Multi-microgrid distributed energy scheduling method and system based on energy flow information entropy

Abstract

The invention discloses a multi-microgrid distributed energy scheduling method and system based on energy flow information entropy. The method comprises the following steps: collecting grid structure information of multiple microgrids; acquiring multi-microgrid operation information; introducing energy flow information entropy to construct a multi-microgrid control target model; constructing a reinforcement learning network of an output scheduling strategy according to the multi-microgrid control target model; training the reinforcement learning network; inputting multi-microgrid operation information collected in real time into the trained reinforcement learning network, and outputting a scheduling strategy; and executing the multi-microgrid system scheduling strategy. According to the invention, three-level balance in the micro-grids, between adjacent micro-grids and the global micro-grid group can be realized; by minimizing the energy flow information entropy, the orderliness of the system is improved, the action frequency of equipment in the multi-microgrid system is reduced, the scheduling requirement is reduced, and the local consumption efficiency of distributed energy is improved; and power transmission channel capacity limitation is considered in an energy storage system capacity reward item, so that the network congestion probability is reduced.

Description

Technical Field

[0001] The present invention belongs to the field of power systems, relates to microgrid dispatching control technology, and specifically relates to a multi-microgrid distributed energy dispatching method and system based on energy flow information entropy. Background Art

[0002] At present, the energy dispatching control mode is gradually evolving from traditional centralized control to decentralized control. The centralized dispatching control architecture relies on the central controller to centrally collect and process the information of the whole network, including data such as the output of distributed power sources and load demands. After global optimization calculation through the centralized energy management system, the dispatching instructions are then sent to each local execution unit. The decentralized dispatching control mode emphasizes more on the autonomy and flexibility of the system. On the premise of ensuring the economic operation of the microgrid, higher decision-making autonomy is given to each control unit. In this mode, real-time information interaction is maintained between the central controller and local controllers. Each local controller not only feeds back the current operating state and future prediction data to the central controller, but also can actively put forward power generation or load regulation requests according to local demands.

[0003] The existing microgrid energy dispatching methods based on centralized control rely on high-frequency global information interaction, and traditional algorithms are limited by high-dimensional control variables and non-convex optimization objectives when solving the load dispatching problem, resulting in poor control effects. There are the following problems: 1. The problem of large communication overhead and the need to frequently exchange global data; 2. The problem that the overall control design needs to be modified for newly added distributed power generation sources; 3. Low energy dispatching efficiency, making it difficult to achieve near-balanced power, resulting in line losses and network congestion caused by long-distance power transmission; 4. Facing the massive devices in multi-microgrids, centralized control needs to process a large amount of information, and the computational load of the central controller is huge.

[0004] Although the existing technology has also proposed a framework for applying reinforcement learning to microgrid distributed control, each microgrid is regarded as an agent, and they are restricted to only interact with agents of adjacent microgrids. After training, each agent works in a distributed manner, enabling the local microgrid to be self-sufficient relying on distributed power generation devices and conduct transactions with the power grid with power demands, achieving local autonomy and privacy protection of energy. However, the existing microgrid distributed control technology only focuses on the operating conditions of each microgrid and ignores the overall operating conditions of the multi-microgrid system, and does not effectively guide the system to evolve towards an orderly state. It may lead to a reduction in the lifespan of power electronic devices in the microgrid due to frequent forward and reverse power interactions; it may be difficult to coordinate the output plans of distributed units in the pre-dispatching stage due to ignoring the overall needs of the multi-microgrid system. Summary of the Invention

[0005] Objective of the Invention: To overcome the deficiencies in the prior art, the present invention provides a multi-microgrid distributed energy scheduling method and system based on energy flow information entropy, which can achieve three-level balance at the levels of within the microgrid, between adjacent microgrids, and globally for the microgrid group; by minimizing the energy flow information entropy, the system orderliness is improved, the action frequency of equipment in the multi-microgrid system is reduced, the scheduling demand is decreased, and the local consumption efficiency of distributed energy is increased; considering the transmission channel capacity limit in the energy storage system capacity reward item, the network congestion probability is reduced.

