A power distribution network voltage out-of-limit control method based on community division and related equipment

Abstract

This invention discloses a two-stage voltage over-limit control method and related equipment for distribution networks based on community partitioning. First, a modified particle competition learning algorithm is used to partition the nodes of a photovoltaic (PV) distribution network into several communities, enhancing the rationality and stability of the partitioning. Second, a voltage over-limit control model based on linearized power flow equations is established within each community. When a node voltage exceeds its limit, rapid autonomous control is achieved by adjusting the reactive power of the SVG (Static Var Generator) and the active and reactive power of the PV within the community, eliminating local voltage over-limits. Third, after autonomous control within the community is completed, power and voltage information of the community boundary nodes is collected. A distributed optimization method is used to coordinate and regulate the SVG and PV across communities, and the power flow equations are used to correct the voltage of the boundary nodes, thereby achieving global optimization of the entire network voltage.

Description

Technical Field

[0001] This invention relates to the field of power distribution networks, and specifically to a method and related equipment for controlling voltage over-limit in power distribution networks based on community division. Background Technology

[0002] Photovoltaic power generation, as a clean and renewable energy source, has become a crucial pillar for my country's energy structure transformation and green, low-carbon development. However, due to the randomness, volatility, and intermittency of photovoltaic power, high-proportion photovoltaic (PV) grid integration brings numerous challenges, including but not limited to voltage exceedance, increased network losses, and disruptions to system protection. Research indicates that the randomness and volatility of PV power generation can exacerbate voltage fluctuations at distribution network nodes, even leading to voltage exceedances, thus affecting user electricity safety and the stable operation of the distribution network. Furthermore, PV grid connection may trigger reverse power flow, further complicating voltage exceedance control. In the future, the penetration rate of distributed PV distribution networks will further increase, leading to more complex voltage exceedance issues that centralized control methods will be insufficient to meet requirements.

[0003] Currently, in the field of voltage overrun control in distribution networks, scholars both domestically and internationally have conducted a series of studies. Existing control methods do not fully utilize the system's controllable resources and have limited voltage regulation capabilities. At the model and algorithm level, centralized control relies on high-precision prediction, resulting in poor robustness under uncertain scenarios; distributed control presents a significant contradiction between convergence speed and communication reliability; and the sensitivity of local control parameters and system stability are difficult to balance. Therefore, the solution to these problems has shifted to a community control approach of "autonomy within the community and coordination between communities." This control model significantly simplifies the optimization process, substantially shortens the problem-solving time, and greatly reduces the difficulty of distribution network operation control, making it suitable for future voltage overrun control and scheduling of distribution networks with a high proportion of distributed power sources. Summary of the Invention

[0004] The technical problem this invention aims to solve is that in distribution networks with a high proportion of photovoltaic (PV) integration, node voltage is prone to exceeding limits. Traditional centralized control methods are computationally complex and have slow response times, making it difficult to balance global and real-time performance, resulting in unsatisfactory voltage optimization control effects. The purpose is to provide a distribution network voltage limit control method and related equipment based on community division. By combining community division, intra-community voltage autonomous control, and inter-community voltage collaborative control, it can quickly eliminate local voltage limits and achieve global optimization of the entire network voltage. This solves the problem that existing voltage control methods struggle to balance local autonomy and global coordination, improves the accuracy and flexibility of voltage limit management, and ensures the safe and stable operation of distribution networks containing PV.

[0005] This invention is achieved through the following technical solution:

[0006] A distribution network voltage over-limit control method based on community division includes:

[0007] Community division: Collect operational data of the photovoltaic power distribution network and divide it into communities. Clustering is performed based on the improved particle competition learning algorithm. The initial position is determined according to the node influence factor. The frog leap optimization algorithm is used to guide the movement of particles and iteratively update until convergence to obtain the community division result.

[0008] Community-based voltage over-limit control: Collect node load, node voltage, and photovoltaic power within the community, set safe operation constraints, and establish a community-based voltage over-limit control model based on linearized power flow balance equations. The voltage over-limit control model aims to minimize the node voltage deviation rate. Set the decision variables as reactive power of SVG, active power of photovoltaics, and reactive power. Update the node voltage according to the solution results and iterate until the change in node voltage between two adjacent times does not exceed the preset threshold, so as to eliminate voltage over-limit within the community.

[0009] Global voltage optimization between communities: Voltage and power data of boundary nodes of each community are collected, and boundary voltage consistency constraints and boundary power balance constraints are added. A global voltage optimization model between communities is established with the goal of minimizing the sum of the active power regulation cost of the photovoltaic system and the distribution network loss. The decision variables are set as the reactive power of SVG, the active power of photovoltaic and the reactive power. Distributed optimization using the alternating direction multiplier method is used for coordinated solution. After each round of alternation, adjacent community units exchange boundary data and residuals, and boundary data correction and voltage compensation are updated based on the power flow equation until the preset convergence condition is met. The voltage regulation control quantity and operation setpoint of each community unit are output.

[0010] Furthermore, the community partitioning uses an influence factor to select the initial position and updates the particles using a leapfrog optimization algorithm to achieve global convergence. The influence factor is represented as:

[0011]

[0012] in, For nodes The initial influence value, For nodes The set of adjacent nodes, It is a node and nodes The weighted adjacency matrix between them and These are the minimum and maximum values ​​of the non-zero elements in the weighted adjacency matrix, respectively. This represents the total number of nodes in the distribution network.

[0013] The update representation of the frog-jump optimization algorithm is as follows:

[0014]

[0015]

[0016] in, For frog jump stride length, and These are the current best and worst particle positions. For the updated particle positions, For the maximum step size, This indicates the generation of random numbers between [0, 1].

