Distribution network distributed collaborative voltage regulation method, device, computer equipment and storage medium

Abstract

The present application relates to a distributed collaborative voltage regulation method, device, computer equipment and storage medium for a distribution network, and relates to the technical field of smart grids, including: determining, from all nodes in the distribution network, nodes to be regulated according to the monitoring data of each node in the distribution network, and obtaining the target reactive power utilization rate thereof according to its monitoring data, a preset sensitivity nonlinear model and a target voltage value; controlling the nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and performing iterative calculation on the target reactive power utilization rates of each adjacent node by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model and the monitoring data of the adjacent nodes until a preset iteration end condition is met and until all nodes in the distribution network are traversed to obtain the target reactive power utilization rates of each node; generating reactive power output instructions for each node according to the target reactive power utilization rates of each node to instruct each node to perform voltage regulation operations. Using this method can improve adaptability.

Description

Technical Field

[0001] The present application relates to the technical field of smart grids, and particularly to a distributed cooperative voltage regulation method, device, computer device, and storage medium for a distribution network. Background Art

[0002] With the rapid development of distributed photovoltaic power generation, more and more photovoltaic power generation systems are connected to the distribution network. High-penetration photovoltaic power generation generates a large amount of active power during the daytime peak period. And due to the volatility and intermittency characteristics of photovoltaic power generation, it may cause problems such as voltage fluctuations in the distribution network and difficulties in reactive power regulation. In particular, the reactive power fluctuations output by photovoltaic power generation have a significant impact on the voltage stability of the distribution network, and further affect the operation safety of the power system. In related technologies, through a distributed cooperative regulation strategy, without relying on a centralized control system, voltage regulation and reactive power optimization can be achieved through the cooperation of local devices, which is particularly suitable for distribution networks with high-penetration photovoltaic power generation.

[0003] However, the distributed voltage regulation methods in related technologies are difficult to ensure voltage stability and the safe operation of the system under rapidly changing voltage fluctuations, and there is a problem of weak adaptability. Summary of the Invention

[0004] Based on this, it is necessary to provide a distributed cooperative voltage regulation method, device, computer device, and computer-readable storage medium for a distribution network that can improve adaptability for the above technical problems.

[0005] In a first aspect, the present application provides a distributed cooperative voltage regulation method for a distribution network, including:

[0006] According to the monitoring data of each node in the distribution network, determine the nodes to be regulated from all the nodes; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device;

[0007] According to the monitoring data of the nodes to be regulated, a preset sensitivity nonlinear model, and a target voltage value, obtain the target reactive power utilization rate of the nodes to be regulated; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data;

[0008] Control the nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes, use a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each adjacent node until a preset iteration end condition is met, and obtain the target reactive power utilization rate of each adjacent node;

[0009] Control each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated. According to the target reactive power utilization rate, the sensitivity non-linear model, and the monitoring data of the new adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rate of each of the new adjacent nodes to obtain the target reactive power utilization rate of each of the new adjacent nodes;

[0010] Take the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated, until all the nodes in the distribution network are traversed to obtain the target reactive power utilization rate of each node;

[0011] Generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node to instruct each node to perform a voltage regulation operation.

[0012] In one embodiment, the monitoring data includes voltage values. Determining the node to be regulated from all the nodes according to the monitoring data of each node in the distribution network includes:

[0013] Determine the comparison result between the voltage value of each node in the distribution network and a preset voltage threshold range; according to the comparison result, determine the nodes whose voltage values exceed the voltage threshold range as the nodes to be regulated.

[0014] In one embodiment, the preset iteration end condition includes any one of that the voltage values of all the nodes are within the voltage threshold range, the difference between the target reactive power utilization rates of two adjacent updates is lower than a preset threshold, and the number of iterations reaches a preset iteration number threshold.

[0015] In one embodiment, obtaining the target reactive power utilization rate of the node to be regulated according to the monitoring data of the node to be regulated, a preset sensitivity non-linear model, and a target voltage value includes:

[0016] According to the voltage value of the node to be regulated and the target voltage value, obtain the voltage change amount of the node to be regulated; input the voltage change amount of the node to be regulated and other monitoring data except the voltage value into the inverse function of the sensitivity non-linear model to obtain the target reactive power utilization rate of the node to be regulated.

[0017] In one embodiment, iteratively calculating the target reactive power utilization rate of each adjacent node using the consensus algorithm model according to the target reactive power utilization rate, the sensitivity non-linear model, and the monitoring data of the adjacent nodes until a preset iteration end condition is met to obtain the target reactive power utilization rate of each adjacent node includes:

[0018] Obtain the voltage change amount of each of the adjacent nodes; input the voltage change amounts of the monitoring data of the adjacent nodes and other monitoring data except the voltage value into the sensitivity non-linear model to obtain the reactive voltage sensitivity of each of the adjacent nodes; input the reactive voltage sensitivity of the adjacent nodes and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied, and obtain the target reactive power utilization rate of each of the adjacent nodes.

[0019] In one embodiment, the step of inputting the reactive voltage sensitivity of the adjacent nodes and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied to obtain the target reactive power utilization rate of each of the adjacent nodes includes:

[0020] For each of the adjacent nodes, input the reactive voltage sensitivity of the adjacent nodes and the target reactive power utilization rate into the model of the consensus algorithm respectively to obtain a candidate reactive power utilization rate; when the candidate reactive power utilization rate satisfies the preset inverter constraint, use the candidate reactive power utilization rate as the target reactive power utilization rate updated in this round of iteration; when the candidate reactive power utilization rate does not satisfy the preset inverter constraint, perform a projection operation on the candidate reactive power utilization rate according to the inverter constraint to obtain the target reactive power utilization rate updated in this round of iteration; return to the step of inputting the reactive voltage sensitivity of the adjacent nodes and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied, and obtain the target reactive power utilization rate of each of the adjacent nodes.

[0021] In one embodiment, the expression corresponding to the consensus algorithm model includes:

[0022]

[0023] where k represents the number of iterations, represents the target reactive power utilization rate, r and respectively represent the learning rate and training parameter of the sensitivity non-linear model, Q i represents the reactive power of the i-th node, is the Chebyshev filter acceleration operator, W represents the state transition matrix composed of the state transition weights of each communication link in the distribution network, represents the reactive voltage sensitivity of the node i, represents the sensitivity weighting matrix, represents the step size, represents the set of adjacent nodes of the node i, and j represents the j-th adjacent node.

[0024] In a second aspect, the present application also provides a distributed cooperative voltage regulation device for a distribution network, including:

[0025] A node determination module, configured to determine, from all the nodes, the nodes to be regulated according to the monitoring data of each node in the distribution network; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device;

[0026] A reactive power determination module, configured to obtain the target reactive power utilization rate of the nodes to be regulated according to the monitoring data of the nodes to be regulated, a preset sensitivity nonlinear model, and a target voltage value; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data;

[0027] An iterative update module, configured to control the nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and perform iterative calculation on the target reactive power utilization rates of the adjacent nodes by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes until a preset iteration end condition is satisfied, so as to obtain the target reactive power utilization rates of the adjacent nodes;

[0028] The iterative update module is further configured to control each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the nodes to be regulated, and perform iterative calculation on the target reactive power utilization rates of the new adjacent nodes by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, so as to obtain the target reactive power utilization rates of the new adjacent nodes;

[0029] The iterative update module is further configured to use the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the nodes to be regulated until all the nodes in the distribution network are traversed, so as to obtain the target reactive power utilization rate of each node;

[0030] An instruction output module, configured to generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node, so as to instruct each node to perform a voltage regulation operation.