[0006] Technical Solution: To achieve the above objective, the present invention provides a multi-microgrid distributed energy scheduling method based on energy flow information entropy, including the following steps:

[0007] S1: Collect the multi-microgrid network structure information;

[0008] S2: Based on the multi-microgrid network structure information, collect the multi-microgrid operation information;

[0009] S3: Based on the multi-microgrid operation information, introduce the energy flow information entropy to construct a multi-microgrid control target model;

[0010] S4: According to the multi-microgrid control target model, construct a reinforcement learning network for outputting a scheduling strategy;

[0011] S5: Train the reinforcement learning network;

[0012] S6: Input the real-time collected multi-microgrid operation information into the trained reinforcement learning network to output a scheduling strategy;

[0013] S7: Execute the multi-microgrid system scheduling strategy.

[0014] Further, the multi-microgrid network structure information in step S1 includes: the number of microgrid nodes in the target multi-microgrid area , the system node impedance matrix of the target multi-microgrid area , the electrical distance matrix between each microgrid node in the target multi-microgrid area , the maximum capacity matrix of the transmission channels between each microgrid node in the target multi-microgrid area , the maximum energy storage capacity matrix of each microgrid node in the target multi-microgrid area .

[0015] Further, the expression of the system node impedance matrix of the target multi-microgrid area in step S1 is:

[0016]

[0017] In the matrix among row column element​ For the system equivalent impedance seen from the port; The electrical distance matrix between each microgrid node in the target multi - microgrid area is obtained according to formula (1)

[0018] , and the expression is: In the matrix

[0019]

[0020] The element in the row and column represents the electrical distance from microgrid node to microgrid node , which is used to characterize the tightness of the electrical connection between two microgrids. , when , the electrical distance ;

[0021] The expression of the maximum capacity matrix of the power transmission channels between each microgrid node in the target multi - microgrid area is:

[0022]

[0023] In the matrix the element in the row and column is the maximum power transmission capacity of the power transmission channel from microgrid node to microgrid node . is a symmetric matrix, that is

[0024] The expression of the maximum energy storage capacity matrix of each microgrid node in the target multi - microgrid area is:

[0025]

[0026] where represents the maximum energy storage capacity of microgrid node .

[0027] Furthermore, the multi - microgrid operation information in step S2 includes the energy storage power status of each microgrid , the interactive power between microgrids , the multi - microgrid user load demand , the output of the multi - microgrid distributed power sources , and the operation status of the power transmission channels between multi - microgrids .

[0028] Further, the energy storage power status of each microgrid in step S2 is expressed as follows:

[0029]

[0030] where, is the percentage of the available remaining capacity of the energy storage in microgrid at the -th collection;

[0031] The interactive power between microgrids is expressed as follows:

[0032]

[0033] where, is the power sent from microgrid to microgrid at the -th collection;

[0034] The load demand of multi - microgrid users is expressed as follows:

[0035]

[0036] where, is the user load demand in microgrid at the -th collection;

[0037] The output of distributed power sources in multi - microgrids is expressed as follows:

[0038]

[0039] where, is the total output of distributed power sources in microgrid i at the -th collection;

[0040] The operating status of the power transmission channels between multi - microgrids is expressed as follows:

[0041]

[0042] where, is whether the transmission line in the direction from microgrid to microgrid is operating at the -th collection. If it is in the operating state, then ; if it is not in the operating state, then .

[0043] Furthermore, the construction of the multi-microgrid control target model in step S3 includes:

[0044] The energy interaction target is set to minimize the energy flow information entropy of each microgrid node, that is:

[0045]

[0046] At the th acquisition moment, the expression of the energy flow information entropy for the microgrid

[0047]

[0048]

[0049]

[0050] Among them, is the absolute value function, is the energy flow probability, indicating that the energy interaction amount between the microgrid and the microgrid accounts for the proportion of the total interaction power of the microgrid , is the logarithmic control parameter.