[0017] Furthermore, the safe operation within the community satisfies node voltage safety constraints and equipment constraints, wherein the node voltage safety constraints are:

[0018]

[0019] The equipment constraints for photovoltaics are:

[0020]

[0021]

[0022]

[0023] The device constraints for SVG are:

[0024]

[0025] in, The maximum power factor angle of the photovoltaic equipment's output power. For nodes The available active power of photovoltaic equipment For nodes Active power regulation of photovoltaic equipment For nodes The reactive power output of photovoltaic equipment For nodes The installed capacity of photovoltaic equipment For SVG reactive power output, and They are nodes The lower and upper limits of reactive power of the SVG.

[0026] Furthermore, the voltage over-limit control model aims to minimize the sum of node voltage deviation rates, and its objective function is expressed as:

[0027]

[0028] in, The value is 1.05 pu. This refers to the node voltage after voltage regulation.

[0029] In the process of modeling the voltage over-limit control model, the linearized power flow equation constraint is:

[0030]

[0031] in,

[0032]

[0033] in, and For nodes The active and reactive power of the load. Represents a node and nodes The reactance value of the line, Represents a node and nodes The resistance value of the circuit. and They are nodes The active and reactive power of the load. For nodes Active power output at the maximum power point of photovoltaic power generation For nodes Photovoltaic active power regulation For nodes The reactive power output of photovoltaic power, For nodes The SVG outputs reactive power.

[0034] Furthermore, the objective function for the inter-community voltage coordinated control is:

[0035]

[0036] in, Indicates community The objective function of the voltage over-limit control model. and These represent the revenue generated by photovoltaic equipment and the on-grid electricity price for active power, respectively. For nodes Photovoltaic equipment active power regulation capacity, For nodes The reactive power output of photovoltaic equipment For nodes The reactive power output of the SVG Represents a node and nodes The resistance value of the circuit. and Representing nodes respectively Inflow branch Active power and reactive power, By node Pointing to node A side road.

[0037] Furthermore, the added boundary voltage consistency constraint and boundary power balance constraint are expressed as follows:

[0038]

[0039] in, As the boundary node of the community, This represents the square of the voltage at the upstream community boundary node. The square of the voltage at the virtual balancing node in the downstream community and These represent inter-community routes. Transmitted active and reactive power, and This corresponds to the virtual load equivalent. Represents boundary nodes Globally consistent variables, and These represent inter-community routes. Globally consistent variables for transmitted active and reactive power.

[0040] Furthermore, the inter-community collaborative optimization employs the alternating direction multiplier method for distributed solution, and the augmented Lagrangian function of the objective function can be expressed as:

[0041]

[0042] After each cycle of alternation, the boundary data are corrected using the power flow equations, and compensation parameters are applied. and Update the node voltage consistency constraints at the community boundary;

[0043] The compensated node voltage safety constraint is expressed as follows:

[0044]

[0045] in, For the community highest voltage With the corresponding node voltage difference, For the community minimum voltage The corresponding node voltages of and difference.

[0046] This invention also provides a distribution network voltage over-limit control system based on community division, used to implement the aforementioned distribution network voltage over-limit control method based on community division, comprising:

[0047] The community division unit collects operation data of the photovoltaic power distribution network and divides it into communities. Clustering is performed based on the improved particle competition learning algorithm. The initial position is determined according to the node influence factor. The frog leap optimization algorithm is used to guide the movement of particles and iteratively update until convergence to obtain the community division result.

[0048] The community-based autonomous control unit collects node loads, node voltages, and photovoltaic power within the community, sets safe operation constraints, and establishes a voltage over-limit control model based on linearized power flow balance equations. The voltage over-limit control model aims to minimize the node voltage deviation rate, and sets the decision variables as the reactive power of SVG, the active power of photovoltaics, and the reactive power. The node voltage is updated according to the solution results and iteratively solved until the voltage change between two adjacent nodes does not exceed the preset threshold, so as to eliminate voltage over-limits within the community.

[0049] The inter-community collaborative control unit collects voltage and power data from the boundary nodes of each community, adds boundary voltage consistency constraints and boundary power balance constraints, and establishes a global voltage optimization model between communities with the goal of minimizing the sum of photovoltaic active power regulation cost and distribution network loss. The decision variables are set as reactive power of SVG, active power and reactive power of photovoltaic. Distributed optimization using the alternating direction multiplier method is used for coordinated solution. After each round of alternation, adjacent communities exchange boundary data and residuals, and perform boundary data correction and voltage compensation update based on power flow equations until the preset convergence condition is met, and output the voltage regulation control quantity and operation setpoint of each community.

[0050] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the aforementioned distribution network voltage over-limit control method based on community division.

[0051] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned community-based distribution network voltage over-limit control method.

[0052] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0053] This invention comprehensively constructs performance indicators by introducing electrical distance indicators and active and reactive power adequacy indicators; and combines an improved particle competition learning method with node influence factors and frog leap optimization algorithm to avoid getting trapped in local optima and improve the accuracy and stability of community division.

[0054] This invention employs an over-limit control model based on linearized power flow equations within the community, aiming to minimize node voltage deviations and quickly solve for the reactive power regulation of the SVG and the active and reactive power regulation schemes of the photovoltaic system. It can promptly eliminate voltage over-limits within the local area, reducing dependence on the global system and improving real-time response capabilities. At the inter-community level, aiming to minimize photovoltaic active power regulation costs and distribution network losses, a distributed optimization method is used to coordinate regulation resources across communities. Through a boundary node voltage correction mechanism, the linearized model errors are corrected using power flow equations, ensuring the accuracy and reliability of boundary voltage and power constraints.

[0055] This invention, through a two-stage control mode of autonomy and cooperation, can solve the problem of local voltage exceeding limits and achieve voltage optimization of the entire network; it reduces the computational complexity of centralized optimization and improves the flexibility and scalability of voltage control; and effectively improves the voltage control accuracy and operational safety of photovoltaic power distribution networks. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0057] Figure 1 This is a flowchart of a distribution network voltage over-limit control method based on community division;

[0058] Figure 2 The flowchart of the improved particle competition learning algorithm based on the frog-jumping strategy in Example 1 is shown.