[0031] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0032] Determine, from all the nodes, the nodes to be regulated according to the monitoring data of each node in the distribution network; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device;

[0033] Based on the monitoring data of the node to be regulated, the preset sensitivity nonlinear model, and the target voltage value, obtain the target reactive power utilization rate of the node to be regulated; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data;

[0034] Control the node to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of the adjacent nodes until the preset iteration end condition is satisfied, to obtain the target reactive power utilization rates of the adjacent nodes; control each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the node to be regulated, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of the new adjacent nodes, to obtain the target reactive power utilization rates of the new adjacent nodes;

[0035] Take the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the node to be regulated, until all nodes in the distribution network are traversed, to obtain the target reactive power utilization rate of each node;

[0036] Generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node, to instruct each node to perform a voltage regulation operation.

[0037] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0038] Based on the monitoring data of each node in the distribution network, determine the node to be regulated from all the nodes; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device;

[0039] Based on the monitoring data of the node to be regulated, the preset sensitivity nonlinear model, and the target voltage value, obtain the target reactive power utilization rate of the node to be regulated; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data;

[0040] Control the node to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of the adjacent nodes;

[0041] Control each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated. According to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rate of each of the new adjacent nodes, and obtain the target reactive power utilization rate of each of the new adjacent nodes;

[0042] Take the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated, until all the nodes in the distribution network are traversed, and obtain the target reactive power utilization rate of each node;

[0043] Generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node to instruct each node to perform voltage regulation operations.

[0044] The above-mentioned distributed collaborative voltage regulation method, device, computer equipment and computer-readable storage medium for a distribution network. The method determines regulated nodes to be regulated from all nodes according to the monitoring data of each node in the distribution network, where a node is a node configured with a miniature circuit breaker and a photovoltaic inverter device, so as to grasp the operating conditions of each node in real time and obtain data support for subsequent voltage regulation decisions. Further, according to the monitoring data of the regulated nodes to be regulated, a preset sensitivity nonlinear model and a target voltage value, the target reactive power utilization rate of the regulated nodes to be regulated is obtained; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data; and controlling the regulated nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and according to the target reactive power utilization rate, the sensitivity nonlinear model and the monitoring data of adjacent nodes, using a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each adjacent node until a preset iteration end condition is satisfied, so as to obtain the target reactive power utilization rate of each adjacent node; in addition, controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the regulated node to be regulated, and according to the target reactive power utilization rate, the sensitivity nonlinear model and the monitoring data of the new adjacent nodes, using a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each new adjacent node to obtain the target reactive power utilization rate of each new adjacent node; taking the new adjacent nodes as the current adjacent nodes, and returning to the step of controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the regulated node to be regulated until all nodes in the distribution network are traversed to obtain the target reactive power utilization rate of each node, realizing information exchange between each node, generating a reactive power output instruction for each node according to the target reactive power utilization rate of each node to instruct each node to perform voltage regulation operations. The sensitivity nonlinear model can dynamically track the nonlinear relationship between node voltage and reactive power, providing a more accurate decision-making basis for regulation, and the use of the consensus algorithm model can efficiently achieve collaborative regulation between nodes. Through information exchange and sensitivity weighting between nodes, it ensures that the system can achieve coordinated voltage regulation of the entire network or local areas in the face of load fluctuations and photovoltaic power generation fluctuations, thereby improving the overall adaptability. Description of the Drawings

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0046] Figure 1 It is an application environment diagram of the distributed collaborative voltage regulation method for a distribution network in an embodiment;

[0047] Figure 2 It is a schematic flow chart of a distributed collaborative voltage regulation method for a distribution network in an embodiment;

[0048] Figure 3 It is a schematic flow chart of the iterative calculation steps of the target reactive power utilization rate in an embodiment;

[0049] Figure 4 It is a schematic flow chart of a distributed self-disciplined collaborative voltage regulation method for a distribution network based on intelligent micro circuit breakers in an embodiment;

[0050] Figure 5 It is a structural block diagram of a distributed collaborative voltage regulation device for a distribution network in an embodiment;

[0051] Figure 6 It is an internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0052] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] As described in the background art, the distributed collaborative voltage regulation method of the related technology has the problem of low adaptability. Through the research of the inventors, it is found that the reason for this problem is that with the rapid development of distributed photovoltaic power generation, more and more photovoltaic power generation systems are connected to the distribution network. Especially in the low-voltage distribution network, the high penetration of photovoltaic power makes the power system face unprecedented challenges. High-penetration photovoltaic power generation generates a large amount of active power during the daytime peak period, and due to the volatility and intermittency characteristics of photovoltaic power generation, it may lead to problems such as voltage fluctuations in the distribution network and difficulties in reactive power regulation. In particular, the reactive power fluctuations of photovoltaic power generation output have a significant impact on the voltage stability of the distribution network, and further affect the operation safety of the power system. To effectively address these challenges, traditional distribution network voltage regulation methods are no longer able to meet the requirements of high-penetration photovoltaic power generation. Most of the existing voltage regulation methods rely on traditional centralized control or static regulation methods, and these methods often cannot respond in real time to the rapidly changing loads and photovoltaic power generation in the distribution network. Moreover, traditional methods usually require relying on detailed power flow analysis models, which are computationally complex and difficult to cope with the rapidly changing grid conditions, resulting in the lag and inaccuracy of the regulation process. To solve the above problems, in recent years, distributed control methods have gradually become a new research direction. This method can achieve voltage regulation and reactive power optimization through local device cooperation without relying on a centralized control system through distributed collaborative regulation strategies, and is particularly suitable for distribution networks with high-penetration photovoltaic power generation. However, there are still some technical problems in the existing distributed voltage regulation methods, which are mainly manifested in the following aspects: insufficient voltage regulation accuracy, some existing distributed control methods have poor voltage regulation accuracy and stability under the conditions of high-penetration photovoltaic power generation, especially in the case of rapidly changing voltage fluctuations, it is difficult to ensure voltage stability and the safe operation of the system; difficult reactive power regulation, due to the volatility and intermittency of the reactive power output of distributed photovoltaic systems, it is difficult to regulate the reactive power in the distribution network. The existing regulation methods have a lag in response to reactive power fluctuations and are difficult to adapt to the changes of the system in real time; low intelligent level, although some existing regulation methods try to introduce distributed collaborative algorithms, most methods rely on fixed models or static control strategies and lack the intelligent regulation ability for different operating conditions and device states, and are difficult to adapt to the dynamically changing grid environment; high system complexity and difficult deployment, the existing control methods usually require complex hardware facilities and algorithm support, and face high costs and technical difficulties in actual deployment. Especially during the on-site installation and commissioning process, secondary commissioning and configuration are required, increasing the operation and maintenance costs of the system.

[0054] For the above reasons, the present application provides a distributed collaborative voltage regulation method for a distribution network, aiming to improve the adaptability to the fluctuations of the distribution network system.

