A power distribution network voltage analysis management and control method, system, device and medium

Abstract

The present application belongs to the technical field of power system, and discloses a power distribution network voltage analysis management and control method, comprising: according to the power distribution network topology, carrying out zoning processing on the power distribution network to obtain active zoning and reactive zoning; constructing a voltage regulation model for each zoning, and constructing an agent according to the initial voltage setting parameters of each zoning; obtaining actual voltage data of the power distribution network, and determining the actual voltage data of each zoning according to the actual voltage data of the power distribution network; updating the strategy network of the agent according to the actual voltage data of each zoning to obtain an actual strategy network; and based on the actual strategy network, regulating the voltage of each zoning through the corresponding voltage regulation model. The present application solves the problem that the power distribution network voltage centralized control model is complex and slow to solve, and cannot achieve real-time decision-making.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a method, system, equipment and medium for voltage analysis and control of distribution networks. Background Technology

[0002] Most of my country's power distribution network is radial, with unidirectional power flow. However, photovoltaic (PV) systems are integrated into the distribution network in a distributed manner, altering the original power flow characteristics and causing problems such as reverse power flow, which in turn lead to voltage limit violations. Therefore, this presents a challenge to voltage control methods for the distribution network.

[0003] Currently, photovoltaic (PV) systems are mostly controlled using centralized methods. Centralized control is suitable when PV penetration in the distribution network is low and the number of nodes in the network is small. However, when a large number of distributed PV systems are randomly and randomly connected to the distribution network, resulting in a large number of installations and dispersed locations, centralized control methods can lead to a high number of variables, a complex solution process, and excessively long optimization times. Summary of the Invention

[0004] This invention provides a method, system, equipment, and medium for voltage analysis and control of power distribution networks, in order to solve the problems of multiple variable dimensions, complex solution process, and excessively long optimization time caused by the centralized photovoltaic control method in the prior art.

[0005] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or to describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0006] According to a first aspect of the present invention, a method for voltage analysis and control of a power distribution network is provided.

[0007] In one embodiment, the power distribution network voltage analysis and control method includes:

[0008] Based on the distribution network topology, the distribution network is divided into active power zones and reactive power zones.

[0009] A voltage regulation model is constructed for each partition, and parameters are set according to the initial voltage of each partition to build an intelligent agent;

[0010] Obtain the actual voltage data of the distribution network, and determine the actual voltage data of each zone based on the actual voltage data of the distribution network;

[0011] The agent's policy network is updated based on the actual voltage data of each partition to obtain the actual policy network;

[0012] Based on the actual policy network, the voltage of each partition is regulated using the corresponding voltage regulation model.

[0013] In one embodiment, partitioning the distribution network according to its topology to obtain active and reactive power partitions includes: partitioning the distribution network according to its topology, based on the Jacobian matrix and spectral clustering algorithm, to obtain active and reactive power partitions.

[0014] In one embodiment, constructing a voltage regulation model for each partition includes: generating an objective function based on the adjustable resource action cost and voltage over-limit penalty coefficient of each partition; and constructing a voltage regulation model for the corresponding partition based on the constraints of each partition and the objective function.

[0015] In one embodiment, the constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

[0016] In one embodiment, constructing an agent based on the initial voltage setting parameters of each partition includes: constructing a voltage state space, a voltage action space, and a voltage reward function for each partition based on the initial voltage setting parameters of each partition; constructing an agent for each partition based on the voltage state space, the voltage action space, and the voltage reward function; and training the agent for each partition based on a multi-agent reinforcement learning method.

[0017] In one embodiment, updating the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network includes: determining the voltage environment state of each partition based on the actual voltage data of each partition; analyzing the voltage environment state of each partition based on the agent's policy network of each partition to obtain the agent's action for each partition; executing the agent's action to obtain the agent's actual state space and reward value; and updating the corresponding agent's policy network based on the agent's action, the agent's state space, and the reward value to obtain the actual policy network.

