Multi-agent fully-distributed micro-grid economic dispatching method based on accurate diffusion strategy

Abstract

The invention discloses a multi-agent fully-distributed microgrid economic dispatching method based on an accurate diffusion strategy, and relates to the field of microgrid economic dispatching. Comprising the following steps: S1, constructing a multi-agent micro-grid system; s2, based on a multi-agent micro-grid system, establishing an economic dispatching optimization model by taking minimization of the operation cost of the micro-grid as a target; and S3, solving the economic dispatching optimization model by adopting an improved precise diffusion algorithm so as to obtain a dispatching scheme. According to the invention, the efficiency and reliability of economic dispatching of the micro-grid can be remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of microgrid economic dispatching, and in particular to a multi-agent fully distributed microgrid economic dispatching method based on precise diffusion strategy. Background Art

[0002] With the widespread application of renewable energy and the development of microgrid technology, how to efficiently and economically dispatch various energy resources in microgrids has become an important research direction. Traditional economic dispatch methods mainly rely on centralized control strategies, which perform well in small-scale or stable environments. However, in the face of large-scale microgrids and frequently changing energy supply and demand, centralized control has the disadvantages of insufficient flexibility, heavy computational burden, and sensitivity to communication failures.

[0003] In recent years, distributed economic dispatch methods have attracted attention due to their high flexibility and scalability. Existing microgrid economic dispatch mostly relies on central controllers or partially distributed strategies, which limits the flexibility and scalability of the system. Existing methods face the challenges of high computational load and low communication efficiency when dealing with large-scale, dynamically changing microgrid systems. These methods can theoretically improve the reliability and efficiency of the system by distributing the calculation and decision-making process among the nodes of the microgrid. However, existing distributed methods often rely on approximate algorithms, which make it difficult to ensure the global optimal solution. In addition, when the communication network changes dynamically, the stability and convergence speed of the algorithm need to be improved, which reduces the reliability of dispatch. Summary of the invention

[0004] In view of the above problems, the present invention provides a multi-agent fully distributed microgrid economic dispatch method based on a precise diffusion strategy, which can achieve fast, accurate and stable economic dispatch.

[0005] The technical solution of the present invention is: a multi-agent fully distributed microgrid economic dispatch method based on a precise diffusion strategy, comprising the following steps:

[0006] Step S1, constructing a multi-agent microgrid system;

[0007] Step S2, based on a multi-agent microgrid system, an economic dispatch optimization model is established with the goal of minimizing the operation cost of the microgrid;

[0008] Step S3, using an improved exact diffusion algorithm to solve the economic dispatch optimization model to obtain a dispatch plan.

[0009] In step S1, the structure of the multi-agent microgrid system includes:

[0010] Microgrid Control Center (MGCC) agent at the upper level;

[0011] The bottom layer is controlled locally by device agents, including battery agents (BA), flexible load agents (FLA), renewable energy generation device agents, and traditional distributed generation device agents;

[0012] Among them, renewable energy power generation equipment agents include photovoltaic agents (PVA) and wind turbine agents (WTA), and traditional distributed power generation equipment agents include micro generator agents (MTA).

[0013] The cost function of the battery agent is as follows:

[0014]

[0015] In the formula, C Bj (P Bj ) is the power cost of battery j, P Bj is the power of battery j, C Bj is the power cost, a Bj , b Bj 、c Bj is the function cost coefficient, SOC represents the current state of charge of the battery;

[0016] The cost function of the traditional distributed generation equipment (DGA) agent is as follows:

[0017]

[0018] In the formula, C DGt (P DGt ) is the power cost of conventional generator t, P DGt is the power of the conventional generator t, C DGt is the power cost, a DGt , b DGt 、c DGt is the function cost coefficient,

[0019] The cost function of the flexible load agent is:

[0020]

[0021] In the formula, C FLn (P FLn ) is the power cost of flexible load n, P FLn is the power of the flexible load n, C FLn is the power cost, a n , b n 、c n is the function cost coefficient,

[0022] The cost function of the microgrid control center agent is as follows:

[0023] C MG (PMG )=C DGt (P DGt )+C Bj (P Bj )-C FLn (P FLn )+ep×P MG (4)

[0024] In the formula, C MG (P MG ) is the cost of exchanging operation between the microgrid and the public grid, P MG is the amount of electricity exchanged between the microgrid and the public grid, C MG is the operating cost of the microgrid, and ep is the electricity price.

