if the reactive adjustment amount is higher than the target adjustment amount, reactive tracking control is triggered;

[0093] The selection of the dominant control node includes: determining the reactive sensitivity and active sensitivity of each node in each target node partition to other nodes in the partition, selecting a node in the partition whose comprehensive sensitivity to other nodes exceeds a preset sensitivity threshold as the dominant control node and configuring a dynamic priority weight;

[0094] The selection of the leading observation node includes: determining a node in each target node partition whose absolute value of voltage fluctuation in a corresponding time period is lower than a preset fluctuation threshold as the leading observation node.

[0095] This implementation uses a hierarchical approach to trigger reactive power tracking control or coordinated control incorporating voltage sensitivity control based on voltage overshooting. The dominant control node is dynamically selected based on the node's combined sensitivity. This mechanism rapidly eliminates overshooting through precise sensitivity-driven regulation. The hierarchical strategy avoids resource waste caused by over-regulation, significantly improving response speed and exception handling efficiency in emergencies.

[0096] The following example illustrates the specific implementation of each embodiment of the present application when combined with an actual application scenario. The method provided by the present application addresses the voltage over-limit and fluctuation problems of a high proportion of photovoltaic access to the distribution network. In a distribution network system containing multiple photovoltaic nodes, dynamic partitioning and coordinated control are used to achieve rapid voltage regulation and stable operation. The method provided by the present application may include the following four parts:

[0097] Step 1: K-means time period clustering. Based on the day-ahead dynamic reactive power optimization results, the K-means algorithm is used to cluster the reactive power optimization time periods.

[0098] Step 2: Determine the initial partitioning scheme. Use voltage sensitivity to characterize the voltage regulation capability of PV reactive output on the distribution network, and determine the initial partitioning scheme based on the criterion that the node voltage is most affected by PV reactive output during the clustering period.

[0099] Step 3: Initial inter-sub-partition cooperative game model. Treat each sub-partition under the initial partitioning scheme as an independent stakeholder, and determine the final partitioning scheme by establishing and solving the initial inter-sub-partition cooperative game model.

[0100] Step 4: Real-time reactive power / voltage zoning control: A real-time reactive power / voltage zoning control method based on voltage sensitivity and reactive power tracking control is used to effectively suppress voltage over-limit and fluctuations.

[0101] Obviously, step 1 can be regarded as the specific implementation of the aforementioned S101 at the day-ahead time scale, that is, clustering time periods based on historical information. Step 4 can be regarded as the specific implementation of application at the real-time time scale after all modeling work is completed, that is, the implementation method corresponding to S103, and steps 2 and 3 are the detailed decomposition of S102 in this application.

[0102] In an embodiment of an improved IEEE 33-node system, in step 1, the K-means time period clustering step considers the serious voltage fluctuations caused by uncertain changes in PV output and load power within a day, necessitating dynamic adjustment of the intraday partitions. To reduce the number and complexity of partition adjustments, the K-means algorithm is used to cluster the 24 time periods into K (K < 24) consecutive time periods. The specific implementation process is as follows:

[0103] Step 1.1: Randomly select K time periods from all 24h time periods within a day as the initial cluster centers , where each cluster center Including photovoltaic output expectations , Active load expectation and reactive load expectation Equal eigenvectors.

[0104] Step 1.2: Find the boundary time periods between adjacent cluster centers in sequence , the boundary period is determined by optimizing the following objective function:

[0105]

[0106] The distance metric Defined as:

[0107]

[0108] Where, Represents the expected photovoltaic output column vector for period t, whose elements is the predicted photovoltaic power generation power of node i in time period t (unit: kW); Indicates time period The expected active load column vector, whose elements is node j in time period The predicted active load (unit: kW); Indicates time period The reactive load expectation column vector, whose elements For nodes In the period The predicted reactive load (unit: kvar); Represents the Euclidean norm (L2 norm) of the vector, which is used to quantify the overall difference between feature vectors.

[0109] Step 1.3: Set adjacent boundary periods and All time periods The final classification of the boundary period is determined by the following criteria:

[0110]

[0111] Step 1.4: Update the cluster center feature vector:

[0112]

[0113] in Represents the set of time periods contained in the k-th cluster.

