Multi-agent cooperative micro-grid cluster optimization method and device

Abstract

The invention relates to the technical field of intelligent power grid regulation and control, in particular to a multi-agent cooperation micro-grid cluster optimization method and device, and the method comprises the steps: constructing a power coordination control strategy based on distributed consistency based on a system architecture of a micro-grid cluster, so as to determine a distribution proportion; a hierarchical control system is designed while the power coordination control strategy is executed, so that a corresponding economic optimization dispatching instruction is generated according to the current frequency modulation stage; according to the method, the lowest system operation cost and the highest renewable energy consumption rate are taken as targets, an optimal power instruction is generated, a preset event trigger mechanism is utilized to perform communication, and the purpose of multi-agent cooperative micro-grid cluster optimization is achieved. Therefore, the problem that communication efficiency and control performance are difficult to consider at the same time under complex conditions due to the fact that interaction between micro-grids is carried out by adopting a communication mechanism with a fixed period in related technologies and the capability of adaptive adjustment on network topology changes and communication conditions is lacked is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of smart grid regulation, and in particular to a multi-agent cooperative micro-grid cluster optimization method and device. BACKGROUND

[0002] In related technologies, a micro-grid cluster is a cluster system composed of multiple distributed micro-grids. Due to differences in geographical location, load characteristics and renewable energy access of each micro-grid, while ensuring the autonomy of each micro-grid, global power balance and economic operation are required. Therefore, a distributed optimization and coordination method is usually used for power control. This method can improve scheduling efficiency, enhance system robustness, and to some extent, protect data privacy.

[0003] However, in related technologies, the distributed power control method often uses a fixed cycle communication mechanism for interaction between micro-grids, lacks adaptive adjustment capability for network topology changes and communication conditions, and in complex situations where communication resources are limited or network topology changes dynamically, it is difficult to balance communication efficiency and control performance at the same time, resulting in a decline in optimization performance. Therefore, it is urgent to solve the problem. SUMMARY

[0004] The present application provides a multi-agent cooperative micro-grid cluster optimization method and device to solve the problem that in related technologies, the distributed power control method often uses a fixed cycle communication mechanism for interaction between micro-grids, lacks adaptive adjustment capability for network topology changes and communication conditions, and in complex situations, it is difficult to balance communication efficiency and control performance at the same time, resulting in a decline in optimization performance.

[0005] The first aspect of the present application provides a multi-agent cooperative micro-grid cluster optimization method, comprising the following steps: establishing a system architecture of a micro-grid cluster; based on the system architecture, constructing a power coordination control strategy based on distributed consistency, so that the power coordination control strategy is used to determine the distribution ratio based on power shortage or surplus in the micro-grid cluster; while executing the power coordination control strategy, designing a hierarchical control system to generate corresponding economic optimization scheduling instructions according to the current frequency modulation stage, and in response to the economic optimization scheduling instructions, generating optimal power instructions with the lowest system operating cost and the highest renewable energy consumption rate as the target, and using a preset event trigger mechanism for communication, achieving the purpose of multi-agent cooperative micro-grid cluster optimization.

[0006] By means of the above technical means, the embodiment of the application can effectively improve the operation performance of the micro-grid cluster containing a high proportion of renewable energy by using the coordinated architecture of system architecture modeling, distributed consistency coordination, hierarchical control system, model predictive rolling optimization and event-triggered communication, so as to cope with the uncertainty of distributed power output and the randomness of load demand, and enhance the stability and economy of the system.

[0007] Optionally, in an embodiment of the application, the power coordination control strategy based on distributed consistency is constructed, and the power coordination control strategy is used to determine the distribution ratio based on power shortage or surplus in the micro-grid cluster, comprising: after determining the power imbalance amount, the power imbalance amount estimation global imbalance amount of each sub-micro-grid of the system architecture is obtained, and the power distribution ratio is calculated; the power adjustment instruction of each sub-micro-grid is generated according to the global imbalance amount and the power distribution ratio, wherein the distributed convergence detection based on Lyapunov is performed until the preset convergence condition is met, and the iteration is stopped, so that after the collaborative optimization of each sub-micro-grid by using the preset value function, the collaborative distribution of power shortage or surplus is realized.

[0008] By means of the above technical means, the embodiment of the application can generate the power adjustment instruction of each sub-micro-grid according to the global imbalance amount and the power distribution ratio, realize the collaborative distribution of power shortage or surplus, maintain the global power balance of the micro-grid cluster under the premise of ensuring the autonomy of each sub-micro-grid, improve the adaptability of the system to the renewable energy output fluctuation and load demand randomness, and thus enhance the stability and robustness of the system.

[0009] Optionally, in an embodiment of the application, the power adjustment amount calculation formula of the power adjustment instruction is:

[0010]

[0011] wherein, represents the power adjustment amount that the i th sub-micro-grid needs to undertake; λ is a differential gain coefficient; represents the global power imbalance amount estimation value of the i th sub-micro-grid obtained by the consistency algorithm; represents the power distribution ratio of the i th sub-micro-grid;

[0012] The preset convergence condition is:

[0013] and

[0014] wherein, ε V is a Lyapunov function change threshold; ε δ is a maximum consistency deviation threshold; N represents the total number of sub-micro-grids in the micro-grid cluster.

[0015] The generation formula of the power adjustment instruction is:

[0016]

[0017] wherein, represents a policy network of the i-th agent, θ i represents a parameter; s i represents local state information of the i-th agent; φ ij is a coordination weight coefficient between neighbor agents; is a neighbor set of the i-th agent.

[0018] Through the above technical means, the power adjustment amount of the power adjustment instruction is quantified, and the power increase / decrease amplitude of each sub-microgrid in each adjustment period can be determined; according to the generation formula of the power adjustment instruction, in combination with the global imbalance and the power distribution ratio, the power adjustment instruction of each sub-microgrid can be dynamically generated; meanwhile, the corresponding convergence condition can be set to determine whether the power adjustment iteration meets the system balance and performance optimization requirements, so as to ensure that the microgrid cluster realizes stable and efficient global power adjustment and optimization under the cooperation of multiple agents.

[0019] Optionally, in an embodiment of the present application, the hierarchical control system is designed to generate corresponding economic optimization scheduling instructions according to the current frequency modulation stage, which comprises: adopting an improved active power-frequency droop characteristic for primary power distribution to realize the design of the droop control layer in the primary frequency modulation stage; adopting an adaptive PI (Proportional-Integral Controller) controller with anti-saturation integration to eliminate the steady-state frequency and voltage deviation to realize the design of the voltage / frequency recovery control in the secondary frequency modulation stage; establishing an optimization problem with the minimum operation cost as the target to realize the design of the economic optimization scheduling layer in the tertiary frequency modulation stage.

[0020] Through the above technical means, the power output of the microgrid or the power system can be hierarchically adjusted according to the droop control of the primary frequency modulation, the voltage / frequency recovery of the secondary frequency modulation, and the economic optimization scheduling of the tertiary frequency modulation, so as to realize the rapid response to load fluctuation, recover the system voltage and frequency stability, and optimize the economic operation on this basis, thereby improving the stability, reliability and operation efficiency of the system.