[0051] Furthermore, in the multi-microgrid control target model of step S3, the power balance constraint is used to ensure the stable operation of the entire system and the matching of power sources and loads. The expression of the power balance constraint is:

[0052]

[0053] Furthermore, the construction of the reinforcement learning network in step S4 includes:

[0054] A1: Construct the multi-microgrid system state space , and the mathematical expression is as follows:

[0055]

[0056] Among them, represents the iteration round, represents the set of microgrid battery states, , is the available state percentage of the remaining capacity of the energy storage in the microgrid at the th iteration; represents the set of interactive power between microgrids, , is the interactive power between microgrids at the th iteration The power output sent to the microgrid ; represents the set of user loads in the microgrid, , is the user load demand in the microgrid at the k-th iteration; represents the set of distributed power generation outputs in the microgrid, , is the output of the distributed power sources in the microgrid at the k-th iteration, including fossil energy and new energy units; represents the set of operating states of the transmission channels between microgrids, , is whether the transmission line in the direction of the microgrid to the microgrid is operating at the k-th iteration. If it is in the operating state, then , and if it is not in the operating state, then ;

[0057] A2: Construct the action space , and the mathematical expression is as follows:

[0058]

[0059] where represents the set of interactive power adjustment plans for the microgrid at the next iteration, , is the power output plan sent from the microgrid to the microgrid at the k-th iteration; represents the set of output plans for the distributed generation devices in each microgrid at the next iteration, , is the output plan of the distributed power sources in the microgrid at the (k + 1)-th iteration; represents the charge and discharge power of the energy storage in each microgrid at the next iteration, , is the charge and discharge power of the energy storage in the microgrid at the k-th iteration. The power is positive during charging and negative during discharging.

[0060] A3: Construct the reward function , and map the multi-microgrid control constraint model to the reward function.

[0061] Furthermore, the reward function The construction includes:

[0062] A3-1: Construct the energy flow information entropy reward term :

[0063]

[0064]

[0065]

[0066] Among them, is the energy flow probability, indicating the between the microgrid The proportion of the energy interaction amount between the microgrid total interaction power ; is the logarithmic control parameter;

[0067] A3-2: Construct the power balance penalty term :

[0068]

[0069]

[0070] Among them, is the th iteration, the total unbalanced power caused by the fluctuations of distributed energy and load output in the microgrid during time; is the user load demand;

[0071] A3-3: Construct the energy storage system capacity reward term :

[0072]

[0073]

[0074]

[0075] This reward term is to avoid the charge and discharge plan exceeding the capacity limit of the energy storage. Among them, the upper bound of the energy storage elasticity range of the microgrid at the th iteration is the maximum charging capacity , and the lower bound of the energy storage elasticity range is the maximum discharge capacity , and they are both determined by the energy storage charge state and the energy storage capacity;

[0076] A3-4: Construct the reward function belonging to the microgrid :

[0077]

[0078] Among them, is the weight coefficient.

[0079] The present invention also provides a multi-microgrid distributed energy scheduling system based on energy flow information entropy, including:

[0080] A multi-microgrid grid structure information acquisition module, which is used to obtain the grid structure information of the scheduling target multi-microgrid area, including the number of microgrid nodes in the target multi-microgrid area, the system node impedance of the target multi-microgrid area, the maximum capacity of the power transmission channels between each microgrid node in the target multi-microgrid area, and the maximum energy storage capacity of each microgrid node in the target multi-microgrid area;

[0081] A multi-microgrid operation information acquisition module, which is used to collect the operation information of the multi-microgrid in real time, including the energy storage power state of each microgrid, the interactive power between microgrids, the load demand of microgrid users, the output of distributed power sources in the multi-microgrid, and the operation state of the power transmission channels between multi-microgrids;

[0082] A multi-microgrid control target model construction module, which is used to construct scheduling control targets and related constraints;

[0083] A reinforcement learning agent module for outputting scheduling strategies, which is used to establish a reinforcement learning framework, train the reinforcement learning agent according to historical data, and output scheduling strategies according to real-time collected data;

[0084] A multi-microgrid system scheduling strategy execution module, which is used to execute the energy scheduling strategy of the multi-microgrid system.

