A method and system for flexible self-healing of grid-connected point voltage with coordinated edge-end

Abstract

The present invention discloses a method and system for flexible self-healing of voltage at a grid-connected point in coordination with an edge end, the method comprising: obtaining measurement data at a node, and determining the total active power and reactive power at the node according to the measurement data; considering the influence of the low-voltage side outlet voltage of the transformer in the substation, determining the rated voltage of the bus where the node is located, and setting the voltage regulation threshold according to the rated voltage of the bus where the node is located; designing the flexible load and the photovoltaic inverter as different controllable intelligent agents, expressing the voltage regulation problem as a Markov game problem with #imgabs0# intelligent agents, and establishing a Markov decision model for voltage adjustment; solving the Markov decision model by the MAPPO algorithm, and obtaining a voltage local control strategy. The overvoltage problem at the grid-connected point is solved, and source-load friendly interaction is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of voltage regulation of active power distribution networks, and in particular relates to a method and system for flexible self-healing of grid-connected point voltage with coordinated edge terminals. Background Art

[0002] The power quality problems of distribution networks are becoming more and more prominent, including voltage over-limit and voltage imbalance. Among them, the reverse flow caused by distributed photovoltaic access, which leads to voltage over-limit, is the main factor limiting its access capacity. At present, the research on voltage over-limit focuses on adjusting photovoltaic active power, reactive power output and increasing energy storage, but these methods have slow adjustment speed and high investment cost.

[0003] With the development of new power systems, the potential of load-side regulation is constantly emerging. By fully tapping the load-side regulation capability to participate in the consumption of new energy, the utilization rate and consumption level of new energy power generation can be improved through voltage control. Therefore, it is urgent to propose a flexible self-healing method and system for grid-connected point voltage with edge-end coordination, which can realize flexible dispatch and rapid response of various resources in the distribution network by coordinating various controllable devices, improve power quality, and realize safe and reliable operation of the distribution network. Summary of the invention

[0004] The present invention provides an edge-end coordinated grid-connected point voltage flexible self-healing method and system, which are used to solve the technical problem of on-site voltage regulation that cannot take into account the dual uncertainty of source and load.

[0005] In a first aspect, the present invention provides an edge-end coordinated grid-connected point voltage flexible self-healing method, comprising:

[0006] Acquire measurement data at the node, and construct a state space of the Markov model according to the measurement data, wherein the measurement data includes adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage;

[0007] Considering the influence of the low-voltage side outlet voltage of the transformer in the substation, the rated voltage of the bus where the node is located is determined, and the voltage regulation threshold is set according to the rated voltage of the bus where the node is located, so as to determine the reward function of the Markov model;

[0008] The flexible load and photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is formulated as a The Markov game problem of the intelligent agent is solved, and the decision model for voltage adjustment is established according to the reward function of the Markov model;

[0009] The decision model is solved by the MAPPO algorithm to obtain a voltage local control strategy.

[0010] In a second aspect, the present invention provides an edge-end coordinated grid-connected point voltage flexible self-healing system, comprising:

[0011] An acquisition module is configured to acquire measurement data at a node, and determine the total active power and reactive power at the node according to the measurement data, wherein the measurement data includes the active power of an adjustable load, the reactive power of an adjustable load, the active output power of a photovoltaic power generation, the reactive output power of a photovoltaic power generation, and the node voltage;

[0012] A setting module is configured to consider the influence of the low-voltage side outlet voltage of the transformer in the substation, determine the rated voltage of the bus where the node is located, and set the voltage regulation threshold according to the rated voltage of the bus where the node is located, thereby determining the reward function of the Markov model;

[0013] The building blocks are configured to design the flexible load and photovoltaic inverter as different controllable agents and formulate the voltage regulation problem as a The Markov game problem of intelligent agents is solved, and a Markov decision model for voltage regulation is established;

[0014] The solution module is configured to solve the Markov decision model through a MAPPO algorithm to obtain a voltage local control strategy.

[0015] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the edge-end coordinated grid-connected point voltage flexible self-healing method of any embodiment of the present invention.

[0016] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor executes the steps of the edge-end coordinated grid-connected point voltage flexible self-healing method of any embodiment of the present invention.