[0021] Optionally, in an embodiment of the present application, the generating the optimal power instruction with the lowest system operation cost and the highest renewable energy consumption rate as the target comprises: establishing a multi-objective optimization function and defining the system operation cost and the renewable energy consumption rate to establish a prediction model and construct a rolling optimization framework; and generating the optimal power instruction based on the prediction model and the rolling optimization framework.

[0022] Through the above technical means, the embodiment of the present application constructs the rolling optimization framework of the system operation cost and the renewable energy consumption rate, and then generates the optimal power instruction, so that the global power balance and the economic and efficient operation of the micro-grid cluster can be realized under the premise of ensuring the autonomy of each sub-micro-grid, the renewable energy consumption rate is improved, the system operation cost is reduced, and the adaptability of the system to load fluctuation and distributed power output uncertainty is enhanced.

[0023] Optionally, in an embodiment of the present application, the utilizing the preset event triggering mechanism to communicate and realizing the optimization purpose of the multi-agent collaborative micro-grid cluster comprises: defining a communication topology structure and designing an event triggering condition to establish the preset event triggering mechanism; defining a communication quality index and designing an adaptive triggering threshold based on reinforcement learning based on the preset event triggering mechanism to establish a communication reliability model; and designing a fault-tolerant communication protocol to count communication traffic and evaluate the reliability index of communication by using the communication reliability model.

[0024] Through the above technical means, the embodiment of the present application establishes the preset event triggering mechanism and designs the fault-tolerant communication protocol, so that efficient and reliable information interaction can be realized in the micro-grid cluster, each sub-micro-grid can quickly respond when a sudden event or communication anomaly occurs, the timely delivery and execution of the power adjustment instruction are ensured, and thus the stability, robustness and adaptability to renewable energy output fluctuation of the system are improved.

[0025] The second aspect embodiment of the present application provides a multi-agent collaborative micro-grid cluster optimization device, comprising: an establishment module configured to establish a system architecture of a micro-grid cluster; a determination module configured to construct a power coordination control strategy based on distributed consistency based on the system architecture, so that the micro-grid cluster is distributed based on the power coordination control strategy. The distribution ratio is determined based on power shortage or surplus; an optimization module configured to generate corresponding economic optimization scheduling instructions according to the current frequency modulation stage while executing the power coordination control strategy, and in response to the economic optimization scheduling instructions, generate an optimal power instruction with the lowest system operation cost and the highest renewable energy consumption rate as the target, and utilize a preset event triggering mechanism to communicate and realize the optimization purpose of the multi-agent collaborative micro-grid cluster.

[0026] Through the above technical means, the embodiments of this application adopt a collaborative architecture of system architecture modeling, distributed consistency coordination, hierarchical control system, model prediction rolling optimization and event-triggered communication, which can effectively improve the operating performance of microgrid clusters with a high proportion of renewable energy, thereby coping with the uncertainty of distributed power output and the randomness of load demand, and enhancing the stability and economy of the system.

[0027] Optionally, in one embodiment of this application, the determining module includes: a calculation unit, configured to, after determining the power imbalance, obtain an estimate of the global imbalance of each sub-microgrid in the system architecture and calculate the power allocation ratio; and a first generation unit, configured to generate power adjustment instructions for each sub-microgrid based on the global imbalance and the power allocation ratio, wherein distributed convergence detection based on Lyapunov is performed until a preset convergence condition is met, and iteration is stopped, so that after each sub-microgrid performs collaborative optimization using a preset value function, collaborative allocation of power deficit or surplus is achieved.

[0028] Through the above technical means, the embodiments of this application can generate power adjustment instructions for each sub-microgrid according to the global imbalance and power allocation ratio, realize the coordinated allocation of power deficit or surplus, maintain the global power balance of the microgrid cluster while ensuring the autonomy of each sub-microgrid, improve the system's adaptability to the fluctuation of renewable energy output and the randomness of load demand, thereby enhancing the system's stability and robustness.

[0029] Optionally, in one embodiment of this application, the power adjustment amount of the power adjustment command is calculated using the following formula:

[0030]

[0031] in, The value represents the power regulation amount that the i-th sub-microgrid needs to undertake; λ is the differential gain coefficient. This represents the estimated global power imbalance value obtained by the consensus algorithm for the i-th sub-microgrid; This represents the power allocation ratio of the i-th sub-microgrid;

[0032] The preset convergence condition is:

[0033] and

[0034] Where, ε V ε is the threshold for the change in the Lyapunov function; δ The maximum consistency deviation threshold; N represents the total number of sub-microgrids in the microgrid cluster;

[0035] The formula for generating the power adjustment command is:

[0036]

[0037] in, Let θ represent the policy network of the i-th agent. i Indicates a parameter; s i φ represents the local state information of the i-th agent; ij The collaborative weighting coefficients among neighboring intelligent agents; Let be the set of neighbors of the i-th agent.

[0038] Through the above technical means, the embodiments of this application quantify the power adjustment amount of the power adjustment command, which can determine the power increase or decrease of each sub-microgrid in each adjustment cycle; based on the power adjustment command generation formula, combined with the global imbalance and power allocation ratio, the power adjustment command of each sub-microgrid can be dynamically generated; at the same time, corresponding convergence conditions are set, which can be used to determine whether the power adjustment iteration meets the system balance and performance optimization requirements, thereby ensuring that the microgrid cluster achieves stable and efficient global power regulation and optimization under the cooperation of multiple agents.

[0039] Optionally, in one embodiment of this application, the optimization module includes: a first design unit for performing primary power allocation using improved active power-frequency droop characteristics to achieve droop control layer design in the primary frequency regulation stage; a second design unit for eliminating steady-state frequency and voltage deviations using an adaptive PI controller with anti-saturation integral to achieve voltage / frequency recovery control design in the secondary frequency regulation stage; and a third design unit for establishing an optimization problem with the goal of minimizing operating costs to achieve economic optimization scheduling layer design in the tertiary frequency regulation stage.

[0040] Through the above technical means, the embodiments of this application can perform hierarchical regulation of the power output of microgrids or power systems based on the droop control of primary frequency regulation, the voltage / frequency recovery of secondary frequency regulation, and the economic optimization scheduling of tertiary frequency regulation. This enables rapid response to load fluctuations, restoration of system voltage and frequency stability, and optimization of economic operation, thereby improving the stability, reliability, and operating efficiency of the system.

[0041] Optionally, in one embodiment of this application, the optimization module includes: a construction unit, used to establish a multi-objective optimization function and define system operating cost and renewable energy absorption rate to establish a prediction model and construct a rolling optimization framework; and a second generation unit, used to generate the optimal power command based on the prediction model and the rolling optimization framework.