[0025] In step S2, the objective function and constraints of the economic dispatch optimization model are:

[0026]

[0027] In the formula, C MG is the microgrid operating cost, q i is the weight coefficient of each agent cost function, and q i >0,P i is the power of the agent, C i is the agent at different power outputs P i The operating cost is

[0028] P L is the normal load demand, P REk is the power of renewable energy, S Bj is the state variable of the battery,

[0029] P FLn,min , P FLn,max is the lower and upper power limits of the flexible load n, P DGt,min , P DGt,max is the lower and upper power limit of generator t; P REk,min , P REk,max is the lower and upper limits of renewable energy power; S Bj,min , S Bj,max Charge the battery to the lower and upper limits.

[0030] In step S3, the specific steps of solving are as follows:

[0031] Step S31, information sharing and initialization, each agent shares state information with neighboring agents through a sparse communication network and sets initial consensus variables;

[0032] Step S32, each agent calculates its own local optimization problem based on the current information and state, while taking into account the constraints of the system;

[0033] Step S33, adopting the precise diffusion strategy, each agent updates its own state variable by adding a correction term after calculating the local optimization result;

[0034] Step S34, the agent updates its own consensus variables and optimization results according to the corrected information and the status of neighboring agents, until the consensus variables of all agents converge to the same value and the system reaches the overall optimal economic dispatch solution;

[0035] In step S35, the system determines whether convergence is achieved according to the preset convergence conditions; if the convergence conditions are met, the entire optimization process ends, and the agent determines the final scheduling decision.

[0036] The method also includes step S4, a microgrid economic dispatch process based on a precise diffusion strategy, including: initialization, calculation of a marginal cost function, optimization solution based on a modified EDA, power allocation considering a line loss model, and verification and adjustment of satisfying various agent constraints.

[0037] In step S4, the microgrid economic dispatch process based on the precise diffusion strategy specifically includes:

[0038] Step S41, compile the adjacency matrix according to the communication topology, and update the element a in the coefficient matrix A ij , a ij represents the weight of information received by agent i from agent j;

[0039] Step S42, randomly select an initial value x i,0 ,make Initial data processing, calculation step size μ i ;

[0040] Step S43, determining whether the microgrid is connected to the public grid;

[0041] If connected to the grid, P MG If it is not equal to 0, the microgrid and the public grid exchange power; otherwise, P MG is 0, distributed generation and load need to maintain power balance;

[0042] Step S44, calculating the marginal cost function of each agent;

[0043] Step S45, obtaining the economic optimal solution of agent i at time k according to the modified EDA;

[0044] Step S46, calculating the active power P of agent i at time k according to whether the line loss model is considered i,k ;

[0045] Step S47, determine whether the power meets the agent's constraints: if not, take the upper or lower limit of the constraint and update x i,k , time k is increased by 1, and the process returns to step 45; otherwise, it is determined whether the absolute value of the net load is less than an infinitesimal random number: if not, time k is increased by 1, and the process returns to step 45; if yes, the economic optimization process ends.

[0046] In step S44, when calculating the marginal cost function of each agent, the marginal cost calculation and equilibrium conditions in the economic dispatch of the microgrid are considered;

[0047] Set the consensus variable x to the marginal cost as follows:

[0048]

[0049] In the formula, C t is the cost function of the power generation equipment, C 1 ……C n is the cost function of each scheduling agent, P 1 ...P n is the power output of each agent, λ is the optimal value of the consensus variable, b t is the linear operating cost of the generator;

[0050] Using the Lagrange multiplier method, it is concluded that when the marginal cost of each dispatchable agent converges to the same value, that is, when the consensus variable x converges to a certain λ value, the microgrid has the greatest economic benefit, and the determined value is the electricity price when connected to the grid.

