Photovoltaic grid-connected point voltage control method and system based on actor-commentator mechanism

Abstract

The invention discloses a photovoltaic grid-connected point voltage control method and system based on an actor-commentator mechanism, and belongs to the technical field of photovoltaic grid-connected point voltage control. The method comprises the following steps: monitoring a photovoltaic grid-connected point in real time to obtain power grid state data and photovoltaic output data of the photovoltaic grid-connected point; generating a control strategy through an actor network according to the real power grid state data and the photovoltaic output data, performing utility value evaluation on the control strategy through a commentator network, updating the control strategy based on an evaluation result, and generating an update strategy; and generating an action mode according to the updating strategy, executing a control action according to the action mode, and controlling the voltage of the photovoltaic grid-connected point. According to the application of the invention, the utilization efficiency of energy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic grid-connected point voltage control, and more specifically, to a photovoltaic grid-connected point voltage control method and system based on an actor-critic mechanism. Background Art

[0002] Photovoltaic power generation system is a renewable energy power generation method currently widely used in the energy field. It uses solar energy to convert light energy into electrical energy. One of the key issues of photovoltaic power generation system is how to connect the generated electricity to the grid, that is, to effectively connect the electricity of the power generation system to the power grid. However, the voltage fluctuation of photovoltaic power generation system is large, which may affect the power grid.

[0003] The voltage at the existing photovoltaic grid-connected points often fluctuates. At this time, voltage fluctuations may cause the quality of the power output of the photovoltaic power generation system to decline, affecting the stable operation of the power grid. Voltage fluctuations may also cause failures in power equipment, thereby affecting the transmission and distribution of electricity. At the same time, voltage fluctuations will cause the working state of photovoltaic cells to be unstable, thereby reducing power generation efficiency. Long-term voltage fluctuations will lead to a reduction in the power generation of the photovoltaic power generation system, increasing the problem of energy waste. Summary of the invention

[0004] In view of the above problems, the present invention proposes a voltage control method for a photovoltaic grid-connected point, comprising:

[0005] Real-time monitoring of photovoltaic grid-connected points to obtain grid status data and photovoltaic output data of the photovoltaic grid-connected points;

[0006] Generate a control strategy according to the real grid state data and photovoltaic output data through an actor network, evaluate the utility value of the control strategy through a critic network, and update the control strategy based on the evaluation result to generate an updated strategy;

[0007] An action mode is generated according to the update strategy, and a control action is executed according to the action mode to control the voltage of the photovoltaic grid-connected point.

[0008] Optionally, grid status data includes: voltage data and power data.

[0009] Optionally, photovoltaic output data, including: inverter status data.

[0010] Optionally, the control strategy is evaluated for utility value through a critic network, including:

[0011] The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

[0012] Optionally, rewards are determined through the status and actions of the PV grid-connected point, and the actor network and the critic network are updated based on the determined rewards.

[0013] Optionally, the reward function formula for determining the reward is as follows:

[0014] R=-α|VV target |-β|P LOSS |-γΔV

[0015] Among them, R is the reward, α, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, ΔV is the voltage fluctuation rate, P LOSS It is the line loss in the station area.

[0016] On the other hand, the present invention also proposes a voltage control system for a photovoltaic grid-connected point, comprising:

[0017] A real-time monitoring unit, used to monitor the photovoltaic grid-connected point in real time to obtain the grid status data and photovoltaic output data of the photovoltaic grid-connected point;

[0018] An actor-critic unit, configured to generate a control strategy according to the real grid state data and the photovoltaic output data through an actor network, evaluate the utility value of the control strategy through a critic network, and update the control strategy based on the evaluation result to generate an updated strategy;

[0019] A control unit is used to generate an action mode according to the update strategy, execute a control action according to the action mode, and control the voltage of the photovoltaic grid-connected point.

[0020] Optionally, grid status data includes: voltage data and power data.

[0021] Optionally, photovoltaic output data, including: inverter status data.

[0022] Optionally, the control strategy is evaluated for utility value through a critic network, including:

[0023] The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

[0024] Optionally, the actor-critic plus unit determines rewards through the status and actions of the PV grid-connected point, and updates the actor network and the critic network based on the determined rewards.

[0025] Optionally, the reward function formula for determining the reward is as follows:

[0026] R=-α|VV target|-β|P LOSS |-γΔV

[0027] Among them, R is the reward, α, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, ΔV is the voltage fluctuation rate, P LOSS It is the line loss in the station area.

