Power distribution network multi-agent voltage control method and system and storage medium

Abstract

The invention discloses a multi-agent voltage control method for a power distribution network. The method comprises the following steps: acquiring data information of a target power distribution network; constructing a multi-agent voltage control model of the target power distribution network; and solving the constructed multi-agent voltage control model by adopting a DDQN algorithm so as to realize multi-agent voltage control of the target power distribution network. The invention also discloses a system for realizing the power distribution network multi-agent voltage control method, and a storage medium comprising the power distribution network multi-agent voltage control method. According to the method, the multi-agent voltage control model of the target power distribution network is constructed, and the DDQN algorithm is adopted to carry out corresponding solution, so that the multi-agent voltage control of the power distribution network can be realized, and the method is higher in reliability and better in accuracy.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, and specifically relates to a multi-agent voltage control method, system and storage medium for power distribution networks. Background Technology

[0002] With economic and technological development and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and daily life, bringing endless convenience. Therefore, ensuring a stable and reliable supply of electricity has become one of the most important tasks of the power system.

[0003] Currently, with increasingly severe environmental problems, more and more distributed renewable energy generation systems are being integrated into the power grid. The widespread integration of these systems has not only altered the unidirectional power flow characteristics of traditional distribution networks but also transformed them into complex, bidirectional interactive systems. This transformation brings numerous benefits to the power system, such as improved energy efficiency, reduced transmission losses, enhanced power supply reliability, and promotion of renewable energy consumption. However, the integration of a high proportion of distributed generation systems also presents new technical challenges, including voltage fluctuations, decreased frequency stability, and difficulties in protection coordination.

[0004] The increasing penetration of distributed renewable energy generation systems in distribution networks has made voltage exceeding limits a pressing issue that needs to be addressed. Furthermore, the large-scale integration of distributed renewable energy generation systems significantly increases the amount of data that needs to be processed during distribution network control, thus exacerbating the system's computational burden. Therefore, existing distribution network voltage control schemes are gradually becoming inadequate to meet current distribution network control requirements. Summary of the Invention

[0005] One of the objectives of this invention is to provide a highly reliable and accurate multi-agent voltage control method for power distribution networks.

[0006] A second objective of this invention is to provide a system for implementing the multi-agent voltage control method for power distribution networks.

[0007] A third objective of this invention is to provide a storage medium on which a computer program is stored; when the computer program is executed by a processor, it implements the aforementioned multi-agent voltage control method for power distribution networks.

[0008] The multi-agent voltage control method for power distribution networks provided by this invention includes the following steps:

[0009] S1. Obtain data information from the target power distribution network;

[0010] S2. Based on the data obtained in step S1, construct a multi-agent voltage control model for the target distribution network;

[0011] S3. The DDQN algorithm is used to solve the multi-agent voltage control model constructed in step S2 in order to achieve multi-agent voltage control of the target distribution network.

[0012] Step S1, which involves obtaining data information from the target distribution network, specifically includes the following steps:

[0013] Acquire data information of new energy power generation systems within the target distribution network;

[0014] The data information of the new energy power generation system includes the output data information of the distributed photovoltaic power generation system, the output data information of the distributed wind power generation system, and the output data information of the distributed energy storage system.

[0015] Step S2, which involves constructing a multi-agent voltage control model for the target distribution network based on the data obtained in step S1, includes the following steps:

[0016] Based on the data obtained in step S1, a multi-agent voltage control model for the target distribution network is constructed with the minimum node voltage deviation as the objective function and the power flow conditions, operating voltage conditions, distributed photovoltaic output conditions, distributed photovoltaic capacity conditions, distributed wind power output conditions, distributed energy storage power conditions, and distributed energy storage state of charge conditions as constraints.

[0017] Step S2 specifically includes the following steps:

[0018] The following formula is used as the objective function:

[0019]

[0020] In the formula, F(x) is the objective function value; Ω n,O V represents the observation area of ​​each intelligent agent within the target distribution network. i Vi represents the voltage of node i within the observation area of ​​each intelligent agent in the target distribution network; V0 represents the rated voltage of the target distribution network.