[0085] Beneficial effects: Compared with the prior art, the present invention introduces the concept of information entropy into the scheduling control of the microgrid in the power system, and can achieve three-level balance at the global level of the interior of the microgrid, between adjacent microgrids, and among microgrid groups; by minimizing the energy flow information entropy, the orderliness of the system is improved, the action frequency of equipment in the multi-microgrid system is reduced, the scheduling demand is reduced, and the local consumption efficiency of distributed energy is improved; the capacity reward item of the energy storage system considers the capacity limitation of the power transmission channel, and the probability of network congestion is reduced. Brief Description of the Drawings

[0086] Figure 1 is a schematic flow chart of the method of the present invention;

[0087] Figure 2 is a schematic structural diagram of the system of the present invention. Detailed Embodiments

[0088] The present invention will be further illustrated below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. After reading the present invention, various equivalent modifications made by those skilled in the art to the present invention fall within the scope defined by the appended claims of this application.

[0089] Embodiment 1:

[0090] As Figure 1 shown, this embodiment provides a multi-microgrid distributed energy scheduling method based on energy flow information entropy, including the following steps:

[0091] S1: According to the design drawing information, collect and obtain the multi-microgrid network structure information;

[0092] The multi-microgrid network structure information includes: the number of microgrid nodes in the target multi-microgrid area 、the system node impedance matrix of the target multi-microgrid area 、the electrical distance matrix between each microgrid node in the target multi-microgrid area 、the maximum capacity matrix of the power transmission channels between each microgrid node in the target multi-microgrid area 、the maximum energy storage capacity matrix of each microgrid node in the target multi-microgrid area .

[0093] The expression of the system node impedance matrix of the target multi-microgrid area is:

[0094]

[0095] In the matrix the element in the th row and th column is the system equivalent impedance seen from the port;

[0096] According to formula (1), the electrical distance matrix between each microgrid node in the target multi-microgrid area is obtained, and the expression is:

[0097]

[0098] In the matrix the meaning of the element in the th row and th column is the electrical distance from microgrid node to microgrid node which is used to characterize the tightness of the electrical connection between two microgrids. When ;

[0099] The maximum capacity matrix of the power transmission channels among the microgrid nodes in the target multi - microgrid area The expression is as follows:

[0100]

[0101] In the matrix among the element in the row and column is the maximum power transmission capacity of the power transmission channel from the microgrid node to the microgrid node ;

[0102] The maximum energy storage capacity matrix of each microgrid node in the target multi - microgrid area The expression is as follows:

[0103]

[0104] Among them, represents the maximum energy storage capacity of the microgrid node

[0105] S2: Through the sensors installed at the energy input and output ports, energy storage, users, and distributed power sources of each microgrid node, collect data in real time, and the collection frequency should be equal to the control frequency , let be the number of collections;

[0106] The multi - microgrid operation information includes the energy storage power status of each microgrid , the interactive power between microgrids , the load demand of multi - microgrid users , the output of multi - microgrid distributed power sources , the operation status of the power transmission channels between multi - microgrids .

[0107] The expression of the energy storage power status of each microgrid is as follows:

[0108]

[0109] Among them, is the percentage of the remaining capacity available state of the energy storage in the microgrid at the th collection;

[0110] The expression of the interactive power between microgrids is as follows:​

[0111]

[0112] Among them, is the power output from the microgrid to the microgrid at the th collection; to the microgrid power output.

[0113] The expression of the multi - microgrid user load demand is as follows:

[0114]

[0115] Among them, is the user load demand in the microgrid at the th collection; in the microgrid.

[0116] The expression of the multi - microgrid distributed power output is as follows:

[0117]

[0118] Among them, is the total output of the distributed power sources in microgrid i at the th collection;

[0119] The expression of the operation state of the transmission channels between multi - microgrids is as follows:

[0120]

[0121] Among them, is whether the transmission line in the direction from microgrid to microgrid to microgrid is operating at the th collection. If it is in the operating state, then .

