Distributed photovoltaic double-layer cluster division method fusing energy storage and photovoltaic regulation and control capability

Abstract

The invention provides a distributed photovoltaic double-layer cluster division method fusing energy storage and photovoltaic regulation and control capability, and belongs to the field of distributed photovoltaic regulation and control of a power system. Firstly, an electrical distance is defined based on an active and reactive sensitivity matrix, and primary clustering is performed on power distribution network nodes; secondly, constructing a comprehensive cluster index fusing photovoltaic reactive power regulation and control capability, energy storage active power regulation and control capability and photovoltaic output consistency, and performing secondary fine clustering on the photovoltaic nodes by adopting an improved Fast unfolding algorithm; and finally, implementing an intra-partition autonomy-inter-partition coordination control strategy according to a cluster result to realize global optimization. According to the method, the comprehensive cluster indexes fusing the electrical distance, the photovoltaic reactive power regulation and control capability, the energy storage active power regulation and control capability and the photovoltaic output consistency are constructed, the distributed photovoltaic cluster division result is optimized, the resource coupling degree and the cooperative regulation and control efficiency in the cluster are improved, and efficient cooperative utilization of energy storage and distributed photovoltaic is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of distributed photovoltaic regulation of power systems, in particular to a distributed photovoltaic double-layer cluster division method fusing the regulation capabilities of energy storage and photovoltaic. BACKGROUND

[0002] At present, new power systems mainly based on new energy systems are developing rapidly. The penetration rate of distributed photovoltaic in distribution networks is increasing day by day, and problems such as insufficient regulation resources and increasing safety risks are becoming increasingly prominent. Large-scale distributed power sources have great differences in type, access location, capacity and regulation characteristics, and their output is volatile and intermittent due to the influence of light. The household capacity of household photovoltaic is relatively small, the number of households is large, the control device has insufficient functions, the automation system is scarce, and the dispatching relationship is loose. It is difficult to operate in a centralized control mode for all distributed photovoltaics, and there are many controllable nodes, which cannot meet the control time scale requirements in the operation stage. Cluster division is performed on distributed photovoltaics, the autonomous characteristics of the cluster are fully utilized, and flexible grid integration is performed, which is an important solution to realize the orderly and efficient access of distributed power generation to distribution networks and complete the mutual coordination and full consumption of power grids and renewable energy power sources.

[0003] Therefore, there is an urgent need for a cluster division method fusing the regulation capabilities of energy storage and photovoltaic inverters to realize more precise distributed photovoltaic cluster management and improve the stability of power system operation and resource utilization. SUMMARY

[0004] The purpose of the present application is to provide a distributed photovoltaic double-layer cluster division method fusing the regulation capabilities of energy storage and photovoltaic. The purpose is to build a comprehensive cluster index fusing electrical distance, photovoltaic reactive power regulation capability, energy storage active power regulation capability and photovoltaic output consistency, optimize the distributed photovoltaic cluster division result, improve the resource coupling degree and collaborative regulation efficiency in the cluster, and realize the efficient collaborative use of energy storage and distributed photovoltaic.

[0005] To achieve the above purpose, the following technical solutions are used.

[0006] The distributed photovoltaic double-layer cluster division method fusing the regulation capabilities of energy storage and photovoltaic includes the following steps: S1, define the electrical distance based on the active and reactive sensitivity matrix, and perform primary clustering on the nodes of the distribution network according to the electrical distance; S2, build a comprehensive cluster index, the comprehensive cluster index fuses photovoltaic reactive power regulation capability, energy storage active power regulation capability and photovoltaic output consistency, and the expression of the comprehensive cluster index is: In the formula, , , The weighting coefficients for different indicators are respectively , , , These are the normalized reactive power regulation capability coefficient of photovoltaic inverter, active power regulation capability coefficient of energy storage, and photovoltaic output consistency coefficient, respectively. S3. The improved Fast unfolding algorithm is used to perform secondary clustering of photovoltaic nodes. Based on the clustering results, the autonomous regulation within the partition and the coordinated mutual control strategy between the partitions are implemented.