[0114] Step 1.5: Repeat steps 1.2 to 1.4 until the cluster center no longer changes (i.e. ) or reaches the maximum number of iterations, and finally outputs K stable clustering periods .

[0115] See also Figure 2 , Figure 2A flow chart of a method for controlling reactive power and voltage in a distribution network based on cooperative game is provided as a flowchart of an embodiment of the present application. Figure 2 Specifically, in an embodiment of an improved IEEE 33-node system, in step 2, during the initial partitioning, the number of initial sub-partitions is set to the number of photovoltaic grid-connected points. The photovoltaic access nodes are used as inherent subordinate nodes of the sub-partitions, and voltage sensitivity is used to characterize the degree of influence of photovoltaic reactive output on the voltage of its characteristic partition nodes. The specific implementation process is as follows:

[0116] Step 2.1: Partition initialization. Set the initial sub-partition number Set as the number of photovoltaic grid-connected points , each photovoltaic grid-connected node As the corresponding sub-partition The core nodes are:

[0117]

[0118] Step 2.2: Voltage sensitivity calculation. For each node to be allocated , calculate its clustering period Internal photovoltaic node Comprehensive impact indicators of reactive power output:

[0119]

[0120] Where, For the period Voltage sensitivity matrix midpoint Inject reactive power to nodes Voltage sensitivity coefficient; For nodes Voltage over-limit risk probability, where and They are the lower and upper voltage limits respectively.

[0121] Step 2.3: Determine the node partition. Divide into the sub-zones of the photovoltaic nodes that have the largest comprehensive impact index:

[0122]

[0123] in is the total number of nodes in the system.

[0124] Step 2.4: Initial partition generation. By iteratively executing steps 2.2-2.3, the partition assignment of all nodes is completed and the initial partitioning scheme is formed. In the IEEE 33-node system embodiment, set The final generated initial partition contains an average of 3-5 nodes.

[0125] Specifically, in an embodiment of an improved IEEE 33-bus system, step 3 sets each initial sub-partition as an independent stakeholder based on the obtained initial partitioning scheme, sets the optimal effect of each initial partition in suppressing voltage fluctuations in the partition as the goal, and then considers merging the initial sub-partitions to form the final partition to achieve coordinated suppression of voltage fluctuations within the partition. The initial partition merging problem is equivalent to solving a cooperative game problem between the initial sub-partitions. The specific implementation process is as follows:

[0126] Step 3.1: Initial sub-partition cooperative game model construction. Establish the initial sub-partition cooperative game model , where the participant set is the initial subpartition, the payment function For partition Voltage control cost:

[0127]

[0128] Where, is the partition deviation penalty coefficient; Voltage fluctuations caused by photovoltaic and load forecast errors; The photovoltaic regulation amount of this zone; The influence of other partitions.

[0129] Step 3.2: Setting constraints.

[0130] Step 3.2.1: PV dynamic reactive power regulation constraints:

[0131]

[0132] in, for Time Node Photovoltaic dynamic reactive power regulation; for Time Node Maximum photovoltaic reactive capacity; is obtained after day-ahead dynamic reactive power optimization Time period node PV reactive power regulation plan value.

[0133] Step 3.2.2: Constraint on the number of partitions:

[0134]

[0135] in, The minimum number of partitions for an IEEE 33-node system.

[0136] Step 3.3: Solve the cooperative game between the initial sub-partitions. Since the additivity constraint in the cooperative game model between the initial sub-partitions is based on the Nash equilibrium of the non-cooperative game in the initial partition, it is necessary to first solve the Nash equilibrium of the non-cooperative game in the initial partition. By proving the existence of the Nash equilibrium, we can verify whether the payoff function of each participant in the non-cooperative game can be minimized. The specific solution process is shown in Figure 2 , the solution process uses the cplex algorithm for iterative calculation:

[0137] Step 3.3.1: Initialize each sub-partition strategy , set the number of iterations ;

[0138] Step 3.3.2: Fix other partitioning strategies and optimize each sub-partition in turn The payment function ;

[0139] Step 3.3.3: Determine the convergence of the payment function: ,in is the convergence threshold;

[0140] Step 3.3.4: If it does not converge, set Return to step 3.3.2, otherwise output the Nash equilibrium strategy .

[0141] Step 3.4: Cooperative game optimization. Based on the Nash equilibrium solution, the optimal merging solution is solved with the goal of minimizing the total partition payment:

[0142]

[0143] in, For partition The payment function.