[0055] The distributed cooperative voltage regulation method provided by the embodiment of the present application can be applied to, for example, Figure 1 the application environment shown in the figure. The application environment includes: a control system 102, a node to be regulated 104 in the distribution network, and adjacent nodes 106 in the distribution network. Among them, both the node to be regulated 104 and the adjacent nodes 106 include a miniature circuit breaker and a photovoltaic inverter. Among them, the control system 102 communicates with the miniature circuit breakers in each node through wireless communication. The control system 102 can be integrated on a server, or placed in the cloud or other network servers. The control system 102 determines the node to be regulated 104 from all nodes according to the monitoring data of each node in the distribution network; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device; the control center 102 obtains the target reactive power utilization rate of the node to be regulated 104 according to the monitoring data of the node to be regulated 104, a preset sensitivity nonlinear model, and a target voltage value; the control system 102 controls the node to be regulated 104 to broadcast the target reactive power utilization rate to the adjacent nodes 106, and according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes 106, uses a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each adjacent node 106 until a preset iteration end condition is met, and obtains the target reactive power utilization rate of each adjacent node 106; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data; the control system 102 controls each adjacent node 106 to broadcast its target reactive power utilization rate to new adjacent nodes 106 in the direction away from the node to be regulated, and according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes 106, uses a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each new adjacent node 106, and obtains the target reactive power utilization rate of each new adjacent node; the control system 102 takes the new adjacent node as the current adjacent node 106, and returns to the step of controlling each adjacent node 106 to broadcast its target reactive power utilization rate to new adjacent nodes 106 in the direction away from the node to be regulated 104 until all nodes in the distribution network are traversed, and the target reactive power utilization rate of each node is obtained; the control center 102 generates a reactive power output instruction for each node according to the target reactive power utilization rate of each node to instruct each node to perform a voltage regulation operation.

[0056] In an exemplary embodiment, as Figure 2 shown, a distributed cooperative voltage regulation method for a distribution network is provided. Taking the control center 102 in Figure 1 as an example, the method includes the following steps S202 to step S212. Among them:

[0057] Step S202: Determine the node to be regulated from all nodes according to the monitoring data of each node in the distribution network.

[0058] Among them, the distribution network is the part of the power system that transmits electric energy from the transmission network or power plant to the end users (such as households, commercial users, and industrial users). It is mainly responsible for reducing high-voltage electricity to low-voltage electricity so that civil and industrial users can use it safely. The distribution network usually includes substations, distribution lines, distribution transformers, and power supply equipment, etc.

[0059] Among them, the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device, which is usually the access point of a distributed photovoltaic power generation system. Among them, the miniature circuit breaker is used for overload and short-circuit protection, isolating the photovoltaic system from the power grid (such as the output end of the inverter or the user's distribution box); among them, the photovoltaic inverter converts the direct current of the photovoltaic array into alternating current and synchronizes it with the power grid.

[0060] Among them, the monitoring data includes data such as voltage value, current value, active power, reactive power, and power factor.

[0061] Optionally, the control system periodically collects the monitoring data of each node in the distribution network. According to the monitoring data of each node, it judges whether there is a situation where the voltage of a node exceeds the limit, and determines the node with the voltage exceeding the limit as the node to be regulated. It can be understood that after determining the node to be regulated, an alarm message can be automatically generated to feedback to the technical personnel that there is an abnormality in the node to be regulated.

[0062] Step S204: Obtain the target reactive power utilization rate of the node to be regulated according to the monitoring data of the node to be regulated, the preset sensitivity nonlinear model, and the target voltage value.

[0063] Among them, the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data. The sensitivity nonlinear model characterizes the nonlinear relationship between the voltage and reactive power of the node. The output of this model is the reactive voltage sensitivity, that is, the response degree of voltage change to reactive power regulation.

[0064] Among them, the target voltage value can be the safe voltage value at which each node can operate normally set in advance.

[0065] Among them, the reactive power utilization rate can be the reactive power output, which refers to the ratio of the reactive power output by the photovoltaic inverter to its reactive power capacity provided.

[0066] Optionally, the control system inputs the monitoring data of the node to be regulated and the target voltage value into the preset sensitivity nonlinear model, obtains the reactive voltage sensitivity of the node to be regulated as the output, and further obtains the target reactive power utilization rate of the node to be regulated according to the reactive voltage sensitivity.

[0067] Step S206: Control the node to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of each adjacent node until the preset iteration end condition is met, and obtain the target reactive power utilization rates of each adjacent node.

[0068] Among them, in the topological structure of the distribution network, adjacent nodes can be nodes that have a direct connection relationship with the node to be regulated.

[0069] Among them, the consensus algorithm model can be an algorithm that coordinates the states of each node through information exchange between nodes to ensure that a consistent goal is achieved among multiple nodes.

[0070] Optionally, the control system controls the node to be regulated to broadcast the calculated target reactive power utilization rate along the topological connection to adjacent nodes directly connected to the node to be regulated, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of each adjacent node until the preset iteration end condition is met, and obtain the target reactive power utilization rates of each adjacent node adjacent to the node to be regulated.

[0071] Step S208: Control each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the node to be regulated, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, use the consensus algorithm model to iteratively calculate the target reactive power utilization rates of each new adjacent node, and obtain the target reactive power utilization rates of each new adjacent node.

[0072] It can be understood that in the topological structure of the distribution network, there are different connection relationships between each node. The node to be regulated has directly connected adjacent nodes, and the adjacent nodes also have other adjacent nodes directly connected to themselves.

[0073] Optionally, the control system controls the adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the node to be regulated. It should be noted that when an adjacent node broadcasts to a new adjacent node, a node that has already iteratively calculated its own target reactive power utilization rate does not need to broadcast. The control system uses the consensus algorithm model to iteratively calculate the target reactive power utilization rates of each new adjacent node based on the target reactive power utilization rate sensitivity nonlinear model and the monitoring data of the new adjacent nodes, and obtains the target reactive power utilization rates of each new adjacent node.

[0074] Step S210: Take the new adjacent node as the current adjacent node, and return to the step of controlling each adjacent node to broadcast its target reactive power utilization rate to the new adjacent node in the direction away from the node to be regulated, until all nodes in the distribution network are traversed to obtain the target reactive power utilization rate of each node.

[0075] Among them, traversal can refer to systematically accessing or checking all elements (or nodes) in a certain structure according to a certain rule or order, ensuring that each element is processed without repetition. In this embodiment, according to the rule of broadcasting the target reactive power utilization rate and the rule of iteratively calculating the target reactive power utilization rate of nodes, each node in the distribution network is traversed.

[0076] Optionally, the control system takes the new adjacent node as the current adjacent node, and returns to the step of controlling each adjacent node to broadcast its target reactive power utilization rate to the new adjacent node in the direction away from the node to be regulated, until all nodes in the distribution network are traversed according to the rule of broadcasting the target reactive power utilization rate and the rule of iteratively calculating the target reactive power utilization rate of nodes to obtain the target reactive power utilization rate of each node.

[0077] Step S212: Generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node to instruct each node to perform a voltage regulation operation.

[0078] Among them, the reactive power output instruction is used to instruct the photovoltaic inverters of each node to adjust the reactive power output according to the target reactive power utilization rate, thereby regulating the voltage of each node.

[0079] Optionally, the control system generates a reactive power output instruction for each node according to the target reactive power utilization rate of each node, and sends it to reactive power devices such as the photovoltaic inverters of each node, and then completes the voltage regulation operation of each node by adjusting the reactive power output of reactive power devices such as photovoltaic inverters.