[0007] Furthermore, the electrical distance described in step S1 is defined as follows: Establish the nodal voltage-active and voltage-reactive sensitivity matrices using the Jacobian matrix: in, , , , These represent the changes in active power, reactive power, voltage phase, and voltage amplitude, respectively, and H, N, J, and L are elements of the Jacobian matrix. After sorting, we get: In the formula, S P S is the voltage / active power sensitivity matrix. Q This is the voltage / reactive power sensitivity matrix, where S P and S QF They are respectively: In the formula, n is the number of nodes, where Let be the change in active power at node j corresponding to the change in voltage amplitude at node i; where The change in reactive power at node j corresponds to the change in voltage amplitude at node i. Therefore, the electrical distance is defined as: In the formula: .

[0008] Furthermore, the clustering process of a node in step S1 includes the following steps: S11. Randomly select k nodes as initial cluster centers; S12. Calculate the electrical distance between each node and the cluster center, and assign each node to the cluster corresponding to the nearest center; S13. Update the cluster centers based on the node distribution density; S14. Repeat steps S12-S13 until the cluster partitioning results converge.

[0009] Furthermore, the quantification method for photovoltaic reactive power regulation capability in step S2 is as follows: In the formula, It is the rated capacity of the grid-connected photovoltaic inverter at node i. It is the active power of the photovoltaic grid-connected node i. It refers to the reactive power regulation capacity of the photovoltaic inverter; The reactive power regulation capability coefficient of the photovoltaic inverter is: In the formula, m1 represents the number of distributed photovoltaic nodes within the cluster. This represents the maximum reactive power regulation capacity of the photovoltaic inverter. Normalize the reactive power regulation capability coefficient of the photovoltaic inverter: In the formula, To find the minimum reactive power regulation coefficient for all photovoltaic inverters, This is to iterate through the maximum reactive power regulation coefficient of all photovoltaic inverters.

[0010] Furthermore, the quantification method for the active power regulation capability of energy storage in step S2 is as follows: Define the active power regulation capability coefficient of energy storage as: In the formula: n1 is the number of energy storage nodes in the cluster, and n2 is the number of nodes in the cluster. The maximum energy storage discharge power of node i; Normalize the active power regulation coefficient of the energy storage cluster: In the formula: To find the minimum active power regulation coefficient for all energy storage systems, This represents the maximum active power regulation coefficient obtained by iterating through all energy storage systems.

[0011] Furthermore, the quantification method for photovoltaic output consistency in step S2 is as follows: The photovoltaic output consistency coefficient is defined as: In the formula: and This represents the power output of the photovoltaic power stations at nodes i and j, where... For the distributed photovoltaic relative coefficients of nodes i and j: In the formula: , For nodes i and nodes j Distributed photovoltaic at t Active power at any given time Representation Nodes i and nodes j In t Consistency of output at any given time: When the output curves of the two nodes are consistent. The value is 1.

[0012] Normalization of consistency coefficients ; Where m is the number of nodes in the cluster, and n represents the total number of nodes.

[0013] Furthermore, the improved Fast unfolding algorithm described in step S3 includes the following steps: Treat each distributed photovoltaic node as a community and calculate its modularity at this point; Iterate through each node, move it to an adjacent community, calculate the sum of the indicators after moving to the adjacent community, and update the community division result with the largest sum of indicators as the best community. Update the community label of the node, check if convergence has been achieved, and exit the loop if the community label no longer changes.

[0014] Furthermore, the autonomous regulation strategy within the partition described in step S3 includes: Distributed photovoltaic inverter cluster autonomous control: Based on the real-time load demand and voltage status within the zone, the cluster controller issues mode commands to the photovoltaic inverters within its own cluster to dynamically adjust the reactive power output. Autonomous regulation of energy storage system within the cluster: Based on the fluctuation of photovoltaic output and load changes within the zone, the energy storage system smooths out power fluctuations through charging and discharging strategies to achieve photovoltaic absorption and load matching. When the photovoltaic output is greater than the load within the zone, the energy storage system absorbs excess energy at the maximum charging power not exceeding its rated charging power.