[0018] In one embodiment, updating the policy network of the corresponding agent based on the agent's actions, the agent's state space, and the reward value to obtain the actual policy network includes: obtaining agent actions, agent state space, and reward values ​​multiple times based on the actual voltage data of each partition; merging the agent actions, agent state space, and reward values ​​obtained each time into round data and storing it in the agent's experience pool; and when the round data in the agent's experience pool reaches a predetermined number, randomly extracting round data from the agent's experience pool to update the agent's policy network to obtain the actual policy network.

[0019] According to a second aspect of the present invention, a power distribution network voltage analysis and control system is provided.

[0020] In one embodiment, the power distribution network voltage analysis and control system includes:

[0021] The voltage partitioning module is used to partition the distribution network according to the distribution network topology to obtain active power partitions and reactive power partitions.

[0022] The partition construction module is used to build a voltage regulation model for each partition and construct an intelligent agent by setting parameters according to the initial voltage of each partition.

[0023] The data acquisition module is used to acquire the actual voltage data of the distribution network and determine the actual voltage data of each zone based on the actual voltage data of the distribution network.

[0024] The policy update module is used to update the policy network of the agent based on the actual voltage data of each partition to obtain the actual policy network.

[0025] The voltage regulation module is used to regulate the voltage of each partition based on the actual policy network and the corresponding voltage regulation model.

[0026] In one embodiment, when the voltage partitioning module partitions the distribution network according to the distribution network topology to obtain active power partitions and reactive power partitions, it partitions the distribution network according to the distribution network topology based on the Jacobian matrix and spectral clustering algorithm to obtain active power partitions and reactive power partitions.

[0027] In one embodiment, the partition construction module includes: a voltage regulation model construction module, used to generate an objective function based on the adjustable resource action cost and voltage over-limit penalty coefficient of each partition; and to construct a voltage regulation model for the corresponding partition based on the constraints of each partition and the objective function.

[0028] In one embodiment, the constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

[0029] In one embodiment, the partition construction module includes: an agent construction module, configured to construct a voltage state space, a voltage action space, and a voltage reward function for each partition based on the initial voltage setting parameters of each partition; construct an agent for each partition based on the voltage state space, the voltage action space, and the voltage reward function; and train the agent for each partition based on a multi-agent reinforcement learning method.

[0030] In one embodiment, when the policy update module updates the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network, it determines the voltage environment state of each partition based on the actual voltage data of each partition; analyzes the voltage environment state of each partition based on the agent's policy network of each partition to obtain the agent's action for each partition; executes the agent's action to obtain the agent's actual state space and reward value; and updates the corresponding agent's policy network based on the agent's action, the agent's state space, and the reward value to obtain the actual policy network.

[0031] In one embodiment, when the policy update module updates the policy network of the corresponding agent based on the agent's actions, the agent's state space, and the reward value to obtain the actual policy network, it obtains the agent's actions, agent's state space, and reward value multiple times based on the actual voltage data of each partition; it merges the agent's actions, agent's state space, and reward value obtained each time into round data and stores it in the agent's experience pool; when the round data in the agent's experience pool reaches a predetermined number, it randomly extracts the round data from the agent's experience pool to update the agent's policy network and obtain the actual policy network.

[0032] According to a third aspect of the present invention, a computer device is provided.

[0033] In some embodiments, the computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0034] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0035] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the above method.

[0036] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0037] This invention divides the distribution network into active and reactive power zones, and constructs a voltage regulation model and intelligent agent for each zone. Based on the intelligent agent and voltage regulation model corresponding to each zone, the voltage of each zone is regulated, thereby enabling rapid voltage regulation of the distribution network under distributed photovoltaic zone autonomy. This solves the problem that the centralized voltage control model of the distribution network is complex and slow to solve, and cannot achieve real-time decision-making.