[0021] The following formula is used as the power flow constraint:

[0022]

[0023] In the formula P D,i P is the active power injected at node i within the target distribution network. L,i U represents the active power of the load at node i within the target distribution network; i N represents the voltage magnitude at node i; node U is the set of nodes in a distribution network. j G represents the voltage magnitude at node j. ijB is the conductance value on the line from node i to node j; ij δ represents the susceptance value on the line from node i to node j; ij Q represents the voltage phase angle difference between node i and node j. D,i Q is the reactive power injected at node i within the target distribution network. L,i The reactive power of the load at node i within the target distribution network;

[0024] The following formula is used as the operating voltage constraint for the distribution network:

[0025] V min ≤V i ≤V max

[0026] In the formula V min The set lower limit value for the node voltage; V max This is the set upper limit value for the node voltage;

[0027] The following formula is used as the active power output constraint for distributed photovoltaic power generation:

[0028] P j,PV,min ≤P j,PV ≤P j,PV,max

[0029] In the formula P j,PV The active power output of the j-th distributed photovoltaic power generation; P j,PV,min P is the set lower limit value of the active power output of the j-th distributed photovoltaic power generation; j,PV,max This is the upper limit of the active power output of the j-th distributed photovoltaic power generation.

[0030] The following formula is used as the capacity constraint for distributed photovoltaic power generation:

[0031] (P j,PV ) 2 +(Q j,PV ) 2 ≤(S j,PV ) 2

[0032] In the formula Q j,PV For the reactive power output of the j-th distributed photovoltaic power generation; S j,PV Let be the capacity of the j-th distributed photovoltaic power generation unit;

[0033] The absolute value of the reactive power output boundary of distributed photovoltaic power generation is calculated using the following formula:

[0034]

[0035] In the formula Q j,PV,boundtanθ is the absolute value of the reactive power output boundary of distributed photovoltaic power generation; tanθ is the power factor that the photovoltaic inverter needs to satisfy for operation.

[0036] The following formula is used as the active power output constraint for wind power generation:

[0037] P WD,min ≤P WD ≤P WD,max

[0038] In the formula P WD For the active power output of wind power generation; P WD,min P is the set lower limit for the active power output of wind power generation; WD,max The set upper limit for the active power output of wind power generation;

[0039] The following formula is used as a constraint on the reactive power regulation capability of the wind turbine generator stator:

[0040]

[0041] In the formula U s X represents the effective value of the stator voltage of the wind turbine generator; s For the stator reactance of the wind turbine generator; X m For the excitation reactance of the wind turbine generator; I rmax Q represents the maximum rotor current of the wind turbine generator. WD For reactive power output from wind power generation;

[0042] The following formula is used as the maximum power constraint for charging and discharging of distributed energy storage:

[0043] P j,B,ch ≤P j,B ≤P j,B,dis

[0044] In the formula P j,B For the active power output of the j-th distributed energy storage; P j,B,ch P represents the maximum charging power of the j-th distributed energy storage device. j,B,dis Let be the maximum discharge power of the j-th distributed energy storage;

[0045] The following formula is used as the state-of-charge constraint for distributed energy storage:

[0046] S OC,min ≤S OC ≤S OC,max

[0047] In the formula S OC State of charge (S) for distributed energy storage; OC,min The set state-of-charge limit for distributed energy storage; S OC,max This is the upper limit of the state of charge for the distributed energy storage.

[0048] Step S3 describes solving the multi-agent voltage control model constructed in step S2 using the DDQN algorithm to achieve multi-agent voltage control of the target distribution network, including the following steps:

[0049] For the multi-agent voltage control model constructed in step S2, the DDQN algorithm is used for solving. During the solution process, the distributed wind power generation control module, distributed photovoltaic power generation control module, and distributed energy storage control module in the target distribution network are used as agents of the DDQN algorithm. The node voltage, node active power, and node reactive power in the target distribution network are used as the state space. The obtained node voltage, node active power, and node reactive power in the target distribution network are used as the input of the DDQN algorithm. The objective function of the multi-agent voltage control model of the target distribution network is used as the reward function.

[0050] After the solution is completed, the multi-agent voltage control of the target distribution network is completed based on the solution results.