[0144] Satisfy the additivity constraint of cooperative game:

[0145]

[0146] Where: For the initial partition This constraint ensures that the partitions after merging The total payoff does not exceed the sum of the payoffs of its included sub-partitions, ensuring the rationality of the merging scheme. In an IEEE 33-bus system implementation, this cooperative game approach was used to merge the initial 11 sub-partitions into three final partitions, reducing the inter-partition voltage coupling by 42.7%, validating the effectiveness of the approach. Figure 2 The complete cooperative game solving process is demonstrated, including two key stages: non-cooperative equilibrium solving and cooperative optimization.

[0147] See also Figure 3 , Figure 3 A flow chart of a method for controlling reactive power and voltage in a distribution network based on cooperative game is provided as a flowchart of an embodiment of the present application. Figure 3 As shown, specifically, in an embodiment of an improved IEEE 33-node system, step 4 includes:

[0148] Step 4.1: Voltage sensitivity control. In real-time reactive power / voltage zone control, voltage sensitivity control is first used to achieve coordinated suppression of voltage over-limit within the zone. The specific implementation process is as follows:

[0149] Step 4.1.1: Select the leading control node. The selection conditions of the internal dominant control node can be expressed as:

[0150]

[0151] Where: Indicates time period node Inject reactive power to nodes Voltage sensitivity coefficient; Indicates time period Node $j$ injects active power to node Voltage sensitivity coefficient; For the period node Voltage regulation weight coefficient; Indicates the cluster time period set to which the current cluster belongs.

[0152] Step 4.1.2: Calculation of PV dynamic reactive power regulation:

[0153]

[0154] in, For the period Voltage-reactive sensitivity coefficient; Represents a node Sub-partitions for the core; voltage control requirements For the moment node The voltage control requirement is defined as:

[0155]

[0156] Step 4.1.3: PV active power reduction control (when reactive power regulation is insufficient):

[0157]

[0158] Where: For the period Voltage-active power sensitivity coefficient; is the calculated photovoltaic reactive power regulation; Representation node The voltage exceeds the upper limit.

[0159] Step 4.2: Reactive power tracking control. In real-time reactive power / voltage zone control, reactive power tracking control is adopted to suppress voltage fluctuations by utilizing the PV surplus reactive power regulation capability while achieving tracking control of the grid connection point voltage to the dominant observation node voltage within the zone.

[0160] Step 4.2.1: Select the leading observation node. In order to minimize the voltage fluctuation amplitude of the grid connection point after control, select the node with the smallest voltage fluctuation severity index within the partition as the leading observation node. Internal nodes Voltage fluctuation severity index:

[0161]

[0162] Where: Indicates time node The voltage fluctuation amount; Indicates the cluster period set to which the current cluster belongs. The selection conditions of the internal dominant observation node can be expressed as:

[0163]

[0164] Step 4.2.2: Calculation of reactive tracking regulation. According to the voltage fluctuation of the dominant observation node, calculate Photovoltaic nodes at all times Reactive power regulation:

[0165]

[0166] in:

[0167]

[0168] Where: R, X are line resistance and reactance respectively; For the moment Leading observation node The voltage amplitude; Represents a photovoltaic node The change in active power output; It is the compensation item for the reactive power variation of the load.

[0169] Step 4.2.3: Control execution. Since reactive power tracking control uses local control, it is easily limited by the maximum reactive power capacity of photovoltaic cells, making it difficult to completely smooth out voltage fluctuations. Therefore, when the voltage exceeds the limit within the partition, the voltage sensitivity control method is preferred. When there is surplus reactive power regulation capacity of photovoltaic cells, the reactive power tracking control method is adopted. The control process can be found in Figure 3 .

[0170] In the IEEE 33-bus system test, this method reduced the voltage fluctuation amplitude at node 17 by 52.4%, and the control response time was less than 1 second, meeting the real-time control requirements.