[0122] S3: Based on the multi - microgrid operation information, introduce the energy flow information entropy to construct a multi - microgrid control target model;

[0123] The control objective is to ensure the stable operation of the entire system and achieve source - load matching while minimizing the energy interaction between multi - microgrids, ultimately achieving the effect of "local autonomy" of multi - microgrid energy.

[0124] The construction of the multi - microgrid control target model includes:

[0125] The energy interaction target is set to minimize the energy flow information entropy of each microgrid node, that is:

[0126]

[0127] At the th collection moment, the expression of the energy flow information entropy for the microgrid is as follows:

[0128]

[0129]

[0130]

[0131] Among them, is the absolute value function; is the energy flow probability, which represents the proportion of the energy interaction amount between the microgrid and the microgrid in the total interaction power of the microgrid ; is the logarithmic control parameter, and its value affects the training step and convergence speed of the reinforcement learning agent. In this embodiment, a fixed value of 10 is taken.

[0132] For a pure output microgrid with only one transmission line, the energy flow probability is zero, that is, the energy flow information entropy is zero. The purpose of defining the energy flow information entropy is not to prevent the energy interaction between microgrids, but to perform reasonable energy interaction to ensure the local balance of electricity.

[0133] In the multi-microgrid control target model, the power balance constraint is used to ensure the stable operation of the entire system and the matching of sources and loads. The expression of the power balance constraint is:

[0134]

[0135] S4: According to the multi-microgrid control target model, construct a reinforcement learning network for the output scheduling strategy;

[0136] In this embodiment, a multi-agent deep Q network is adopted, with a DNN (Deep Neural Neworks, DNN) network as the neural network kernel. In order to achieve a distributed energy scheduling effect, a reinforcement learning agent is set at each microgrid node. The agent perceives information from the state space, calculates the reward function, and adjusts the control operations of each agent in the action space by minimizing the reward function.

[0137] The construction of the reinforcement learning network includes:

[0138] A1: Construct the state space of the multi-microgrid system , and the mathematical expression is as follows:

[0139]

[0140] Among them, represents the iteration round, represents the set of microgrid battery states, , is the available state percentage of the remaining capacity of energy storage in the microgrid at the th iteration; represents the set of interactive power between microgrids, , is the power output plan from the microgrid to the microgrid to the microgrid at the th iteration; represents the set of user loads in the microgrid, is the user load demand in the microgrid at the th iteration; represents the set of distributed power generation outputs in the microgrid, , is the output of distributed power generation in the microgrid at the th iteration, including fossil energy and new energy units; represents the set of operating states of the transmission channels between microgrids, , is whether the transmission line in the direction from the microgrid to the microgrid to the microgrid is operating at the th iteration. If it is in the operating state, then ; It should be noted that when the reinforcement learning agent is training, the historical operation data of the target multi - microgrid system is used in the state space, rather than real - time collected data;

[0141] A2: Construct the action space , and the mathematical expression is as follows:

[0142]

[0143] Among them, represents the set of interactive power adjustment plans for the microgrid at the next iteration, , is the power output plan from the microgrid to the microgrid to the microgrid at the th iteration; represents the set of output plans for distributed power generation devices in each microgrid at the next iteration, is the microgrid at the k+1th iteration Output plan of distributed power generation; Indicates the charging and discharging power of each microgrid energy storage in the next iteration, , For the Microgrid The charging and discharging power of the medium energy storage is positive when charging and negative when discharging.

[0144] A3: Constructing the Reward Function , mapping the multi-microgrid control constraint model into a reward function;

[0145] The core role of the reward function is to guide the agent to achieve the preset goal. The agent performs actions in a specific state according to the current strategy, and the environment generates state transitions based on the action feedback and simultaneously provides an immediate reward signal. The essence of this reward signal is to build an incentive mechanism, and the goal is to maximize the cumulative sum of these reward signals. In the present invention, each agent has its own reward function.