[0059] Figure 3 This is the model framework for the improved particle competition learning based on the frog-jumping strategy in Example 1;

[0060] Figure 4 This is a diagram of the IEEE 33-node distribution network model in Example 1;

[0061] Figure 5 shows the community partitioning results of the IEEE 33-node system under three schemes; 5-a shows the community partitioning results of the IEEE 33-node system under scheme 1; 5-b shows the community partitioning results of the IEEE 33-node system under scheme 2; and 5-c shows the community partitioning results of the IEEE 33-node system under scheme 3.

[0062] Figure 6 This is the voltage control logic diagram within and between communities in Example 1;

[0063] Figure 7-a and 7-b The figures are as follows: 7-a is the voltage optimization result diagram of the two-stage voltage over-limit control of the distribution network in Example 1; 7-b is the voltage optimization diagram of resource self-governance within the community; and 7-c is the voltage optimization diagram of collaborative operation between communities.

[0064] Figure 8-a and 8-b The results are respectively the voltage over-limit control and economic cost optimization results under different scenarios in Example 1; Figure 8-a 8-b is a comparison chart of node voltages; 8-c is a comparison chart of economic costs. Detailed Implementation

[0065] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0066] Example 1

[0067] A distribution network voltage over-limit control method based on community division, such as Figure 1 As shown, it includes:

[0068] Community division: Collect operational data of the photovoltaic power distribution network and divide it into communities. Clustering is performed based on the improved particle competition learning algorithm. The initial position is determined according to the node influence factor. The frog leap optimization algorithm is used to guide the movement of particles and iteratively update until convergence to obtain the community division result.

[0069] Community-based voltage over-limit control: Collect node load, node voltage, and photovoltaic power within the community, set safe operation constraints, and establish a community-based voltage over-limit control model based on linearized power flow balance equations. The voltage over-limit control model aims to minimize the node voltage deviation rate. Set the decision variables as reactive power of SVG, active power of photovoltaics, and reactive power. Update the node voltage according to the solution results and iterate until the change in node voltage between two adjacent times does not exceed the preset threshold, so as to eliminate voltage over-limit within the community.

[0070] Global voltage optimization between communities: Voltage and power data of boundary nodes of each community are collected, and boundary voltage consistency constraints and boundary power balance constraints are added. A global voltage optimization model between communities is established with the goal of minimizing the sum of the active power regulation cost of the photovoltaic system and the distribution network loss. The decision variables are set as the reactive power of SVG, the active power of photovoltaic and the reactive power. Distributed optimization using the alternating direction multiplier method is used for coordinated solution. After each round of alternation, adjacent community units exchange boundary data and residuals, and boundary data correction and voltage compensation are updated based on the power flow equation until the preset convergence condition is met. The voltage regulation control quantity and operation setpoint of each community unit are output.

[0071] like Figure 2 The flowchart shown below details the improved particle competition learning algorithm based on the frog-leaping strategy. The community partitioning of nodes in a photovoltaic distribution network based on the improved particle competition learning algorithm (IPCL) includes:

[0072] The initial position of particles is determined based on the node influence factor;

[0073] The frog-jump optimization algorithm is used to guide the direction of particle movement and the particle state is updated iteratively;

[0074] When the convergence condition is met, output the community partitioning result.

[0075] The particle competition learning algorithm simulates the competition and cooperation process between particles to gradually approach the optimal solution and is used to determine the community division result.

[0076] The improved particle competition learning model framework based on the frog-leaping strategy addresses the deficiency in particle movement tendency by designing an influence factor index (in this invention, the node influence factor) and adding a regularization determination mechanism to initialize the initial position of the particles. The main idea is to select the top K nodes in terms of influence index (IP) based on node degree and node power flow relationship as the starting position of the particles, as shown in the following equation:

[0077]

[0078] in, For nodes The initial influence value, For nodes The set of adjacent nodes, It is a node and nodes The weighted adjacency matrix between them and These are the minimum and maximum values ​​of the non-zero elements in the weighted adjacency matrix, respectively. It represents the total number of nodes in the distribution network.

[0079] The particle distribution optimization method based on the spatial guidance mechanism can effectively avoid non-ideal states of local clustering or discrete distribution in the search space, and simultaneously improve convergence efficiency and iterative stability. This improved mechanism not only optimizes the computational performance of the particle competition learning mechanism, but also significantly enhances the robustness and interpretability of community partitioning in complex networks, and can accurately analyze the community structure characteristics of network topology.

[0080] Although the improved Particle Competition Learning (IPCL) algorithm guides a more uniform particle distribution through influence factors, thus improving the convergence speed, it cannot avoid the algorithm's tendency to get trapped in local optima. Therefore, this invention enhances the global search capability through the Frog Leaping Algorithm (SFLA) and proposes an improved IPCL optimization algorithm based on SFLA, significantly improving the effectiveness of community partitioning in power systems.

[0081] The frog-jumping algorithm simulates the process of a frog constantly jumping to find food, thus representing the search for feasible solutions. This algorithm boasts efficient and stable computational performance. By actively pushing disadvantaged particles towards advantageous regions, the frog-jumping algorithm balances the depth of local search and global information search, making it suitable for solving discrete and combinatorial problems. Therefore, the frog-jumping algorithm exhibits advantages such as ease of implementation, few parameters, and good global optimization capabilities. The update representation of the frog-jumping optimization algorithm is as follows:

[0082]

[0083]

[0084] in, For frog jump stride length, and These are the current best and worst particle positions. For the updated particle positions, For the maximum step size, This indicates the generation of random numbers between [0, 1].

[0085] Compared to the original algorithm, which relies solely on random exploration and preference shifting, the improved IPCL algorithm adds directional jumps, strengthens global information exchange, transmits optimal position information, and enhances global optimization capabilities.