[0171] The embodiment of the present application further provides a distribution network reactive power and voltage control device based on cooperative game, the device comprising:

[0172] A processing module is used to dynamically aggregate the distribution network operation time periods to form multiple continuous time period clusters based on the day-ahead time scale;

[0173] The processing module is further configured to generate, within each time period cluster, an initial node partition based on how the node voltage corresponding to each node in the distribution network is affected by photovoltaic reactive output, optimize each of the initial node partitions using a cooperative game model, and determine a plurality of target node partitions based on voltage control costs;

[0174] The processing module is further configured to select a corresponding dominant control node in each target node partition based on a real-time time scale, and control the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control.

[0175] This application forms continuous time period clusters by dynamically aggregating operating time periods on a day-ahead time scale, and generates initial node partitions within each cluster based on the impact of photovoltaic reactive output on node voltage. It further uses a cooperative game model to optimize the partitioning strategy, and ultimately selects a dominant control node on a real-time scale to achieve precise regulation of voltage fluctuations. This method solves the response lag problem caused by communication delays and computational burdens in traditional centralized control, as well as the control oscillation caused by coupling effects between partitions in distributed control. Through dynamic time period clustering, the system can adapt to the time-varying characteristics of photovoltaic output and improve flexibility in the time dimension; the cooperative game model optimizes the partitioning strategy, effectively reducing mutual interference in reactive regulation between partitions and reducing the global voltage control cost; the real-time dominant node control mechanism accurately locates key regulation nodes through sensitivity analysis, and combines reactive tracking with collaborative control strategies to quickly respond to voltage fluctuations and over-limit problems. Ultimately, the stability, response efficiency, and robustness of distribution network voltage control are significantly improved, providing an efficient solution for voltage management in high-proportion photovoltaic scenarios.

[0176] As an optional implementation manner, the processing module dynamically aggregates the distribution network operation time periods to form multiple continuous time period clusters based on the day-ahead time scale, including:

[0177] Based on the day-ahead time scale, determining a preset number of operating time periods corresponding to the intraday time, determining the number of clusters and the initial cluster center corresponding to each cluster, and determining a multidimensional feature space;

[0178] The multidimensional feature space is used to indicate the operating characteristics corresponding to each of the operating time periods and the cluster centers obtained by real-time iterative optimization, including photovoltaic output characteristics, active load characteristics, and reactive load characteristics;

[0179] By iteratively optimizing the optimal segmentation boundaries of adjacent cluster centers, updating the feature vectors corresponding to the real-time cluster centers based on the multidimensional feature space until the feature vectors corresponding to two adjacent iterative optimizations meet the preset convergence conditions, and outputting the time period cluster division results with time period continuity according to the number of clusters;

[0180] The convergence condition includes that the characteristic fluctuation of the real-time cluster center corresponding to two adjacent iterative optimizations is lower than the convergence threshold, or the preset number of iterations is reached.

[0181] This implementation iteratively optimizes cluster centers in a multidimensional feature space and dynamically updates optimal segmentation boundaries based on feature differences, ultimately outputting cluster partitioning results with time-period continuity. This design ensures that the time-period clustering accurately reflects the dynamic changes in PV output, avoiding control deviations caused by traditional static partitioning that ignores temporal continuity. This improves the coordination and adaptability of the voltage control strategy within the cluster, laying a solid foundation for subsequent partitioning optimization.

[0182] As an optional implementation manner, the processing module determines the optimal segmentation boundary in the following manner:

[0183] Calculating the feature differences between each of the running time periods and adjacent cluster centers according to the multidimensional feature space, and selecting the time period with the largest sum of the differences as the current optimal segmentation boundary;

[0184] The characteristic difference is calculated by weighted accumulation of the Euclidean distances of the photovoltaic output characteristic, the active load characteristic, and the reactive load characteristic.

[0185] This implementation calculates the feature differences between the operating time period and the cluster center using weighted Euclidean distance accumulation, and determines the optimal segmentation boundary by maximizing the sum of these differences. This method enhances the discriminability of time-period clustering, accurately capturing the correlation between PV output fluctuations and load changes, and reduces clustering ambiguity caused by feature overlap, further improving the clarity of time-period cluster boundaries and the targeted nature of control strategies.

[0186] As an optional implementation manner, the processing module generates a specific method of generating the initial node partition according to the influence of photovoltaic reactive output on the node voltage corresponding to each node in the distribution network, including:

[0187] Taking the photovoltaic grid-connected nodes in the distribution network as partition master nodes and determining the number of partitions;

[0188] In the time period cluster, a voltage sensitivity coefficient is determined according to each node to be assigned in the distribution network and each partition master node, and node attribution is determined based on the voltage sensitivity coefficient to generate an initial node partition.