[0051] Step S3 specifically includes the following steps:

[0052] The state space is represented as S n ={V i ,P i Q i |i∈Ω c}, where S n V is the state space of the agent. i P is the node voltage. i Q represents the active power of the node. i For node reactive power, Ω c The set of nodes within the partition where the agent resides;

[0053] The action space of an agent is represented as A. n ={Q c,WD,PV ,P c,B |c∈Ω c,WD,PV,B}, where A n Q represents the action space of the agent. c,WD,PV For the reactive power output of the control devices within the intelligent body, P c,B For the active power output of the control devices within the intelligent body, Ω c,WD,PV,B A collection of control devices within a zoned intelligent agent area;

[0054] The following formula is used as the reward function:

[0055]

[0056] In the formula r n This is the reward function value.

[0057] This invention also provides a system for implementing the multi-agent voltage control method for the distribution network, comprising a data acquisition module, a model building module, and a voltage control module; the data acquisition module, model building module, and voltage control module are connected in series; the data acquisition module is used to acquire data information of the target distribution network and upload the data information to the model building module; the model building module is used to construct a multi-agent voltage control model of the target distribution network based on the received data information and the acquired data information, and upload the data information to the voltage control module; the voltage control module is used to solve the constructed multi-agent voltage control model using the DDQN algorithm based on the received data information to achieve multi-agent voltage control of the target distribution network.

[0058] A third objective of this invention is to provide a storage medium on which a computer program is stored; when the computer program is executed by a processor, it implements the aforementioned multi-agent voltage control method for power distribution networks.

[0059] The multi-agent voltage control method, system, and storage medium for distribution networks provided by this invention, through the construction of a multi-agent voltage control model of the target distribution network and the corresponding solution using the DDQN algorithm, not only achieves voltage control of the multi-agent distribution network, but also has higher reliability and better accuracy. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0061] Figure 2 This is a schematic diagram of the functional modules of the system of the present invention. Detailed Implementation

[0062] like Figure 1 The diagram shown is a flowchart of the method of the present invention: The multi-agent voltage control method for power distribution networks provided by the present invention includes the following steps:

[0063] S1. Obtain data information from the target distribution network; specifically including the following steps:

[0064] Acquire data information of new energy power generation systems within the target distribution network;

[0065] The data information of the new energy power generation system includes the output data information of the distributed photovoltaic power generation system, the output data information of the distributed wind power generation system, and the output data information of the distributed energy storage system.

[0066] S2. Based on the data obtained in step S1, construct a multi-agent voltage control model for the target distribution network; including the following steps:

[0067] Based on the data obtained in step S1, a multi-agent voltage control model for the target distribution network is constructed with the minimum node voltage deviation as the objective function and the power flow conditions, operating voltage conditions, distributed photovoltaic output conditions, distributed photovoltaic capacity conditions, distributed wind power output conditions, distributed energy storage power conditions, and distributed energy storage state of charge conditions as constraints.

[0068] The specific implementation includes the following steps:

[0069] The following formula is used as the objective function:

[0070]

[0071] In the formula, F(x) is the objective function value; Ω n,O V represents the observation area of ​​each intelligent agent within the target distribution network. i Vi represents the voltage of node i within the observation area of ​​each intelligent agent in the target distribution network; V0 represents the rated voltage of the target distribution network.

[0072] The following formula is used as the power flow constraint:

[0073]

[0074] In the formula P D,i P is the active power injected at node i within the target distribution network. L,i U represents the active power of the load at node i within the target distribution network; i N represents the voltage magnitude at node i; node U is the set of nodes in a distribution network. j G represents the voltage magnitude at node j. ij B is the conductance value on the line from node i to node j; ij δ represents the susceptance value on the line from node i to node j; ij Q represents the voltage phase angle difference between node i and node j. D,i Q is the reactive power injected at node i within the target distribution network. L,i The reactive power of the load at node i within the target distribution network;

[0075] The following formula is used as the operating voltage constraint for the distribution network:

[0076] V min ≤V i ≤V max

[0077] In the formula V min The set lower limit value for the node voltage; V max This is the set upper limit value for the node voltage;

[0078] Constraints related to distributed photovoltaic (PV) power generation; Distributed PV power generation participates in the voltage regulation of the distribution network through grid-connected inverters. When distributed PV power generation is connected to the grid and voltage exceeds the limit, the distributed PV inverter provides reactive power regulation.