[0042] Through the above technical means, the embodiments of this application construct a rolling optimization framework for system operating costs and renewable energy absorption rate, and then generate the optimal power command. Under the premise of ensuring the autonomy of each sub-microgrid, the global power balance and economical and efficient operation of the microgrid cluster can be achieved, while improving the renewable energy absorption rate, reducing system operating costs, and enhancing the system's adaptability to load fluctuations and uncertainties in distributed power output.

[0043] Optionally, in one embodiment of this application, the optimization module includes: a first establishment unit, configured to define a communication topology and design event triggering conditions to establish the preset event triggering mechanism; a second establishment unit, configured to define communication quality indicators based on the preset event triggering mechanism, design an adaptive triggering threshold based on reinforcement learning, and establish a communication reliability model; and an evaluation unit, configured to design a fault-tolerant communication protocol, statistically analyze communication traffic, and evaluate the communication reliability indicators using the communication reliability model.

[0044] Through the above technical means, the embodiments of this application establish a preset event triggering mechanism and design a fault-tolerant communication protocol, which can realize efficient and reliable information interaction in the microgrid cluster, enabling each sub-microgrid to respond quickly when a sudden event or communication anomaly occurs, ensuring the timely transmission and execution of power adjustment commands, thereby improving the system's stability, robustness and adaptability to fluctuations in renewable energy output.

[0045] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the multi-agent cooperative microgrid cluster optimization method as described in the above embodiments.

[0046] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described multi-agent cooperative microgrid cluster optimization method.

[0047] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, is used to implement the above-described multi-agent cooperative microgrid cluster optimization method.

[0048] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0049] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0050] Figure 1 This is a flowchart of a multi-agent cooperative microgrid cluster optimization method provided according to an embodiment of this application;

[0051] Figure 2 This is a schematic diagram illustrating the process of optimizing a microgrid cluster through multi-agent cooperation, according to one embodiment of this application.

[0052] Figure 3 This is a block diagram of a multi-agent cooperative microgrid cluster optimization device according to an embodiment of this application;

[0053] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0054] Figure label:

[0055] 10 - Microgrid cluster optimization device for multi-agent collaboration; 100 - Establishment module, 200 - Determination module, 300 - Optimization module; 401 - Memory, 402 - Processor, 403 - Communication interface. Detailed Implementation

[0056] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0057] The following description, with reference to the accompanying drawings, illustrates a multi-agent cooperative microgrid cluster optimization method and apparatus according to embodiments of this application. Addressing the technical problem mentioned in the background art, where distributed power control methods often employ fixed-period communication mechanisms for microgrid interaction, lacking adaptive adjustment capabilities to network topology changes and communication conditions, making it difficult to simultaneously balance communication efficiency and control performance under complex conditions, thus resulting in decreased optimization performance, this application provides a multi-agent cooperative microgrid cluster optimization method. In this method, a microgrid cluster architecture is first constructed. Then, a collaborative architecture combining distributed consensus coordination, hierarchical control systems, model prediction rolling optimization, and event-triggered communication is used to generate optimal power commands with the goals of minimizing system operating costs and maximizing renewable energy absorption, thereby achieving multi-agent... Collaborative microgrid cluster optimization can effectively address the uncertainty of distributed power output and the randomness of load demand. It coordinates the automatic proportional allocation of power deficits or surpluses among sub-microgrids, and simultaneously suppresses power disturbances, restores voltage and frequency, and coordinates economic dispatch to ensure system stability and high response speed through a hierarchical control system. Furthermore, the event-triggered communication mechanism can dynamically adjust trigger thresholds and intervals based on communication quality, reducing redundant communication traffic and improving system reliability and convergence. This provides an efficient, reliable, and economical power control solution for microgrid clusters with a high proportion of renewable energy, achieving stable system operation, efficient energy utilization, and multi-agent collaborative optimization. This solves the problem that related technologies often use fixed-period communication mechanisms for microgrid interaction, lacking adaptive adjustment capabilities to network topology changes and communication conditions, making it difficult to simultaneously achieve communication efficiency and control performance in complex situations.

[0058] Specifically, Figure 1 This is a flowchart illustrating a multi-agent cooperative microgrid cluster optimization method provided in an embodiment of this application.

[0059] like Figure 1 As shown, the multi-agent cooperative microgrid cluster optimization method includes the following steps:

[0060] In step S101, the system architecture of the microgrid cluster is established.

[0061] It can be explained that a microgrid cluster consists of multiple sub-microgrids, each of which may include, but is not limited to, distributed power sources, energy storage units, and loads. Distributed power sources can provide controllable active and reactive power to meet local load demands, energy storage units can regulate power balance and store energy through charging and discharging, and loads can consume power as time fluctuates. While operating autonomously, each sub-microgrid can achieve economic operation and stable regulation of the cluster as a whole through power exchange.

[0062] Specifically, embodiments of this application may include, but are not limited to, the following steps:

[0063] (1) Define the microgrid cluster topology.

[0064] In this embodiment, a microgrid cluster can be configured to contain N sub-microgrids, forming an undirected connected graph. in For a set of nodes, Let A be the edge set, representing the power interaction links between sub-microgrids; define the adjacency matrix A = [a ij ] N×N , where a ij =1 indicates that nodes i and j can exchange power; otherwise, a ij =0.

[0065] (2) Establish a distributed power source model.

[0066] In the embodiments of this application, the output model of distributed power sources (such as photovoltaics and wind turbines) in the i-th sub-microgrid can be expressed as:

[0067]

[0068] in, To contribute to distributed power sources To predict the output, Let α be the prediction error, representing a Gaussian distribution with a mean of zero. i The climbing power factor. This is the power compensation term for hill climbing.

[0069] (3) Establish an energy storage system model

[0070] As a concrete example, the charging and discharging power constraint of the energy storage system in the i-th sub-microgrid can be expressed as:

[0071]

[0072] in, This refers to the charging and discharging power. and These are the minimum charge / discharge power and the maximum charge / discharge power, respectively. For the energy storage capacity in the next moment; The current energy storage capacity is given by η, where Δt is the time interval and η is the energy storage capacity. c η d These are the charging and discharging efficiencies, respectively. This is the rated capacity.

[0073] (4) Establish a load model.

[0074] In the embodiments of this application, the load power of the i-th sub-microgrid can be expressed as:

[0075]

[0076] in, It is a rigid load; For adjustable loads, the following can be satisfied:

[0077]

[0078] in, This represents the total adjustable energy of the adjustable load in the i-th sub-microgrid during the scheduling cycle.

[0079] (5) Define power interaction constraints.

[0080] It can be noted that, in the embodiments of this application, the interaction power between sub-microgrids i and j It can satisfy:

[0081]

[0082] in, Let i be the set of nodes adjacent to sub-microgrid i; This represents the relationship between the i-th sub-microgrid and all its neighboring sub-microgrids. The power commutative algebraic sum; This represents the line power transmission limit; a ij Denotes the adjacency matrix, if a ij A value of 1 indicates that nodes i and j can exchange power; a value of 0 indicates that they cannot exchange power.