[0015] Furthermore, the mode command allows for setting different control modes, including: Maximum power mode: When the load demand in the zone is high and the voltage is stable, the photovoltaic inverter operates in maximum power point tracking mode, with reactive power output of 0, prioritizing active power supply. Constant Proportional Reactive Power Mode: When the voltage within a zone deviates from the rated value, reactive power output is distributed according to a preset ratio based on the inverter's rated capacity and real-time active power output, as shown in the following formula: In the formula, The reactive power output of the photovoltaic inverter at node i is... The reactive power distribution factor (0 < ≤1), which is dynamically adjusted by the cluster controller based on the voltage deviation.

[0016] Furthermore, the triggering condition for the inter-regional coordinated and mutually controlling control described in step S3 is: The cluster controller in each partition collects data in real time and executes autonomous control strategies within the partition. When autonomous control within a partition fails to eliminate power deficit or voltage deviation, the inter-partition coordination mechanism is triggered. The main controller of the distribution network receives information from each zone, solves the global optimization model and issues coordination instructions. Each zone executes the coordination instructions and adjusts its local control strategy. The main controller of the distribution network updates the control effect in real time until the global indicators are met. The global optimization model takes the minimization of photovoltaic active power reduction and node voltage deviation within the region as its objective functions. In the formula: It is a collection of distributed photovoltaic nodes. Let k be the active power reduction of the photovoltaic node. For nodes i Voltage, Where N is the node reference voltage, and N is the set of nodes; The global optimization model satisfies the following constraints: (1) System power flow constraints In the formula: Let J be the power flowing from node j to node i. Let be the power flowing from node i to node k. Let i be the upstream node of node i. Let i be the downstream node of node i. , These represent the line resistance and reactance between nodes ij, respectively. , Let represent the active power and reactive power of node i, respectively. This represents the current at node ij. , These represent the active and reactive power outputs of the distributed power source connected to node i. , These represent the active and reactive power flowing between nodes ij, respectively. (2) Node voltage constraints The voltage amplitude at each load node is within the normal operating range: In the formula: , They are nodes i Maximum and minimum allowable voltages; Network separation is achieved using a decomposition and coordination method. The boundary node voltage equality constraints between adjacent clusters and the corresponding line transmission power equality constraints between adjacent clusters are as follows: In the formula: Indicates the upstream cluster boundary node a The square value of the voltage, Represents the virtual balancer node of the downstream cluster. a* The square value of the voltage, and These represent the upstream cluster boundary nodes respectively. a The virtual load active and reactive power, and These represent inter-group lines. am The transmitted active and reactive power, Represents boundary nodes a Voltage squared, and These represent inter-group lines. am Global values ​​for transmitted active and reactive power. This represents the set of inter-group lines.

[0017] The advantages of this invention are: Traditional cluster division only considers a single dimension such as electrical distance or output characteristics. This invention integrates the active power regulation capability of energy storage, the reactive power regulation capability of photovoltaics, and the output consistency into a comprehensive index, and the index weight is adjustable, which can be adapted to different distribution network scenarios. It adopts a two-layer control architecture of intra-regional autonomy and inter-regional mutual governance, combined with the dynamic switching mode of photovoltaic inverters, and uses a global optimization model to trigger coordination to improve the response speed of centralized control and solve the problem of cross-regional resource imbalance. The improved Fast unfolding algorithm reduces time complexity and its computational efficiency is more suitable for large-scale power grids. Attached Figure Description

[0018] Figure 1 This is a flowchart of the distributed photovoltaic two-layer cluster partitioning method that integrates energy storage and photovoltaic regulation capabilities according to the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0020] This embodiment proposes a distributed photovoltaic two-layer cluster partitioning method that integrates energy storage and photovoltaic regulation capabilities. The principle diagram is shown below. Figure 1 As shown, firstly, based on the active and reactive power sensitivity matrix, electrical distance is defined, and the distribution network nodes are clustered to achieve preliminary partitioning at the network topology level. Secondly, a comprehensive clustering index integrating photovoltaic reactive power regulation capability, energy storage active power regulation capability, and photovoltaic output consistency is constructed. An improved Fast unfolding algorithm is used to perform secondary fine-grained clustering of photovoltaic nodes, taking into account both resource regulation characteristics and output synergy. Finally, a control strategy of intra-partition autonomy and inter-partition coordination is implemented based on the clustering results. Within each partition, the distributed photovoltaic and energy storage output of each cluster is regulated, and between partitions, global optimization is achieved through coordinated control and resource complementarity.