[0086] This invention is based on the division of distribution network communities that include photovoltaics, such as Figure 3 As shown, the improved particle competition learning model framework based on the frog-jumping strategy can be seen, which comprehensively considers the community division index based on the modularity of electrical distance and the sufficiency of active and reactive power for community division.

[0087] The comprehensive performance index comprehensively considers the community division index based on modularity and active and reactive power adequacy. Specifically, the comprehensive performance index in community division includes:

[0088] Modularity based on electrical distance is used to characterize the electrical topological connections between nodes;

[0089] The relative reactive power support capacity between nodes;

[0090] The relative active power support capacity between nodes;

[0091] The comprehensive performance index is obtained through weighted calculation and used to determine the effectiveness of community division.

[0092] The comprehensive performance index proposed in this invention The expression is as follows:

[0093]

[0094] in, , It is a weighting indicator. It is based on modularity according to electrical distance. There are two nodes. and The relative reactive power support capacity between them There are two nodes. and The greater the relative support capacity of active power between them, the better the overall performance index, and the better the structural and electrical connectivity of the community.

[0095] like Figure 5-a , 5-b As shown in 5-c, by Figure 4 This paper analyzes the division results of the IEEE 33-node distribution network under different community division schemes.

[0096] Option 1 (5-a): Community division based on traditional modularity indicators;

[0097] Option 2 (5-b): Community partitioning based on improved modularity index of active and reactive power adequacy;

[0098] Scheme 3 (5-c): Community partitioning based on the improved preference migration adjustment parameter x value of Scheme 2;

[0099] Scheme 1 only considers the structural connections between nodes for community division, resulting in a modularity index of 0.6335. The four communities divided by this method focus on the coupling degree between nodes. Scheme 2 incorporates the coordination and matching degree of active and reactive power sources within the community. Due to the effect of functional indicators, the power nodes tend to be more evenly distributed, and the maximum improved modularity index in this scheme is 0.684. Compared with Scheme 1, Scheme 2 improves the modularity index by 7.97%. Scheme 3, based on Scheme 2 which considers both structural and functional indicators, reselects the preference migration adjustment parameter x=0.7. Compared with Scheme 2, due to the consideration of power balance of sources and loads within the community, its modularity index increases by 2.91%, with slightly weaker structural strength but a more significant improvement in functional indicators.

[0100] Based on the results of the three community division schemes mentioned above, the modularity index of the distribution network community divided using the comprehensive performance index of this invention is 0.7039, which is greater than the first two schemes. This indicates that the scheme of this invention can objectively take into account both structural and functional characteristics, ensuring the electrical connection requirements between community nodes and leveraging the ability to coordinate source and load power within the community. This verifies the feasibility of the community division method that considers the adequacy of active and reactive power.

[0101] The integration of photovoltaic (PV) power into the distribution network can cause a systemic voltage rise, with nodes connected to PV experiencing a relatively larger rise. A high proportion of PV in the system can lead to some nodes exceeding their voltage limits. Therefore, voltage exceeding limits can be mitigated by adjusting PV power output to reduce node voltage amplitude, and PV units can reduce node voltage by absorbing reactive power. Furthermore, when SVG (Static Var Generator) is integrated into nodes with severe voltage exceeding limits, reactive power compensation significantly reduces the node voltage.

[0102] This invention primarily aims to provide a solution for voltage optimization control in distribution networks with high-proportion distributed photovoltaic (PV) grid integration, effectively addressing voltage limit exceedance issues caused by PV access and improving the safety and economy of distribution network operation. This invention proposes a two-stage (intra-community and inter-community) voltage limit exceedance control strategy for PV-integrated distribution networks. The intra-community and inter-community voltage limit exceedance control logic is as follows: Figure 6 As shown.

[0103] In the first stage, utilizing the community division results, the goal is to minimize the node voltage deviation rate, with power flow equality constraints, Static Var Generator (SVG) and photovoltaic (PV) safe operation constraints as conditions. The problem of node voltage exceeding limits within the community is solved by adjusting SVG reactive power compensation and PV active and reactive power. In the second stage, the goal is to achieve optimal economic efficiency, i.e., minimizing the sum of PV active power regulation and distribution network losses. Based on the power flow equality constraints, SVG and PV safe operation constraints, community boundary node voltage consistency constraints and boundary transmission power balance constraints are added. Upstream community boundary nodes are equivalent to downstream virtual balancing nodes to account for the impact of local community voltage regulation resources on the voltage of adjacent community nodes. Finally, an alternating optimization mechanism is used to coordinately adjust SVG and PV power. During the community autonomy stage, virtual node parameters are fixed to obtain local optimum. During the inter-community coordination stage, boundary information is updated until the overall network voltage converges and the system achieves optimal economic efficiency. Simulation verification shows that this method can efficiently eliminate voltage exceeding limits while maintaining control accuracy.

[0104] like Figure 1 As shown, in this embodiment, the first stage of the community-based voltage autonomy control includes:

[0105] A voltage over-limit control model is established based on linear power flow equations;

[0106] Under the conditions of satisfying the upper and lower limits of node voltage and the capacity limitations of distributed power sources and compensation equipment, the reactive power regulation of SVG and the active and reactive power regulation scheme of photovoltaic system are obtained by solving.

[0107] The SVG and photovoltaic systems within the community are adjusted according to the aforementioned regulation scheme to eliminate voltage over-limits.

[0108] The optimization objective of the voltage over-limit control model is to minimize the deviation between the voltage of each node in the community and the reference voltage. The voltage over-limit control model includes the following constraints:

[0109] The node voltage amplitude must be kept within the preset safety range;

[0110] The reactive power of photovoltaic systems shall not exceed the rated capacity.

[0111] The active power regulation of photovoltaic systems shall not exceed the allowable regulation range;

[0112] The reactive power of the SVG shall not exceed the equipment limit.