[0038] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0040] Figure 1 This is a flowchart illustrating a distribution network voltage analysis and control method according to an exemplary embodiment;

[0041] Figure 2 This is a schematic diagram of the structure of a power distribution network voltage analysis and control system according to an exemplary embodiment;

[0042] Figure 3 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0043] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some embodiments may include or substitute parts and features of other embodiments. The scope of the embodiments herein encompasses the entire scope of the claims and all available equivalents thereof. Throughout this document, the terms “first,” “second,” etc., are used only to distinguish one element from another without requiring or implying any actual relationship or order between the elements. Indeed, a first element can also be referred to as a second element, and vice versa. Furthermore, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a structure, apparatus, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a structure, apparatus, or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the structure, apparatus, or device that includes said element. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0044] The terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer" used in this document to indicate orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings. They are used solely for the convenience of describing the document and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description herein, unless otherwise specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two elements; they can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0045] In this document, unless otherwise stated, the term "multiple" means two or more.

[0046] In this article, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.

[0047] In this article, the term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.

[0048] It should be understood that although the steps in the flowchart are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order constraint on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the diagram may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0049] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0050] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0051] Figure 1An embodiment of the distribution network voltage analysis and control method of the present invention is shown.

[0052] In this optional embodiment, the power distribution network voltage analysis and control method includes:

[0053] Step S101: Based on the distribution network topology, the distribution network is partitioned to obtain active power partitions and reactive power partitions.

[0054] Step S103: Construct a voltage regulation model for each partition, and set parameters according to the initial voltage of each partition to construct an intelligent agent;

[0055] Step S105: Obtain the actual voltage data of the distribution network, and determine the actual voltage data of each zone based on the actual voltage data of the distribution network.

[0056] Step S107: Update the policy network of the agent according to the actual voltage data of each partition to obtain the actual policy network;

[0057] Step S109: Based on the actual policy network, the voltage of each partition is regulated using the corresponding voltage regulation model.

[0058] In one embodiment, when partitioning the distribution network according to its topology to obtain active and reactive power partitions, the partitioning can be performed based on the Jacobian matrix and spectral clustering algorithm to obtain active and reactive power partitions.

[0059] Specifically, based on the Jacobian matrix in power flow calculation, the following formula can be obtained:

[0060]

[0061] In the formula, ΔP represents the change in active power injected into the node, ΔQ represents the change in reactive power injected into the node, Δδ represents the change in the phase angle of the node voltage, ΔU represents the change in the magnitude of the node voltage, and A, B, C, and D are matrices that reflect the relationship between the changes in active and reactive power and the changes in the phase angle and magnitude of the voltage.

[0062] By transforming the above matrix, the voltage sensitivity matrix can be obtained:

[0063]

[0064] In the formula, S1 and S2 reflect the relationship between the change in voltage phase angle and the change in injected power, while S3 and S4 reflect the relationship between the change in voltage amplitude and the change in injected power.

[0065] The distribution network is partitioned based on spectral clustering. A fully connected adjacency matrix W is constructed using the adjacency method, where the element in the i-th row and j-th column of the adjacency matrix is ​​w. ijSpecifically, it is expressed as:

[0066]

[0067] In the formula, x i With x j σ represents the elements of the i-th and j-th rows of the voltage sensitivity matrix, σ is the width coefficient of the adjacency matrix, and N is the number of nodes in the distribution network.

[0068] Then, the degree matrix D is calculated. The degree matrix is ​​a diagonal matrix, and the formula for calculating the elements of the i-th row is as follows:

[0069]

[0070] Therefore, the Laplace matrix L = DW can be calculated.

[0071] At this point, the clustering problem can be transformed into a graph partitioning problem for solution, and the objective function can be expressed as:

[0072]

[0073] In the formula, k is the number of clusters formed, and C i For the i-th cluster in the clustering results, C i The complement of vol(C) i Let be the weighted sum of all edges. Based on the degree matrix D, we maximize the similarity within each cluster, transforming the maximization problem into a minimization problem:

[0074]

[0075] stF T F = I

[0076] In the formula, I is the identity matrix, tr(·) is the trace of the matrix, and F is the characteristic moment. Let be the eigenvalues ​​and st be the constraints.