[0080] In the above-mentioned distributed collaborative voltage regulation method for the distribution network, the method determines the nodes to be regulated from all nodes according to the monitoring data of each node in the distribution network. Among them, the nodes are the nodes equipped with miniature circuit breakers and photovoltaic inverter devices, so as to grasp the operating conditions of each node in real time and obtain the data support for subsequent voltage regulation decisions. Further, according to the monitoring data of the nodes to be regulated, the preset sensitivity nonlinear model, and the target voltage value, the target reactive power utilization rate of the nodes to be regulated is obtained; the sensitivity nonlinear model is obtained by training the preset neural network with historical monitoring data; and the nodes to be regulated are controlled to broadcast the target reactive power utilization rate to adjacent nodes, and according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of adjacent nodes, the consensus algorithm model is used to iteratively calculate the target reactive power utilization rate of each adjacent node until the preset iteration end condition is met, and the target reactive power utilization rate of each adjacent node is obtained; in addition, each adjacent node is controlled to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the nodes to be regulated, and according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, the consensus algorithm model is used to iteratively calculate the target reactive power utilization rate of each new adjacent node, and the target reactive power utilization rate of each new adjacent node is obtained; the new adjacent nodes are used as the current adjacent nodes, and the step of controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in the direction away from the nodes to be regulated is returned until all nodes in the distribution network are traversed, and the target reactive power utilization rate of each node is obtained, realizing the information exchange between each node. According to the target reactive power utilization rate of each node, a reactive power output instruction for each node is generated to instruct each node to perform voltage regulation operations. The sensitivity nonlinear model can dynamically track the nonlinear relationship between node voltage and reactive power, providing a more accurate decision-making basis for regulation, and the use of the consensus algorithm model can efficiently achieve the collaborative regulation between nodes. Through information exchange and sensitivity weighting between nodes, it ensures that the system can achieve coordinated voltage regulation for the whole network or local areas in the face of load fluctuations and photovoltaic power generation fluctuations, thereby improving the overall adaptability.

[0081] In an exemplary embodiment, the monitoring data includes voltage values. Step S202 determines the nodes to be regulated from all nodes according to the monitoring data of each node in the distribution network, including:

[0082] Determine the comparison results between the voltage values of each node in the distribution network and the preset voltage threshold range; according to the comparison results, determine the nodes with voltage values exceeding the voltage threshold range as the nodes to be regulated.

[0083] Among them, the comparison results include that the voltage value is within the preset voltage threshold range, or the voltage value exceeds the preset voltage threshold range. Among them, the preset voltage threshold range can be a voltage value range that can ensure the normal operation of each node set in advance.

[0084] Optionally, the control system compares the voltage values in the monitoring data of each node in the distribution network with a preset voltage threshold range respectively to obtain the comparison result of each node. According to the comparison result of each node, the nodes with the comparison result that the voltage value exceeds the voltage threshold range are determined as the nodes to be regulated.

[0085] In this embodiment, by real-time monitoring the voltage values of each node in the distribution network and comparing them with a preset voltage threshold range, it is possible to determine whether there is an abnormality in the voltage values of each node, timely grasp the abnormal situation of the distribution network, and use the nodes with abnormalities as the nodes to be regulated to participate in the collaborative role as information sources in subsequent distributed regulation.

[0086] In an exemplary embodiment, step S204 obtains the target reactive power utilization rate of the node to be regulated according to the monitoring data of the node to be regulated, a preset sensitivity nonlinear model, and a target voltage value, including:

[0087] According to the voltage value of the node to be regulated and the target voltage value, the voltage change amount of the node to be regulated is obtained; the voltage change amount of the node to be regulated and other monitoring data except the voltage value are input into the inverse function of the sensitivity nonlinear model to obtain the target reactive power utilization rate of the node to be regulated.

[0088] Among them, the voltage change amount can be the difference between the voltage value and the target voltage value. Among them, the inverse function can be the reverse operation of a function.

[0089] Optionally, the control system calculates the difference between the voltage value of the node to be regulated and the target voltage value to obtain the voltage change amount of the node to be regulated, and inputs the voltage change amount of the node to be regulated and other monitoring data except the voltage value into the inverse function of the sensitivity nonlinear model to obtain the target reactive power utilization rate of the node to be regulated. The corresponding calculation method is:

[0090]

[0091] Among them, k represents the number of iterations, r is the learning rate of the sensitivity nonlinear model, ΔV i is the voltage change amount between the current voltage of node i and the target voltage, is the inverse function of the sensitivity nonlinear model, indicating the required reactive power regulation amount calculated according to the voltage deviation. It should be noted that the number of iterations of k here represents the number of voltage regulation times of the node to be regulated. Before performing the voltage regulation operation, the voltage change amount between the voltage value in the current monitoring data and the target voltage is used to calculate the target reactive power utilization rate.

[0092] It should be noted that the method for determining the sensitivity non - linear model is as follows: Determine the neural network of the sum of elements and its input and output. At each miniature circuit breaker, based on the locally collected voltage, current, active power, and reactive power data, construct a lightweight neural network. This neural network is used to capture the non - linear relationship between voltage changes and reactive power regulation. The inputs of the sensitivity non - linear model are the voltage change amount, reactive power, active power, and power factor of the node, and the output is the reactive voltage sensitivity of the node, that is, the degree of response of voltage change to reactive power regulation. Specifically, the structure of the network model is: Input layer: Input vector ; Hidden layer: A shallow multi - layer perceptron with ReLU activation function, having 3 hidden layers; Output layer: The output of the reactive voltage sensitivity model, representing the response degree of voltage to reactive power regulation , that is, the reactive voltage sensitivity. Among them, ΔV i is the voltage change amount of node i, I i is the current of node i, P i is the active power of node i, Q i is the reactive power of node i, PF i is the power factor of node i. The expression of the initial sensitivity non - linear model includes:

[0093]

[0094] Among them, represents the training parameters of the neural network, represents the voltage change amount, is the reactive power, is the active power, is the power factor. Further, the training process of the sensitivity non - linear model is: Optimize the model parameters by minimizing the prediction error , that is, try to reduce the difference between the reactive voltage sensitivity predicted by the model and the actual observed value. The training objective is:

[0095]

[0096] Among them, is the actual reactive power-voltage sensitivity obtained from historical data, and T is the total amount of training data. To cope with the dynamic changes in the electrical state of nodes and load fluctuations, the training of the initial sensitivity nonlinear model adopts an online recursive training method, that is, when each new data point arrives, the model parameters are updated in real time. This training method does not require storing a large amount of historical data or performing offline training, but dynamically optimizes the model through incremental updates (or recursive updates). The core of the online recursive training method is to adjust the model parameters using the current sample without traversing all historical data. In this embodiment, the recursive least squares method is used to achieve this goal. The update process of this method can be expressed by the following formula:

[0097]

[0098] In the formula, are the network parameters of node i after the t-th iteration, is the covariance matrix, which is used to adjust the speed of the learning process, is the learning rate, which controls the step size of model update, is the regularization factor, which is used to balance the weights of historical information and current data. is the prediction error, which reflects the deviation between the predicted value and the actual value of the current model. When each new sampling period t + 1 arrives, the node calculates the current voltage change and reactive power and other parameters, inputs them into the neural network, and obtains a new predicted value of reactive power-voltage sensitivity. Then, the node compares this predicted value with the actual observed value to calculate the error . Using the recursive least squares method, the network weights are adjusted through the error , so that the model can more accurately predict the relationship between the new voltage change and reactive power regulation.