[0021] S1, Distribution network node cluster based on electrical distance.

[0022] The purpose of clustering is to optimize voltage control; therefore, the electrical distance between distribution network nodes is defined based on voltage-active power sensitivity and voltage-reactive power sensitivity. Nodes with close electrical distances, i.e., nodes with high voltage sensitivity to each other, are clustered together to achieve high coupling within the cluster and low coupling between nodes in the cluster.

[0023] First, consider the Jacobian matrix in the Newton-Raphson power flow method to establish the node voltage-active and voltage-reactive sensitivity matrices.

[0024] in, , , , These represent the changes in active power, reactive power, voltage phase, and voltage amplitude, respectively. H, N, J, and L are Jacobian matrix elements, reflecting the relationships between active power, phase angle, reactive power, and voltage in the system.

[0025] Expanding the above equation, we get: voltage phase Eliminate to get: In the formula, S P S is the voltage / active power sensitivity matrix. Q This is the voltage / reactive power sensitivity matrix, where S P and SQF They are respectively: In the formula, n is the number of nodes, where Let be the change in active power at node j corresponding to the change in voltage amplitude at node i; where The change in reactive power at node j corresponds to the change in voltage amplitude at node i. Therefore, the electrical distance is defined as: In the formula: .

[0026] A single clustering process for nodes includes the following steps: S11. Randomly select k nodes as initial cluster centers; S12. Calculate the electrical distance between each node and the cluster center, and assign each node to the cluster corresponding to the nearest center; S13. Update the cluster centers based on the node distribution density; S14. Repeat steps S12-S13 until the cluster partitioning results converge, the clustering ends, and the clustering results are output.

[0027] S2. Distributed photovoltaic clusters considering energy storage and photovoltaic inverter control capabilities.

[0028] By considering electrical distance, the distribution network nodes are clustered. Based on this, the reactive power regulation capability of the photovoltaic inverters in each cluster, the active power regulation capability of the energy storage, and the consistency of the distributed photovoltaic output within the cluster are considered to divide the distributed photovoltaic and energy storage into clusters.

[0029] S21, reactive power regulation capability of cluster photovoltaic inverters.

[0030] Distributed photovoltaic (PV) systems provide active power to the grid while also having the ability to flexibly adjust reactive power. By absorbing or generating reactive power, they can regulate the voltage at the PV grid connection point. The adjustable reactive power capacity is related to the rated capacity of the inverter and the active power generated by the PV grid connection.

[0031] In the formula, It is the rated capacity of the grid-connected photovoltaic inverter at node i. It is the active power of grid-connected photovoltaic system at node i. It refers to the reactive power regulation capacity of the photovoltaic inverter.

[0032] The reactive power regulation capability coefficient of the photovoltaic inverter is: In the formula, m1 represents the number of distributed photovoltaic nodes within the cluster. This represents the maximum reactive power regulation capacity of the photovoltaic inverter. Because multiple index clusters are required in the final stage, the reactive power regulation capability coefficient of the photovoltaic inverter is normalized to avoid differences in the dimensions of different cluster characteristic indicators: In the formula, To find the minimum reactive power regulation coefficient for all photovoltaic inverters, This is to iterate through the maximum reactive power regulation coefficient of all photovoltaic inverters.

[0033] Distributed photovoltaic (PV) systems, in terms of reactive power reserve capacity, can provide necessary reactive power support to the distribution network during emergencies such as power grid failures. While many residential distributed PV systems are connected to low-voltage distribution networks, their individual capacities are small and their distribution nodes are numerous. Traditional substations cannot directly perform real-time scheduling and control of residential distributed PV systems, resulting in a waste of reactive power resources.