[0017] The edge-coordinated grid-connected point voltage flexible self-healing method and system of the present application uses a grid-connected intelligent measurement and control terminal device to collect grid-connected point load power, photovoltaic operation data, and historical data of node voltage, to establish a Markov decision model for voltage adjustment, and solves the Markov decision model based on the reactive adjustable capacity and load adjustable capacity of the photovoltaic inverter to obtain a control strategy for the inverter and the adjustable load. Through edge-end collaborative flexible autonomy, the node adjustable load and the reactive adjustable capacity of the photovoltaic inverter can be fully mobilized, and a flexible on-site control strategy for the grid-connected point voltage is formulated based on the terminal voltage adjustment threshold to solve the overvoltage problem at the grid-connected point and achieve source-load friendly interaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A flow chart of a method for flexible self-healing of grid-connected point voltage with edge-end coordination provided by one embodiment of the present invention;

[0020] Figure 2 A structural block diagram of an edge-end coordinated grid-connected point voltage flexible self-healing system provided by an embodiment of the present invention;

[0021] Figure 3 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1 , which shows a flow chart of an edge-end coordinated grid-connected point voltage flexible self-healing method of the present application.

[0024] like Figure 1 As shown, the edge-end coordinated grid-connected point voltage flexible self-healing method specifically includes the following steps:

[0025] Step S101, obtaining measurement data at a node, and determining the total active power and reactive power at the node based on the measurement data, wherein the measurement data includes adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage.

[0026] In this step, the historical data and real-time measurement data of the adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage are collected through the grid-connected intelligent measurement and control terminal equipment, so as to determine the total active power and reactive power at the node. The expression is:

[0027] ,

[0028] In the formula, for Time Node The total active power at for Time Node Active power at normal load, for Time Node Active power of flexible adjustable load, for Time Node The active output power of photovoltaic power generation, for Time Node The total reactive power at for Time Node Reactive power of conventional load, for Time Node The reactive power of the flexible adjustable load, for Time Node The reactive output power of photovoltaic power generation.

[0029] Step S102, considering the influence of the low-voltage side outlet voltage of the substation transformer, determining the rated voltage of the bus where the node is located, and setting the voltage regulation threshold according to the rated voltage of the bus where the node is located, thereby determining the reward function of the Markov model.

[0030] In this step, the power flow constraint is determined by the active and reactive power of the node, the rated voltage of the bus where the node is located is determined, and the voltage regulation threshold is set.

[0031] Specifically, the power flow constraint of the distribution network is:

[0032] ,

[0033] In the formula, is the active power of node i at time t, is the reactive power of node i at time t, is the voltage value of node i at time t, is the voltage value of node j at time t, is the admittance between nodes i and j, is the phase angle difference between nodes i and j, is the susceptance between node i and node j;

[0034] The safety voltage constraint is:

[0035] ,

[0036] In the formula, is the voltage lower limit of node i, is the voltage upper limit of node i, is the rated voltage value;

[0037] Taking the photovoltaic inverter as a controllable device, the voltage is adjusted by adjusting the reactive output of the inverter. The adjustable reactive output of the inverter is constrained by the active output and the inverter capacity as follows:

[0038] ,

[0039] In the formula, for Time Node The active output power of photovoltaic power generation, for Time Node The reactive output power of photovoltaic power generation is for Time Node The photovoltaic power generation is at power;

[0040] The adjustable flexible load is used as a controllable device, and the voltage is adjusted by adjusting the active power of the load. The adjustable load is subject to the load capacity constraint as follows:

[0041] ,

[0042] In the formula, For Node The lower limit of active power of flexible adjustable load, for Time Node Active power of flexible adjustable load, is the increase in active power of the adjustable load at node n at time t, For Node The active power upper limit of the flexible adjustable load.

[0043] Step S103, the flexible load and the photovoltaic inverter are designed as different controllable intelligent entities, and the voltage regulation problem is expressed as A Markov game problem of three intelligent agents is solved, and a Markov decision model for voltage regulation is established.