[0146] Reward Function The build includes:

[0147] A3-1: Constructing Energy Flow Information Entropy Bonus :

[0148]

[0149]

[0150]

[0151] in, is the energy flow probability, representing the microgrid With Microgrid The energy interaction between microgrids accounts for Total interactive power The proportion of is a logarithmic control parameter, and in this embodiment, a fixed value of 10 is taken;

[0152] A3-2: Constructing power balance penalty items :

[0153]

[0154]

[0155] in, For the In the first iteration, the microgrid exist The total unbalanced power caused by the fluctuations of distributed energy and load output within a certain period of time, using the user load demand The penalty term is corrected;

[0156] A3-3: Construct the energy storage system capacity reward term :

[0157]

[0158]

[0159]

[0160] This reward term is to avoid the charge and discharge plan exceeding the capacity limit of the energy storage. Among them, at the th iteration, the upper bound of the energy storage elasticity range of the microgrid is the maximum charging capacity The lower bound of the energy storage elasticity range is the maximum discharge capacity Both of them are determined by the state of charge of the energy storage and the energy storage capacity;

[0161] A3-4: Construct the reward function belonging to the microgrid :

[0162]

[0163] Among them, is the weight coefficient, which is set to in this embodiment.

[0164] S5: Train the reinforcement learning network;

[0165] S6: Input the real-time collected multi-microgrid operation information into the trained reinforcement learning network and output the scheduling strategy;

[0166] S7: Execute the multi-microgrid system scheduling strategy.

[0167] Each microgrid node agent takes actions according to the scheduling strategy generated in step S6, including acting on the input and output port controllers of the microgrid node to adjust the on-off state of the power transmission channels between microgrids at the next moment; acting on the distributed generation device controllers within the microgrid to adjust the output plan of each microgrid's distributed generation device at the next moment; acting on the energy storage system controller to set the energy storage charge and discharge power of each microgrid at the next moment.

[0168] The multi - microgrid system is composed of multiple microgrids, which are connected by transmission tie - lines and operate at the same voltage level. During the actual operation of the multi - microgrid, due to the volatility of the output of distributed generation devices within it, the magnitude and even direction of the line power flow change frequently, which poses challenges to the stable and orderly operation of the multi - microgrid system. In order to guide the multi - microgrid system to evolve towards an orderly state, the method provided in this embodiment introduces the concept of information entropy into the dispatching control of the microgrid in the power system. Information entropy is the expected value of the amount of information of all possible events and measures the uncertainty of the system. The larger the information entropy, the more uncertain and complex the system is. For the multi - microgrid system, the larger the information entropy of the interactive energy flow, the more difficult it is for the system to achieve local energy autonomy. The more orderly a system is, the lower the information entropy. The method of the present invention uses a reinforcement learning algorithm to achieve the minimum control objective of the energy flow information entropy, and the ultimate goal is to make the multi - microgrid system an orderly system.

[0169] Based on the multi - microgrid network structure information obtained by collection, in the target multi - microgrid dispatching control area, the method of the present invention uses the historical multi - microgrid operation information for reinforcement learning training. Under the guidance of the set reward function, it interacts with the environmental state to learn the optimal strategy to train the agent network, where the dispatching control objectives and constraints: the energy flow information entropy, power balance constraint, and energy storage system capacity constraint are reflected by the reward function. After the training is completed, the real - time collected multi - microgrid operation information is input into the reinforcement learning agent, and the action information of the multi - microgrid system at the next moment is output, including the on - off state of the transmission channels between microgrids adjusted by the multi - microgrid system at the next moment; the output of the distributed generation devices within each sub - microgrid node and the charge - discharge power of the energy storage at the next moment are set at each sub - microgrid node.