[0083] In step S102, based on the system architecture, a power coordination control strategy based on distributed consistency is constructed to utilize the power coordination control strategy to determine the allocation ratio within the microgrid cluster based on power deficit or surplus.

[0084] It can be explained that in a microgrid cluster, in order to ensure the power balance of the entire cluster, when a sub-microgrid experiences a power deficit (supply falling short of demand) or a power surplus (excess energy), these power gaps or excess power need to be rationally allocated among the microgrids. A power coordination control strategy based on distributed consensus can, through local communication between the sub-microgrids, enable each microgrid to automatically adjust its output or energy storage behavior according to a certain proportion (usually related to its capacity or preset weight), thereby achieving gradual power consistency and balance for the entire cluster without relying on a centralized control center.

[0085] Optionally, in one embodiment of this application, a power coordination control strategy based on distributed consensus is constructed to enable the allocation ratio within a microgrid cluster to be determined based on power deficit or surplus. This includes: after determining the power imbalance, obtaining the power imbalance of each sub-microgrid in the system architecture to estimate the global imbalance and calculating the power allocation ratio; generating power adjustment instructions for each sub-microgrid based on the global imbalance and the power allocation ratio, wherein distributed convergence detection based on Lyapunov is performed until a preset convergence condition is met, and iteration is stopped, so that after each sub-microgrid performs collaborative optimization using a preset value function, the collaborative allocation of power deficit or surplus is achieved.

[0086] Among them, the Lyapunov method can measure the degree of deviation of the system state from equilibrium by constructing a positive definite function and analyzing its changing trend over time. In the distributed power coordination of microgrids, this function can be used to represent the power imbalance of each sub-microgrid, and a local control law can be designed to make the Lyapunov function monotonically decrease, thereby ensuring that the entire cluster gradually converges under the condition of relying only on neighbor communication, achieving power consistency and stable system operation.

[0087] It can be explained that the preset convergence condition can be that the derivative of the Lyapunov function is less than zero, i.e., the system state decreases monotonically with time error; or it can be that the absolute value of the power imbalance of each sub-microgrid is less than a certain threshold, i.e., when the power difference is within the allowable range, the system is considered to have reached an approximately convergent state. The preset value function can be used to measure the quality of the system state or control strategy. It can be a power balance-related indicator, such as the power imbalance of the entire cluster or the sum of squares of the power deviations of each sub-microgrid, used to minimize the supply and demand difference; or it can be an economic operation indicator, such as the total generation cost, energy storage usage cost, or renewable energy utilization rate, used to optimize operating costs and energy efficiency while ensuring stability. It can be set by those skilled in the art according to the actual situation, and no specific restrictions are imposed here.

[0088] As one possible implementation, embodiments of this application may include the following steps:

[0089] (1) Define the power imbalance.

[0090] In this embodiment of the application, for the i-th sub-microgrid, the local power imbalance ΔP i (t) can be defined as:

[0091]

[0092] Wherein, ΔP i (t) represents the power deficit or power surplus of the i-th sub-microgrid at time t.

[0093] (2) Construct a consistency coordination algorithm.

[0094] In the embodiments of this application, each sub-microgrid exchanges power imbalance information through a distributed communication network. The global imbalance can be estimated using a consensus update protocol with a forgetting factor, which can be expressed as:

[0095]

[0096] in, This represents the estimated global average power imbalance of the i-th sub-microgrid at the k-th iteration; denoted as the estimated global average power imbalance of the j-th sub-microgrid in the k-th iteration; κ is the forgetting factor, which can be used to balance historical information with new information; is the time-varying communication weight; l is the local feedback gain coefficient; Let represent the set of sub-microgrids adjacent to the i-th sub-microgrid.

[0097] (3) Calculate the power distribution ratio.

[0098] Furthermore, each sub-microgrid can determine its power allocation ratio based on its adjustable capacity:

[0099]

[0100] in, Total adjustable capacity; ζ represents the reputation weight based on historical response performance; ζ represents the adjustment sensitivity index; and N represents the total number of sub-microgrids in the microgrid cluster.

[0101] (4) Generate power adjustment command.

[0102] As a specific example, the power adjustment amount of the i-th sub-microgrid root can be calculated as follows:

[0103]

[0104] in, λ represents the amount of power regulation that the i-th sub-microgrid needs to undertake; λ is the differential gain coefficient, used to suppress overshoot; This represents the estimated global power imbalance value obtained by the consensus algorithm for the i-th sub-microgrid.

[0105] (5) Implement distributed convergence detection based on Lyapunov.

[0106] In the embodiments of this application, iteration can be stopped when the following convergence condition is met:

[0107] and

[0108] Where, ε V ε is the threshold for the change in the Lyapunov function; δ This is the threshold for the maximum consistency deviation.

[0109] (6) Perform power redistribution.

[0110] This application's embodiments can employ a multi-agent reinforcement learning framework to achieve collaborative allocation of power deficits or surpluses. Specifically, in this application's embodiments, each sub-microgrid can act as an agent, generating power adjustment commands through a distributed policy network:

[0111]

[0112] in: Let θ represent the policy network of the i-th agent. i ;s i This represents the local state information of the i-th agent, including but not limited to distributed power output, energy storage state of charge, load power, local power imbalance, voltage amplitude, and system frequency; φ ij The collaborative weighting coefficients among neighboring intelligent agents; Let be the set of neighbors of the i-th agent, which includes all neighboring sub-microgrids with which it has a physical connection and reliable communication quality.

[0113] Furthermore, the agents can perform collaborative optimization through the following value function:

[0114]

[0115] Where χ is the discount factor, r i Let be the instantaneous reward function for the i-th agent. For expectation operator; a t For the joint action of the system at time t; s t Let be the system state at time t; s0 = s indicates that the calculation of the value function is based on the premise that the system starts from a specific initial state s.

[0116] In step S103, while executing the power coordination control strategy, a hierarchical control system is designed to generate corresponding economic optimization dispatch instructions based on the current frequency regulation stage. In response to the economic optimization dispatch instructions, with the goal of minimizing system operating costs and maximizing renewable energy consumption, the optimal power instruction is generated. Furthermore, a preset event triggering mechanism is used for communication to achieve the microgrid cluster optimization objective of multi-agent collaboration.

[0117] It can be explained that the hierarchical control system may include, but is not limited to: droop control of primary frequency regulation, which can be used to quickly respond to load fluctuations and achieve local power balance; voltage / frequency recovery of secondary frequency regulation, which can bring the system voltage and frequency back to the rated values ​​by adjusting energy storage or controllable power sources; and economic optimization scheduling of tertiary frequency regulation, which can minimize operating costs and maximize the utilization of renewable energy through power optimization and energy storage scheduling while ensuring system stability, thereby forming a complete control closed loop from rapid response to long-term optimization.

[0118] The preset event triggering mechanism can trigger control actions based on power imbalance, frequency deviation, voltage deviation, or energy storage status, etc. It can be set by those skilled in the art according to the actual situation, and no specific restrictions are imposed here.