[0044] In this step, the state space is set , the node voltage is related to the active power and reactive power of the distribution network bus node, then the state of the intelligent agent is defined as , is the active power of the distribution network bus node at node n at time t, is the reactive power of the distribution network bus node at node n at time t, is the voltage at node n at time t;

[0045] Setting the Action Space , which consists of the action instructions of the adjustable load and the inverter. By coordinating the reactive output of the inverter and the active power of the adjustable load, the node voltage is adjusted to a safe range, while reducing the active power reduction of the photovoltaic. The action of the intelligent agent is defined as , where is the increase in active power of the adjustable load at node n at time t, is the active power reduction of the photovoltaic at node n at time t, is the reactive compensation amount of the photovoltaic inverter at node n at time t;

[0046] Setting the reward function ,express The reward obtained after executing the action at the moment, the reward function Includes a reward function related to the parent deviation , the reward function related to PV active power reduction And the reward function related to flexible load regulation , where the agent’s reward function is The expression is:

[0047] ,

[0048] In the formula, is the weighted coefficient of photovoltaic active power reduction cost, It is the weighted coefficient of flexible load control cost.

[0049] It should be noted that the reward function related to the mother line deviation is The expression is:

[0050] ,

[0051] In the formula, is the bus voltage value at time t, is the upper limit of the bus voltage, is the lower limit of bus voltage;

[0052] Reward function related to PV active power curtailment The expression is:

[0053] ,

[0054] In the formula, is the active power reduction of the photovoltaic at node n at time t;

[0055] Reward function related to flexible load regulation The expression is:

[0056] ,

[0057] In the formula, is the discomfort cost coefficient of the flexible load, It is the increase in active power of the adjustable load at node n at time t.

[0058] Specifically, the Bellman expectation equation describes the relationship between the value function and the action value function, as well as the relationship between the current state value function and the subsequent state value function:

[0059] Using reward function and the decay coefficient represent the Markov reward process:

[0060] ,

[0061] In the formula, Indicates that at time t, Next action After the reward, It represents the sum of all reward decay benefits from time t on a Markov reward chain. is the discount factor;

[0062] Determine the agent used to evaluate the current state The state value function of each agent is based on its own strategy function The Markov chain starting from the current state harvests the reward Expectations:

[0063] ,

[0064] In the formula, is the expected reward from the state at time t, To implement the strategy The expected value of It represents the sum of all reward decay benefits from time t on a Markov reward chain. The state at time t is , , is the state of agent n at time t;

[0065] Determine the agent used to evaluate the action The action value function of the agent The action value function of is input as the states and actions of all agents, where the agent The expression of the action value function is:

[0066] ,

[0067] In the formula, To implement the strategy Lower Agent The action value function is To represent the action at time t , , is the action of agent n at time t, for The reward function at that moment;

[0068] In the Markov decision model, there is an optimal strategy Make strategy The state value function of any state under the strategy is greater than the state value function under other strategies, that is, the Bellman optimal equation:

[0069] ,

[0070] ,

[0071] In the formula, the state value function Indicates that the state The long-term optimal value of Indicates the environment according to the state and actions The immediate reward returned to the agent, is the state transition probability, which satisfies the Markov property, that is, the state at the next moment is only related to the state and action at this moment, but not to the previous state. The action value function Indicates that the state and perform an action The long-term optimal value brought by Status The next state.

[0072] Step S104, solving the Markov decision model by using the MAPPO algorithm to obtain a voltage local control strategy.

[0073] In this step, the MAPPO algorithm is used to solve the Markov decision model to obtain the optimal control strategy for the reactive power of each photovoltaic inverter and the active power of the adjustable load, and the voltage is regulated locally in the simulation model according to the optimal control strategy.

[0074] Specifically, the MAPPO algorithm adopts an Actor-Critic network structure. Each agent has an Actor-Critic network. The Critic network learns the central value function based on the global state. When the input is the environmental state, it can evaluate the value of the current state. The Actor network selects actions based on local observations and the central value function of the Critic network to maximize the discounted cumulative reward. It can evaluate the quality of the action in the current state and adjust the strategy. The optimization goal of the Actor network is:

[0075] ,

[0076] ,

[0077] ,

[0078] In the formula, is a representative of a network parameter. represents the batch size, represents the number of agents, represents the scale factor of importance sampling, Represents the advantage function, which is used to measure a state Next, select the i-th action rationality, Indicates that the MAPPO algorithm uses the clip function to calculate parameters , represents the cropping factor, represents the entropy of the policy, is a hyperparameter that controls the entropy coefficient, represents local observation information, is the policy function in the current state, is the strategy function in the previous state;

[0079] The optimization goal of the Critic network is:

[0080] ,

[0081] In the formula, is another representative of network parameters, Represents a discount reward, represents the value function in the current state, is the value function in the previous state.