[0113] Specifically, the objective function for community self-governance control in the first stage is:

[0114]

[0115] in, The value is the initial value of 1.05 pu. Let be the node voltage after voltage regulation. The objective function is the sum of the node voltage deviation rates across the entire distribution network. This represents the optimal ability of the system to maintain the node voltages within a specified range and restore them to a stable state through voltage regulation measures after a voltage exceedance problem occurs.

[0116] In voltage optimization modeling, the traditional DistFlow branch power flow equality constraint is:

[0117]

[0118] in,

[0119]

[0120] in, and For nodes The active and reactive power of the load; Represents a node and nodes The reactance value of the line; and They are nodes The active and reactive power of the load; For nodes Active power output at the photovoltaic maximum power point. Node Actual active power output of photovoltaic for With photovoltaic active power regulation The difference.

[0121] However, while the traditional DistFlow branch power flow equations accurately describe the power flow characteristics of distribution networks, they contain squared and fractional terms, resulting in non-convex and nonlinear characteristics, making fast and efficient solutions difficult. Considering the typical operating conditions of relatively small line losses and small voltage deviations between nodes in distribution networks, this invention adopts a linearized approximation of the DistFlow equations (LinDistFlow model). While maintaining computational accuracy, this transforms the optimization problem into a convex problem, significantly reducing computational complexity and facilitating rapid voltage limit control within the community. Specifically:

[0122]

[0123] in,

[0124]

[0125] The community operates in accordance with node voltage safety constraints and node device constraints. The node voltage safety constraints are as follows:

[0126]

[0127] The constraints on photovoltaic equipment are:

[0128]

[0129]

[0130]

[0131] The constraints of SVG are:

[0132]

[0133] in, It is 0.95 pu. It is 1.05 pu. The maximum power factor angle of the photovoltaic equipment's output power. For nodes The available active power of photovoltaic equipment For nodes Active power regulation of photovoltaic equipment For nodes The reactive power output of photovoltaic equipment For nodes The installed capacity of photovoltaic equipment For SVG reactive power output, and They are nodes The lower and upper limits of reactive power for SVG.

[0134] The solution is obtained based on the LinDistFlow model. and The node voltage is updated according to the solution results, and the solution is iterated until the voltage change of two adjacent nodes does not exceed the preset threshold, so as to eliminate the voltage exceeding the limit within the community.

[0135] like Figure 1 As shown, after completing the autonomous voltage control within the community, in order to further optimize the overall network operation, this invention proposes a community-based voltage collaborative control method, the core contents of which include the following aspects:

[0136] 1) With optimal economic efficiency as the objective, the following objective function is established:

[0137] Photovoltaic active power regulation cost: This refers to the economic cost incurred by photovoltaic units in regulating active power.

[0138] Distribution network loss: refers to the active power loss of lines during system operation;

[0139] Overall objective: To minimize the sum of photovoltaic active power regulation costs and distribution network losses, thereby achieving economical operation while ensuring voltage safety.

[0140] 2) The voltage global optimization model must simultaneously satisfy the following constraints during the solution process:

[0141] Power flow equation constraints: ensure that the power balance relationship of each node satisfies the power flow equation of the distribution network;

[0142] Equipment constraints: The active and reactive power of the photovoltaic system shall not exceed the rated capacity, and the reactive power compensation of the SVG shall not exceed the equipment limit;

[0143] Boundary voltage consistency constraint: The voltage of adjacent communities at the boundary nodes remains consistent to avoid voltage deviation across communities;

[0144] Boundary power balance constraint: Ensure that the power exchange of boundary nodes is balanced between different communities.

[0145] 3) To account for voltage coupling between communities, this invention introduces a virtual balancing node:

[0146] The boundary nodes of the upstream community are equivalent to the virtual balancing nodes of the downstream community;

[0147] This virtual node allows the modeling to reflect the impact of local community regulation measures on the voltage of adjacent communities, thereby embedding global coupling information across communities into local optimization.

[0148] 4) Distributed solution using the Alternating Direction Multiplier Method (ADMM):

[0149] During the community self-governance phase, the virtual node parameters are fixed, and the local optimum solution for the community is obtained independently.

[0150] During the inter-community coordination phase, the boundary node information (voltage and power) is updated and transmitted to neighboring communities;

[0151] Through alternating iterations, the boundary node data is continuously corrected until the entire network voltage converges and the economic objective is optimized.

[0152] The second stage of inter-community voltage coordination control aims to minimize the sum of photovoltaic active power regulation costs and distribution network losses. Under the constraints of power flow equations, SVG and photovoltaic safe operation, community boundary node voltage consistency, and boundary power balance, upstream community boundary nodes are equivalent to downstream virtual balance nodes, and a distributed optimization method is used to allocate regulation resources to each community. Coordinated regulation of SVG and photovoltaic power is implemented in each community. During the intra-community autonomy phase, the virtual balance node parameters are fixed to obtain local optimum; during the inter-community coordination phase, boundary information is updated until the entire network voltage converges and economic optimum is achieved.

[0153] The objective function for inter-community collaborative control considers the economic optimal control result that minimizes the sum of active power regulation and total network loss of distributed photovoltaic power generation. Based on network separation and power flow constraint equations, the objective function is:

[0154]

[0155] in, Indicates community The objective function of the voltage over-limit control model; and These represent the revenue generated by photovoltaic equipment and the on-grid electricity price for active power, respectively. For nodes The active power regulation capacity of photovoltaic equipment; For nodes The reactive power output of photovoltaic equipment For nodes The reactive power output of the SVG; Represents a node and nodes The resistance value of the circuit; and Representing nodes respectively Inflow branch Active power and reactive power; Represented as relationships between nodes; For nodes The voltage amplitude.