[0189] The voltage sensitivity coefficient is used to indicate the influence of the photovoltaic reactive power output of each partition master node on each of the nodes to be allocated, and is determined according to the voltage over-limit risk probability and the voltage sensitivity coefficient.

[0190] This implementation combines the risk probability of overshooting with voltage sensitivity, determining node ownership through the voltage sensitivity coefficient and generating an initial partition centered around the PV grid-connected node. This mechanism quantifies the differential impact of PV reactive output on each node, avoiding the waste of control resources caused by subjective partitioning. It ensures the physical rationality and regulation efficiency of the initial partitioning, providing high-precision input for cooperative game optimization.

[0191] As an optional implementation manner, the process of constructing the cooperative game model executed by the processing module specifically includes:

[0192] Determine each of the initial node partitions as a participant set, and determine the voltage control cost corresponding to each of the initial node partitions as a payment function;

[0193] Determine the constraints of photovoltaic dynamic reactive power regulation and the number of zones;

[0194] The voltage control cost includes a linear combination of the local voltage deviation, the reactive power regulation voltage fluctuation in the partition, the reactive power regulation voltage fluctuation of other partitions on the partition, and the partition deviation penalty coefficient.

[0195] This implementation constructs a cooperative game model using voltage control cost as the payment function and introduces dynamic PV regulation and constraints on the number of zones. This design incorporates inter-zone coupling effects into the cost optimization objective and coordinates the regulation strategies of each zone through a game equilibrium solution, significantly reducing global control costs while avoiding voltage oscillations caused by independent zone regulation.

[0196] As an optional implementation manner, the processing module optimizes each of the initial node partitions through a cooperative game model, and determines a specific manner of multiple target node partitions based on voltage control costs, including:

[0197] Verify the existence of Nash equilibrium solutions in a non-cooperative game framework;

[0198] Under the condition that there is a Nash equilibrium solution, the optimization goal is to minimize the payment function of each partition. For each adjusted node partition, the partition adjustment plan is determined under the condition that the total payment function of the adjusted node partition does not exceed the payment function of the node partition before adjustment. The partition strategy and topology structure are alternately optimized to reduce the global voltage control cost and determine the final target node partition.

[0199] This implementation alternately optimizes partitioning strategies and topology structures, assuming a Nash equilibrium solution exists, with the goal of minimizing the payoff function. This process ensures that total costs do not exceed pre-optimization levels through partition adjustment, achieving a dynamic balance between partition coordination and resource allocation. Ultimately, it achieves the optimal balance between cost control and stability, improving the system's economic efficiency and robustness.

[0200] As an optional implementation manner, the reactive power tracking control or the coordinated control is determined based on whether a voltage exceeding a limit occurs in each target node partition, and the coordinated control includes voltage sensitivity control and reactive power tracking control executed in sequence;

[0201] The processing module selects the corresponding dominant control node in each target node partition and controls the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control. The specific method includes:

[0202] Selecting the corresponding dominant control node in each target node partition;

[0203] For each target node partition, when the voltage fluctuation exceeds the set fluctuation threshold, reactive power tracking control is triggered, and the photovoltaic reactive power output of each node is dynamically adjusted according to the fluctuation of the leading observation node;

[0204] Furthermore, when voltage over-limit is detected, voltage sensitivity control is triggered to perform reactive power regulation, and based on the sensitivity of the reactive power regulation of the leading control node to the voltage, reactive power regulation is allocated until the voltage over-limit phenomenon in the corresponding target node partition is resolved;

[0205] If the reactive adjustment amount is lower than the target adjustment amount corresponding to the voltage over-limit phenomenon being resolved, active adjustment is triggered; if the reactive adjustment amount is higher than the target adjustment amount, reactive tracking control is triggered;

[0206] The selection of the dominant control node includes: determining the reactive sensitivity and active sensitivity of each node in each target node partition to other nodes in the partition, selecting a node in the partition whose comprehensive sensitivity to other nodes exceeds a preset sensitivity threshold as the dominant control node and configuring a dynamic priority weight;

[0207] The selection of the leading observation node includes: determining a node in each target node partition whose absolute value of voltage fluctuation in a corresponding time period is lower than a preset fluctuation threshold as the leading observation node.