[0082] In summary, the method of the present application is to collect historical data of grid-connected point load power, photovoltaic operation data, and node voltage by the grid-connected intelligent measurement and control terminal equipment, establish a Markov decision model for voltage adjustment, and solve the Markov decision model according to the reactive adjustable capacity and load adjustable capacity of the photovoltaic inverter to obtain the control strategy of the inverter and the adjustable load. Through the flexible autonomy of edge collaboration, the node adjustable load and the reactive adjustable capacity of the photovoltaic inverter can be fully mobilized. According to the terminal voltage adjustment threshold, a flexible on-site control strategy for the grid-connected point voltage is formulated to solve the overvoltage problem of the grid-connected point and realize source-load friendly interaction.

[0083] See also Figure 2 , which shows a structural block diagram of an edge-coordinated grid-connected point voltage flexible self-healing system of the present application.

[0084] like Figure 2 As shown, the grid-connected point voltage flexible self-healing system 200 includes an acquisition module 210 , a setting module 220 , a construction module 230 and a solution module 240 .

[0085] Among them, the acquisition module 210 is configured to obtain the measurement data at the node, and determine the total active power and reactive power at the node according to the measurement data, wherein the measurement data includes the active power of the adjustable load, the reactive power of the adjustable load, the active output power of the photovoltaic power generation, the reactive output power of the photovoltaic power generation and the node voltage; the setting module 220 is configured to consider the influence of the low-voltage side outlet voltage of the substation transformer, determine the rated voltage of the bus where the node is located, and set the voltage regulation threshold according to the rated voltage of the bus where the node is located, so as to determine the reward function of the Markov model; the construction module 230 is configured to design the flexible load and the photovoltaic inverter as different controllable intelligent agents, and express the voltage regulation problem as having A Markov game problem of an intelligent agent is solved to establish a Markov decision model for voltage adjustment; a solution module 240 is configured to solve the Markov decision model through a MAPPO algorithm to obtain a voltage local control strategy.

[0086] It should be understood that Figure 2 Modules and references documented in Figure 1 Therefore, the operations and features described above for the method and the corresponding technical effects are also applicable to Figure 2 The modules in it will not be described in detail here.

[0087] In some other embodiments, the embodiments of the present invention further provide a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor executes the edge-end coordinated grid-connected point voltage flexible self-healing method in any of the above method embodiments;

[0088] As an implementation mode, the computer-readable storage medium of the present invention stores computer-executable instructions, and the computer-executable instructions are configured as follows:

[0089] Acquire measurement data at the node, and construct a state space of the Markov model according to the measurement data, wherein the measurement data includes adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage;

[0090] Considering the influence of the low-voltage side outlet voltage of the transformer in the substation, the rated voltage of the bus where the node is located is determined, and the voltage regulation threshold is set according to the rated voltage of the bus where the node is located, so as to determine the reward function of the Markov model;

[0091] The flexible load and photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is formulated as a The Markov game problem of the intelligent agent is solved, and the decision model for voltage adjustment is established according to the reward function of the Markov model;

[0092] The decision model is solved by the MAPPO algorithm to obtain a voltage local control strategy.

[0093] The computer-readable storage medium may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created according to the use of the edge-coordinated grid-connected point voltage flexible self-healing system, etc. In addition, the computer-readable storage medium may include a high-speed random access memory, and may also include a memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the edge-coordinated grid-connected point voltage flexible self-healing system via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0094] Figure 3 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 3 As shown, the device includes: a processor 310 and a memory 320. The electronic device may also include: an input device 330 and an output device 340. The processor 310, the memory 320, the input device 330 and the output device 340 may be connected via a bus or other means. Figure 3In the example, the connection through the bus is taken as an example. The memory 320 is the above-mentioned computer-readable storage medium. The processor 310 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 320, that is, the flexible self-healing method of the grid-connected point voltage coordinated by the edge end of the above-mentioned method embodiment is implemented. The input device 330 can receive input digital or character information, and generate key signal input related to user settings and function control of the flexible self-healing system of the grid-connected point voltage coordinated by the edge end. The output device 340 may include a display device such as a display screen.