[0170] Embodiment 2:

[0171] Based on the method of Embodiment 1, as Figure 2 shown, this embodiment provides a multi - microgrid distributed energy scheduling system based on energy flow information entropy, including:

[0172] A multi - microgrid network structure information acquisition module, which is used to obtain the network structure information of the scheduling target multi - microgrid area, including the number of microgrid nodes in the target multi - microgrid area, the system node impedance of the target multi - microgrid area, the maximum capacity of the transmission channels between each microgrid node in the target multi - microgrid area, and the maximum energy storage capacity of each microgrid node in the target multi - microgrid area;

[0173] A multi - microgrid operation information acquisition module, which is used to collect the operation information of the multi - microgrid in real - time, including the energy storage power state of each microgrid, the interactive power between microgrids, the load demand of microgrid users, the output of distributed power sources in the multi - microgrid, and the operation state of the transmission channels between multi - microgrids;

[0174] A multi - microgrid control target model construction module, which is used to construct the dispatching control objectives and related constraints;

[0175] The reinforcement learning agent module for the output scheduling strategy is used to establish a reinforcement learning framework, train the reinforcement learning agent according to historical data, and output the scheduling strategy according to the real-time collected data;

[0176] The multi-microgrid system scheduling strategy execution module is used to execute the energy scheduling strategy of the multi-microgrid system.

Claims

1. A multi-microgrid distributed energy scheduling method based on energy flow information entropy, characterized in that: The steps include: S1: Collecting multi-microgrid grid structure information; S2: Based on the multi-microgrid grid structure information, collect multi-microgrid operation information; S3: Based on the multi-microgrid operation information, the energy flow information entropy is introduced to construct the multi-microgrid control target model; S4: Based on the multi-microgrid control target model, a reinforcement learning network for output scheduling strategy is constructed; S5: training the reinforcement learning network; S6: Input the real-time collected multi-microgrid operation information into the trained reinforcement learning network and output the scheduling strategy; S7: Execute the multi-microgrid system scheduling strategy.

2. According to claim 1, a multi-microgrid distributed energy scheduling method based on energy flow information entropy is characterized in that: The multi-microgrid grid structure information in step S1 includes: the number of microgrid nodes in the target multi-microgrid area , the system node impedance matrix of the target multi-microgrid area , the electrical distance matrix between microgrid nodes in the target multi-microgrid area , the maximum capacity matrix of the transmission channels between microgrid nodes in the target multi-microgrid area , the maximum energy storage capacity matrix of each microgrid node in the target multi-microgrid area .

3. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 2 is characterized in that: The system node impedance matrix of the target multi-microgrid area in step S1 The expression is: ； In the matrix middle OK Elements of a column For , The system equivalent impedance seen by the port; According to formula (1), the electrical distance matrix between microgrid nodes in the target multi-microgrid area is obtained: , the expression is: ； In the matrix middle OK Elements of a column The meaning of microgrid node To microgrid node The electrical distance is used to characterize the closeness of the electrical connection between two microgrids. ; The maximum capacity matrix of the transmission channels between microgrid nodes in the target multi-microgrid area The expression is: ； In the matrix middle OK Elements of a column Microgrid Node To microgrid node The maximum transmission capacity of the transmission channel, is a symmetric matrix, that is ; Maximum energy storage capacity matrix of each microgrid node in the target multi-microgrid area The expression is: ； in, Represents microgrid node Maximum energy storage capacity.

4. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 3 is characterized in that: The multi-microgrid operation information in step S2 includes the energy storage power status of each microgrid 、Interactive electricity between microgrids 、Multi-microgrid user load requirements 、Multi-microgrid distributed power output , Operation status of transmission channels between multiple microgrids .

5. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 4 is characterized in that: The energy storage state of each microgrid in step S2 The expression is as follows: ； in, For the Microgrid The percentage of available capacity remaining in the energy storage system; Interactive electricity between microgrids The expression is as follows: ； in, For the Microgrid Micronet The amount of electricity delivered; Multi-microgrid user load requirements The expression is as follows: ； in, For the Microgrid Medium user load demand; Multi-microgrid distributed power output The expression is as follows: ； in, For the The total output of distributed generation in microgrid i at the time of collection; Operation status of transmission channels between multiple microgrids The expression is as follows: ； in, For the The transmission direction of the second collection is microgrid Micronet Whether the transmission line in the direction is running, if it is running, , if not in operation .

6. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 5 is characterized in that: The construction of the multi-microgrid control target model in step S3 includes: The energy interaction target is set to minimize the energy flow information entropy of each microgrid node, that is: ； exist Microgrid The expression of energy flow information entropy is: ； ； ； in, To find the absolute value function, is the energy flow probability, representing the microgrid With Microgrid The energy interaction between microgrids accounts for Total interactive power The proportion of is the logarithmic control parameter.

7. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 6 is characterized in that: In the multi-microgrid control target model of step S3, the power balance constraint is used to ensure the stable operation of the entire system and the matching of source and load. The expression of the power balance constraint is: 。 8. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 7 is characterized in that: The construction of the reinforcement learning network in step S4 includes: A1: Constructing the state space of multi-microgrid system , the mathematical expression is as follows: ； in, represents the iteration round, Represents the microgrid battery status set, , For the Microgrid The percentage of available capacity remaining in the energy storage system; represents the interactive electricity collection between microgrids, , For the Microgrid Micronet The amount of electricity delivered; represents the user load set in the microgrid, , For the Microgrid Medium user load demand; represents the output of distributed generation in the microgrid, , For the Microgrid The output of distributed power sources, including fossil energy and new energy units; represents the set of operating states of the transmission channels between microgrids, , For the The transmission direction of the second iteration is the microgrid Micronet Whether the transmission line in the direction is running, if it is running, , if not in operation ; A2: Constructing the action space , the mathematical expression is as follows: ； in, represents the interactive power adjustment plan set of the microgrid in the next iteration, , For the Microgrid Micronet The power delivery plan; represents the output plan set of each microgrid distributed generation device in the next iteration, , For the Microgrid Output plan of distributed generation; Indicates the charging and discharging power of each microgrid energy storage in the next iteration, , For the Microgrid The charging and discharging power of the medium energy storage; A3: Constructing the Reward Function , mapping the multi-microgrid control constraint model into a reward function.

9. The multi-microgrid distributed energy scheduling method based on energy flow information entropy according to claim 8 is characterized in that: The reward function in step A3 The build includes: A3-1: Constructing Energy Flow Information Entropy Bonus : ； ； ； in, is the energy flow probability, representing the microgrid With Microgrid The energy interaction between microgrids accounts for Total interactive power The proportion of is the logarithmic control parameter; A3-2: Constructing power balance penalty items : ； ； in, For the In the first iteration, the microgrid exist The unbalanced total power caused by the fluctuation of distributed energy and load output within a certain period of time; For user load demand; A3-3: Building Energy Storage System Capacity Incentive : ； ； ； Among them, Microgrid The upper limit of the energy storage elastic range is the maximum charging capacity , the lower limit of the energy storage elastic range is the maximum discharge capacity ; A3-4: Build a microgrid The reward function is: ； in, is the weight coefficient.

10. A multi-microgrid distributed energy dispatching system based on energy flow information entropy, characterized in that: include: A multi-microgrid grid structure information acquisition module is used to obtain grid structure information of the target multi-microgrid area, including the number of microgrid nodes in the target multi-microgrid area, the system node impedance of the target multi-microgrid area, the maximum capacity of the transmission channel between the microgrid nodes in the target multi-microgrid area, and the maximum energy storage capacity of each microgrid node in the target multi-microgrid area; The multi-microgrid operation information collection module is used to collect the operation information of multiple microgrids in real time, including the energy storage power status of each microgrid, the interactive power between microgrids, Microgrid user load demand, multi-microgrid distributed power output, and the operating status of transmission channels between microgrids; Multi-microgrid control target model building module, used to build dispatch control targets and related constraints; The reinforcement learning agent module that outputs the scheduling strategy is used to establish a reinforcement learning framework, train the reinforcement learning agent based on historical data, and output the scheduling strategy based on real-time collected data; The multi-microgrid system scheduling strategy execution module is used to execute the multi-microgrid system energy scheduling strategy.

Images