[0028] In yet another aspect, the present invention further provides a computing device, comprising: one or more processors;

[0029] a processor for executing one or more programs;

[0030] When the one or more programs are executed by the one or more processors, the above-described method is implemented.

[0031] In yet another aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed, the method described above is implemented.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention provides a voltage control method for photovoltaic grid-connected points, comprising: real-time monitoring of photovoltaic grid-connected points to obtain grid state data and photovoltaic output data of the photovoltaic grid-connected points; generating a control strategy according to the real grid state data and photovoltaic output data through an actor network, evaluating the utility value of the control strategy through a critic network, updating the control strategy based on the evaluation result, and generating an updated strategy; generating an action mode with the updated strategy, executing a control action with the action mode, and controlling the voltage of the photovoltaic grid-connected points. The application of the present invention improves the utilization efficiency of energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the method of the present invention;

[0035] Figure 2 It is the principle diagram of the method of the present invention;

[0036] Figure 3 is a voltage control flow chart of the method of the present invention;

[0037] Figure 4 is a flowchart of the actor-critic algorithm of the method of the present invention;

[0038] Figure 5 A schematic diagram for defining states, actions, and rewards for the method of the present invention;

[0039] Figure 6 It is a structural diagram of the system of the present invention. DETAILED DESCRIPTION

[0040] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0041] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0042] Embodiment 1:

[0043] The present invention proposes a voltage control method for photovoltaic grid-connected points, such as Figure 1 As shown, including:

[0044] Step 1: Real-time monitoring of the photovoltaic grid-connected point to obtain grid status data and photovoltaic output data of the photovoltaic grid-connected point;

[0045] Step 2: Generate a control strategy based on the real grid state data and photovoltaic output data through the actor network, evaluate the utility value of the control strategy through the critic network, and update the control strategy based on the evaluation result to generate an updated strategy;

[0046] Step 3: Generate an action mode according to the update strategy, execute a control action according to the action mode, and control the voltage of the photovoltaic grid-connected point.

[0047] Among them, the power grid status data includes: voltage data and power data.

[0048] Among them, photovoltaic output data includes: inverter status data.

[0049] The utility value of the control strategy is evaluated by a critic network, including:

[0050] The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

[0051] Among them, the rewards are determined by the status and actions of the photovoltaic grid-connected points, and the actor network and the critic network are updated based on the determined rewards.

[0052] Among them, the reward function formula for determining the reward is as follows:

[0053] R=-α|VV target |-β|P LOSS |-γΔV

[0054] Among them, R is the reward, α, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, and ΔV is the voltage fluctuation rate.

[0055] The present invention will be further described below in conjunction with specific embodiments and drawings of the present invention:

[0056] The implementation principle of the present invention is as follows Figure 2 As shown in the figure, the system monitors the grid status and the output of the photovoltaic power generation system in real time, and intelligently regulates the power generation to achieve optimal control of the grid voltage. This method mainly uses the actor-critic algorithm; the actor module is responsible for generating specific power regulation strategies. This module receives the current grid status and photovoltaic output as input, and outputs specific instructions for power regulation. The critic module evaluates the effectiveness of the actor's output strategy. By comparing the real-time grid status and the expected grid status, the critic calculates the utility value of the current strategy, provides feedback to the actor, and helps it optimize its decision. Through the continuous collection and analysis of real-time data, the system continuously learns and adjusts strategies to adapt to changes in grid conditions.

[0057] The technical solution of the present invention not only optimizes the operation efficiency of the power grid by intelligently regulating the photovoltaic power generation, but also improves the energy utilization efficiency, reduces energy waste, effectively stabilizes the grid voltage, and adapts to the changing grid load conditions. The actor-critic algorithm enables the system to achieve optimal performance in any given situation, ensuring the reliability of the power system and the economic benefits of photovoltaic power generation.

[0058] The present invention realizes the real-time collection of voltage data, power data and inverter status by installing a photovoltaic collection and monitoring unit at the grid-connected point. The photovoltaic collection and monitoring unit can generate an action mode through real-time status judgment and the actor-critic algorithm by considering multiple indicators of voltage regulation, thereby realizing stable voltage regulation of the photovoltaic grid-connected point.