[0156] The additional community boundary node voltage consistency constraints and boundary transmission power balance constraints are as follows:

[0157]

[0158]

[0159]

[0160] in, As the boundary node of the community, This represents the square of the voltage at the upstream community boundary node. The square of the voltage at the virtual balancing node in the downstream community and These represent inter-community routes. Transmitted active and reactive power, and This corresponds to the virtual load equivalent. Represents boundary nodes Globally consistent variables, and These represent inter-community routes. Globally consistent variables for transmitted active and reactive power.

[0161] The Alternating Direction Method of Multipliers (ADMM) is an efficient distributed optimization algorithm suitable for solving convex optimization problems with separable structures. This method solves the optimization problem through two key steps: First, based on the separability of the problem, ADMM decomposes the original optimization problem into several relatively simple subproblems. This decomposition strategy allows each subproblem to be solved independently, significantly reducing computational complexity. Second, by introducing Lagrange multipliers and coordination variables, ADMM establishes a coupling mechanism between the subproblems. During the iteration process, the algorithm alternately updates the solutions of each subproblem and synchronously adjusts the coordination variables, ultimately ensuring that the solutions of all subproblems converge to the optimal solution of the original problem.

[0162] make , and , , and Representing communities The virtual balancing node a* voltage, the active and reactive power transmitted by the inter-community line am, and the boundary node Voltage, boundary node Given the Lagrange multipliers of the virtual load's active and reactive power, the augmented Lagrange function of the objective function can be expressed as:

[0163]

[0164] in,

[0165]

[0166] in, To augment the penalty coefficient of the Lagrange function and To ensure the convergence of boundary variables between adjacent communities during the distributed optimization process, the LinDistFlow linearization model, which ignores line losses and their resulting voltage deviations, may lead to a decrease in power flow calculation accuracy for distribution network systems with long lines and high transmission power. To compensate for the voltage limit control error caused by the LinDistFlow model, this invention introduces a post-processing-based correction strategy: after obtaining the optimization solution, each community can use the DistFlow equation to correct the boundary data exchanged between communities, and add peak-valley voltage compensation parameters to the node voltage constraints. and The compensated node voltage safety constraints are shown in the following equation:

[0167]

[0168] in, Community obtained for DistFlow equation highest voltage Corresponding node voltages calculated by the LinDistFlow equation difference, Community obtained for DistFlow equation minimum voltage Corresponding node voltages calculated by the LinDistFlow equation difference.

[0169] like Figure 1 As shown, it specifically includes:

[0170] S1. Acquisition and Initialization: Acquire power distribution network operation data and complete parameter initialization. Set the iteration count k=0 and initialize the local variable set. With boundary variables Initialize consistent global variables and Lagrange multipliers, given a penalty coefficient. and convergence threshold ;

[0171] S2, in the Round 1, according to the augmented Lagrange decomposition

[0172]

[0173] Solve the sub-problems for each community and update the local decisions and boundary variables: ;

[0174] S3. Boundary Correction and Compensation Update: The DistFlow equation is used to perform physical consistency correction on the community boundary voltage and power data to obtain the corrected boundary quantities, and the voltage compensation parameters are updated accordingly.

[0175] S4. Exchange boundary data with neighboring communities by exchanging corrected boundary data.

[0176] S5. Update global variables and multipliers in-situ based on the local solution of round k+1, and update the consistent global variables and Lagrange multipliers in-situ at the current site according to the ADMM rule (illustrated):

[0177]

[0178] S6. Calculate the original residuals and dual residuals to ensure boundary consistency:

[0179]

[0180] S7, Exchange residuals with neighboring communities and Convergence is determined by collaborative judgment.

[0181] S8. Convergence Criterion and Iteration If

[0182]

[0183] Then proceed to S9; otherwise, let And return to S2.

[0184] After the S9 output meets the threshold, it obtains the most economical reactive and active voltage regulation settings, and outputs the equipment output and operation settings of each community unit.

[0185] This invention employs the IEEE 33-node system after community division, such as Figure 5-c The simulation analysis is shown below. Each community first uses its own resources to optimize the voltage. The optimization results are as follows: Figure 7-aAs shown, when photovoltaic power is connected, the voltage of nodes in each community collectively rises, with some exceeding the safety threshold. Communities with nodes exceeding the voltage limit autonomously utilize their own voltage regulation resources to absorb reactive power and regulate active power until the voltage amplitude of all nodes in the community is less than the safety threshold of 1.05 pu.

[0186] Because community-based autonomous voltage optimization lacks coordination between communities, the community-based optimization method only utilizes resources within the community to solve voltage limit exceedance problems, resulting in insufficient active power regulation from local photovoltaic systems. To analyze the combined voltage regulation effect of reactive power compensation and active power regulation in this community on adjacent communities, this invention considers the global values ​​of community boundary data and locally updated boundary data to achieve global optimization control of distributed power generation output power and reactive power compensation values. Voltage regulation is achieved using SVG and PV active power regulation, with results as follows: Figure 7-b As shown, after the system was connected to distributed photovoltaic (PV) power, the voltage peak reached 1.115 pu. To suppress the voltage over-limit problem, a coordinated control strategy of PV active power reduction and SVG reactive power compensation was adopted, and the over-limit voltage of communities 1, 2, and 4 was reduced to within the safe threshold.

[0187] To better meet the needs of voltage regulation and economical operation, this invention is based on Figure 4 The IEEE 33-node system example sets the penalty coefficient for the augmented Lagrangian function. =100. Considering system safety and economy, the active power regulation of photovoltaic (PV) power and the reactive power compensation capacity of the connected SVG (Static Var Generator) are optimized. Three scenarios are set: only photovoltaic renewable energy is connected, active power regulation of PV power is used, and active power regulation of SVG and PV unit reactive power compensation are used in conjunction with PV power regulation. Voltage over-limit control and economic cost optimization are performed for each scenario, and the results are as follows. Figure 8-a As shown in Figure-b.

[0188] Figure 8-a The voltage distribution at each node is shown under different control scenarios. It can be seen that:

[0189] When only photovoltaic power is connected, the voltage of some nodes exceeds the upper limit, resulting in voltage exceeding the limit.