[0099] In this embodiment, the sensitivity nonlinear model realizes local self-modeling of the distribution network, avoids relying on traditional complex power flow calculation models, realizes dynamic tracking of the relationship between node voltage and reactive power changes, and reversely solves the target reactive power utilization rate of the nodes to be regulated through the inverse function of the sensitivity nonlinear model, improving the calculation efficiency and further enhancing the adaptability of the control system.

[0100] In an exemplary embodiment, as Figure 3 shown, the content of iteratively calculating the target reactive power utilization rate of each adjacent node according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of adjacent nodes in step S206 until the preset iteration end condition is met to obtain the target reactive power utilization rate of each adjacent node includes:

[0101] Step S302: Obtain the voltage change of each adjacent node.

[0102] Optionally, the control system obtains the target voltage value of each adjacent node, and determines the difference between the current voltage value and the target voltage value in the monitoring data of each adjacent node as the voltage change of each adjacent node.

[0103] Step S304: Input the voltage change of the monitoring data of each adjacent node and other monitoring data except the voltage value into the sensitivity non-linear model to obtain the reactive voltage sensitivity of each adjacent node.

[0104] Optionally, the control system inputs the voltage change of the monitoring data of the adjacent node and other monitoring data except the voltage value, including: current, active power, reactive power, and power factor, into the sensitivity non-linear model to obtain the reactive voltage sensitivity of each adjacent node.

[0105] Step S306: Input the reactive voltage sensitivity of the adjacent node and the target reactive utilization rate into the consensus algorithm model until the preset iteration end condition is met, and obtain the target reactive utilization rate of each adjacent node.

[0106] Among them, the expression corresponding to the consensus algorithm model includes:

[0107]

[0108] Among them, k represents the number of iterations, represents the target reactive utilization rate, r and respectively represent the learning rate and training parameter of the sensitivity non-linear model, Q i represents the reactive power of the i-th node, is the Chebyshev filter acceleration operator, W represents the state transition matrix composed of the state transition weights of each communication link in the distribution network, represents the reactive voltage sensitivity of node i, represents the sensitivity weighting matrix, represents the step size, represents the set of adjacent nodes of node i, and j represents the j-th adjacent node.

[0109] Among them, the preset iteration end condition includes any one of the voltage values of all nodes being within the voltage threshold range, the difference between the target reactive utilization rates of two adjacent updates being lower than the preset threshold, and the number of iterations reaching the preset iteration number threshold.

[0110] Optionally, the control system inputs the reactive voltage sensitivity and target reactive power utilization rate of adjacent nodes into the consensus algorithm model for iterative calculation until any one of the voltage values of all nodes satisfies the preset iterative end condition within the voltage threshold range, the difference between the target reactive power utilization rates of two adjacent updates is lower than the preset threshold, and the number of iterations reaches the preset iteration number threshold, so as to obtain the target reactive power utilization rate of each adjacent node.

[0111] In this embodiment, in order to accelerate the convergence speed of the consensus algorithm model, especially in application scenarios where the system requires high precision and fast response, a Chebyshev polynomial filter is introduced. This filter can accelerate the convergence process of the consensus algorithm and avoid the problem of excessive oscillation caused by high-frequency fluctuations or network disturbances. By designing a suitable filtering function, the Chebyshev filter controls the step size of each state update, making the iterative process smoother, thereby improving the voltage regulation efficiency of the distribution network.

[0112] In an exemplary embodiment, step S306 inputs the reactive voltage sensitivity and target reactive power utilization rate of adjacent nodes into the consensus algorithm model until the preset iterative end condition is satisfied, and obtains the target reactive power utilization rate of each adjacent node, including:

[0113] For each adjacent node, the reactive voltage sensitivity and target reactive power utilization rate of the adjacent node are respectively input into the model of the consensus algorithm to obtain the candidate reactive power utilization rate; when the candidate reactive power utilization rate satisfies the preset inverter constraint, the candidate reactive power utilization rate is used as the target reactive power utilization rate after this round of iterative update; when the candidate reactive power utilization rate does not satisfy the preset inverter constraint, a projection operation is performed on the candidate reactive power utilization rate according to the inverter constraint to obtain the target reactive power utilization rate after this round of iterative update; return to the step of inputting the reactive voltage sensitivity and target reactive power utilization rate of adjacent nodes into the consensus algorithm model until the preset iterative end condition is satisfied, and obtain the target reactive power utilization rate of each adjacent node.

[0114] Among them, the candidate reactive power utilization rate is the reactive power utilization rate obtained by iterative calculation without judging whether it conforms to the inverter constraint.

[0115] Among them, the inverter constraint can be the physical constraint of the photovoltaic inverter, including:

[0116]

[0117]

[0118] Among them, Q i (k) represents the reactive power obtained by the i-th node in the k-th iteration, Q min represents the minimum reactive power capacity of the photovoltaic inverter, Q maxRepresents the maximum reactive power capacity of the PV inverter, cosφ i Represents the power factor of the i-th node.

[0119] Optionally, for each adjacent node, the control system inputs the reactive voltage sensitivity and the target reactive power utilization rate of the adjacent node into the model of the consensus algorithm respectively to obtain the candidate reactive power utilization rate; updates the current reactive power and power factor with the candidate reactive power utilization rate, and compares and analyzes the updated reactive power and the updated power factor with the preset inverter constraints respectively. When the candidate reactive power utilization rate meets the preset inverter constraints, the control system uses the candidate reactive power utilization rate as the target reactive power utilization rate updated in this round of iteration; when the candidate reactive power utilization rate does not meet the preset inverter constraints, a projection operation is performed on the candidate reactive power utilization rate according to the inverter constraints to obtain the target reactive power utilization rate updated in this round of iteration, and the corresponding projection operation includes:

[0120]

[0121]

[0122] Uses the reactive power utilization rate corresponding to the reactive power and power factor projected into the inverter constraint range as the updated target reactive power utilization rate. The control system returns to the step of inputting the reactive voltage sensitivity and the target reactive power utilization rate of the adjacent node into the consensus algorithm model until the preset iteration end condition is met, and obtains the target reactive power utilization rate of each adjacent node.

[0123] In this embodiment, after each round of iteration, a feasibility check is performed on the calculated candidate reactive power utilization rate, and the local sensitivity model and the inverter constraints are considered to correct the updated state. When each node updates the state, over-regulation is avoided when the voltage is already close to the target interval, thereby improving safety and stability.

[0124] In an exemplary embodiment, as Figure 4 shown, a distributed autonomous collaborative voltage regulation method for a distribution network based on an intelligent miniature circuit breaker is provided, including:

[0125] Step 1: Self-detection of the node state of the intelligent miniature circuit breaker. The intelligent miniature circuit breaker is used to periodically collect data such as the voltage, current, active power, reactive power, and power factor of each node (monitoring data), monitor the node voltage state in real time, judge whether there is a voltage over-limit situation, and automatically generate an alarm message, and mark the abnormal node as a "node to be regulated".