[0119] Optionally, in one embodiment of this application, a hierarchical control system is designed to generate corresponding economically optimized scheduling instructions based on the current frequency regulation stage. This includes: using improved active power-frequency droop characteristics for primary power allocation to achieve droop control layer design for the primary frequency regulation stage; using an adaptive PI controller with anti-saturation integral to eliminate steady-state frequency and voltage deviations to achieve voltage / frequency recovery control design for the secondary frequency regulation stage; and establishing an optimization problem with the goal of minimizing operating costs to achieve economically optimized scheduling layer design for the tertiary frequency regulation stage.

[0120] The following examples illustrate the hierarchical control system of this application. The embodiments of this application may include the following steps:

[0121] (1) Design of the droop control layer for primary frequency modulation.

[0122] The embodiments of this application can achieve primary power distribution by using an improved active power-frequency (Pf) droop characteristic:

[0123]

[0124] Where, ω i (t) represents the output angular frequency of the i-th distributed source at time t; ω nom K is the system's rated angular frequency. p,i K is the active power-frequency droop factor. d,i The active power differential damping coefficient; P represents the actual output active power of the i-th distributed power source at time t. ref,i (t) represents the active power reference value of the i-th distributed power source.

[0125] Furthermore, its reactive power-voltage droop characteristic can be expressed as:

[0126]

[0127] Among them, V i (t) represents the output voltage amplitude of the i-th distributed power source at time t; V nom The system's rated voltage; K q,i K is the reactive power-voltage droop factor. v,i The reactive power differential damping coefficient; Q represents the actual reactive power output of the i-th distributed power source at time t. ref,i (t) represents the reactive power reference value of the i-th distributed power source.

[0128] (2) Design of voltage / frequency recovery control for secondary frequency modulation.

[0129] The embodiments of this application can use an adaptive PI controller with anti-saturation integral to eliminate steady-state frequency and voltage deviation.

[0130] Specifically, in the embodiments of this application, frequency recovery control can be expressed as:

[0131]

[0132] Where, Δω corr,i (t) represents the frequency correction amount of the i-th distributed source at time t; e ω,i (t)=ω nom -ω i (t) represents the frequency deviation; K p,ω,i K is the frequency proportional control coefficient. i,ω,i σ is the frequency integral control coefficient; ω This is the anti-saturation adjustment coefficient.

[0133] Meanwhile, voltage recovery control can be expressed as:

[0134]

[0135] Where, ΔV corr,i (t) represents the voltage correction amount of the i-th distributed source at time t; e v,i (t)=V nom -V i (t) represents the voltage deviation; K p,v,i K i,v,i σ is the voltage proportional control coefficient. v This is the anti-saturation adjustment coefficient.

[0136] (3) Design of economically optimized scheduling layer for three-frequency modulation.

[0137] The embodiments of this application can establish an optimization problem with the objective of minimizing operating costs, which can be expressed as:

[0138]

[0139] in, Let represent the generation cost function of the i-th distributed power source; This represents the operating cost function of the i-th energy storage system; Denotes the cost function of the curtailment penalty for the i-th distributed power source; This represents the amount of power abandoned by the i-th distributed power source.

[0140] (4) Embodiments of this application may also include power balance constraints:

[0141]

[0142] Among them, P loss This represents the total active power loss of the system.

[0143] Operational safety constraints:

[0144]

[0145] in, Let represent the lower and upper limits of the active power output of the i-th distributed power source, respectively; V represents the lower and upper limits of the reactive power output of the i-th distributed power source, respectively; i min V i max P represents the lower and upper limits of the allowable voltage for the i-th node, respectively; line,ij This represents the active power flow of line ij; This represents the active power transmission limit of line ij.

[0146] Optionally, in one embodiment of this application, the optimal power command is generated with the goal of minimizing system operating costs and maximizing renewable energy absorption rate. This includes: establishing a multi-objective optimization function and defining system operating costs and renewable energy absorption rate to establish a prediction model and construct a rolling optimization framework; and generating the optimal power command based on the prediction model and the rolling optimization framework.

[0147] As one possible implementation, embodiments of this application can generate optimal power commands in real time based on a rolling optimization algorithm based on model predictive control, with the goal of minimizing system operating costs and maximizing renewable energy absorption. Embodiments of this application may include the following steps:

[0148] (1) Establish a multi-objective optimization function.

[0149] First, within each prediction time domain, the embodiments of this application can construct the following multi-objective optimization function:

[0150]

[0151] in, H represents the local objective function of the i-th sub-microgrid at time t; P To predict the time domain length; C sys,i (t) represents the operating cost of the i-th sub-microgrid at time t; η ren,i This represents the renewable energy absorption rate of the i-th sub-microgrid at time t; u represents the power adjustment of the i-th sub-microgrid at time t; i (t) represents the control input vector of the i-th sub-microgrid at time t; α i β i γ i δ i These are the weighting coefficients for operating costs, renewable energy consumption, power smoothing, and control input changes, respectively; R is the penalty matrix for control input changes; Φ i (g) is the terminal cost function; ρ i This is the terminal cost weighting coefficient.

[0152] (2) Define the system operating cost.

[0153] Secondly, the system operating cost can be defined as follows in the embodiments of this application:

[0154]

[0155] in, This represents the generation cost of the i-th distributed power source at time k. This represents the operating cost of the i-th energy storage system at time k; Let $\frac{i}{i}$ represent the cost of energy forfeiture penalty for the $i$-th distributed power source at time $k$. This represents the amount of power abandoned by the i-th distributed power source at time k.

[0156] (3) Define the renewable energy absorption rate.

[0157] Next, the renewable energy integration rate can be defined in this embodiment as:

[0158]

[0159] in, This represents the actual power utilized by the i-th renewable energy source at time k; This represents the available power of the i-th renewable energy source at time k; To ensure the fraction is a very small number, avoid having a denominator of 0.

[0160] (4) Establish a prediction model.

[0161] Furthermore, this application can employ an autoregressive integral moving average model for power prediction:

[0162]

[0163] in, φ represents the predicted power of the i-th renewable energy source at time k+1; p represents the autoregression order; q represents the moving average order; φ j θ represents the j-th autoregressive coefficient; j denoted as the coefficient of the j-th moving average; k is a dimensionless number; ∈ represents a white noise sequence.

[0164] (5) Construct a rolling optimization framework.

[0165] As one possible approach, embodiments of this application can generate an optimal control sequence by solving a finite-time optimization problem at each sampling time, and implement only the first control variable, thereby enabling the prediction and optimization of the future behavior of the system while handling various operational constraints.