[0095] The electronic device can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided by the embodiment of the present invention.

[0096] As an implementation mode, the electronic device is applied to a grid-connected point voltage flexible self-healing system coordinated by an edge end, and is used for a client, and includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can:

[0097] Acquire measurement data at the node, and construct a state space of the Markov model according to the measurement data, wherein the measurement data includes adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage;

[0098] Considering the influence of the low-voltage side outlet voltage of the transformer in the substation, the rated voltage of the bus where the node is located is determined, and the voltage regulation threshold is set according to the rated voltage of the bus where the node is located, so as to determine the reward function of the Markov model;

[0099] The flexible load and photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is formulated as a The Markov game problem of the intelligent agent is solved, and the decision model for voltage adjustment is established according to the reward function of the Markov model;

[0100] The decision model is solved by the MAPPO algorithm to obtain a voltage local control strategy.

[0101] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for flexible self-healing of grid-connected point voltage with edge coordination, characterized in that: include: Acquire measurement data at the node, and construct a state space of the Markov model according to the measurement data, wherein the measurement data includes adjustable load active power, adjustable load reactive power, photovoltaic power generation active output power, photovoltaic power generation reactive output power and node voltage; Considering the influence of the low-voltage side outlet voltage of the transformer in the substation, the rated voltage of the bus where the node is located is determined, and the voltage regulation threshold is set according to the rated voltage of the bus where the node is located, so as to determine the reward function of the Markov model; The flexible load and photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is formulated as a A Markov game problem of multiple intelligent agents is proposed, and a decision model for voltage regulation is established based on the reward function of the Markov model, wherein the flexible load and the photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is expressed as a problem with The Markov game problem of the intelligent agent is solved, and the decision model for voltage adjustment is established based on the reward function of the Markov model, including: Setting the state space , the node voltage is related to the active power and reactive power of the distribution network bus node, then the state space of the intelligent agent is defined as , is the active power of the distribution network bus node at node n at time t, is the reactive power of the distribution network bus node at node n at time t, is the voltage at node n at time t; Setting the Action Space , which consists of the action instructions of the adjustable load and the inverter. By coordinating the reactive output of the inverter and the active power of the adjustable load, the node voltage is adjusted to a safe range while reducing the active power reduction of the photovoltaic. The action space of the intelligent agent is defined as , where is the increase in active power of the adjustable load at node n at time t, is the active power reduction of the photovoltaic at node n at time t, is the reactive compensation amount of the photovoltaic inverter at node n at time t; Setting the reward function ,express The instant reward obtained after executing the action at the moment, the reward function Includes a reward function related to the parent deviation , the reward function related to PV active power reduction And the reward function related to flexible load regulation , where the agent’s reward function is The expression is: ， In the formula, is the weighted coefficient of photovoltaic active power reduction cost, is the weighted coefficient of flexible load control cost; The reward function related to the parent line deviation The expression is: ， In the formula, is the bus voltage value at time t, is the upper limit of the bus voltage, is the lower limit of bus voltage; The reward function related to PV active power reduction is The expression is: ， In the formula, is the active power reduction of the photovoltaic power at node n at time t, which is limited by the grid connection point voltage and transformer capacity; The reward function related to flexible load regulation The expression is: ， In the formula, is the discomfort cost coefficient of the flexible load, is the increase in active power of the adjustable load at node n at time t, which is limited by the capacity of the adjustable load; The decision model is solved by the MAPPO algorithm to obtain a voltage local control strategy.

2. The method for flexible self-healing of grid connection point voltage with edge coordination according to claim 1, characterized in that: in, The expression for determining the measurement data at the node is: ， In the formula, for Time Node The total active power at for Time Node Active power at normal load, for Time Node Active power of flexible adjustable load, for Time Node The active output power of photovoltaic power generation, for Time Node The total reactive power at for Time Node Reactive power of conventional load, for Time Node The reactive power of the flexible adjustable load, for Time Node The reactive output power of photovoltaic power generation.