[0126] Exemplarily, in the distributed self-regulating and cooperative voltage regulation method of this embodiment, voltage state self-detection is the first step, mainly responsible for real-time monitoring of the voltage and power states of each node, promptly detecting and marking the nodes with voltage over-limit, so as to initiate the subsequent regulation process. During the periodic acquisition process, each intelligent miniature circuit breaker will obtain multiple electrical parameters of its own node, including voltage (Vi), current (Ii), active power (Pi), reactive power (Qi), power factor (PFi), etc. Through the local sensors and calculation modules, these data can be obtained in real time within each sampling period T. At the end of each period, the intelligent miniature circuit breaker will analyze the collected voltage data to determine whether the voltage exceeds the preset upper and lower voltage limits (voltage threshold range). When it is determined that the voltage of a certain node is over-limit, the intelligent miniature circuit breaker will automatically generate an alarm message and mark this node as a "node to be regulated". This node will become one of the leading nodes in the subsequent regulation process, responsible for initiating regulation operations, or participating in the collaborative role as an information source in the subsequent distributed regulation.

[0127] Step 2: Local self-modeling of intelligent miniature circuit breakers. A local "reactive power-voltage" sensitivity nonlinear model based on a lightweight neural network is proposed, and an online recursive training method is used to update the parameters in real time, realizing the dynamic tracking of the relationship between node voltage and reactive power change, without relying on the distribution network power flow model.

[0128] Step 3: Determination of the regulation target of the leading node. For the abnormal nodes (nodes to be regulated) identified in Step 1, mark them as leading nodes, and according to the local sensitivity model obtained in Step 2, calculate the target reactive power utilization rate or target reactive power output required for the node voltage to return to the safe range, providing an initial reference basis for the subsequent distributed cooperative control.

[0129] Step 4: Construct an improved consensus self-regulating and cooperative regulation model. On the basis of the traditional consensus algorithm, the Chebyshev polynomial filtering method is introduced for algorithm improvement. By fusing the local reactive power-voltage sensitivity information of the nodes, an improved consensus iterative equation (consensus algorithm model) with "sensitivity weighting" is constructed, making the distributed iteration more accurate, fast and stable.

[0130] Exemplarily, the basic principle of the consensus algorithm is to coordinate the states of each node through information exchange between nodes to ensure that a consistent goal is achieved among multiple nodes. In the present invention, the target reactive power utilization rate of each node i is updated according to the target states of its neighbor nodes (adjacent nodes) j in each iteration. The iterative formula of the traditional classic consensus algorithm is as follows:

[0131]

[0132] Wherein, is the node i the target reactive power utilization rate or output at the k-th iteration is the step size is the set of neighbor nodes of node i

[0133] Traditional consensus algorithms do not consider the non-linear response characteristics between nodes, while the present invention introduces the reactive power-voltage sensitivity model of each node , so that the state update of each node is not only affected by the state difference of neighbor nodes, but also weighted and adjusted according to local sensitivity information. Specifically, the target state update formula of each node i is modified as:

[0134]

[0135] where is the reactive power-voltage sensitivity function obtained by training based on the lightweight neural network in step 2, which dynamically reflects the non-linear impact of node voltage changes on reactive power regulation. Nodes with high sensitivity values will update their states more actively during the regulation process, while nodes with lower sensitivity will be more conservative in reactive power adjustment, thus avoiding over-regulation of the system

[0136] To accelerate the convergence speed of the consensus algorithm, especially in application scenarios where the system requires high precision and fast response, the present invention introduces a Chebyshev polynomial filter. This filter can accelerate the convergence process of the consensus algorithm and avoid the over-oscillation problem caused by high-frequency fluctuations or network disturbances. The Chebyshev filter controls the step size of each state update by designing an appropriate filtering function, making the iterative process smoother. Specifically, the form of the Chebyshev filter is as follows:

[0137]

[0138]

[0139] where is the core function of the Chebyshev filter, cosh and sech represent the hyperbolic cosine function and hyperbolic secant function respectively, which can generate appropriate acceleration curves and control the smoothness and acceleration effect during the convergence process by adjusting k and x is the Chebyshev filter acceleration operator, W represents the state transition matrix composed of the state transition weights of each communication link in the distribution network is the control parameter of the filter, which determines the filtering accuracy and response speed. It can be understood that there are communication links between each node in the distribution network, and weights are pre-set for the transmission of control state variables between two nodes

[0140] The iterative model of the improved consensus algorithm for the sub-invention is as follows:

[0141]

[0142] where represents the system state variable vector at the (k + 1)-th iteration accelerates the convergence process of the consensus algorithm through a Chebyshev filter is the reactive power-voltage sensitivity function is the sensitivity weighting matrix, representing the contribution of each node to the state update in the iteration

[0143] Step 5: Distributed consensus collaborative voltage regulation iteration in the distribution network. After the leading node initiates control, each node exchanges the current target reactive power utilization rate or output state with its neighbor nodes, and performs distributed iterative updates through the improved consensus algorithm proposed in Step 4. When each node performs iterative updates, it simultaneously considers the constraints of the local sensitivity model to ensure that the iterative process meets the physical limit conditions of the inverter until the whole network or local nodes reach collaborative control convergence

[0144] Exemplarily, for state exchange and initial setting, the leading node first broadcasts the target reactive power utilization rate calculated at the current moment to all nodes within its neighborhood, where:

[0145]

[0146] is the weighting coefficient in the state transition matrix is the target reactive power utilization rate of node j in the previous iteration. After receiving this information, the neighboring nodes initialize their own states according to the improved consensus algorithm proposed in Step 4, combining local sensitivity information and the filtering acceleration matrix. At the same time, set the maximum number of iterations k max , the upper and lower voltage limits and are 1.05 p.u. and 0.95 p.u. respectively, the voltage deviation threshold and the state change threshold

[0147] Furthermore, for distributed consensus iterative updates, in each iteration cycle k, each node updates the target reactive power utilization rate or reactive power output according to the states of its neighboring nodes and the local sensitivity model

[0148]

[0149] where is the Chebyshev filter acceleration operator is the reactive power-voltage sensitivity function of node i, represents the sensitivity weighting matrix, is the step size. Considering the local sensitivity model and inverter constraints, the updated state is corrected. When updating the state of each node, it is necessary to combine the reactive power-voltage nonlinear characteristics reflected in the local sensitivity model to avoid excessive adjustment when the voltage is already close to the target interval. Therefore, after each iteration, the feasibility of the updated reactive power utilization or reactive power output is verified to ensure that the physical constraints of the inverter are met.

[0150] Judge the iteration termination condition and convergence situation. When the system repeats the above state exchange and local update steps, the target reactive power utilization or output of each node will continuously approach a certain stable point. When any of the following conditions is met, it can be determined that the iteration converges and the update is terminated: Voltage deviation threshold: The voltage of each node has entered the interval, and the voltage fluctuation is less than the preset threshold; State change threshold: The change in reactive power utilization between two adjacent iterations has been lower than the set threshold; Maximum number of iterations: Reach the preset maximum number of iterations k max .

[0151] Step 6: Reactive power control instruction execution and closed-loop feedback. After the iteration converges, each intelligent miniature circuit breaker sends specific reactive power output instructions to reactive power devices such as local PV inverters according to the collaborative iteration results, completes the voltage regulation action, and monitors the voltage status in real time to form a control closed-loop and continuously optimize and adjust the voltage regulation effect.