[0051] In step S46, if the line loss model is not considered, the active power P of agent i at time k is calculated based on the optimized power output of each agent. i,k ;

[0052] The optimized power of each agent is:

[0053]

[0054] Where b FLn is the linear cost of flexible load power output, c FLn is the secondary cost of flexible load power output.

[0055] In step S46, if the line loss model is considered, the active power P of agent i at time k is calculated based on the optimized power output of each agent after considering the line loss. i,k ;

[0056] The optimized power output of each agent after considering line loss is:

[0057]

[0058] Where, L t is the loss coefficient of the generator when considering the line loss model, L n It is the loss coefficient of the flexible load when considering the line loss model.

[0059] In operation, the present invention constructs a multi-agent microgrid system; establishes an economic dispatch optimization model with the goal of minimizing the operating cost of the microgrid; and uses an improved precise diffusion algorithm to solve the economic dispatch optimization model to obtain a dispatch solution. Through the precise diffusion strategy, the system output can be guaranteed to converge quickly to the optimal dispatch solution even under conditions of limited communication connection. Compared with traditional centralized control methods, the present invention can reduce communication load, improve the system's ability to resist faults, and is more suitable for the development needs of future distributed smart grids.

[0060] The present invention can significantly improve the efficiency and reliability of economic dispatching of microgrids. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the following briefly introduces the drawings required for the specific implementation or the prior art description. In the drawings, each part is not necessarily drawn according to the actual scale.

[0062] Figure 1 is a flow chart of the present invention,

[0063] Figure 2 is a schematic diagram of the consensus algorithm used in the embodiment,

[0064] Figure 3 This is a schematic diagram of the method of the present invention used in the embodiment. Figure 1 ,

[0065] Figure 4 This is a schematic diagram of the method of the present invention used in the embodiment. Figure 2 . DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0067] The present invention Figure 1 As shown in FIG. 1 , the multi-agent fully distributed microgrid economic dispatch method based on the precise diffusion strategy includes the following steps:

[0068] Step S1, construct a multi-agent microgrid system; design a hierarchical control framework, including the upper-layer microgrid control center agent (MGCC) and the lower-layer device agent, and define the cost function of each agent to determine its optimal power generation or load under a given power demand.

[0069] Step S2, based on a multi-agent microgrid system, an economic dispatch optimization model is established with the goal of minimizing the operation cost of the microgrid;

[0070] Step S3, using an improved exact diffusion algorithm to solve the economic dispatch optimization model to obtain a dispatch plan.

[0071] The present invention solves the problems of heavy communication load and susceptibility to single point failure in the centralized control method in the prior art. By introducing the precise diffusion strategy, accurate convergence under sparse communication network is achieved, the active power of all dispatchable agents in the microgrid is optimized, the economic utilization of resources is achieved, and the stability and scalability of the system are guaranteed.

[0072] The present invention realizes fast and accurate economic dispatch through an improved algorithm, while ensuring the robustness of the algorithm in a dynamic communication network.

[0073] EDA is the abbreviation of "Estimation of Distribution Algorithm" in a broad sense, and the precise diffusion strategy is a method to implement this algorithm. This strategy achieves the goal of economic dispatch by sharing information in a distributed environment of multiple agents, gradually optimizing their respective cost functions, and finally converging to the global optimal solution.

[0074] In step S1, the structure of the multi-agent microgrid system includes:

[0075] Microgrid Control Center (MGCC) agent at the upper level;

[0076] The bottom layer is controlled locally by device agents, including battery agents (BA), flexible load agents (FLA), renewable energy generation device agents, and traditional distributed generation device agents;

[0077] Among them, renewable energy power generation equipment agents include photovoltaic agents (PVA) and wind turbine agents (WTA), and traditional distributed power generation equipment agents include micro generator agents (MTA).

[0078] The cost function of the battery agent is as follows:

[0079]

[0080] In the formula, C Bj (P Bj) is the power cost of battery j, P Bj is the power of battery j, which is positive when discharging and negative when charging, C Bj is the power cost, a Bj , b Bj 、c Bj is the function cost coefficient, SOC represents the current state of charge of the battery or energy storage device, which is used to describe the current charge level of the battery as a percentage of its maximum capacity;

[0081] Battery power constraints such as: P Bj,min ≤P Bj ≤P Bj,max , P Bj,min , P Bj,max are the lower and upper limits of battery power.