[0208] This implementation uses a hierarchical approach to trigger reactive power tracking control or coordinated control incorporating voltage sensitivity control based on voltage overshooting. The dominant control node is dynamically selected based on the node's combined sensitivity. This mechanism rapidly eliminates overshooting through precise sensitivity-driven regulation. The hierarchical strategy avoids resource waste caused by over-regulation, significantly improving response speed and exception handling efficiency in emergencies.

[0209] It should be noted that it should be understood that the division of the various modules of the above device is merely a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. Moreover, these modules can all be implemented in the form of software called by a processing element; or they can all be implemented in the form of hardware; or some modules can be implemented in the form of software called by a processing element, and some modules can be implemented in the form of hardware. For example, the processing module can be a separately established processing element, or it can be integrated into a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called by a processing element of the above device to perform the functions of the above-mentioned module. The implementation of other modules is similar. In addition, these modules can all or partly be integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each of the above modules can be completed by the hardware integrated logic circuit in the processor element or by instructions in the form of software.

[0210] Schematically, as Figure 4 As shown, Figure 4 This is a schematic diagram of the internal structure of a computer device provided in an embodiment of the present application. The computer device 300 can be provided as a server. Figure 4 Computer device 300 includes a processing component 302, which further includes one or more processors, and a memory resource represented by memory 301 for storing instructions executable by processing component 302, such as an application. The application stored in memory 301 may include one or more modules, each corresponding to a set of instructions. In addition, processing component 302 is configured to execute the instructions to perform the method of any of the above embodiments.

[0211] The computer device 300 may further include a power supply component 303 configured to perform power management of the computer device 300, a wired or wireless network interface 304 configured to connect the computer device 300 to a network, and an input / output (I / O) interface 305. The computer device 300 may operate based on an operating system stored in the memory 301, such as Windows Server™, Mac OS X™, Unix™, Linux™, Free BSD™, or the like.

[0212] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0213] An embodiment of the present application provides a storage medium storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors execute a method as provided in any embodiment.

[0214] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0215] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referenced to each other.

[0216] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A distribution network reactive power and voltage control method based on cooperative game, characterized in that: The method comprises: Based on the day-ahead time scale, the distribution network operation periods are dynamically aggregated to form multiple continuous period clusters; In each of the time period clusters, initial node partitions are generated based on how the node voltage corresponding to each node in the distribution network is affected by photovoltaic reactive output, each of the initial node partitions is optimized using a cooperative game model, and multiple target node partitions are determined based on voltage control costs; Based on the real-time time scale, the corresponding dominant control node in each target node partition is selected, and the voltage fluctuation of each target node partition is controlled through reactive power tracking control or collaborative control.

2. The method according to claim 1, characterized in that Based on the day-ahead time scale, the distribution network operation time periods are dynamically aggregated to form multiple continuous time period clusters, including: Based on the day-ahead time scale, determining a preset number of operating time periods corresponding to the intraday time, determining the number of clusters and the initial cluster center corresponding to each cluster, and determining a multidimensional feature space; The multidimensional feature space is used to indicate the operating characteristics corresponding to each of the operating time periods and the cluster centers obtained by real-time iterative optimization, including photovoltaic output characteristics, active load characteristics, and reactive load characteristics; By iteratively optimizing the optimal segmentation boundaries of adjacent cluster centers, updating the feature vectors corresponding to the real-time cluster centers based on the multidimensional feature space until the feature vectors corresponding to two adjacent iterative optimizations meet the preset convergence conditions, and outputting the time period cluster division results with time period continuity according to the number of clusters; The convergence condition includes that the characteristic fluctuation of the real-time cluster center corresponding to two adjacent iterative optimizations is lower than the convergence threshold, or the preset number of iterations is reached.

3. The method according to claim 2, characterized in that The method for determining the optimal segmentation boundary includes: Calculating the feature differences between each of the running time periods and adjacent cluster centers according to the multidimensional feature space, and selecting the time period with the largest sum of the differences as the current optimal segmentation boundary; The characteristic difference is calculated by weighted accumulation of the Euclidean distances of the photovoltaic output characteristic, the active load characteristic, and the reactive load characteristic.