[0034] S22, Cluster energy storage active power regulation capability.

[0035] The active power regulation capability of energy storage is characterized by its role in supporting the active power of the distribution network. Building upon the existing reactive power regulation of distributed photovoltaic systems, it considers the active power regulation of energy storage devices within the cluster. This means that each energy storage device is rationally distributed within each cluster, ensuring sufficient active power regulation capability within each cluster. The active power regulation capability coefficient of energy storage is: In the formula: n1 is the number of energy storage nodes in the cluster, and n2 is the number of nodes in the cluster. The maximum energy storage discharge power of node i; Normalize the active power regulation coefficient of the energy storage cluster: In the formula: To find the minimum active power regulation coefficient for all energy storage systems, This represents the maximum active power regulation coefficient obtained by iterating through all energy storage systems.

[0036] S23, Consistency of output of clustered distributed photovoltaic power.

[0037] Distributed photovoltaic (PV) consistency parameters are used to group distributed PV power plants with similar dynamic characteristics into the same cluster based on their daily output curves, thereby evaluating the output and reactive power regulation capabilities of the distributed PV clusters. The PV output consistency coefficient is: In the formula: and This represents the power output of the photovoltaic power stations at nodes i and j, where... For the distributed photovoltaic relative coefficients of nodes i and j: In the formula: , For nodes i and nodes j Distributed photovoltaic at t Active power at any given time Representation Nodes i and nodes j In t Consistency of output at any given time: When the output curves of the two nodes are consistent. The value is 1.

[0038] Normalization of the consistency coefficient: ; Where m is the number of nodes in the cluster, and n represents the total number of nodes.

[0039] If the output curves within the cluster are highly consistent and the overall output fluctuations are uniform, it is possible to coordinate with energy storage for scheduling optimization and to provide reasonable energy storage and buffering for distributed photovoltaics.

[0040] S24. Construct comprehensive cluster indicators.

[0041] Considering factors such as the reactive power regulation capability of distributed photovoltaic (PV) systems, the active power regulation capability of energy storage systems, and the consistency of distributed PV systems, the following comprehensive index is established for clustering distributed PV systems: In the formula, , , The weighting coefficients for different indicators are respectively In this embodiment, we take , , , These are the normalized reactive power regulation capability coefficient of photovoltaic inverters, the active power regulation capability coefficient of energy storage, and the photovoltaic output consistency coefficient, respectively.

[0042] S3. Cluster partitioning method using the Fast unfolding algorithm.

[0043] Developing a distributed photovoltaic (PV) cluster can be equivalent to dividing the distributed PV nodes into communities. Each node attempts to traverse the community labels of all its neighbors and selects the community label that maximizes the modularity increment. The final optimization objective is to maximize the modularity index of the entire community network. The community modularity index reflects the quality of community division. Furthermore, the Fast unfolding algorithm has low time complexity and is suitable for large-scale low-voltage distribution network networks.

[0044] By improving the traditional modular index that considers node connections, a comprehensive cluster index is adopted as the improved modular index.

[0045] S31. Treat each photovoltaic node as a community and calculate the modularity at this time; S32. Iterate through each node and try to move it to an adjacent community. Calculate the sum of the indicators after moving to an adjacent community, and update the best community, which is the community with the largest sum of indicators. S33. Update the community label of the node. Check if convergence has occurred. If the community label no longer changes, exit the loop.

[0046] S4. Based on the cluster partitioning results, implement autonomous regulation within the partition and coordinated mutual control strategy between partitions.

[0047] Within each zone, a cluster controller is used to achieve localized autonomous control, fully leveraging the reactive power control capabilities of the photovoltaic inverters and the active power control capabilities of the energy storage within the cluster, prioritizing power balance and voltage stability within the zone. Based on the distributed photovoltaic cluster results in S1 and S3, the distributed photovoltaic systems of each cluster within the zone are regulated, changing their operating strategies, while simultaneously regulating the energy storage output within each zone.

[0048] Autonomous regulation strategies within a region include: S41, Autonomous control of distributed photovoltaic inverter clusters.