[0077] Find The eigenvalues ​​are obtained, and then the first k eigenvectors of the eigenvalues ​​are standardized to transform them into a feature matrix F. The feature matrix F is then clustered using the k-means clustering method to obtain the zoning results of the power distribution network.

[0078] Using the above method, the distribution network is divided into active and reactive power partitions. The resulting partitions have high coupling between nodes within each partition and low coupling between nodes in different partitions. Therefore, the centralized optimization problem can be decomposed into an active and reactive power partition optimization problem.

[0079] In one embodiment, when constructing a voltage regulation model for each partition, an objective function can be generated based on the adjustable resource action cost and voltage over-limit penalty coefficient of each partition; and a voltage regulation model for the corresponding partition can be constructed based on the constraints of each partition and the objective function.

[0080] Specifically, the formula for generating the objective function based on the adjustable resource action cost and voltage over-limit penalty coefficient for each partition is as follows:

[0081]

[0082]

[0083]

[0084]

[0085]

[0086] In the formula, F is the objective function, w1 is the action cost coefficient, w2 is the voltage over-limit penalty coefficient, and w DR w PV w CB w ESS These are the demand response cost coefficient, photovoltaic output cost coefficient, capacitor cost coefficient, and energy storage charging and discharging cost coefficient, respectively. For demand response costs, To contribute to cost reduction in photovoltaics, For capacitor switching costs, To reduce the cost of energy storage charging and discharging, For voltage over-limit penalties, C pun U max U is the upper limit of the allowable voltage. min U is the lower limit of the allowable voltage. t,max Let be the voltage of all nodes at time t. These are demand response power, photovoltaic power, capacitor level, and energy storage power, respectively.

[0087] In one embodiment, the constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

[0088] Specifically, the formula for calculating the power balance constraint is as follows:

[0089]

[0090]

[0091] In the formula, and For load power, and In order to obtain active and reactive power from the large power grid, and For photovoltaic power, For energy storage power, This represents the capacitor's power.

[0092] The equipment operation constraints include photovoltaic power constraints and capacitor switching constraints. The calculation formula for the photovoltaic power constraints is as follows:

[0093]

[0094]

[0095] In the formula, and The lower and upper limits of adjustable active power for photovoltaic systems. and The upper and lower limits of adjustable reactive power for photovoltaic systems. Active photovoltaic power; This refers to reactive photovoltaic power.

[0096] The formula for calculating the capacitor switching constraint is as follows:

[0097]

[0098]

[0099] In the formula, This is the capacitor setting. N is the reactive power that a group of capacitors can provide. CB This represents the number of times the capacitor operates. This represents the maximum number of times the capacitor can operate. This represents the capacitor's power.

[0100] The energy storage charging and discharging constraints include charging and discharging power constraints and state of charge constraints, and their calculation formulas are as follows:

[0101]

[0102] SOC min ≤SOCt≤SOC max

[0103]

[0104] In the formula, This is the upper limit of charging power. This is the upper limit of discharge power. For energy storage power, SOC t State of SOC min and SOC max η represents the upper and lower limits of the SOC state. ESS This refers to the charge / discharge efficiency.

[0105] The formula for calculating the demand response constraints is as follows:

[0106]

[0107] In the formula, and The lower and upper limits of the demand response load. This refers to the demand response power.

[0108] In one embodiment, when constructing an agent based on the initial voltage setting parameters of each partition, the voltage state space, voltage action space, and voltage reward function of each partition can be constructed based on the initial voltage setting parameters of each partition; the agent of each partition is constructed based on the voltage state space, the voltage action space, and the voltage reward function, and the agent of each partition is trained based on a multi-agent reinforcement learning method.

[0109] Specifically, when constructing the voltage state space, voltage action space, and voltage reward function, the voltage state space of the reactive power partitioning agent is... The definition is as follows:

[0110]

[0111] In the formula, For the predicted reactive power of photovoltaics, Reactive power zone load power.