[0190] After adopting photovoltaic active power regulation, the node voltage has been improved to some extent, but some nodes are still close to the upper limit.

[0191] When SVG reactive power compensation is introduced and works in conjunction with photovoltaic active and reactive power regulation, the voltage of all nodes is effectively controlled within the allowable range, significantly improving voltage stability.

[0192] This indicates that the two-stage voltage over-limit control method proposed in this invention can efficiently eliminate the voltage over-limit problem.

[0193] Figure 8-bThe paper presents a comparison of economic indicators under three scenarios, where the indicator is the sum of photovoltaic active power regulation costs and distribution network losses. It can be seen that:

[0194] The system has the highest economic cost when photovoltaics are not introduced.

[0195] Relying solely on photovoltaic active power regulation can reduce some costs;

[0196] By adopting SVG reactive power compensation in conjunction with photovoltaic active power regulation, the overall economic cost has been significantly reduced, the operating curve has become smoother, and a balance between voltage control and economic optimization has been achieved.

[0197] Comprehensive economic analysis reveals that while a power supply mode solely connected to photovoltaics theoretically achieves the lowest operating cost, the node voltage under this mode will remain in an over-limit state for an extended period. To balance economy and safety, a comparative study of two typical voltage regulation schemes reveals that while a single active power resource voltage regulation strategy (PV power regulation only) can suppress the voltage to a safe range, it increases the total operating cost due to curtailment. In contrast, the active and reactive power coordinated voltage regulation scheme (PV power regulation and SVG dynamic reactive power compensation) not only precisely controls the voltage within the safe range but also further reduces the total system operating cost compared to a single PV active power regulation scheme through spatiotemporal optimization of reactive power. This result verifies the superiority of the voltage regulation strategy—under voltage safety constraints, economic losses are minimized through active and reactive power coupling optimization.

[0198] In summary, simulation results show that the two-stage voltage over-limit control method for distribution networks based on community division proposed in this invention can eliminate voltage over-limits and has economic value. The method has high accuracy and reliability.

[0199] This invention comprehensively constructs performance indicators by introducing electrical distance indicators and active and reactive power adequacy indicators; and combines an improved particle competition learning method with node influence factors and frog leap optimization algorithm to avoid getting trapped in local optima and improve the accuracy and stability of community division.

[0200] This invention employs a voltage limit control model based on linearized power flow equations within the community, aiming to minimize node voltage deviations and quickly solve for the reactive power regulation of the SVG and the active and reactive power regulation schemes of the photovoltaic system. It can promptly eliminate voltage limits locally, reducing dependence on the global system and improving real-time response capabilities. At the inter-community level, aiming to minimize photovoltaic active power regulation costs and distribution network losses, a distributed optimization method is used to coordinate regulation resources across communities. Through a boundary node voltage correction mechanism, the linearized model errors are corrected using power flow equations, ensuring the accuracy and reliability of boundary voltage and power constraints.

[0201] This invention, through a two-stage control mode of autonomy and cooperation, can solve the problem of local voltage exceeding limits and achieve voltage optimization of the entire network; it reduces the computational complexity of centralized optimization and improves the flexibility and scalability of voltage control; and effectively improves the voltage control accuracy and operational safety of photovoltaic power distribution networks.

[0202] Example 2

[0203] This invention also provides a distribution network voltage over-limit control system based on community division, used to implement the distribution network voltage over-limit control method based on community division described in the previous embodiment, including:

[0204] The community partitioning unit is used to partition the distribution network containing photovoltaic systems into communities. An improved particle competition learning algorithm is used for clustering. The initial position is determined based on the node influence factor. The leapfrog optimization algorithm is used to guide the particle movement and iteratively update until convergence to obtain the community partitioning result.

[0205] The community-based autonomous control unit is used to establish a community-based voltage over-limit control model based on the linearized power flow balance equation under node voltage safety constraints and node equipment constraints. With the goal of minimizing the node voltage deviation rate, it adjusts the reactive power of the SVG and the active and reactive power of the photovoltaic equipment set in each community to eliminate voltage over-limit within the community.

[0206] The inter-community collaborative control unit is used to collect voltage and power data of each community boundary node. The upstream community boundary node is equivalent to the downstream virtual balance node. Boundary voltage consistency constraints and boundary power balance constraints are set. Under the power flow balance equation and equipment operation constraints, with the goal of minimizing the sum of the active power regulation cost of the photovoltaic system and the distribution network loss, the distributed optimization of the alternating direction multiplier method is used to coordinate and regulate the SVG and photovoltaic equipment power of each community, and perform consistency correction on the boundary data based on the power flow equation.

[0207] Example 3

[0208] This application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method of embodiment 1.

[0209] Example 4

[0210] This application provides a computer-readable storage medium (non-transitory) that stores computer instructions that cause a computer to execute the method of embodiment 1.

[0211] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0212] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0213] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0214] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A community partition-based voltage out-of-limit control method for a power distribution network, characterized in that, include: Operational data of photovoltaic power distribution networks are collected and communities are divided. Clustering is performed based on an improved particle competition learning algorithm. The initial position is determined according to the node influence factor. The frog leap optimization algorithm is used to guide particle movement and iteratively update until convergence to obtain the community division results. Data on node load, node voltage, and photovoltaic power within the community are collected. Safety operation constraints are set, and a voltage over-limit control model is established within the community based on the linearized power flow balance equation. The voltage over-limit control model aims to minimize the node voltage deviation rate. The decision variables are set as the reactive power of SVG, the active power and reactive power of photovoltaic power. The node voltage is updated according to the solution results and iteratively solved until the voltage change between two adjacent nodes does not exceed the preset threshold, so as to eliminate voltage over-limit within the community. Voltage and power data of each community boundary node are collected, and boundary voltage consistency constraints and boundary power balance constraints are added. A global voltage optimization model between communities is established with the goal of minimizing the sum of the active power regulation cost of the photovoltaic system and the distribution network loss. The decision variables are set as the reactive power of SVG, the active power of photovoltaic and the reactive power. Distributed optimization using the alternating direction multiplier method is used for coordinated solution. After each round of alternation, adjacent community units exchange boundary data and residuals, and boundary data correction and voltage compensation are updated based on the power flow equation until the preset convergence condition is met. The voltage regulation control quantity and operation setpoint of each community unit are output.

