Medium and low voltage AC / DC power distribution network distributed control method and system based on partition

Abstract

The invention discloses a partition-based distributed control method and system for a medium and low voltage AC / DC power distribution network, and relates to the field of power distribution network control, and the method comprises the steps: on the basis of partitions of the medium and low voltage AC / DC power distribution network, setting an agent as a controller in each partition, and carrying out the distributed control through employing a multi-agent technology. The method comprises the following steps: firstly, proposing a distributed control model of the medium and low voltage AC / DC power distribution network, and performing optimization control on adjustable resources in a region to reduce network loss in the region; then, converting the multi-time-scale and multi-device coordinated optimization model into a Markov process, and defining a state space, an action space and a reward function; and finally, solving the Markov process by using an MAPPO algorithm to realize partition-based distributed control of the medium and low voltage AC / DC power distribution network. According to the invention, distributed parallel control can be realized for reducing network loss and eliminating voltage out-of-limit problems, and compared with centralized control, the control complexity of the medium and low voltage AC / DC power distribution network is reduced.

Description

Technical Field

[0001] The present invention relates to the field of distribution network control, and in particular to a partition-based distributed control method and system for medium and low voltage AC and DC distribution networks. Background Art

[0002] With the access of a high proportion of renewable energy and the increase of new power-consuming equipment, medium and low voltage coordinated AC and DC distribution networks will become the development trend of distribution networks. The use of AC and DC distribution networks can improve the carrying capacity of new sources and loads. However, distributed power sources have a wide distribution range and strong random output fluctuations. A large number of access will lead to voltage limit problems.

[0003] Traditional centralized voltage optimization control methods face challenges such as many control devices and low computational efficiency. Distributed control methods have begun to be studied to reduce the complexity of distribution network optimization control, but the characteristics of medium and low voltage AC and DC distribution networks across voltage levels and AC and DC levels make traditional distributed control methods difficult to adapt, and the coordination and interaction at the medium and low voltage AC and DC levels are limited. Summary of the invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: how to effectively realize the optimization control of medium and low voltage AC and DC distribution networks.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, an embodiment of the present invention provides a partition-based distributed control method for a medium and low voltage AC and DC distribution network, comprising:

[0008] Divide the medium and low voltage AC and DC distribution networks into zones, and set an intelligent agent as a controller in each zone;

[0009] Based on the operating status and physical characteristics of the partitioned power grid, a distributed control model for medium and low voltage AC and DC distribution networks is constructed;

[0010] Convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition;

[0011] Solve the Markov process to obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize the partition-based distributed control of the medium and low voltage AC and DC distribution network.

[0012] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0013] The step of setting up an agent as a controller in each partition includes:

[0014] The intelligent agents work in parallel and complete tasks collaboratively. Each intelligent agent has autonomy, local perspective, and can operate without central control. A single intelligent agent is set in each medium and low voltage AC and DC distribution network area to control the equipment in the area. The single intelligent agent collects voltage and power information of nodes in the area. It optimizes the area through its own optimization process based on the received information and transmits information with other intelligent agents.

[0015] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0016] The construction of a distributed control model for a medium and low voltage AC and DC distribution network includes:

[0017] The objective function of the distributed control model of medium and low voltage AC / DC distribution networks is assumed to be as follows:

[0018]

[0019] Among them, r ij ,I ij are the resistance and current between node i and node j respectively; N is the total number of nodes in the region;

[0020] The power flow constraint after second-order cone relaxation is:

[0021]

[0022] Among them, v i is the square of the voltage at node i; l ij is the square of the current flowing between node i and node j; U i is the voltage of node i;

[0023]

[0024] Among them, p j The active power injected into node j is numerically equal to the active power of the generator minus the load power; is the total active power flowing from node j to the downstream nodes; is the total active power injected from upstream node i into node j;

[0025]

[0026] Among them, q j The reactive power injected into node j is equal to the reactive power of the generator minus the load power; is the total reactive power flowing from node j to the downstream nodes; is the total reactive power injected into node j from upstream node i;

[0027]

[0028] Among them, x ij is the reactance between node i and node j;

[0029]

[0030] Among them, P ij is the active power flowing between node i and node j; Q ij is the active power flowing between node i and node j.