[0166] Specifically, in the embodiments of this application, the finite-time optimization problem to be solved can be expressed as:

[0167]

[0168] Meanwhile, the embodiments of this application need to meet the following constraints:

[0169]

[0170] Where U(t) is the control input sequence in the control time domain, U(t) = [u(t), u(t+1), ..., u(t+H)]. c -1)];u(τ) is the control input vector at time τ, including but not limited to the power commands of each distributed power source, the charging and discharging power of energy storage, etc.; X(τ) is the system state vector at time τ, including but not limited to the voltage, frequency, and energy storage state of each node, etc.; W(τ) is the disturbance vector at time τ, including load changes, renewable energy fluctuations, etc.; α is the weight coefficient of the operating cost term, which is a normal number; β is the weight coefficient of the curtailment penalty term, which is a normal number; ρ is the weight coefficient of the terminal cost term, which is a normal number; H p For predicting the system parameters of the time domain length, H represents the number of future time steps considered in the optimization problem; c For system parameters controlling the time domain length, represents the future time steps of the optimization control input; A, B, and D are the coefficient matrices of the system state-space model; G, H, and M are the coefficient matrices and boundary vectors of the inequality constraints; E, F, and N are the coefficient matrices and boundary vectors of the equality constraints; U min U maxThese are the lower and upper bound vectors controlling the input, respectively; X min X max C represents the lower and upper bound vectors of the state variables, respectively; sys (·) represents the system operating cost function; P curt (τ) is the power abandonment vector at time τ; Φ(·) is the terminal cost function, which can be used to ensure the stability of the optimization problem.

[0171] (6) Implement feedback correction.

[0172] Furthermore, at the end of each control cycle, embodiments of this application can use actual measured values ​​to correct the prediction model:

[0173]

[0174] in, This represents the corrected predicted power; K represents the correction gain coefficient. This represents the actual power measurement at time k.

[0175] (7) Generate the optimal power command.

[0176] Finally, the optimal control sequence can be obtained by solving an optimization problem in the embodiments of this application:

[0177] u * (t)={u * (t|t),u * (t+1|t),…,u * (t+H-1|t)},

[0178] It can be noted that, in the embodiments of this application, the first control quantity u can be... * (t|t) is the optimal power command for the current moment, which is then sent to each distributed power source.

[0179] Optionally, in one embodiment of this application, a preset event triggering mechanism is used for communication to achieve the optimization of a microgrid cluster with multi-agent cooperation. This includes: defining a communication topology and designing event triggering conditions to establish a preset event triggering mechanism; defining communication quality indicators based on the preset event triggering mechanism, designing an adaptive triggering threshold based on reinforcement learning, and establishing a communication reliability model; designing a fault-tolerant communication protocol to statistically analyze communication traffic and using the communication reliability model to evaluate the reliability indicators of communication.

[0180] Specifically, embodiments of this application can employ an event-triggered mechanism based on a communication network model to reduce communication traffic and improve system reliability. Embodiments of this application may include the following steps:

[0181] (1) Define the communication topology.

[0182] The embodiments of this application can construct a directed graph as the communication network for a microgrid cluster. in, This represents a set of communication nodes, with each node corresponding to a sub-microgrid. This represents a set of communication links; a communication adjacency matrix A can be defined. c =[a c,ij ] N×N , where a c,ij =1 indicates that node i can send information to node j, otherwise it is 0.

[0183] (2) Design event triggering conditions.

[0184] As one possible approach, the local event triggering condition of the i-th sub-microgrid can be based on the state error covariance matrix, which can be expressed as:

[0185]

[0186] Among them, e i (t) is the state error vector; Σ i (t) is the time-varying state estimation error covariance matrix; κ i This is the trigger sensitivity coefficient; For dynamic triggering threshold, δ i,min and δ i,max η represents the minimum and maximum values ​​of the trigger threshold, respectively. i For threshold adaptation rate, λ is the gradient of the local objective function. i The exponential decay coefficient; This indicates the k-th trigger time of the i-th sub-microgrid.

[0187] (3) Establish a hybrid triggering mechanism.

[0188] Furthermore, embodiments of this application can design a time-varying hybrid triggering strategy to dynamically adjust the triggering interval and event judgment threshold according to the system operating status and communication quality, which can be expressed as:

[0189]

[0190] in, The trigger interval is a fixed time. ρ is the event trigger interval; i (t) represents the network congestion index; ρ thresh This is the congestion threshold.

[0191] In the embodiments of this application, the event trigger interval calculation can be expressed as:

[0192]

[0193] Where σ(·) is the Sigmoid function; α i To adjust the parameters.

[0194] (4) Define communication quality indicators.

[0195] In some cases, embodiments of this application may use PRR (Packet Reception Rate) to evaluate communication quality, which can be expressed as:

[0196]

[0197] Among them, PRR ij This represents the packet reception rate from node i to node j; This represents the number of data packets that node j successfully received from node i within the statistical time period; This represents the total number of data packets sent from node i to node j within the statistical time period.

[0198] (5) Design an adaptive triggering threshold based on reinforcement learning.

[0199] Furthermore, embodiments of this application can dynamically adjust the trigger threshold according to communication quality, raising the threshold when communication quality is poor and lowering it when quality is good, thereby ensuring system stability while also considering response speed; embodiments of this application can be represented as follows:

[0200]

[0201] in, This represents the adaptive triggering threshold of the i-th sub-microgrid at time t; These represent the minimum and maximum values ​​of the trigger threshold, respectively; PRR avg (t) represents the average packet reception rate at time t.

[0202] (6) Establish a communication reliability model.

[0203] As one possible implementation, embodiments of this application can use a Markov chain model to characterize the reliability of the communication link, dividing the link state into three modes: normal, congested, and interrupted, and describing its transition characteristics over time through a state transition probability matrix; the Markov chain model used in embodiments of this application to describe communication state transitions can be expressed as follows:

[0204] P{S(t+1)=j|S(t=i}=p ij ,

[0205] Where S(t) represents the communication state at time t; p ijThis represents the probability of transitioning from state i to state j. Communication states are divided into three types: normal, congested, and interrupted.

[0206] (7) Design a fault-tolerant communication protocol.

[0207] In actual implementation, when a communication failure is detected, the embodiments of this application can design a local backup control strategy to ensure that the sub-microgrid can still maintain supply and demand balance and basic stable operation even in the event of loss of external communication; the embodiments of this application can be represented as follows:

[0208]

[0209] in, K represents the backup control command for the i-th sub-microgrid; i Represents the local control gain matrix; This indicates the rated control command.

[0210] (8) Implement communication traffic statistics:

[0211] This application embodiment can also count the amount of communication data in each control cycle. By recording the number and size of data packets sent and received by each node in that cycle, the total amount of communication data is calculated, and its relationship with the control frequency, network topology, and communication protocol is further analyzed to assess the communication load level and provide data support for bandwidth planning, protocol optimization, and distributed control strategy adjustment. This application embodiment can be represented as follows:

[0212]

[0213] Among them, D total Indicates the total amount of communication data; d ij f represents the amount of data transmitted in a single communication from node i to node j; ij This represents the communication frequency from node i to node j.

[0214] (9) Assess system reliability.