[0079] The following formula is used as the active power output constraint for distributed photovoltaic power generation:

[0080] P j,PV,min ≤P j,PV ≤P j,PV,max

[0081] In the formula P j,PV The active power output of the j-th distributed photovoltaic power generation; P j,PV,min P is the set lower limit value of the active power output of the j-th distributed photovoltaic power generation; j,PV,max This is the upper limit of the active power output of the j-th distributed photovoltaic power generation.

[0082] The following formula is used as the capacity constraint for distributed photovoltaic power generation:

[0083] (P j,PV ) 2 +(Q j,PV ) 2 ≤(S j,PV ) 2

[0084] In the formula Q j,PV For the reactive power output of the j-th distributed photovoltaic power generation; S j,PV Let be the capacity of the j-th distributed photovoltaic power generation unit;

[0085] The absolute value of the reactive power output boundary of distributed photovoltaic power generation is calculated using the following formula:

[0086]

[0087] In the formula Q j,PV,bound tanθ is the absolute value of the reactive power output boundary of distributed photovoltaic power generation; tanθ is the power factor that the photovoltaic inverter needs to satisfy for operation.

[0088] Wind turbine generator set related constraints: The stator of the wind turbine generator set has a certain reactive power regulation capability. Unlike its grid-side converter, the stator reactive power regulation is not easy to cause power fluctuations. When distributed resources are connected to the grid for power generation, causing voltage over-limit phenomenon, the wind turbine generator set can perform reactive voltage regulation.

[0089] The following formula is used as the active power output constraint for wind power generation:

[0090] P WD,min ≤P WD ≤P WD,max

[0091] In the formula P WD For the active power output of wind power generation; P WD,min P is the set lower limit for the active power output of wind power generation; WD,max The set upper limit for the active power output of wind power generation;

[0092] The following formula is used as a constraint on the reactive power regulation capability of the wind turbine generator stator:

[0093]

[0094] In the formula U s X represents the effective value of the stator voltage of the wind turbine generator; s For the stator reactance of the wind turbine generator; X m For the excitation reactance of the wind turbine generator; I rmax Q represents the maximum rotor current of the wind turbine generator. WD For reactive power output from wind power generation;

[0095] Constraints related to distributed energy storage: During peak periods of distributed energy output, distributed energy storage can absorb the active power output of distributed resources through charging, alleviate voltage rise at distribution network nodes, and enable the distribution network to operate safely.

[0096] The following formula is used as the maximum power constraint for charging and discharging of distributed energy storage:

[0097] P j,B,ch ≤P j,B ≤P j,B,dis

[0098] In the formula P j,B For the active power output of the j-th distributed energy storage; P j,B,ch P represents the maximum charging power of the j-th distributed energy storage device. j,B,dis Let be the maximum discharge power of the j-th distributed energy storage;

[0099] The following formula is used as the state-of-charge constraint for distributed energy storage:

[0100] S OC,min ≤S OC ≤S OC,max

[0101] In the formula S OC State of charge (S) for distributed energy storage; OC,min The set state-of-charge limit for distributed energy storage; S OC,max The upper limit of the state of charge of the distributed energy storage is set.

[0102] S3. For the multi-agent voltage control model constructed in step S2, the DDQN algorithm is used to solve it in order to achieve multi-agent voltage control of the target distribution network; including the following steps:

[0103] For the multi-agent voltage control model constructed in step S2, the DDQN algorithm is used for solving. During the solution process, the distributed wind power generation control module, distributed photovoltaic power generation control module, and distributed energy storage control module in the target distribution network are used as agents of the DDQN algorithm. The node voltage, node active power, and node reactive power in the target distribution network are used as the state space. The obtained node voltage, node active power, and node reactive power in the target distribution network are used as the input of the DDQN algorithm. The objective function of the multi-agent voltage control model of the target distribution network is used as the reward function.