[0082] The cost function of the traditional distributed generation equipment (DGA) agent is as follows:

[0083]

[0084] In the formula, C DGt (P DGt ) is the power cost of conventional generator t, P DGt is the power of the conventional generator t, C DGt is the power cost, a DGt , b DGt 、c DGt is the function cost coefficient,

[0085] The cost function of the flexible load agent is:

[0086]

[0087] In the formula, C FLn (P FLn ) is the power cost of flexible load n, P FLn is the power of the flexible load n, C FLn is the power cost, a n , b n 、c n is the function cost coefficient,

[0088] The cost function of the microgrid control center agent is as follows:

[0089] C MG (P MG )=C DGt (P DGt )+C Bj (P Bj )-C FLn (P FLn )+ep×P MG(4)

[0090] In the formula, C MG (P MG ) is the cost of exchanging operation between the microgrid and the public grid, P MG is the amount of electricity exchanged between the microgrid and the public grid, C MG is the operating cost of the microgrid, and ep is the electricity price.

[0091] PVA and WTA are both renewable energy generators. When the grid operates in island mode, from an economic point of view, priority should be given to utilization, so it makes no sense to participate in economic dispatch. Here, renewable energy entities are regarded as non-dispatchable entities and do not participate in economic dispatch.

[0092] In step S2, a mathematical model is formulated to represent the economic dispatch problem of the microgrid, including an objective function and constraints. The operating cost function of each agent given in step S1 above constitutes the economic dispatch objective function shown below, which is convex and differentiable.

[0093] The objective function and constraints of the economic dispatch optimization model are:

[0094]

[0095] In the formula, C MG is the microgrid operating cost, q i is the weight coefficient of each agent cost function, and q i >0,P i is the power of the agent, C i is the agent at different power outputs P i The operating cost is

[0096] P L is the normal load demand, P REk is the power of renewable energy, S Bj It is a state variable related to the battery or energy storage device, usually the charge state of the battery.

[0097] The inequality constraints are upper and lower bounds on the power output of each agent.

[0098] P FLn,min , P FLn,max is the lower and upper power limits of the flexible load n, P DGt,min , P DGt,max is the lower and upper power limit of generator t; P REk,min , P REk,max is the lower and upper limits of renewable energy power; S Bj,min , S Bj,max Charge the battery to the lower and upper limits.

[0099] In step S3, agents share information through a sparse communication network, and each agent calculates its own local solution to the optimization problem (i.e., through multiple iterations, each agent optimizes locally, gradually adjusts based on feedback from neighboring agents, and finally achieves a global optimization result). An accurate diffusion strategy is adopted, and correction steps are added during the iteration process to ensure that the algorithm can accurately converge to the global optimal solution in a distributed environment. Each agent updates its own state based on the information of its neighbors until the entire multi-agent microgrid system reaches the optimal economic dispatch solution.

[0100] The specific steps of solving are as follows:

[0101] Step S31, information sharing and initialization, each agent shares state information with neighboring agents through a sparse communication network and sets initial consensus variables;

[0102] Step S32, each agent calculates its own local optimization problem based on the current information and status, while considering the constraints of the multi-agent microgrid system;

[0103] Step S33, adopting the precise diffusion strategy, each agent updates its own state variable by adding a correction term (usually the difference of the result of the previous iteration) after calculating the local optimization result;

[0104] Step S34, the agent updates its own consensus variables and optimization results according to the corrected information and the status of the neighboring agents, until the consensus variables of all agents converge to the same value, and the multi-agent microgrid system reaches the overall optimal economic dispatch solution;

[0105] In step S35, the multi-agent microgrid system determines whether convergence has been achieved according to a preset convergence condition; if the convergence condition is met, the entire optimization process ends, and the agent determines the final scheduling decision.