[0112] The action space of the intelligent agent without functional partitioning. Including photovoltaic reactive power regulation and capacitor settings, it can be expressed as:

[0113]

[0114] In the formula, This refers to reactive photovoltaic power. This is the capacitor setting.

[0115] State space of active zone agent Incorporate observations of reactive actions, defined as follows:

[0116]

[0117] In the formula, For the predicted active power of photovoltaics, SOCt This is the SOC state of energy storage. For active load power, This refers to reactive photovoltaic power. This is the capacitor setting.

[0118] Actions of intelligent agents in functional zones Including photovoltaic active power regulation, demand-side response power, and energy storage charging and discharging power, it can be expressed as:

[0119]

[0120] In the formula, For active photovoltaic power, For demand response power, This refers to energy storage capacity.

[0121] The reward r is taken as the negative of the objective function, transforming the problem into a maximization problem. Specifically, it can be represented as:

[0122] r = -F

[0123] In application, maximum entropy is incorporated into the multi-agent reinforcement learning method. Its objective, besides maximizing cumulative reward, is to maximize entropy, enabling the agent to explore as many policies as possible. The policy π(a|s) represents the probability of choosing action a in state s, defined as:

[0124]

[0125] In the formula, E represents the expectation. For state s t The following action a t The gains obtained; H(π(·|s t H(π(·|s)) is the entropy function. t ))=-logπ(a t |s t ); α is the weight of the entropy function.

[0126] In one embodiment, when updating the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network, the voltage environment state of each partition is determined based on the actual voltage data of each partition; the voltage environment state of each partition is analyzed based on the agent's policy network of each partition to obtain the agent's action for each partition; the agent's action is executed to obtain the agent's actual state space and reward value; the agent's action, agent state space, and reward value are obtained multiple times based on the actual voltage data of each partition; the agent's action, agent state space, and reward value obtained each time are merged into round data and stored in the agent's experience pool; when the round data in the agent's experience pool reaches a predetermined number, the round data in the agent's experience pool is randomly extracted to update the agent's policy network to obtain the actual policy network.

[0127] Specifically, for each partition's agent, the policy network (φ), state-value network Ψ, and target state-value network (Ψ) are randomly initialized. * The Q-network (θ) and the initial experience pool D are set. The total number of training epochs M are set. sum Set the update frequency K of the target state value network. targetV .

[0128] Get the current environmental status of the reactive power partition Each reactive power zone agent obtains actions according to the policy network. Then, based on the current environmental status of the active zone... Each functional zone agent obtains actions based on the policy network.

[0129] Execute action and Interacting with the environment yields a reward value R and an updated environment. and Store round data Go to experience pool D.

[0130] Repeat the above steps until the amount of data in the experience pool D exceeds the set value. Then, randomly select a certain amount of data from the experience pool D to update the network and proceed to the next step.

[0131] Update the policy network. Update the current state value network. Q network, Where λ is the learning efficiency, J π (φ), J V (Ψ), J Q (θ) represents the policy network loss function, the state value network loss function, and the Q network loss function.

[0132] Figure 2 An embodiment of the distribution network voltage analysis and control system of the present invention is shown.

[0133] In this optional embodiment, the power distribution network voltage analysis and control system includes:

[0134] Voltage partitioning module 201 is used to partition the distribution network according to the distribution network topology to obtain active power partitioning and reactive power partitioning.

[0135] The partition construction module 203 is used to construct a voltage regulation model for each partition and construct an intelligent agent by setting parameters according to the initial voltage of each partition.

[0136] The data acquisition module 205 is used to acquire the actual voltage data of the distribution network and determine the actual voltage data of each zone based on the actual voltage data of the distribution network.

[0137] The policy update module 207 is used to update the policy network of the agent according to the actual voltage data of each partition to obtain the actual policy network;

[0138] The voltage regulation module 209 is used to regulate the voltage of each partition based on the actual strategy network and the corresponding voltage regulation model.