[0059] The present invention defines state actions and rewards in the following way:

[0060] Status (S): current voltage (V), current power output (P), inverter control percentage (N), current voltage fluctuation rate (ΔV');

[0061] Action (A): Power regulation rate, ranging from -1 to 1, where 1 means increasing power output by 100% and -1 means decreasing power output by 100%:

[0062] Reward (R): The reward function is based on maintaining voltage stability and minimizing power loss, using target voltage, voltage fluctuation rate, and photovoltaic curtailment as rewards:

[0063] R=-α|VV target |-β|P LOSS |-γΔV

[0064] Where V is the current voltage value, α, β, and γ are weight coefficients used to adjust the relative importance of each target in the reward function, ΔV is the voltage fluctuation rate term, which is used to measure the degree of voltage fluctuation within a certain period of time, and P LOSS It is the line loss in the station area.

[0065] Design an actor network and a critic network, where the actor network proposes actions and the critic network evaluates these actions and guides the learning of the actor network.

[0066] The actor network is a specific type of neural network whose main function is to generate control actions based on the current state (including voltage level, power output, inverter state, voltage fluctuation rate). These actions guide the power adjustment of the photovoltaic power generation system to optimize the voltage quality of the power grid. The actor network helps achieve voltage stability and efficiency improvement by receiving real-time status information of the power grid and outputting the best power adjustment strategy. This network is trained to make fast and accurate decisions under various power grid conditions to ensure voltage stability and economic operation of the system.

[0067] The critic network is also a neural network whose role is to evaluate the effects of the control actions generated by the actor network. By analyzing the long-term impact of a given action, the critic network provides feedback to the actor network to help it optimize its action strategy. This feedback is based on predicted future rewards, that is, evaluating how the current action will affect the grid state (including voltage stability, power loss, and voltage fluctuation rate) in the future. The existence of the critic network enables the entire system to self-learn and adjust, and continuously improve its action strategy to achieve more efficient and stable grid management.

[0068] In practice, the two networks learn and adapt to each other through repeated interaction and training to form an efficient closed-loop control system. By continuously optimizing the parameters of the two networks, our voltage control system can adapt to the complex and changing power grid environment and achieve high-efficiency and high-reliability power grid operation. This neural network-based actor-critic framework is the core of realizing advanced voltage regulation technology.

[0069] Parameter update scheme,The actor network updates the policy using a policy gradient method (such as deterministic policy gradient,DPG). The update is guided by the gradient of the critic's value function. The critic,uses the mean squared error (MSE) loss, and the optimization algorithm can use Adam to adjust the prediction of the value function.

[0070] The present invention optimizes the stability of the grid voltage:

[0071] Precise control: Through real-time monitoring and intelligent adjustment of the power output of the photovoltaic power generation system, the system can accurately control the grid voltage to adapt to the changing grid load conditions. This control helps maintain the voltage within an ideal range, thereby enhancing the stability and reliability of the grid.

[0072] The present invention improves energy utilization efficiency:

[0073] Reduce energy waste: The system dynamically adjusts power output to maximize the use of photovoltaic power generation and reduce energy waste caused by excess or insufficient power. This not only improves the economic benefits of photovoltaic power generation, but also optimizes the energy efficiency of the entire power grid.

[0074] The present invention enhances the system responsiveness:

[0075] Rapid response: Reinforcement learning algorithms allow the system to respond instantly to changes in the state of the power grid and adjust strategies in real time. This rapid response capability is difficult to match with traditional control systems, especially when dealing with emergencies in large-scale, complex power grid environments.

[0076] The present invention is capable of self-learning and self-adaptation:

[0077] Continuous Optimization: Over time, the system continuously optimizes its decision-making process by continuously learning and self-adjusting. This learning capability enables the system to adapt to new operating environments and changing grid conditions, thereby continuously providing the optimal control strategy.

[0078] Figure 3 is the voltage control flow chart, Figure 3 The process starts with system initialization, including setting initial parameters and preparing monitoring equipment. Next, the system monitors voltage and power data in real time and inputs this data into the actor network. The actor network generates specific control instructions based on real-time data, indicating how to adjust the power output of the photovoltaic power generation system. The critic network evaluates the effect of these control instructions, calculates the reward value, and feeds the evaluation results back to the actor network to optimize its decision-making strategy. The system continuously updates the strategy, repeatedly monitors, evaluates, and adjusts to ensure the stability of the grid voltage and the efficient operation of photovoltaic power generation.

[0079] Figure 4Actor-Critic Algorithm Flowchart,The actor-critic algorithm flowchart starts with inputting the current state, including voltage level, power output, inverter state, and voltage fluctuation rate. The actor network receives this state information and generates corresponding control actions to guide the power adjustment of the photovoltaic power generation system. After the system executes these control actions, the critic network evaluates their effects and determines the utility of the current strategy by calculating the reward value. The feedback from the critic network is used to update the parameters of the actor network and the critic network to optimize the control strategy and prediction accuracy.