4. The method according to claim 1, wherein The generating of initial node partitions according to the influence of photovoltaic reactive output on node voltages corresponding to nodes in the distribution network includes: Taking the photovoltaic grid-connected nodes in the distribution network as partition master nodes and determining the number of partitions; In the time period cluster, according to each node to be assigned in the distribution network and each partition master node, a voltage sensitivity coefficient is determined, and node attribution is determined based on the voltage sensitivity coefficient to generate an initial node partition; The voltage sensitivity coefficient is used to indicate the influence of the photovoltaic reactive power output of each partition master node on each of the nodes to be allocated, and is determined according to the voltage over-limit risk probability and the voltage sensitivity coefficient.

5. The method according to claim 1, wherein The construction process of the cooperative game model includes: Determine each of the initial node partitions as a participant set, and determine the voltage control cost corresponding to each of the initial node partitions as a payment function; Determine the constraints of photovoltaic dynamic reactive power regulation and the number of zones; The voltage control cost includes a linear combination of the local voltage deviation, the reactive power regulation voltage fluctuation in the partition, the reactive power regulation voltage fluctuation of other partitions on the partition, and the partition deviation penalty coefficient.

6. The method according to claim 5, characterized in that Optimizing each of the initial node partitions through a cooperative game model and determining a plurality of target node partitions based on voltage control costs includes: Verify the existence of Nash equilibrium solutions in a non-cooperative game framework; Under the condition that there is a Nash equilibrium solution, the optimization goal is to minimize the payment function of each partition. For each adjusted node partition, the partition adjustment plan is determined under the condition that the total payment function of the adjusted node partition does not exceed the payment function of the node partition before adjustment. The partition strategy and topology structure are alternately optimized to reduce the global voltage control cost and determine the final target node partition.

7. The method according to claim 1, characterized in that The reactive power tracking control or the coordinated control is determined according to whether a voltage exceeding a limit occurs in each target node partition, and the coordinated control includes voltage sensitivity control and reactive power tracking control executed in sequence; The selecting of the corresponding dominant control node in each target node partition and controlling the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control specifically includes: Selecting a corresponding dominant control node in each target node partition; For each target node partition, when the voltage fluctuation exceeds the set fluctuation threshold, reactive power tracking control is triggered, and the photovoltaic reactive power output of each node is dynamically adjusted according to the fluctuation of the leading observation node; Furthermore, when voltage over-limit is detected, voltage sensitivity control is triggered to perform reactive power regulation, and based on the sensitivity of the reactive power regulation of the leading control node to the voltage, reactive power regulation is allocated until the voltage over-limit phenomenon in the corresponding target node partition is resolved; If the reactive adjustment amount is lower than the target adjustment amount corresponding to the voltage over-limit phenomenon being resolved, active adjustment is triggered; if the reactive adjustment amount is higher than the target adjustment amount, reactive tracking control is triggered; The selection of the dominant control node includes: determining the reactive sensitivity and active sensitivity of each node in each target node partition to other nodes in the partition, selecting a node in the partition whose comprehensive sensitivity to other nodes exceeds a preset sensitivity threshold as the dominant control node and configuring a dynamic priority weight; The selection of the leading observation node includes: determining a node in each target node partition whose absolute value of voltage fluctuation in a corresponding time period is lower than a preset fluctuation threshold as the leading observation node.

8. A reactive power and voltage control device for a distribution network based on cooperative game, characterized in that: The device comprises: A processing module is used to dynamically aggregate the distribution network operation time periods to form multiple continuous time period clusters based on the day-ahead time scale; The processing module is further configured to generate, within each time period cluster, an initial node partition based on how the node voltage corresponding to each node in the distribution network is affected by photovoltaic reactive output, optimize each of the initial node partitions using a cooperative game model, and determine a plurality of target node partitions based on voltage control costs; The processing module is further configured to select a corresponding dominant control node in each target node partition based on a real-time time scale, and control the voltage fluctuation of each target node partition through reactive power tracking control or collaborative control.

9. A computer device, characterized in that: The method comprises one or more processors and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the one or more processors, the steps of the method according to any one of claims 1 to 7 are performed.

10. A storage medium, characterized in that: The storage medium stores computer-readable instructions, which, when executed by one or more processors, enable the one or more processors to perform the steps of the method according to any one of claims 1 to 7.

Images