[0031] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0032] The construction of the distributed control model of the medium and low voltage AC and DC distribution network also includes:

[0033] The node voltage constraint is expressed as:

[0034] U min <U i,t <U max

[0035] Among them, U min , U max They are the upper and lower limits of node voltage respectively, and the medium voltage system takes ±7% of the reference voltage;

[0036] The branch current carrying capacity constraint is expressed as:

[0037] I ij,min <I ij <I ij,max

[0038] Among them, I ij,min ,I ij,max They are the upper and lower limits of the branch's allowable current.

[0039] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0040] The construction of the distributed control model of the medium and low voltage AC and DC distribution network also includes:

[0041] The operating constraints of distributed photovoltaic equipment are expressed as:

[0042] P G,i,min ≤P G,i ≤P G,i,max

[0043] Among them, P G,i,min , PG,i,max is the upper and lower limits of the active output of unit i; P G,i is the active power output of unit i;

[0044]

[0045] Among them, S G,i N is the capacity of unit i; Q G,i is the reactive power output of unit i; S G,i is the complex power output of unit i.

[0046] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0047] The method of converting the distributed control model of the medium and low voltage AC and DC distribution network into a Markov process and clarifying the state space, action space, and reward function of the Markov process includes:

[0048] The distribution network status information includes: power flow distribution, voltage distribution, and active and reactive power output of distributed photovoltaics within the distribution network partition; the state space expression of the Markov process is clarified by the distribution network status information:

[0049] S={P ij ,Q ij ,U i ,P G,i ,Q G,i}

[0050] The action space of the Markov process contains the active and reactive outputs of all regulating devices in the partition within the adjustable range. The agent finds the best action by learning the reward function. Assuming that there are m adjustable devices in the partition where the agent is located, the action vector of each agent is:

[0051] A=[P G,1 ,Q G,1 ,...,P G,m ,Q G,m ]

[0052] Each action vector corresponds to a device adjustment scheme.

[0053] As a preferred solution of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions, wherein:

[0054] The method of converting the distributed control model of the medium and low voltage AC and DC distribution network into a Markov process and clarifying the state space, action space, and reward function of the Markov process also includes:

[0055] The reward function obtained by the system executing a decision after the interval t is defined as:

[0056] R=-w1F-w2N U,ex

[0057] Among them, w1 is the network loss penalty coefficient; w2 is the voltage over-limit penalty coefficient; N U,ex is the number of nodes where voltage exceeds the limit.

[0058] In a second aspect, an embodiment of the present invention provides a partition-based medium and low voltage AC and DC distribution network distributed control system, including:

[0059] A partitioning module is used to partition the medium and low voltage AC and DC distribution network, and set an intelligent agent as a controller in each partition;

[0060] Modeling module, used to build a distributed control model of medium and low voltage AC and DC distribution networks based on the operating status and physical characteristics of the partitioned power grid;

[0061] The conversion module is used to convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition;

[0062] The solution module is used to solve the Markov process, obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize the partition-based distributed control of the medium and low voltage AC and DC distribution network.

[0063] In a third aspect, an embodiment of the present invention provides a computing device, including:

[0064] Memory and processor;

[0065] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the partition-based distributed control method for medium and low voltage AC and DC distribution networks as described in any embodiment of the present invention.

[0066] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the partition-based distributed control method for medium and low voltage AC and DC distribution networks.

[0067] Beneficial effects of the present invention: The present invention proposes a distributed control model for medium and low voltage AC and DC distribution networks. The model is applied to a certain partition of the medium and low voltage AC and DC distribution networks, and realizes optimized control to reduce network losses by adjusting the controllable devices in the area; based on the distributed control model of the medium and low voltage AC and DC distribution networks, the multi-time scale and multi-device coordination problem is converted into a Markov process, and the MAPPO deep reinforcement learning method is used to solve the Markov process, which can realize distributed parallel control for reducing network losses and eliminating voltage over-limit problems; for the medium and low voltage AC and DC distribution networks, the present invention sets an intelligent agent as a controller for each partition on the basis of partition, and uses multi-agent technology for distributed control. After centralized training of all intelligent agents, each intelligent agent can give the optimal action plan for the controllable devices in the partition according to the changes in local information in the area under its jurisdiction, so as to realize distributed optimized control of the medium and low voltage AC and DC distribution networks; compared with centralized control, the control complexity of the medium and low voltage AC and DC distribution networks is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0069] Figure 1 It is an overall flow chart of the distributed control method of medium and low voltage AC and DC distribution networks based on partitions according to the first embodiment of the present invention;

[0070] Figure 2 It is a schematic diagram of a multi-agent system deployed in different regions in a distribution network of a distributed control method for medium and low voltage AC and DC distribution networks based on partitions according to the first embodiment of the present invention;