[0080] Figure 5 The state, action and reward definition diagram details the key variables used in the voltage control system and their functions. The state (S) includes the current voltage (V), the current power output (P), the inverter regulation percentage (N) and the current voltage fluctuation rate. These parameters are used to describe the current state of the grid and photovoltaic power generation system. The action (A) represents the power adjustment rate, which ranges from -1 to 1, indicating the adjustment range from reducing the power output by 100% to increasing the power output by 100%. The reward (R) is calculated based on maintaining voltage stability and minimizing power loss through the target voltage, voltage fluctuation rate and photovoltaic abandonment, aiming to encourage the system to maintain the optimal grid voltage and power output.

[0081] The specific implementation steps are as follows:

[0082] 1. System initialization:

[0083] Install a photovoltaic collection monitoring unit at the grid-connected point of each photovoltaic power generation system and configure sensors to monitor voltage and power data in real time. Initialize the parameters of the actor network and the critic network, including network weights and bias values.

[0084] 2. Real-time data collection:

[0085] The photovoltaic collection and monitoring unit collects data on voltage (V), power output (P), inverter regulation percentage (N) and voltage fluctuation rate every second and transmits this data to the control center.

[0086] 3. Actor network generation control strategy:

[0087] The actor network in the control center receives real-time data input (V, P, N,) and generates a power regulation rate (A) based on the current grid status. The power regulation rate ranges from -1 to 1, indicating a decrease in power output by 100% to an increase in power output by 100%.

[0088] 4. Perform actions and provide feedback:

[0089] The photovoltaic power generation system adjusts the power output of the inverter according to the control instructions generated by the actor network.

[0090] 5. Critic network evaluation effect:

[0091] The critic network evaluates the effect of the control instructions generated by the actor network and calculates the reward value (R). The reward value is calculated based on the following formula:

[0092] R=-(λ1(VV target ) 2 +λ2ΔV+λ3Q)

[0093] Among them, λ1, λ2, λ3 are weight coefficients, V is the current voltage value, V target is the target voltage value, ΔV is the voltage fluctuation rate, and Q is the amount of photovoltaic abandoned light.

[0094] 6. Update network parameters:

[0095] The policy gradient method is used to update the actor network’s policy based on the feedback from the critic network. The critic network uses the mean squared error (MSE) loss function and the Adam optimization algorithm is used to adjust the prediction of the value function.

[0096] 7.Continuous optimization:

[0097] The system forms a closed-loop control by continuously collecting data, generating strategies, evaluating effects and updating parameters, continuously optimizing the voltage control strategy to ensure grid voltage stability and photovoltaic power generation efficiency.

[0098] At noon one day, the intensity of solar radiation increased, and the power output of the photovoltaic power generation system increased significantly. This caused the grid voltage to rise, exceeding the preset target voltage value. At this time, the photovoltaic collection monitoring unit detected the voltage change and transmitted the real-time data to the control center. The actor network received the input of the current state (V = 250V, P = 300W, N = 75%, ΔV = 5V), generated a control instruction to reduce the power output (A = -0.5), and instructed the inverter to reduce the power output by 50%.

[0099] The photovoltaic power generation system reduces its power output to 150W according to the instruction. The critic network evaluates the effect of this adjustment and calculates the reward value. Since the voltage fluctuation is controlled, the system obtains positive feedback and further optimizes the control strategy for the next round.

[0100] In the next few hours, the control center continuously receives and processes real-time data, generates and executes corresponding power adjustment instructions, and maintains the grid voltage within the target range (220V±10V). At the same time, the system gradually improves its response speed and control accuracy to voltage fluctuations through self-learning and parameter updates.

[0101] Embodiment 2:

[0102] The present invention also proposes a voltage control system 200 for a photovoltaic grid-connected point, such as Figure 6 As shown, including:

[0103] A real-time monitoring unit 201 is used to monitor the photovoltaic grid-connected point in real time to obtain the grid status data and photovoltaic output data of the photovoltaic grid-connected point;

[0104] The actor critic unit 202 is used to generate a control strategy according to the real grid state data and photovoltaic output data through an actor network, evaluate the utility value of the control strategy through a critic network, and update the control strategy based on the evaluation result to generate an updated strategy;

[0105] The control unit 203 is used to generate an action mode according to the update strategy, execute a control action according to the action mode, and control the voltage of the photovoltaic grid-connected point.