[0152] Exemplarily, when the reactive power regulation targets of all nodes converge through the consensus algorithm, the nodes will calculate and determine their respective reactive power utilization rates according to the final collaborative iteration results. This target reactive power output is the necessary adjustment amount to ensure that the voltage is restored to the safe range and the system is stable. The intelligent miniature circuit breaker will send specific reactive power control instructions to reactive power devices such as local PV inverters through its control module. These devices will adjust their reactive power output according to the received instructions, thereby regulating the grid voltage. Once the reactive power control instructions are issued and executed, the intelligent miniature circuit breaker will monitor the voltage status in real time. By periodically collecting voltage data, the system will detect whether the voltage has been restored to the target voltage range and , and ensure that no node in the system continues to exceed the limit.

[0153] In this embodiment, precise and efficient voltage regulation can be achieved. By using the distributed cooperative regulation method of local self-modeling and sensitivity weighting, the voltage of each node can be precisely regulated. Especially in the case of large fluctuations in photovoltaic power generation, it can respond to voltage changes in real time to ensure that the grid voltage remains within a safe range. The sensitivity model can dynamically track the non-linear relationship between node voltage and reactive power, providing a more accurate decision-making basis for regulation and avoiding the problems of lag and insufficient accuracy in traditional voltage regulation methods. It does not rely on complex power flow models, reducing the computational complexity. Through intelligent miniature circuit breakers and local self-modeling, the dependence on traditional complex power flow calculation models is avoided. In a distribution network with high-penetration photovoltaic power generation, traditional power flow analysis methods have a large computational load and lag in updating. However, the local self-modeling method provided in this embodiment can dynamically update the model in real time according to on-site data, reducing the computational burden and improving the system response speed. It supports distributed cooperative control, enhancing the system robustness and flexibility. The introduced distributed consensus algorithm combined with Chebyshev filtering acceleration can efficiently achieve cooperative regulation among nodes. Through information exchange and sensitivity weighting among nodes, it ensures that the system can achieve coordinated voltage regulation for the entire network or local areas in the face of load fluctuations and photovoltaic power generation fluctuations. It reduces on-site commissioning and operation and maintenance costs and increases the intelligent regulation ability. A lightweight neural network is used to train the reactive power-voltage sensitivity model, and the model parameters are updated in real time through an online recursive training method. In this way, the system can dynamically optimize the regulation strategy according to the real-time state of the distribution network, enhancing the system's adaptive ability. It supports plug-and-play and remote operation and maintenance. After on-site installation, no secondary commissioning is required, reducing manual intervention in the system deployment and commissioning process and lowering the operation and maintenance costs.

[0154] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown in sequence according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order limit for the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same moment, but can be executed at different moments. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or in turn with at least some of the steps or stages in other steps or other steps.

[0155] Based on the same inventive concept, an embodiment of the present application further provides a distribution network distributed collaborative voltage regulation device for implementing the distribution network distributed collaborative voltage regulation method involved above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the distribution network distributed collaborative voltage regulation device provided below can refer to the limitations on the distribution network distributed collaborative voltage regulation method in the above text, and will not be repeated here.

[0156] In an exemplary embodiment, as Figure 5 shown, a distribution network distributed collaborative voltage regulation device 500 is provided, including: a node determination module 501, a reactive power determination module 502, an iterative update module 503, and an instruction output module 504, where:

[0157] The node determination module 501 is configured to determine, from all nodes, the nodes to be regulated according to the monitoring data of each node in the distribution network; the nodes are the nodes configured with miniature circuit breakers and photovoltaic inverter devices;

[0158] The reactive power determination module 502 is configured to obtain the target reactive power utilization rate of the nodes to be regulated according to the monitoring data of the nodes to be regulated, a preset sensitivity nonlinear model, and a target voltage value, and the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data.

[0159] The iterative update module 503 is configured to control the nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and perform iterative calculation on the target reactive power utilization rates of each adjacent node by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes until a preset iteration end condition is met, so as to obtain the target reactive power utilization rates of each adjacent node.

[0160] The iterative update module 503 is further configured to control each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes along the direction away from the nodes to be regulated, and perform iterative calculation on the target reactive power utilization rates of each new adjacent node by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, so as to obtain the target reactive power utilization rates of each new adjacent node.

[0161] The iterative update module 503 is further configured to use the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes along the direction away from the nodes to be regulated until all nodes in the distribution network are traversed, so as to obtain the target reactive power utilization rate of each node.

[0162] The instruction output module 504 is configured to generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node, so as to instruct each node to perform a voltage regulation operation.

[0163] Further, in one embodiment, the node determination module 501 is further configured to determine the comparison result between the voltage value of each node in the distribution network and a preset voltage threshold range; according to the comparison result, determine the nodes whose voltage values exceed the voltage threshold range as the nodes to be regulated.

[0164] Further, in one embodiment, the reactive power determination module 502 is further configured to obtain the voltage change amount of the node to be regulated according to the voltage value and the target voltage value of the node to be regulated; input the voltage change amount of the node to be regulated and other monitoring data except the voltage value into the inverse function of the sensitivity non-linear model to obtain the target reactive power utilization rate of the node to be regulated.

[0165] Further, in one embodiment, the iterative update module 503 is further configured to obtain the voltage change amount of each adjacent node; input the voltage change amounts of the monitoring data of each adjacent node and other monitoring data except the voltage value into the sensitivity non-linear model to obtain the reactive power-voltage sensitivity of each adjacent node; input the reactive power-voltage sensitivity of the adjacent node and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is met, so as to obtain the target reactive power utilization rate of each adjacent node.

[0166]

[0167] Further, in one embodiment, the iterative update module 503 is further configured to, for each adjacent node, input the reactive power-voltage sensitivity and the target reactive power utilization rate of the adjacent node into the model of the consensus algorithm respectively to obtain a candidate reactive power utilization rate; in the case where the candidate reactive power utilization rate meets the preset inverter constraint, use the candidate reactive power utilization rate as the target reactive power utilization rate after this round of iterative update; in the case where the candidate reactive power utilization rate does not meet the preset inverter constraint, perform a projection operation on the candidate reactive power utilization rate according to the inverter constraint to obtain the target reactive power utilization rate after this round of iterative update; return to the step of inputting the reactive power-voltage sensitivity and the target reactive power utilization rate of the adjacent node into the consensus algorithm model until a preset iteration end condition is met, so as to obtain the target reactive power utilization rate of each adjacent node.

[0167] Each module in the above distribution network distributed collaborative voltage regulation device 500 can be implemented in whole or in part by software, hardware and their combination. The above modules can be embedded in the processor in the computer device in the form of hardware or be independent of the processor, or can be stored in the memory in the computer device in the form of software, so as to facilitate the processor to call and execute the operations corresponding to the above modules.

[0168] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structural diagram may be as shown in Figure 6 . The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store data such as monitoring data and target reactive power utilization rate. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a method for distributed collaborative voltage regulation of a distribution network.

[0169] Those skilled in the art can understand that Figure 6 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0170] In an embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0171] In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0172] In an embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0173] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0174] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered to be within the scope recorded in the present application.