[0071] Figure 3 It is a schematic diagram of regions after decoupling of each region of the distributed control method for medium and low voltage AC and DC distribution networks based on partitions according to the first embodiment of the present invention;

[0072] Figure 4 It is a multi-agent training framework diagram of a partition-based medium and low voltage AC and DC distribution network distributed control method according to the first embodiment of the present invention;

[0073] Figure 5 It is a distributed control framework diagram based on multi-agents of the partition-based medium and low voltage AC and DC distribution network distributed control method described in the first embodiment of the present invention;

[0074] Figure 6It is a network topology diagram in a simulation example of a distributed control method for medium and low voltage AC and DC distribution networks based on partitions according to the second embodiment of the present invention;

[0075] Figure 7 It is a schematic diagram of system voltage distribution of three control cases in a simulation example of a partition-based distributed control method for medium and low voltage AC and DC distribution networks according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0077] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0078] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0079] Example 1

[0080] Reference Figure 1-Figure 5 , which is the first embodiment of the present invention, and provides a partition-based distributed control method for medium and low voltage AC and DC distribution networks, comprising:

[0081] S1: Divide the medium and low voltage AC and DC distribution networks into zones, and set an intelligent agent as a controller in each zone;

[0082] In an embodiment of the present application, setting an agent as a controller in each partition includes:

[0083] The intelligent agents work in parallel and complete tasks collaboratively. Each intelligent agent has autonomy, local perspective, and can operate without central control. A single intelligent agent is set in each medium and low voltage AC and DC distribution network area to control the equipment in the area. The single intelligent agent collects voltage and power information of nodes in the area. It optimizes the area through its own optimization process based on the received information and transmits information with other intelligent agents.

[0084] For example, a schematic diagram of a multi-agent system deployed in different regions in a distribution network is shown in FIG. Figure 2 As shown in the figure. Assume a 4-node radial distribution network, nodes {1,2} belong to area 1, nodes {3,4} belong to area 2, and deploy an agent in each area. The agent collects node voltage and node power information in the area to understand the power status in the area. Then the agent controls the adjustable equipment in the area based on the information obtained, including managing the switching of reactive compensation devices and the reactive power output of photovoltaic inverters, so as to optimize the operating status in the area.

[0085] In order to implement distributed control based on partitions, each area should be decoupled. The schematic diagram of the decoupled area is as follows: Figure 3 As shown in Figure 1, the interconnected branch between area 1 and area 2 is branch 2. For area 1, branch 2 should be equivalent to a load node; for area 2, branch 2 should be equivalent to a balance node. Agent 1 and agent 2 communicate to transmit the power flowing through branch 2 and the voltage of node 2.

[0086] In the distributed optimization process, Agent 2 optimizes according to the initial parameters, calculates the power flowing through Branch 2, and passes it to Agent 1 for optimization; Agent 1 passes the optimized voltage of Node 3 to Agent 2, updates its parameters and optimizes again. Through continuous iteration, the power of Branch 2 and the voltage of Node 3 can be converged, thus completing the distributed optimization of the entire network.

[0087] S2: Based on the operating status and physical characteristics of the partitioned power grid, a distributed control model of medium and low voltage AC and DC distribution networks is constructed;

[0088] In the embodiment of the present application, constructing a distributed control model for a medium and low voltage AC and DC distribution network includes:

[0089] The objective function of the distributed control model of medium and low voltage AC / DC distribution networks is assumed to be as follows:

[0090]

[0091] Among them, r ij ,I ij are the resistance and current between node i and node j respectively; N is the total number of nodes in the region;

[0092] The power flow constraint after second-order cone relaxation is:

[0093]

[0094] Among them, v i is the square of the voltage at node i; l ij is the square of the current flowing between node i and node j; U iis the voltage of node i;

[0095]

[0096] Among them, p j The active power injected into node j is numerically equal to the active power of the generator minus the load power; is the total active power flowing from node j to the downstream nodes; is the total active power injected from upstream node i into node j;

[0097]

[0098] Among them, q j The reactive power injected into node j is equal to the reactive power of the generator minus the load power; is the total reactive power flowing from node j to the downstream nodes; is the total reactive power injected into node j from upstream node i;

[0099]

[0100] Among them, x ij is the reactance between node i and node j;

[0101]

[0102] Among them, P ij is the active power flowing between node i and node j; Q ij is the active power flowing between node i and node j.