[0215] Ultimately, the embodiments of this application can use reliability indicators to evaluate the performance of the communication system, quantify the stability and availability of the communication system under different operating conditions, thereby providing a basis for the optimization of distributed control strategies and fault-tolerant design of microgrid clusters; the embodiments of this application can be represented as follows:

[0216]

[0217] Among them, R system Indicates the overall reliability of the system; R i R represents the reliability of the i-th communication node; ij This indicates the reliability of the communication link (i, j).

[0218] The following is a specific example, such as Figure 2 The flowchart illustrates the multi-agent cooperative microgrid cluster optimization method according to an embodiment of this application. The embodiments of this application may include the following steps:

[0219] In step S201, a system architecture for a microgrid cluster is established, including distributed power sources, energy storage units, and loads of each sub-microgrid, which are interconnected through power and communication networks to achieve local autonomy and global collaboration.

[0220] In step S202, a power coordination control strategy based on distributed consensus is constructed. Through local communication among the sub-microgrids, the power deficit or surplus is automatically allocated proportionally according to capacity / weight, ensuring that the entire cluster can quickly and reliably achieve power consistency without relying on central control, thereby improving the system's reliability and response speed.

[0221] In step S203, a hierarchical control system is designed, including droop control of primary frequency regulation to quickly suppress power disturbances, voltage / frequency recovery of secondary frequency regulation to bring the system operating point back to the rated value, and economic optimization scheduling of tertiary frequency regulation to minimize operating costs and maximize the absorption of renewable energy while ensuring system safety and stability, thereby taking into account both real-time performance and economy.

[0222] In step S204, a rolling optimization algorithm based on model predictive control is proposed. With the goal of minimizing system operating costs and maximizing renewable energy absorption rate, the algorithm generates the optimal power command in real time within each control cycle and considers energy storage constraints, power limits, and network constraints to achieve dynamic optimization and global coordination of the system.

[0223] In step S205, a communication network model is established and an event triggering mechanism is introduced. The triggering threshold and triggering interval are dynamically adjusted according to communication quality, load fluctuation and power imbalance, so as to reduce redundant communication, reduce bandwidth pressure, and at the same time ensure the convergence of distributed control and system stability, thereby improving the reliability, robustness and communication efficiency of the entire microgrid cluster.

[0224] The microgrid cluster optimization method based on multi-agent cooperation proposed in this application first constructs a microgrid cluster architecture. Then, it combines a collaborative architecture of distributed consistency coordination, hierarchical control system, model prediction rolling optimization, and event-triggered communication. With the goal of minimizing system operating costs and maximizing renewable energy absorption, it generates optimal power commands, thereby achieving multi-agent collaborative microgrid cluster optimization. This effectively addresses the uncertainty of distributed power output and the randomness of load demand, coordinating the automatic proportional allocation of power deficits or surpluses among sub-microgrids. Simultaneously, the hierarchical control system can quickly suppress power disturbances, restore voltage and frequency, and coordinate economic dispatch, ensuring system stability and high response speed. Furthermore, the event-triggered communication mechanism dynamically adjusts the trigger threshold and interval based on communication quality, reducing redundant communication traffic and improving system reliability and convergence. Therefore, it provides an efficient, reliable, and economical power control solution for microgrid clusters with a high proportion of renewable energy, achieving stable system operation, efficient energy utilization, and multi-agent collaborative optimization.

[0225] Next, referring to the accompanying drawings, a microgrid cluster optimization device for multi-agent cooperation proposed according to an embodiment of this application is described.

[0226] Figure 3 This is a block diagram of a microgrid cluster optimization device for multi-agent cooperation according to an embodiment of this application.

[0227] like Figure 3 As shown, the multi-agent collaborative microgrid cluster optimization device 10 includes: a setup module 100, a determination module 200, and an optimization module 300.

[0228] Among them, module 100 is used to establish the system architecture of the microgrid cluster.

[0229] The determination module 200 is used to construct a power coordination control strategy based on distributed consensus based on the system architecture, so as to use the power coordination control strategy to determine the allocation ratio within the microgrid cluster based on power deficit or surplus.

[0230] The optimization module 300 is used to design a hierarchical control system while executing the power coordination control strategy. It generates corresponding economic optimization dispatch instructions based on the current frequency regulation stage, and responds to the economic optimization dispatch instructions to generate the optimal power instructions with the goal of minimizing system operating costs and maximizing renewable energy consumption rate. It also uses a preset event triggering mechanism for communication to achieve the microgrid cluster optimization objective of multi-agent collaboration.

[0231] Optionally, in one embodiment of this application, the determining module 200 includes a calculation unit and a first generation unit.

[0232] The calculation unit is used to obtain the power imbalance of each sub-microgrid in the system architecture after determining the power imbalance, estimate the global imbalance, and calculate the power allocation ratio.

[0233] The first generation unit is used to generate power adjustment instructions for each sub-microgrid based on the global imbalance and power allocation ratio. It performs distributed convergence detection based on Lyapunov until the preset convergence condition is met, at which point it stops iterating. This allows for the coordinated allocation of power deficit or surplus after each sub-microgrid performs collaborative optimization using a preset value function.

[0234] Optionally, in one embodiment of this application, the formula for calculating the power adjustment amount of the power adjustment command is:

[0235]

[0236] in, The value represents the power regulation amount that the i-th sub-microgrid needs to undertake; λ is the differential gain coefficient. This represents the estimated global power imbalance value obtained by the consensus algorithm for the i-th sub-microgrid; This represents the power allocation ratio of the i-th sub-microgrid.

[0237] The preset convergence condition can be expressed as:

[0238] and

[0239] Where, ε V ε is the threshold for the change in the Lyapunov function; δ The maximum consistency deviation threshold is denoted as N; N represents the total number of sub-microgrids in the microgrid cluster.

[0240] The formula for generating power adjustment commands can be expressed as:

[0241]

[0242] in, Let θ represent the policy network of the i-th agent. i Indicates a parameter; s i φ represents the local state information of the i-th agent; ij The collaborative weighting coefficients among neighboring intelligent agents; Let be the set of neighbors of the i-th agent.

[0243] Optionally, in one embodiment of this application, the optimization module 300 includes: a first design unit, a second design unit, and a third design unit.

[0244] The first design unit is used to perform primary power allocation using improved active power-frequency droop characteristics to achieve the droop control layer design in the primary frequency modulation stage.

[0245] The second design unit is used to eliminate steady-state frequency and voltage deviations using an adaptive PI controller with anti-saturation integral, so as to realize the voltage / frequency recovery control design in the secondary frequency modulation stage.

[0246] The third design unit is used to establish an optimization problem with the goal of minimizing operating costs, so as to realize the economic optimization scheduling layer design of the third frequency regulation stage.

[0247] Optionally, in one embodiment of this application, the optimization module 300 includes: a construction unit and a second generation unit.

[0248] The building unit is used to establish a multi-objective optimization function and define the system operating cost and renewable energy absorption rate in order to establish a predictive model and build a rolling optimization framework.

[0249] The second generation unit is used to generate the optimal power command based on the prediction model and the rolling optimization framework.