[0104] After the solution is completed, multi-agent voltage control of the target distribution network is performed based on the solution results;

[0105] The specific implementation includes the following steps:

[0106] The state space is represented as S n ={V i ,P i Q i |i∈Ω c}, where S n V is the state space of the agent. i P is the node voltage. i Q represents the active power of the node. i For node reactive power, Ω c The set of nodes within the partition where the agent resides;

[0107] Each zone's intelligent agent controls voltage by controlling its internal regulating devices, such as distributed energy storage, distributed photovoltaic, or wind turbines; the agent's action space is represented as A. n ={Q c,WD,PV ,P c,B |c∈Ω c,WD,PV,B}, where A n Q represents the action space of the agent. c,WD,PV For the reactive power output of the control devices within the intelligent body, P c,B For the active power output of the control devices within the intelligent body, Ω c,WD,PV,B A collection of control devices within a zoned intelligent agent area;

[0108] The DDQN algorithm is an improvement on the DQN algorithm, which solves the problems of overestimation and instability in DQN, so as to achieve better performance in reinforcement learning. Unlike the classic DQN algorithm, DDQN uses two independent neural networks: one for selecting the agent's actions and the other for evaluating the value of the selected actions. The two networks can correct each other, reducing the risk of overestimation.

[0109] The voltage control method uses the negative of the absolute value of the voltage deviation as the reward for the agent. The smaller the voltage deviation in the distribution network system, the greater the reward for the agent, thus training the agent to take actions that reduce voltage deviation. Therefore, the following formula is used as the reward function:

[0110]

[0111] In the formula r n This is the reward function value.

[0112] Each agent senses and collects electrical information of nodes within its corresponding region, using it as elements of its state space. Based on the voltage control strategy deployed by the agent using the DDQN algorithm, it performs corresponding operations by controlling controllable devices within its region, i.e., selecting actions from its action space to minimize node voltage deviation and achieve voltage regulation. The agent records the observed electrical information, corresponding control actions, electrical information observed after control, and reward values ​​calculated based on the post-control observations, and uploads this information to the experience replay pool. Subsequently, the network parameters of the agent are updated using the sampled experience, making the agent's strategy more inclined to select actions that yield higher rewards, i.e., actions more conducive to mitigating voltage deviation. The updated strategy is redeployed to the agent, and through continuous interaction with the distribution network environment, the agent's action strategy is gradually optimized. Through optimized control implemented by each agent within its respective region, the multi-agent system ultimately achieves the voltage control target across the entire network.

[0113] The method of this invention can alleviate the problem of voltage exceeding the limit in the distribution network caused by large-scale distributed resource grid-connected power generation, and has certain guiding significance for the voltage control problem of distribution network with distributed resource access.

[0114] The method of the present invention will be described below with reference to an embodiment:

[0115] Taking the IEEE 33-node system as an example, distributed photovoltaic (PV) systems are connected at nodes 6, 9, 10, 14, 18, 20, 23, 26, 30, and 32; wind turbines are connected at node 5; and distributed energy storage is connected at nodes 2, 12, 25, and 29. The apparent capacity of each PV system is 1 MV·A; the rated power of the wind turbines is 2 MV·A; the state of charge (SOC) of the distributed energy storage ranges from 0.2 to 0.9, with an initial SOC setting of 0.6; the rated capacity of each energy storage unit is 2 MWh; and the PV operating power factor ranges from 0.95 leading to 0.95 lagging.

[0116] The traditional centralized voltage control method based on particle swarm optimization and the method of this invention were used to optimize the voltage over-limit problem caused by distributed resource access in the system. The results are shown in Table 1.

[0117] Table 1. Comparative Analysis of Voltage Control Effects

[0118] Average voltage offset / pu Maximum voltage peak Voltage conditions without control 0.0132 11.07kV Particle Swarm Optimization Centralized Control 0.0101 10.49kV Method of the present invention 0.0096 10.38kV

[0119] As shown in Table 1, without voltage control, the average voltage deviation of the distribution network system is large, and the maximum voltage peak exceeds the upper limit of safe voltage operation, threatening the safe and stable operation of the distribution network system. After voltage control using the two methods, the maximum voltage peak is limited to the upper limit of safe voltage operation, and the average voltage deviation is suppressed. Among them, compared with the traditional centralized voltage control method using particle swarm optimization, the method of this invention further reduces the average voltage deviation and the maximum voltage peak, with the optimization effect of average voltage deviation improved by 4.95% and the maximum voltage peak significantly reduced.