[0139] Correspondingly, in one embodiment, when the voltage partitioning module 201 partitions the distribution network according to the distribution network topology to obtain active power partitions and reactive power partitions, it partitions the distribution network according to the distribution network topology based on the Jacobian matrix and spectral clustering algorithm to obtain active power partitions and reactive power partitions.

[0140] Correspondingly, in one embodiment, the partition construction module 203 includes: a voltage regulation model construction module (not shown in the figure), used to generate an objective function based on the adjustable resource action cost and voltage over-limit penalty coefficient of each partition; and to construct a voltage regulation model for the corresponding partition based on the constraints of each partition and the objective function.

[0141] Correspondingly, in one embodiment, the constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

[0142] Correspondingly, in one embodiment, the partition construction module 203 includes: an agent construction module (not shown in the figure), used to construct a voltage state space, a voltage action space, and a voltage reward function for each partition according to the initial voltage setting parameters of each partition; construct an agent for each partition according to the voltage state space, the voltage action space, and the voltage reward function; and train the agent for each partition based on a multi-agent reinforcement learning method.

[0143] Correspondingly, in one embodiment, when the policy update module 207 updates the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network, it determines the voltage environment state of each partition based on the actual voltage data of each partition; analyzes the voltage environment state of each partition based on the agent's policy network of each partition to obtain the agent's action for each partition; executes the agent's action to obtain the agent's actual state space and reward value; and updates the corresponding agent's policy network based on the agent's action, the agent's state space, and the reward value to obtain the actual policy network.

[0144] Correspondingly, in one embodiment, when the policy update module 207 updates the policy network of the corresponding agent based on the agent action, the agent state space, and the reward value to obtain the actual policy network, it obtains the agent action, agent state space, and reward value multiple times based on the actual voltage data of each partition; it merges the agent action, agent state space, and reward value obtained each time into round data and stores it in the agent experience pool; when the round data in the agent experience pool reaches a predetermined number, it randomly extracts the round data from the agent experience pool and updates the agent's policy network to obtain the actual policy network.

[0145] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and 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 stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0146] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0147] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0148] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the method embodiments described above.

[0149] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. 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.

[0150] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A method for voltage analysis and control in a power distribution network, characterized in that, include: Based on the distribution network topology, the distribution network is divided into active power zones and reactive power zones. A voltage regulation model is constructed for each partition, and parameters are set according to the initial voltage of each partition to build an intelligent agent; Obtain the actual voltage data of the distribution network, and determine the actual voltage data of each zone based on the actual voltage data of the distribution network; The agent's policy network is updated based on the actual voltage data of each partition to obtain the actual policy network; Based on the actual policy network, the voltage of each partition is regulated using the corresponding voltage regulation model; The process of constructing an agent based on the initial voltage setting parameters of each partition includes: constructing a voltage state space, a voltage action space, and a voltage reward function for each partition based on the initial voltage setting parameters of each partition; constructing an agent for each partition based on the voltage state space, the voltage action space, and the voltage reward function; and training the agent for each partition based on a multi-agent reinforcement learning method. The process of updating the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network includes: determining the voltage environment state of each partition based on the actual voltage data of each partition; analyzing the voltage environment state of each partition based on the agent's policy network of each partition to obtain the agent's action for each partition; executing the agent's action to obtain the agent's actual state space and reward value; and updating the corresponding agent's policy network based on the agent's action, the agent's state space, and the reward value to obtain the actual policy network. The process of updating the policy network of the corresponding agent based on the agent's actions, the agent's state space, and the reward value to obtain the actual policy network includes: obtaining agent actions, agent state space, and reward values ​​multiple times based on the actual voltage data of each partition; merging the agent actions, agent state space, and reward values ​​obtained each time into round data and storing it in the agent's experience pool; and when the round data in the agent's experience pool reaches a predetermined number, randomly extracting round data from the agent's experience pool to update the agent's policy network to obtain the actual policy network.