[0250] Optionally, in one embodiment of this application, the optimization module 300 includes: a first establishment unit, a second establishment unit, and an evaluation unit.

[0251] The first establishment unit is used to define the communication topology and design event triggering conditions to establish a preset event triggering mechanism.

[0252] The second establishment unit is used to define communication quality indicators, design adaptive triggering thresholds based on reinforcement learning, and establish a communication reliability model based on a preset event triggering mechanism.

[0253] The evaluation unit is used to design fault-tolerant communication protocols, collect statistics on communication traffic, and evaluate communication reliability metrics using a communication reliability model.

[0254] It should be noted that the explanation of the aforementioned embodiment of the microgrid cluster optimization method for multi-agent cooperation also applies to the microgrid cluster optimization device for multi-agent cooperation in this embodiment, and will not be repeated here.

[0255] According to the multi-agent collaborative microgrid cluster optimization device proposed in this application, a microgrid cluster architecture is first constructed. Then, a collaborative architecture combining distributed consistency coordination, hierarchical control system, model prediction rolling optimization, and event-triggered communication is used to generate optimal power commands with the goal of minimizing system operating costs and maximizing renewable energy absorption rate. This achieves multi-agent collaborative microgrid cluster optimization, which can effectively cope with the uncertainty of distributed power output and the randomness of load demand. It coordinates the automatic proportional allocation of power deficits or surpluses in each sub-microgrid. At the same time, the hierarchical control system can quickly suppress power disturbances, restore voltage and frequency, and coordinate economic dispatch to ensure system stability and high response speed.

[0256] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0257] The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0258] When the processor 402 executes the program, it implements the microgrid cluster optimization method for multi-agent cooperation provided in the above embodiments.

[0259] Furthermore, electronic devices also include:

[0260] Communication interface 403 is used for communication between memory 401 and processor 402.

[0261] The memory 401 is used to store computer programs that can run on the processor 402.

[0262] The memory 401 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0263] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be divided into address buses, data buses, control buses, etc. For ease of representation, Figure 4The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0264] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0265] Processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0266] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described multi-agent cooperative microgrid cluster optimization method.

[0267] This application also provides a computer program product, including a computer program that can run computer instructions. When the computer instructions are executed by a processor, they implement the multi-agent cooperative microgrid cluster optimization method provided in this application.

[0268] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0269] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0270] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0271] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0272] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0273] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0274] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0275] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A multi-agent cooperative microgrid cluster optimization method, characterized in that, Includes the following steps: Establish a system architecture for microgrid clusters; Based on the system architecture, a power coordination control strategy based on distributed consistency is constructed to enable the allocation ratio within the microgrid cluster to be determined based on power deficit or surplus using the power coordination control strategy. While executing the power coordination control strategy, a hierarchical control system is designed to generate corresponding economically optimized dispatch instructions based on the current frequency regulation stage. In response to the economically optimized dispatch instructions, the system generates the optimal power instruction with the goal of minimizing system operating costs and maximizing renewable energy absorption rate. Furthermore, a preset event triggering mechanism is used for communication to achieve the microgrid cluster optimization objective of multi-agent collaboration.

2. The method according to claim 1, characterized in that, The construction of a power coordination control strategy based on distributed consensus, to utilize this strategy to determine the allocation ratio within the microgrid cluster based on power deficit or surplus, includes: After determining the power imbalance, the power imbalance of each sub-microgrid in the system architecture is obtained to estimate the global imbalance, and the power allocation ratio is calculated. Power adjustment instructions for each sub-microgrid are generated based on the global imbalance and the power allocation ratio. Lyapunov-based distributed convergence detection is performed until a preset convergence condition is met, at which point iteration stops. This allows for the coordinated allocation of power deficits or surpluses among the sub-microgrids after they have undergone collaborative optimization using a preset value function.

3. The method according to claim 2, characterized in that, in, The formula for calculating the power adjustment amount of the power adjustment command is as follows: in, The value represents the power regulation amount that the i-th sub-microgrid needs to undertake; λ is the differential gain coefficient. This represents the estimated global power imbalance value obtained by the consensus algorithm for the i-th sub-microgrid; This represents the power allocation ratio of the i-th sub-microgrid; The preset convergence condition is: and Where, ε V ε is the threshold for the change in the Lyapunov function; δ The maximum consistency deviation threshold; N represents the total number of sub-microgrids in the microgrid cluster; The formula for generating the power adjustment command is: in, Let θ represent the policy network of the i-th agent. i Indicates a parameter; s i φ represents the local state information of the i-th agent; ij The collaborative weighting coefficients among neighboring intelligent agents; Let be the set of neighbors of the i-th agent.

4. The method according to claim 1, characterized in that, The aforementioned hierarchical control system is designed to generate corresponding economically optimized scheduling instructions based on the current frequency regulation phase, including: An improved active power-frequency droop characteristic is used for primary power allocation to realize the droop control layer design in the primary frequency modulation stage. An adaptive PI controller with anti-saturation integral is used to eliminate steady-state frequency and voltage deviations, so as to realize the voltage / frequency recovery control design in the secondary frequency modulation stage; An optimization problem with the goal of minimizing operating costs is established to achieve an economically optimized scheduling layer design for the third frequency regulation stage.

5. The method according to claim 1, characterized in that, The process of generating optimal power commands, with the goal of minimizing system operating costs and maximizing renewable energy absorption, includes: A multi-objective optimization function is established, and the system operating cost and renewable energy absorption rate are defined to build a predictive model and construct a rolling optimization framework. Based on the prediction model and the rolling optimization framework, the optimal power command is generated.

6. The method according to claim 1, characterized in that, The method of using a preset event-triggered mechanism for communication to achieve the goal of multi-agent collaborative microgrid cluster optimization includes: Define the communication topology and design event triggering conditions to establish the preset event triggering mechanism; Based on the preset event triggering mechanism, a communication quality index is defined, an adaptive triggering threshold based on reinforcement learning is designed, and a communication reliability model is established. Design a fault-tolerant communication protocol to statistically analyze communication traffic and use the communication reliability model to evaluate communication reliability metrics.

7. A microgrid cluster optimization device for multi-agent cooperation, characterized in that, Includes the following steps: Establish modules for building the system architecture of microgrid clusters; The determination module is used to construct a power coordination control strategy based on the system architecture, so as to use the power coordination control strategy to determine the allocation ratio within the microgrid cluster based on power deficit or surplus. The optimization module is used to design a hierarchical control system while executing the power coordination control strategy, to generate corresponding economic optimization scheduling instructions according to the current frequency regulation stage, and to generate the optimal power instruction in response to the economic optimization scheduling instructions with the goal of minimizing system operating costs and maximizing renewable energy consumption rate, and to communicate using a preset event triggering mechanism to achieve the microgrid cluster optimization objective of multi-agent cooperation.

8. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the multi-agent cooperative microgrid cluster optimization method as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the microgrid cluster optimization method for multi-agent cooperation as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the microgrid cluster optimization method for multi-agent cooperation as described in any one of claims 1-6.

Images