[0120] Therefore, the method of the present invention can alleviate the problem of voltage exceeding the limit in the distribution network caused by large-scale distributed resource grid-connected power generation, and has good significance for the voltage control problem of distribution network with distributed resource access.

[0121] like Figure 2 The diagram shows the functional modules of the system of the present invention: The system for implementing the multi-agent voltage control method of the distribution network disclosed in this invention includes a data acquisition module, a model building module, and a voltage control module; the data acquisition module, model building module, and voltage control module are connected in series; the data acquisition module is used to acquire data information of the target distribution network and upload the data information to the model building module; the model building module is used to construct a multi-agent voltage control model of the target distribution network based on the received data information and the acquired data information, and upload the data information to the voltage control module; the voltage control module is used to solve the constructed multi-agent voltage control model using the DDQN algorithm based on the received data information to achieve multi-agent voltage control of the target distribution network.

Claims

1. A multi-agent voltage control method for a power distribution network, comprising the following steps: S1. Obtain data information from the target power distribution network; S2. Based on the data obtained in step S1, construct a multi-agent voltage control model for the target distribution network; S3. The DDQN algorithm is used to solve the multi-agent voltage control model constructed in step S2 in order to achieve multi-agent voltage control of the target distribution network.

2. The multi-agent voltage control method for power distribution networks according to claim 1, characterized in that... Step S1, which involves obtaining data information from the target distribution network, specifically includes the following steps: Acquire data information of new energy power generation systems within the target distribution network; The data information of the new energy power generation system includes the output data information of the distributed photovoltaic power generation system, the output data information of the distributed wind power generation system, and the output data information of the distributed energy storage system.

3. The multi-agent voltage control method for power distribution networks according to claim 2, characterized in that... Step S2, which involves constructing a multi-agent voltage control model for the target distribution network based on the data obtained in step S1, includes the following steps: Based on the data obtained in step S1, a multi-agent voltage control model for the target distribution network is constructed with the minimum node voltage deviation as the objective function and the power flow conditions, operating voltage conditions, distributed photovoltaic output conditions, distributed photovoltaic capacity conditions, distributed wind power output conditions, distributed energy storage power conditions, and distributed energy storage state of charge conditions as constraints.