[0103] The node voltage constraint is expressed as:

[0104] U min <U i,t <U max

[0105] Among them, U min , U max They are the upper and lower limits of node voltage respectively, and the medium voltage system takes ±7% of the reference voltage;

[0106] The branch current carrying capacity constraint is expressed as:

[0107] I ij,min <I ij <I ij,max

[0108] Among them, I ij,min ,I ij,max They are the upper and lower limits of the branch's allowable current.

[0109] The operating constraints of distributed photovoltaic equipment are expressed as:

[0110] P G,i,min ≤P G,i ≤P G,i,max

[0111] Among them, P G,i,min , P G,i,max is the upper and lower limits of the active output of unit i; P G,i is the active power output of unit i;

[0112]

[0113] Among them, S G,i N is the capacity of unit i; Q G,i is the reactive power output of unit i; S G,i is the complex power output of unit i.

[0114] It should be noted that the active and reactive output of resources can be adjusted in each partition control area to optimize the network losses in the area.

[0115] S3: Convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition;

[0116] In the embodiment of the present application, the distributed control model of the medium and low voltage AC and DC distribution network is converted into a Markov process, and the state space, action space, and reward function of the Markov process are clarified to include:

[0117] The distribution network status information includes: power flow distribution, voltage distribution, and active and reactive power output of distributed photovoltaics within the distribution network partition; the state space expression of the Markov process is clarified by the distribution network status information:

[0118] S={P ij ,Q ij ,U i ,P G,i ,Q G,i}

[0119] The action space of the Markov process contains the active and reactive outputs of all regulating devices in the partition within the adjustable range. The agent finds the best action by learning the reward function. Assuming that there are m adjustable devices in the partition where the agent is located, the action vector of each agent is:

[0120] A=[P G,1 ,Q G,1 ,...,P G,m ,Q G,m ]

[0121] Each action vector corresponds to a device adjustment scheme.

[0122] The reward function obtained by the system executing a decision after the interval t is defined as:

[0123] R=-w1F-w2N U,ex

[0124] Among them, w1 is the network loss penalty coefficient; w2 is the voltage over-limit penalty coefficient; N U,ex is the number of nodes where voltage exceeds the limit.

[0125] It should be noted that the Markov decision process is a mathematical model used to simulate the random decisions and rewards that can be achieved by an agent. It is built based on the agent and the environment, and its elements include state, action, strategy, and reward. In the simulation of the Markov decision process, the agent perceives the current system state, implements actions according to the strategy, thereby changing the state of the environment and obtaining rewards. The agent continuously collects data through interaction with the environment to achieve specific task goals.

[0126] In the present invention, the Markov decision process consists of four elements, namely, the state space (S), which represents the set of all possible states of the system. Each state is a specific situation of the system at a certain moment; the action space (A), which represents the set of all actions that the agent can choose. Each action is an operation that the agent can perform in a certain state; the state transition probability (P), which represents the probability that the system transitions from one state to another under a given state and action; the reward function (R), which represents the immediate reward obtained by the agent under a given state and action. The reward function can vary depending on the state and action.

[0127] S4: Solve the Markov process to obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize partition-based distributed control of medium and low voltage AC and DC distribution networks.

[0128] In the embodiment of the present application, the MAPPO algorithm is used to solve the Markov process. MAPPO is a deep reinforcement learning algorithm used to solve the Markov algorithm decision process in multi-agent optimization. It adopts the classic Actor-Critic architecture and is applied to the scenario of full cooperation of agents. It guides the agents to make decisions in the environment by learning a value function to maximize the reward. The present invention uses a hybrid framework, namely: centralized training and decentralized execution. The model is initially trained in an environment of global perception and global control, and the optimal distributed control strategy obtained is passed to each agent. The multi-agent training framework is as follows: Figure 4As shown in Figure 1. In this framework, each agent has a separate Actor-Critic network. During training, the agent's own Actor network must be scored by each Critic. After training, each agent can call a local control strategy to generate the optimal control action based on its own local observations and information transmitted through communication. The distributed control framework based on multi-agents is shown in Figure 1. Figure 5 shown.

[0129] The above is a schematic scheme of the distributed control method for medium and low voltage AC and DC distribution networks based on partitions of this embodiment. It should be noted that the technical scheme of the distributed control system for medium and low voltage AC and DC distribution networks based on partitions and the technical scheme of the distributed control method for medium and low voltage AC and DC distribution networks based on partitions above belong to the same concept, and the details of the technical scheme of the distributed control system for medium and low voltage AC and DC distribution networks based on partitions in this embodiment that are not described in detail can all be referred to the description of the technical scheme of the distributed control method for medium and low voltage AC and DC distribution networks based on partitions above.