3. The method for flexible self-healing of grid connection point voltage with edge coordination according to claim 1, characterized in that: The flexible load and photovoltaic inverter are designed as different controllable intelligent agents, and the voltage regulation problem is expressed as The Markov game problem of multiple agents includes: The Bellman expectation equation describes the relationship between the value function and the action value function, as well as the relationship between the current state value function and the subsequent state value function: Using reward function and the decay coefficient represent the Markov reward process: ， In the formula, Indicates that at time t, Next action After the reward, It represents the sum of all reward decay benefits from time t on a Markov reward chain. is the discount factor; Determine the agent used to evaluate the current state The state value function of each agent is based on its own strategy function The Markov chain starting from the current state harvests the reward Expectations: ， In the formula, is the expected reward from the state at time t, To implement the strategy The expected value of It represents the sum of all reward decay benefits from time t on a Markov reward chain. The state at time t is , , is the state of agent n at time t; Determine the agent used to evaluate the action The action value function of the agent The action value function of is input as the states and actions of all agents, where the agent The expression of the action value function is: ， In the formula, To implement the strategy Lower Agent The action value function is To represent the action at time t , , is the action of agent n at time t, and is the reward function at time t; In the Markov decision model, there is an optimal strategy Make strategy The state value function of any state under the strategy is greater than the state value function under other strategies, that is, the Bellman optimal equation: ， ， In the formula, the state value function Indicates that the state The long-term optimal value of Indicates the environment according to the state and actions The immediate reward returned to the agent, is the state transition probability, which satisfies the Markov property, that is, the state at the next moment is only related to the state and action at this moment, but not to the previous state. The action value function Indicates that the state and perform an action The long-term optimal value brought by Status The next state.

4. The method for flexible self-healing of grid connection point voltage with edge coordination according to claim 1, characterized in that: The voltage local control strategy obtained by solving the Markov decision model through the MAPPO algorithm includes: The MAPPO algorithm uses an Actor-Critic network structure. Each agent has an Actor-Critic network. The Critic network learns the central value function based on the global state. When the input is the environmental state, it can evaluate the value of the current state. The Actor network selects actions based on local observations and the central value function of the Critic network to maximize the discounted cumulative reward. It can evaluate the quality of the action in the current state and adjust the strategy. The optimization goal of the Actor network is: ， ， ， In the formula, is a representative of a network parameter, represents the batch size, represents the number of agents, represents the scale factor of importance sampling, Represents the advantage function, which is used to measure a state Next, select the i-th action rationality, Indicates that the MAPPO algorithm uses the clip function to calculate parameters , represents the cropping factor, represents the entropy of the policy, is a hyperparameter that controls the entropy coefficient, represents local observation information, is the policy function in the current state, is the strategy function in the previous state; The optimization goal of the Critic network is: ， In the formula, is another representative of network parameters, Represents a discount reward, represents the value function in the current state, is the value function in the previous state.

5. A grid-connected point voltage flexible self-healing system with edge coordination according to any one of the methods described in claims 1-4, characterized in that: include: An acquisition module is configured to acquire measurement data at a node, and determine the total active power and reactive power at the node according to the measurement data, wherein the measurement data includes the active power of an adjustable load, the reactive power of an adjustable load, the active output power of a photovoltaic power generation, the reactive output power of a photovoltaic power generation, and the node voltage; A setting module is configured to consider the influence of the low-voltage side outlet voltage of the transformer in the substation, determine the rated voltage of the bus where the node is located, and set the voltage regulation threshold according to the rated voltage of the bus where the node is located, thereby determining the reward function of the Markov model; The building blocks are configured to design the flexible load and photovoltaic inverter as different controllable agents and formulate the voltage regulation problem as a The Markov game problem of intelligent agents is solved, and a Markov decision model for voltage regulation is established; The solution module is configured to solve the Markov decision model through a MAPPO algorithm to obtain a voltage local control strategy.

6. An electronic device, characterized in that: include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images