[0049] The cluster controller issues mode commands to the photovoltaic inverters within its own cluster to dynamically adjust reactive power output. The main different control modes are as follows: Maximum power mode: When the load demand in the zone is high and the voltage is stable, the photovoltaic inverter operates in maximum power point tracking mode, with reactive power output of 0, prioritizing active power supply. Constant Proportional Reactive Power Mode: When the voltage within a zone deviates from the rated value, reactive power output is distributed according to a preset ratio based on the inverter's rated capacity and real-time active power output, as shown in the following formula: In the formula, The reactive power output of the photovoltaic inverter at node i is... The reactive power distribution factor (0 < ≤1), which is dynamically adjusted by the cluster controller based on the voltage deviation.

[0050] S42, Autonomous regulation of energy storage systems within the cluster.

[0051] Based on the fluctuations in photovoltaic output and load changes within the zone, the energy storage system smooths out power fluctuations through charging and discharging strategies, achieving photovoltaic absorption and load matching. When the photovoltaic output exceeds the load within the zone, the energy storage system absorbs excess energy at a maximum charging power not exceeding its rated charging power.

[0052] S43, Inter-regional coordination and mutual governance strategy.

[0053] While cluster autonomous optimization control can quickly eliminate voltage exceedances within the cluster, it cannot dispatch reactive power resources from other zones, potentially leading to unnecessary photovoltaic power generation losses. When autonomous regulation within a single zone cannot meet the demand—for example, due to limited photovoltaic regulation capacity or sudden load surges causing low voltage levels—cross-zone coordination is achieved through inter-cluster communication. This leverages the resource complementarity of different zones to achieve global power balance and voltage stability. Control strategy implementation process: (1) The cluster controller in each partition collects data within the partition in real time and executes autonomous control strategies within the partition; (2) When the autonomous control within the zone fails to eliminate the power deficit or voltage deviation, the inter-zone coordination mechanism is triggered; (3) The main controller of the distribution network receives information from each zone, solves the global optimization model, and issues coordination instructions; (4) Each zone executes coordination instructions, adjusts local control strategies, and the main controller of the distribution network updates the control effect in real time until the global indicators meet the standards.

[0054] The method for solving the global optimization model is as follows: The interval-based global optimization model takes the minimization of photovoltaic active power reduction and node voltage deviation within the interval as its objective functions. In the formula: It is a collection of distributed photovoltaic nodes. Let k be the active power reduction of the photovoltaic node. For nodes i Voltage, Let N be the node reference voltage, and N be the set of nodes.

[0055] To ensure the normal operation of the system, the global optimization model must satisfy the following constraints: (1) System power flow constraints In the formula: Let J be the power flowing from node j to node i. Let be the power flowing from node i to node k. Let i be the upstream node of node i. Let i be the downstream node of node i. , These represent the line resistance and reactance between nodes ij, respectively. , Let represent the active power and reactive power of node i, respectively. This represents the current at node ij. , These represent the active and reactive power flowing between nodes ij, respectively. This represents the connection state of branch ij. A value of 1 indicates that the line is closed, and a value of 0 indicates that it is open.

[0056] (2) Node voltage constraints The voltage amplitude at each load node is within the normal operating range: In the formula: , They are nodes i Maximum and minimum allowable voltages; Network separation is the foundation for autonomous optimization within partitions and distributed coordination optimization between groups. A decomposition coordination method is used to achieve network separation. The boundary nodes of the upstream cluster are "copied" to the downstream cluster as virtual balancing nodes, while the power transmitted on the inter-group lines serves as the virtual load power of the upstream boundary nodes.

[0057] Inter-cluster distributed coordinated optimization requires the addition of equality constraints on boundary node voltages and inter-cluster line power to allow each cluster to perform independent parallel optimization and ensure the convergence of the inter-cluster distributed optimization. The following equations represent the boundary node voltage equality constraints for adjacent clusters and the corresponding inter-cluster line transmission power equality constraints, respectively: In the formula: Indicates the upstream cluster boundary node a The square value of the voltage, Represents the virtual balancer node of the downstream cluster. a* The square value of the voltage, and These represent the upstream cluster boundary nodes respectively. a The virtual load active and reactive power, and These represent inter-group lines. am Transmitted active and reactive power, Represents boundary nodes a Voltage squared, and These represent inter-group lines. am Global values ​​for transmitted active and reactive power. This represents the set of inter-group lines.