4. The multi-agent voltage control method for power distribution networks according to claim 3, characterized in that... Step S2 specifically includes the following steps: The following formula is used as the objective function: In the formula, F(x) is the objective function value; Ω n,O V represents the observation area of ​​each intelligent agent within the target distribution network. i Vi represents the voltage of node i within the observation area of ​​each intelligent agent in the target distribution network; V0 represents the rated voltage of the target distribution network. The following formula is used as the power flow constraint: In the formula P D,i P is the active power injected at node i within the target distribution network. L,i U represents the active power of the load at node i within the target distribution network; i N represents the voltage magnitude at node i; node U is the set of nodes in a distribution network. j G represents the voltage magnitude at node j. ij B is the conductance value on the line from node i to node j; ij Let be the susceptance value on the line from node i to node j; δ ij Q represents the voltage phase angle difference between node i and node j. D,i Q is the reactive power injected at node i within the target distribution network. L,i The reactive power of the load at node i within the target distribution network; The following formula is used as the operating voltage constraint for the distribution network: In min ≤V i ≤V max In the formula V min The set lower limit value for the node voltage; V max This is the set upper limit value for the node voltage; The following formula is used as the active power output constraint for distributed photovoltaic power generation: P j,PV,min ≤P j,PV ≤P j,PV,max In the formula P j,PV The active power output of the j-th distributed photovoltaic power generation; P j,PV,min P is the set lower limit value of the active power output of the j-th distributed photovoltaic power generation; j,PV,max This is the upper limit of the active power output of the j-th distributed photovoltaic power generation. The following formula is used as the capacity constraint for distributed photovoltaic power generation: (P j,PV ) 2 +(Q j,PV ) 2 ≤(S j,PV ) 2 In the formula Q j,PV For the reactive power output of the j-th distributed photovoltaic power generation; S j,PV Let be the capacity of the j-th distributed photovoltaic power generation unit; The absolute value of the reactive power output boundary of distributed photovoltaic power generation is calculated using the following formula: In the formula Q j,PV,bound tanθ is the absolute value of the reactive power output boundary of distributed photovoltaic power generation; tanθ is the power factor that the photovoltaic inverter needs to satisfy for operation. The following formula is used as the active power output constraint for wind power generation: P WD,min ≤P WD ≤P WD,max In the formula P WD For the active power output of wind power generation; P WD,min P is the set lower limit for the active power output of wind power generation; WD,max The set upper limit for the active power output of wind power generation; The following formula is used as a constraint on the reactive power regulation capability of the wind turbine generator stator: In the formula U s X represents the effective value of the stator voltage of the wind turbine generator; s For the stator reactance of the wind turbine generator; X m For the excitation reactance of the wind turbine generator; I rmax Q represents the maximum rotor current of the wind turbine generator. WD For reactive power output from wind power generation; The following formula is used as the maximum power constraint for charging and discharging of distributed energy storage: P j,B,ch ≤P j,B ≤P j,B,dis In the formula P j,B For the active power output of the j-th distributed energy storage; P j,B,ch P represents the maximum charging power of the j-th distributed energy storage device. j,B,dis Let be the maximum discharge power of the j-th distributed energy storage; The following formula is used as the state-of-charge constraint for distributed energy storage: S OC,min ≤S OC ≤S OC,max In the formula S OC State of charge (S) for distributed energy storage; OC,min The set state-of-charge limit for distributed energy storage; S OC,max This is the upper limit of the state of charge for the distributed energy storage.

5. The multi-agent voltage control method for power distribution networks according to claim 4, characterized in that... Step S3 describes solving the multi-agent voltage control model constructed in step S2 using the DDQN algorithm to achieve multi-agent voltage control of the target distribution network, including the following steps: For the multi-agent voltage control model constructed in step S2, the DDQN algorithm is used for solving. During the solution process, the distributed wind power generation control module, distributed photovoltaic power generation control module, and distributed energy storage control module in the target distribution network are used as agents of the DDQN algorithm. The node voltage, node active power, and node reactive power in the target distribution network are used as the state space. The obtained node voltage, node active power, and node reactive power in the target distribution network are used as the input of the DDQN algorithm. The objective function of the multi-agent voltage control model of the target distribution network is used as the reward function. After the solution is completed, the multi-agent voltage control of the target distribution network is completed based on the solution results.

6. The multi-agent voltage control method for power distribution networks according to claim 5, characterized in that... Step S3 specifically includes the following steps: The state space is represented as S n ={V i ,P i Q i |i∈Ω c }, where S n V is the state space of the agent. i P is the node voltage. i Q represents the active power of the node. i For node reactive power, Ω c The set of nodes within the partition where the agent resides; The action space of an agent is represented as A. n ={Q c,WD,PV ,P c,B |c∈Ω c,WD,PV,B }, where A n Q represents the action space of the agent. c,WD,PV For the reactive power output of the control devices within the intelligent body, P c,B For the active power output of the control devices within the intelligent body, Ω c,WD,PV,B A collection of control devices within a zoned intelligent agent area; The following formula is used as the reward function: In the formula r n This is the reward function value.

7. A system for implementing the multi-agent voltage control method for power distribution networks as described in any one of claims 1 to 6, characterized in that... It includes a data acquisition module, a model building module, and a voltage control module, which are connected in series. The data acquisition module acquires data information from the target distribution network and uploads it to the model building module. The model building module constructs a multi-agent voltage control model of the target distribution network based on the received and acquired data information and uploads the model information to the voltage control module. The voltage control module solves the constructed multi-agent voltage control model using the DDQN algorithm based on the received data information to achieve multi-agent voltage control of the target distribution network.

8. A storage medium having a computer program stored thereon; when the computer program is executed by a processor, it implements the multi-agent voltage control method for power distribution networks as described in any one of claims 1 to 6.

Images