2. The distribution network voltage over-limit control method based on community division according to claim 1, characterized in that, Community partitioning uses an influence factor to select initial positions and updates particles using a leapfrog optimization algorithm to achieve global convergence. The influence factor is represented as: in, For nodes The initial influence value, For nodes The set of adjacent nodes, It is a node and nodes The weighted adjacency matrix between them and These are the minimum and maximum values ​​of the non-zero elements in the weighted adjacency matrix, respectively. This represents the total number of nodes in the distribution network. The update representation of the frog-jump optimization algorithm is as follows: in, For frog jump stride length, and These are the current best and worst particle positions. For the updated particle positions, For the maximum step size, This indicates the generation of random numbers between [0, 1].

3. The distribution network voltage over-limit control method based on community division according to claim 1, characterized in that, The safe operation within the community satisfies both node voltage safety constraints and equipment constraints. The node voltage safety constraints are as follows: The equipment constraints for photovoltaics are: The device constraints for SVG are: in, The maximum power factor angle of the photovoltaic equipment's output power. For nodes The available active power of photovoltaic equipment For nodes Active power regulation of photovoltaic equipment For nodes The reactive power output of photovoltaic equipment For nodes The installed capacity of photovoltaic equipment For SVG reactive power output, and They are nodes The lower and upper limits of reactive power of the SVG.

4. The distribution network voltage over-limit control method based on community division according to claim 1, characterized in that, The voltage over-limit control model aims to minimize the sum of node voltage deviation rates, and its objective function is expressed as: in, The value is the initial value. This refers to the node voltage after voltage regulation. In the process of modeling the voltage over-limit control model, the linearized power flow equation constraint is: in, in, and For nodes The active and reactive power of the load. Represents a node and nodes The reactance value of the line, Represents a node and nodes The resistance value of the circuit. and They are nodes The active and reactive power of the load. For nodes Active power output at the maximum power point of photovoltaic power generation. For nodes Photovoltaic active power regulation For nodes The reactive power output of photovoltaic power, For nodes The SVG outputs reactive power.

5. The distribution network voltage over-limit control method based on community division according to claim 3 or 4, characterized in that, The objective function for inter-community voltage coordination control is: in, Indicates community The objective function of the voltage over-limit control model. and These represent the revenue generated by photovoltaic equipment and the on-grid electricity price for active power, respectively. For nodes Photovoltaic equipment active power regulation capacity, For nodes The reactive power output of photovoltaic equipment For nodes The reactive power output of the SVG Represents a node and nodes The resistance value of the circuit. and Representing nodes respectively Inflow branch Active power and reactive power, By node Pointing to node A side road.

6. The distribution network voltage over-limit control method based on community division according to claim 1, characterized in that, The added boundary voltage consistency constraint and boundary power balance constraint are expressed as follows: in, As the boundary node of the community, This represents the square of the voltage at the upstream community boundary node. The square of the voltage at the virtual balancing node in the downstream community and These represent inter-community routes. Transmitted active and reactive power, and This corresponds to the virtual load equivalent. Represents boundary nodes Globally consistent variables, and These represent inter-community routes. Globally consistent variables for transmitted active and reactive power.

7. The distribution network voltage over-limit control method based on community division according to claim 1, characterized in that, Inter-community collaborative optimization employs the alternating direction multiplier method for distributed solution, and the augmented Lagrangian function of the objective function can be expressed as: After each cycle of alternation, the boundary data are corrected using the power flow equations, and compensation parameters are applied. and Update the node voltage consistency constraints at the community boundary; The compensated node voltage safety constraint is expressed as: in, For the community highest voltage With corresponding node voltage difference, For the community minimum voltage The corresponding node voltages of and difference.

8. A distribution network voltage over-limit control system based on community division, characterized in that, The method for implementing the distribution network voltage over-limit control method based on community division as described in any one of claims 1-7 includes: The community division unit collects operation data of the photovoltaic power distribution network and divides it into communities. Clustering is performed based on the improved particle competition learning algorithm. The initial position is determined according to the node influence factor. The frog leap optimization algorithm is used to guide the movement of particles and iteratively update until convergence to obtain the community division result. The community-based autonomous control unit collects node loads, node voltages, and photovoltaic power within the community, sets safe operation constraints, and establishes a voltage over-limit control model based on linearized power flow balance equations. The voltage over-limit control model aims to minimize the node voltage deviation rate, and sets the decision variables as the reactive power of SVG, the active power of photovoltaics, and the reactive power. The node voltage is updated according to the solution results and iteratively solved until the voltage change between two adjacent nodes does not exceed the preset threshold, so as to eliminate voltage over-limits within the community. The inter-community collaborative control unit collects voltage and power data from the boundary nodes of each community, adds boundary voltage consistency constraints and boundary power balance constraints, and establishes a global voltage optimization model between communities with the goal of minimizing the sum of photovoltaic active power regulation cost and distribution network loss. The decision variables are set as reactive power of SVG, active power and reactive power of photovoltaic. Distributed optimization using the alternating direction multiplier method is used for coordinated solution. After each round of alternation, adjacent communities exchange boundary data and residuals, and perform boundary data correction and voltage compensation update based on power flow equations until the preset convergence condition is met, and output the voltage regulation control quantity and operation setpoint of each community.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the distribution network voltage over-limit control method based on community division as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the distribution network voltage over-limit control method based on community division as described in any one of claims 1 to 7.

Images