[0175] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A distributed collaborative voltage regulation method for a distribution network, characterized in that, The method includes: Based on the monitoring data of each node in the distribution network, determining, from all the nodes, the nodes to be regulated; the nodes are the nodes configured with miniature circuit breakers and photovoltaic inverter devices; Based on the voltage value and the target voltage value in the monitoring data of the nodes to be regulated, obtaining the voltage change amount of the nodes to be regulated; inputting the voltage change amount of the nodes to be regulated and other monitoring data except the voltage value into the inverse function of a preset sensitivity nonlinear model to obtain the target reactive power utilization rate of the nodes to be regulated; the target voltage value is the safe voltage value at which each of the nodes can operate normally; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data, the sensitivity nonlinear model characterizes the nonlinear relationship between the voltage and reactive power of the nodes, and the output is the reactive voltage sensitivity; the reactive voltage sensitivity characterizes the response degree of voltage change to reactive power regulation; the expression corresponding to the inverse function of the sensitivity nonlinear model includes: ； Among them, represents the target reactive power utilization rate, k represents the number of iterations, r is the learning rate of the sensitivity nonlinear model, and ΔV i is the voltage change of node i, is the inverse function of the sensitivity nonlinear model, representing the required reactive power regulation amount calculated according to the voltage deviation; Controlling the nodes to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes, using a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each adjacent node until a preset iteration end condition is met, to obtain the target reactive power utilization rate of each adjacent node; Controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the nodes to be regulated, and based on the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the new adjacent nodes, using a consensus algorithm model to iteratively calculate the target reactive power utilization rate of each new adjacent node, to obtain the target reactive power utilization rate of each new adjacent node; Taking the new adjacent nodes as the current adjacent nodes, and returning to the step of controlling each adjacent node to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the nodes to be regulated, until all the nodes in the distribution network are traversed, to obtain the target reactive power utilization rate of each node; Generating a reactive power output instruction for each node according to the target reactive power utilization rate of each node, to instruct each node to perform a voltage regulation operation.

2. The method according to claim 1, wherein The determining, from all the nodes, the nodes to be regulated based on the monitoring data of each node in the distribution network includes: Determining the comparison results between the voltage values of each node in the distribution network and a preset voltage threshold range; Based on the comparison results, determining the nodes whose voltage values exceed the voltage threshold range as the nodes to be regulated.

3. The method according to claim 2, wherein The preset iteration end condition includes any one of that the voltage values of all the nodes are within the voltage threshold range, the difference between the target reactive power utilization rates updated twice adjacent to each other is lower than a preset threshold, and the number of iterations reaches a preset iteration number threshold.

4. The method according to claim 1, wherein Performing iterative calculations on the target reactive power utilization rates of the adjacent nodes by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity nonlinear model, and the monitoring data of the adjacent nodes until a preset iteration end condition is satisfied, to obtain the target reactive power utilization rates of the adjacent nodes, including: Obtaining the voltage change amount of each adjacent node; Inputting the voltage change amounts of the monitoring data of the adjacent nodes and other monitoring data except the voltage value into the sensitivity nonlinear model to obtain the reactive power-voltage sensitivity of each adjacent node; Inputting the reactive power-voltage sensitivity of the adjacent node and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied, to obtain the target reactive power utilization rates of the adjacent nodes.

5. The method according to claim 4, wherein The step of inputting the reactive power-voltage sensitivity of the adjacent node and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied, to obtain the target reactive power utilization rates of the adjacent nodes, includes: For each adjacent node, inputting the reactive power-voltage sensitivity of the adjacent node and the target reactive power utilization rate into the model of the consensus algorithm respectively to obtain the candidate reactive power utilization rate; When the candidate reactive power utilization rate satisfies the preset inverter constraint, using the candidate reactive power utilization rate as the target reactive power utilization rate updated in this round of iteration; When the candidate reactive power utilization rate does not satisfy the preset inverter constraint, performing a projection operation on the candidate reactive power utilization rate according to the inverter constraint to obtain the target reactive power utilization rate updated in this round of iteration; Returning to the step of inputting the reactive power-voltage sensitivity of the adjacent node and the target reactive power utilization rate into the consensus algorithm model until a preset iteration end condition is satisfied, to obtain the target reactive power utilization rates of the adjacent nodes.

6. The method according to any one of claims 1 to 5, characterized in that, The expression corresponding to the consensus algorithm model includes: ； where k represents the number of iterations, represents the target reactive power utilization rate, r and respectively represent the learning rate and training parameters of the sensitivity nonlinear model, Q i represents the reactive power of the i-th node, is the Chebyshev filter acceleration operator, W represents the state transition matrix composed of the state transition weights of each communication link in the distribution network, represents the reactive voltage sensitivity of node i, represents the sensitivity weighting matrix, represents the step size, represents the set of adjacent nodes of node i, and j represents the j-th adjacent node.

7. A distributed collaborative voltage regulation device for a distribution network, characterized in that, The device includes: A node determination module, configured to determine a node to be regulated from all nodes according to the monitoring data of each node in the distribution network; the node is a node configured with a miniature circuit breaker and a photovoltaic inverter device; A reactive power determination module, configured to obtain the voltage change amount of the node to be regulated according to the voltage value and the target voltage value in the monitoring data of the node to be regulated; inputting the voltage change amount of the node to be regulated and other monitoring data except the voltage value into the inverse function of a preset sensitivity nonlinear model to obtain the target reactive power utilization rate of the node to be regulated; the target voltage value is a safe voltage value at which each node can operate normally preset; the sensitivity nonlinear model is obtained by training a preset neural network with historical monitoring data, the sensitivity nonlinear model characterizes the nonlinear relationship between the voltage and reactive power of the node, and the output is the reactive power-voltage sensitivity; the reactive power-voltage sensitivity characterizes the response degree of voltage change to reactive power regulation; the expression corresponding to the inverse function of the sensitivity nonlinear model includes: ； Among them, represents the target reactive power utilization rate, k represents the number of iterations, r is the learning rate of the sensitivity nonlinear model, and ΔV i is the voltage change of node i, is the inverse function of the sensitivity nonlinear model, representing the required reactive power regulation amount calculated according to the voltage deviation; An iterative update module, configured to control the node to be regulated to broadcast the target reactive power utilization rate to adjacent nodes, and perform iterative calculation on the target reactive power utilization rates of the adjacent nodes by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity non-linear model, and the monitoring data of the adjacent nodes until a preset iterative end condition is satisfied, so as to obtain the target reactive power utilization rates of the adjacent nodes; The iterative update module is further configured to control each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated, and perform iterative calculation on the target reactive power utilization rates of the new adjacent nodes by using a consensus algorithm model according to the target reactive power utilization rate, the sensitivity non-linear model, and the monitoring data of the new adjacent nodes, so as to obtain the target reactive power utilization rates of the new adjacent nodes; The iterative update module is further configured to use the new adjacent nodes as the current adjacent nodes, and return to the step of controlling each of the adjacent nodes to broadcast its target reactive power utilization rate to new adjacent nodes in a direction away from the node to be regulated until all the nodes of the distribution network are traversed, so as to obtain the target reactive power utilization rate of each node; An instruction output module, configured to generate a reactive power output instruction for each node according to the target reactive power utilization rate of each node, so as to instruct each node to perform a voltage regulation operation.

8. The distributed cooperative voltage regulation device for a distribution network according to claim 7, wherein The node determination module is further configured to determine the comparison result between the voltage value of each node in the distribution network and a preset voltage threshold range; according to the comparison result, determine the nodes whose voltage values exceed the voltage threshold range as the nodes to be regulated.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images