[0130] The partition-based medium and low voltage AC and DC distribution network distributed control system in this embodiment includes:

[0131] A partitioning module is used to partition the medium and low voltage AC and DC distribution network, and set an intelligent agent as a controller in each partition;

[0132] Modeling module, used to build a distributed control model of medium and low voltage AC and DC distribution networks based on the operating status and physical characteristics of the partitioned power grid;

[0133] The conversion module is used to convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition;

[0134] The solution module is used to solve the Markov process, obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize the partition-based distributed control of the medium and low voltage AC and DC distribution network.

[0135] This embodiment further provides a computing device, which is applicable to the case of a distributed control method for a medium and low voltage AC and DC distribution network based on partitions, including:

[0136] Memory and processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the partition-based medium and low voltage AC and DC distribution network distributed control method proposed in the above embodiment.

[0137] This embodiment further provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the distributed control method for medium and low voltage AC and DC distribution networks based on partitioning as proposed in the above embodiment is implemented.

[0138] The storage medium proposed in this embodiment and the partition-based distributed control method for medium and low voltage AC and DC distribution networks proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0139] Example 2

[0140] Reference Figure 6-Figure 7 , which is an embodiment of the present invention, provides a partition-based distributed control method for medium and low voltage AC and DC distribution networks. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0141] The proposed distributed control method for medium and low voltage AC and DC power distribution systems based on partitions is verified by using an improved IEEE33 node example. The system has been divided into regions, and the network topology is as follows: Figure 6 As shown. The voltage level of the medium voltage layer of the system is 10kV, the voltage level of the low voltage AC layer is 380V, and the voltage level of the low voltage DC layer is ±375V; nodes 8, 13, 18, 21, and 25 are connected to the photovoltaic power station; nodes 17 and 28 are low voltage AC and DC stations, which are connected to distributed photovoltaics through DC bus confluence. The entire system is divided into three areas, area ① consists of nodes {1-7, 19-22, 23-25}, area ② consists of nodes {26-33}, and area ③ consists of nodes {8-18}. Each area is equipped with an intelligent agent to control the reactive power output of the photovoltaic inverter in the area.

[0142] The distribution network is subjected to distributed optimization control based on partitions, and no optimization and centralized optimization based on linear programming are set for comparison. The system voltage distribution of the three control cases is as follows: Figure 7 shown.

[0143] Before the optimization control, the voltage distribution between the nodes of the system fluctuated greatly, and the network loss power was high, reaching 104kW. After the optimization control, the photovoltaic inverter provided reactive support for the system, which alleviated the problem of excessive voltage difference between nodes, and the voltage distribution became smoother. Among them, the system network loss after centralized optimization was 55.42kW, and after adopting the distributed optimization method proposed in the present invention, the system network loss was 54.79kW, and the loss reduction rate reached 47.3%. The optimization results of the method of the present invention are similar to those of the centralized optimization, which proves the effectiveness of the method of the present invention.

[0144] Centralized control requires obtaining power and voltage information from all nodes in the system and communicating with all controllable devices in the system. It requires large computational scale and communication requirements. Once the centralized optimization controller fails due to a fault, the operation control of the entire network cannot be carried out. The distributed control method proposed in the present invention is based on the regional division of the distribution network. Multiple intelligent controllers are set to control the network in different regions. Each intelligent agent only faces the elements in its region, and the intelligent agents transmit limited information to each other. When a distributed controller fails, only the equipment in this region cannot be optimized and controlled, and the optimization control of the regions where other controllers are located can continue.

[0145] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A distributed control method for medium and low voltage AC and DC distribution networks based on partitioning, characterized in that: include: Divide the medium and low voltage AC and DC distribution networks into zones, and set an intelligent agent as a controller in each zone; Based on the operating status and physical characteristics of the partitioned power grid, a distributed control model for medium and low voltage AC and DC distribution networks is constructed; Convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition; Solve the Markov process to obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize the partition-based distributed control of the medium and low voltage AC and DC distribution network.

2. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 1 is characterized in that: The step of setting up an agent as a controller in each partition includes: The intelligent agents work in parallel and complete tasks collaboratively. Each intelligent agent has autonomy, local perspective, and can operate without central control. A single intelligent agent is set in each medium and low voltage AC and DC distribution network area to control the equipment in the area. The single intelligent agent collects voltage and power information of nodes in the area. It optimizes the area through its own optimization process based on the received information and transmits information with other intelligent agents.

3. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 2 is characterized in that: The construction of a distributed control model for a medium and low voltage AC and DC distribution network includes: The objective function of the distributed control model of medium and low voltage AC / DC distribution networks is assumed to be as follows: Among them, r ij ,I ij are the resistance and current between node i and node j respectively; N is the total number of nodes in the region; The power flow constraint after second-order cone relaxation is: Among them, v i is the square of the voltage at node i; l ij is the square of the current flowing between node i and node j; U i is the voltage of node i; Among them, p j The active power injected into node j is numerically equal to the active power of the generator minus the load power; is the total active power flowing from node j to the downstream nodes; is the total active power injected from upstream node i into node j; Among them, q j The reactive power injected into node j is equal to the reactive power of the generator minus the load power; is the total reactive power flowing from node j to the downstream nodes; is the total reactive power injected into node j from upstream node i; Among them, x ij is the reactance between node i and node j; Among them, P ij is the active power flowing between node i and node j; Q ij is the active power flowing between node i and node j.

4. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 3 is characterized in that: The construction of the distributed control model of the medium and low voltage AC and DC distribution network also includes: The node voltage constraint is expressed as: IN min <In i,t <In max Among them, U min , U max They are the upper and lower limits of node voltage respectively, and the medium voltage system takes ±7% of the reference voltage; The branch current carrying capacity constraint is expressed as: I ij,min <I ij <I ij,max Among them, I ij,min ,I ij,max They are the upper and lower limits of the branch's allowable current.

5. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 4 is characterized in that: The construction of the distributed control model of the medium and low voltage AC and DC distribution network also includes: The operating constraints of distributed photovoltaic equipment are expressed as: P G,i,min ≤P G,i ≤P G,i,max Among them, P G,i,min , P G,i,max is the upper and lower limits of the active output of unit i; P G,i is the active power output of unit i; in, is the capacity of unit i; Q G,i is the reactive power output of unit i; S G,i is the complex power output of unit i.

6. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 5 is characterized in that: The method of converting the distributed control model of the medium and low voltage AC and DC distribution network into a Markov process and clarifying the state space, action space, and reward function of the Markov process includes: The distribution network status information includes: power flow distribution, voltage distribution, and active and reactive power output of distributed photovoltaics within the distribution network partition; the state space expression of the Markov process is clarified by the distribution network status information: S={P ij ,Q ij ,U i ,P G,i ,Q G,i } The action space of the Markov process contains the active and reactive outputs of all regulating devices in the partition within the adjustable range. The agent finds the best action by learning the reward function. Assuming that there are m adjustable devices in the partition where the agent is located, the action vector of each agent is: A=[P G,1 ,Q G,1 ,...,P G,m ,Q G,m ] Each action vector corresponds to a device adjustment scheme.

7. The distributed control method for medium and low voltage AC and DC distribution networks based on partitioning according to claim 6 is characterized in that: The method of converting the distributed control model of the medium and low voltage AC and DC distribution network into a Markov process and clarifying the state space, action space, and reward function of the Markov process also includes: The reward function obtained by the system executing the decision after the interval t is defined as: R=-w1F-w2N U,ex Among them, w1 is the network loss penalty coefficient; w2 is the voltage over-limit penalty coefficient; N U,ex is the number of nodes with voltage exceeding the limit.

8. A system using the partition-based distributed control method for medium and low voltage AC and DC distribution networks as claimed in any one of claims 1 to 7, characterized in that: include: A partitioning module is used to partition the medium and low voltage AC and DC distribution network, and set an intelligent agent as a controller in each partition; Modeling module, used to build a distributed control model of medium and low voltage AC and DC distribution networks based on the operating status and physical characteristics of the partitioned power grid; The conversion module is used to convert the distributed control model of medium and low voltage AC and DC distribution networks into a Markov process, and build the state space, action space, and reward function of the corresponding intelligent agent according to the source-load characteristics and equipment characteristics of each partition; The solution module is used to solve the Markov process, obtain the optimal action reward in the corresponding environment, and output the corresponding device action decision; each partition intelligent agent controls the corresponding device action based on the device action decision to realize the partition-based distributed control of the medium and low voltage AC and DC distribution network.

9. A computing device comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the partition-based medium and low voltage AC and DC distribution network distributed control method are implemented as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the partition-based medium and low voltage AC and DC distribution network distributed control method as described in any one of claims 1 to 7.

Images