[0106] Among them, the power grid status data includes: voltage data and power data.

[0107] Among them, photovoltaic output data includes: inverter status data.

[0108] The utility value of the control strategy is evaluated by a critic network, including:

[0109] The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

[0110] Among them, the actor-critic plus unit determines the reward through the status and actions of the photovoltaic grid-connected point, and updates the actor network and the critic network based on the determined reward.

[0111] Among them, the reward function formula for determining the reward is as follows:

[0112] R=-α|VV target |-β|P LOSS |-γΔV

[0113] Among them, R is the reward, α, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, ΔV is the voltage fluctuation rate, P LOSS It is the line loss in the station area.

[0114] The application of the present invention improves the utilization efficiency of energy.

[0115] Embodiment 3:

[0116] Based on the same inventive concept, the present invention also provides a computer device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, and is specifically suitable for loading and executing one or more instructions in the computer storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of the method in the above embodiment.

[0117] Embodiment 4:

[0118] Based on the same inventive concept, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device for storing programs and data. It is understandable that the computer-readable storage medium here can include both a built-in storage medium in a computer device and an extended storage medium supported by the computer device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the steps of the method in the above embodiment.

[0119] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0120] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0121] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0122] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0123] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0124] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A photovoltaic grid-connected point voltage control method based on an actor-critic mechanism, characterized in that: include: Real-time monitoring of photovoltaic grid-connected points to obtain grid status data and photovoltaic output data of the photovoltaic grid-connected points; Generate a control strategy according to the real grid state data and photovoltaic output data through an actor network, evaluate the utility value of the control strategy through a critic network, and update the control strategy based on the evaluation result to generate an updated strategy; An action mode is generated according to the update strategy, and a control action is executed according to the action mode to control the voltage of the photovoltaic grid-connected point.

2. The method according to claim 1, characterized in that The grid status data includes voltage data and power data.

3. The method according to claim 1, characterized in that: The photovoltaic output data includes: inverter status data.

4. The method according to claim 1, characterized in that: The utility value evaluation of the control strategy through the critic network includes: The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

5. The method according to claim 1, characterized in that The rewards are determined by the status and actions of the PV grid-connected points, and the actor network and critic network are updated based on the determined rewards.

6. The method according to claim 5, characterized in that The reward function formula for determining the reward is as follows: R=-α|VV target |-β|P LOSS |-γΔV Among them, R is the reward, Δ, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, ΔV is the voltage fluctuation rate, P LOSS It is the line loss in the station area.

7. A photovoltaic grid-connected point voltage control system based on an actor-critic mechanism, characterized in that: include: A real-time monitoring unit, used to monitor the photovoltaic grid-connected point in real time to obtain the grid status data and photovoltaic output data of the photovoltaic grid-connected point; An actor-critic unit, configured to generate a control strategy according to the real grid state data and the photovoltaic output data through an actor network, evaluate the utility value of the control strategy through a critic network, and update the control strategy based on the evaluation result to generate an updated strategy; A control unit is used to generate an action mode according to the update strategy, execute a control action according to the action mode, and control the voltage of the photovoltaic grid-connected point.

8. The system according to claim 7, characterized in that The grid status data includes voltage data and power data.

9. The system according to claim 7, characterized in that The photovoltaic output data includes: inverter status data.

10. The system according to claim 7, characterized in that The utility value evaluation of the control strategy through the critic network includes: The real-time power grid state and the expected power grid state are compared through the critic network to obtain a comparison result, and the utility value of the control strategy is determined based on the comparison result.

11. The system according to claim 7, characterized in that The actor-critic adding unit determines rewards through the status and actions of the photovoltaic grid-connected point, and updates the actor network and the critic network based on the determined rewards.

12. The system according to claim 11, characterized in that The reward function formula for determining the reward is as follows: R=-α|VV target |-β|P LOSS |-γΔV Among them, R is the reward, α, β, γ are weight coefficients, V is the current voltage of the photovoltaic grid-connected point, V target is the target voltage value, ΔV is the voltage fluctuation rate, P LOSS It is the line loss in the station area.

13. A computer device, characterized in that: include: one or more processors; a processor for executing one or more programs; When the one or more programs are executed by the one or more processors, the method according to any one of claims 1 to 6 is implemented.

14. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed, the method according to any one of claims 1 to 6 is implemented.

Images