[0058] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for partitioning a distributed photovoltaic two-layer cluster that integrates energy storage and photovoltaic regulation capabilities, characterized in that, Including the following steps: S1. Define the electrical distance based on the active and reactive power sensitivity matrix, and perform a clustering of distribution network nodes according to the electrical distance; S2. Construct a comprehensive cluster index, which integrates photovoltaic reactive power control capability, energy storage active power control capability, and photovoltaic output consistency. The expression for the comprehensive cluster index is: In the formula, , , The weighting coefficients for different indicators are respectively , , , These are the normalized reactive power regulation capability coefficient of photovoltaic inverter, active power regulation capability coefficient of energy storage, and photovoltaic output consistency coefficient, respectively. S3. The improved Fast unfolding algorithm is used to perform secondary clustering of photovoltaic nodes. Based on the clustering results, the autonomous regulation within the partition and the coordinated mutual control strategy between the partitions are implemented.

2. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The electrical distance described in step S1 is defined as follows: Establish the nodal voltage-active and voltage-reactive sensitivity matrices using the Jacobian matrix: in, , , , These represent the changes in active power, reactive power, voltage phase, and voltage amplitude, respectively, and H, N, J, and L are elements of the Jacobian matrix. After sorting, we get: In the formula, S P S is the voltage / active power sensitivity matrix. Q This is the voltage / reactive power sensitivity matrix, where S P and S QF They are respectively: In the formula, n is the number of nodes, where Let be the change in active power at node j corresponding to the change in voltage amplitude at node i; where The change in reactive power at node j corresponds to the change in voltage amplitude at node i. Therefore, the electrical distance is defined as: In the formula: .

3. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The clustering process of a node in step S1 includes the following steps: S11. Randomly select k nodes as initial cluster centers; S12. Calculate the electrical distance between each node and the cluster center, and assign each node to the cluster corresponding to the nearest center; S13. Update the cluster centers based on the node distribution density; S14. Repeat steps S12-S13 until the cluster partitioning results converge.

4. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The quantification method for photovoltaic reactive power regulation capability in step S2 is as follows: In the formula, It is the rated capacity of the grid-connected photovoltaic inverter at node i. It is the active power of the photovoltaic grid-connected node i. It refers to the reactive power regulation capacity of the photovoltaic inverter; The reactive power regulation capability coefficient of the photovoltaic inverter is: In the formula, m1 represents the number of distributed photovoltaic nodes within the cluster. This represents the maximum reactive power regulation capacity of the photovoltaic inverter. Normalize the reactive power regulation capability coefficient of the photovoltaic inverter: In the formula, To find the minimum reactive power regulation coefficient for all photovoltaic inverters, This is to iterate through the maximum reactive power regulation coefficient of all photovoltaic inverters.

5. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The quantification method for the active power regulation capability of energy storage in step S2 is as follows: Define the active power regulation capability coefficient of energy storage as: In the formula: n1 is the number of energy storage nodes in the cluster, and n2 is the number of nodes in the cluster. The maximum energy storage discharge power of node i; Normalize the active power regulation coefficient of the energy storage cluster: In the formula: To find the minimum active power regulation coefficient for all energy storage systems, This represents the maximum active power regulation coefficient obtained by iterating through all energy storage systems.

6. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The quantification method for photovoltaic power output consistency in step S2 is as follows: The photovoltaic output consistency coefficient is defined as: In the formula: and This represents the power output of the photovoltaic power stations at nodes i and j, where... For the distributed photovoltaic relative coefficients of nodes i and j: In the formula: , For nodes i and nodes j Distributed photovoltaic In t Active power at any given time Representation Nodes i and nodes j In t Consistency of output at any given time: When the output curves of the two nodes are consistent. The value is 1. Normalization of consistency coefficients Where m is the number of nodes in the cluster, and n represents the total number of nodes.

7. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The improved Fast unfolding algorithm described in step S3 includes the following steps: Treat each distributed photovoltaic node as a community and calculate its modularity at this point; Iterate through each node, move it to an adjacent community, calculate the sum of the indicators after moving to the adjacent community, and update the community division result with the largest sum of indicators as the best community. Update the community label of the node, check if convergence has been achieved, and exit the loop if the community label no longer changes.

8. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The autonomous regulation strategy within the partition mentioned in step S3 includes: Distributed photovoltaic inverter cluster autonomous control: Based on the real-time load demand and voltage status within the zone, the cluster controller issues mode commands to the photovoltaic inverters within its own cluster to dynamically adjust the reactive power output. Autonomous regulation of energy storage system within the cluster: Based on the fluctuation of photovoltaic output and load changes within the zone, the energy storage system smooths out power fluctuations through charging and discharging strategies to achieve photovoltaic absorption and load matching. When the photovoltaic output is greater than the load within the zone, the energy storage system absorbs excess energy at the maximum charging power not exceeding its rated charging power.

9. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 8, characterized in that, The mode command is for setting different control modes, including: Maximum power mode: When the load demand in the zone is high and the voltage is stable, the photovoltaic inverter operates in maximum power point tracking mode, with reactive power output of 0, prioritizing active power supply. Constant Proportional Reactive Power Mode: When the voltage within a zone deviates from the rated value, reactive power output is distributed according to a preset ratio based on the inverter's rated capacity and real-time active power output, as shown in the following formula: In the formula, The reactive power output of the photovoltaic inverter at node i is... The reactive power distribution factor (0 < ≤1), which is dynamically adjusted by the cluster controller based on the voltage deviation.

10. The distributed photovoltaic two-layer cluster partitioning method integrating energy storage and photovoltaic regulation capabilities according to claim 1, characterized in that, The triggering condition for the inter-regional coordinated and mutually controlling control described in step S3 is: The cluster controller in each partition collects data in real time and executes autonomous control strategies within the partition. When autonomous control within a partition fails to eliminate power deficit or voltage deviation, the inter-partition coordination mechanism is triggered. The main controller of the distribution network receives information from each zone, solves the global optimization model and issues coordination instructions. Each zone executes the coordination instructions and adjusts its local control strategy. The main controller of the distribution network updates the control effect in real time until the global indicators are met. The global optimization model takes the minimization of photovoltaic active power reduction and node voltage deviation within the region as its objective functions. In the formula: It is a collection of distributed photovoltaic nodes. Let k be the active power reduction of the photovoltaic node. For nodes i Voltage, Where N is the node reference voltage, and N is the set of nodes; The global optimization model satisfies the following constraints: (1) System power flow constraints In the formula: Let J be the power flowing from node j to node i. Let be the power flowing from node i to node k. Let i be the upstream node of node i. Let i be the downstream node of node i. , These represent the line resistance and reactance between nodes ij, respectively. , Let represent the active power and reactive power of node i, respectively. This represents the current at node ij. , These represent the active and reactive power outputs of the distributed power source connected to node i. , These represent the active and reactive power flowing between nodes ij, respectively. This represents the connection status of branch ij. A value of 1 indicates that the line is closed, and a value of 0 indicates that it is open. (2) Node voltage constraints The voltage amplitude at each load node is within the normal operating range: In the formula: , They are nodes i Maximum and minimum allowable voltages; Network separation is achieved using a decomposition and coordination method. The boundary node voltage equality constraints between adjacent clusters and the corresponding line transmission power equality constraints between adjacent clusters are as follows: In the formula: Indicates the upstream cluster boundary node a The square value of the voltage, Represents the virtual balancer node of the downstream cluster. a* The square value of the voltage, and These represent the upstream cluster boundary nodes respectively. a The virtual load active and reactive power, and These represent inter-group lines. am The transmitted active and reactive power, Represents boundary nodes a Voltage squared, and These represent inter-group lines. am Global values ​​for transmitted active and reactive power. This represents the set of inter-group lines.

Images