2. The distribution network voltage analysis and control method according to claim 1, characterized in that, Based on the distribution network topology, the distribution network is divided into active power zones and reactive power zones, including: Based on the distribution network topology, the distribution network is partitioned using the Jacobian matrix and spectral clustering algorithm to obtain active power partitions and reactive power partitions.

3. The distribution network voltage analysis and control method according to claim 1, characterized in that, Building a voltage regulation model for each partition includes: The objective function is generated based on the adjustable resource action cost and voltage over-limit penalty coefficient for each partition; Based on the constraints of each partition and the objective function, a voltage regulation model for the corresponding partition is constructed.

4. The distribution network voltage analysis and control method according to claim 3, characterized in that, The constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

5. A power distribution network voltage analysis and control system, characterized in that, include: The voltage partitioning module is used to partition the distribution network according to the distribution network topology to obtain active power partitions and reactive power partitions. The partition construction module is used to build a voltage regulation model for each partition and construct an intelligent agent by setting parameters according to the initial voltage of each partition. The data acquisition module is used to acquire the actual voltage data of the distribution network and determine the actual voltage data of each zone based on the actual voltage data of the distribution network. The policy update module is used to update the policy network of the agent based on the actual voltage data of each partition to obtain the actual policy network. The voltage regulation module is used to regulate the voltage of each partition based on the actual policy network and the corresponding voltage regulation model. The partition construction module includes: an agent construction module, used to construct a voltage state space, a voltage action space, and a voltage reward function for each partition based on the initial voltage setting parameters of each partition; to construct an agent for each partition based on the voltage state space, the voltage action space, and the voltage reward function; and to train the agent for each partition based on a multi-agent reinforcement learning method. When the policy update module updates the agent's policy network based on the actual voltage data of each partition to obtain the actual policy network, it determines the voltage environment state of each partition based on the actual voltage data of each partition; it analyzes the voltage environment state of each partition based on the agent's policy network of each partition to obtain the agent's action for each partition; it executes the agent's action to obtain the agent's actual state space and reward value; and it updates the corresponding agent's policy network based on the agent's action, the agent's state space, and the reward value to obtain the actual policy network. When the policy update module updates the policy network of the corresponding agent based on the agent's actions, the agent's state space, and the reward value to obtain the actual policy network, it repeatedly obtains the agent's actions, agent's state space, and reward value based on the actual voltage data of each partition. It then merges the agent's actions, state space, and reward value obtained each time into round data and stores it in the agent's experience pool. When the round data in the agent's experience pool reaches a predetermined number, it randomly extracts round data from the agent's experience pool to update the agent's policy network, thus obtaining the actual policy network.

6. The power distribution network voltage analysis and control system according to claim 5, characterized in that, When the voltage partitioning module partitions the distribution network according to the distribution network topology to obtain active and reactive power partitions, it further partitions the distribution network based on the Jacobian matrix and spectral clustering algorithm to obtain active and reactive power partitions.

7. The power distribution network voltage analysis and control system according to claim 5, characterized in that, The partition construction module includes: a voltage regulation model construction module, used to generate an objective function based on the adjustable resource action cost and voltage over-limit penalty coefficient of each partition; and to construct a voltage regulation model for the corresponding partition based on the constraints of each partition and the objective function.

8. The power distribution network voltage analysis and control system according to claim 7, characterized in that, The constraints include: power balance constraints, equipment operation constraints, energy storage charging and discharging constraints, and demand response constraints.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 1.

10. A computer device according to claim 9, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 2.

11. A computer device according to claim 9, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 3.

12. A computer device according to claim 9, characterized in that, When the processor executes the computer program, it implements the steps of the method of claim 4.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 1.

14. A computer-readable storage medium according to claim 13, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 2.

15. A computer-readable storage medium according to claim 13, characterized in that, When the computer program is executed by the processor, it implements the steps of the method of claim 3.

16. A computer-readable storage medium according to claim 13, characterized in that, When the computer program is executed by the processor, it implements the steps of the method of claim 4.

Images