[0106] The method also includes step S4, a microgrid economic dispatch process based on a precise diffusion strategy, including: initialization, calculation of a marginal cost function, optimization solution based on a modified EDA, power allocation considering a line loss model, and verification and adjustment of satisfying various agent constraints.

[0107] Step 4 describes in detail the actual application process of the optimization solution obtained in step 3, including initialization, power allocation, and verification and adjustment of the results.

[0108] In step S4, the microgrid economic dispatch process based on the precise diffusion strategy specifically includes:

[0109] Step S41, compile the adjacency matrix according to the communication topology, and update the element a in the coefficient matrix A ij , a ijrepresents the weight of information received by agent i from agent j;

[0110] Adjacency matrix: used to describe the communication topology between agents. If two agents can communicate, the corresponding element in the matrix is ​​1, otherwise it is 0. Coefficient matrix A: represents the communication weight between each agent.

[0111] Step S41 is to build the communication topology between agents and clarify the transmission path and weight of information;

[0112] Step S42, randomly select an initial value x i,0 ,make Initial data processing, calculation step size μ i ;

[0113] x i,0 The initial consensus variable of agent i; is the initial intermediate variable, μ i is the step size, which determines how much the agent state is updated at each iteration.

[0114] Step S42 sets the initial state for each agent to ensure that the distributed optimization process can be started smoothly and eventually converged.

[0115] Step S43, determining whether the microgrid is connected to the public grid;

[0116] If connected to the grid, P in equation (5) MG If it is not equal to 0, the microgrid and the public grid exchange power; otherwise, P MG is 0, distributed generation and load need to maintain power balance;

[0117] Step S44, calculating the marginal cost function of each agent according to formula (6);

[0118] Step S45, obtaining the economic optimal solution of agent i at time k according to the modified EDA;

[0119] Step S46: Depending on whether the line loss model is considered, select equation (7) or equation (8) to calculate the active power P of agent i at time k. i,k ;

[0120] Step S47, determine whether the power meets the agent's constraints: if not, take the upper or lower limit of the constraint and update x i,k , time k is increased by 1, and the process returns to step 45; otherwise, it is determined whether the absolute value of the net load is less than an infinitesimal random number: if not, time k is increased by 1, and the process returns to step 45; if yes, the economic optimization process ends.

[0121] In step S44, when calculating the marginal cost function of each agent, the marginal cost calculation and equilibrium conditions in the economic dispatch of the microgrid are considered;

[0122] Set the consensus variable x to the marginal cost as follows:

[0123]

[0124] In the formula, C t is the cost function of the power generation equipment, C 1 ……C n is the cost function of each scheduling agent, P 1 ...P n is the power output of each agent, λ is the optimal value of the consensus variable, b t is the linear operating cost of the generator;

[0125] Using the Lagrange multiplier method, it is concluded that when the marginal cost of each dispatchable agent converges to the same value, that is, when the consensus variable x converges to a certain λ value, the microgrid has the greatest economic benefit, and the determined value is the electricity price when connected to the grid.

[0126] The consensus variable refers to a key variable that is gradually agreed upon among agents through a distributed optimization algorithm (such as a diffusion strategy) in the microgrid economic dispatch problem. Here, it represents the unified marginal cost of all dispatchable resources in the multi-agent microgrid system.

[0127] This result is consistent with the equal-increment principle of economic dispatch of traditional power systems.

[0128] The equal increment principle is a basic principle in the economic dispatch of traditional power systems. Its core idea is: in order to achieve the economic optimum of the system, each generator unit adjusts its output power within its constraints so that the marginal cost (i.e., incremental cost) of each unit to generate one unit of power is the same. This means that when the marginal costs of all generators tend to be consistent, the total power generation cost of the system is the lowest. Mathematically, the equal increment principle is usually implemented through the Lagrange multiplier method, that is, constructing a Lagrangian function, combining the cost function of each generator and its power output constraints, so as to solve the optimal power allocation plan that meets the conditions.

[0129] According to the equal increment principle, when the cost of each device to generate one unit of active power is the same, the multi-agent microgrid system has the best economic performance. Therefore, the optimized active power of each agent can be calculated by formula (7).

[0130]

[0131] Where b FLn is the linear cost of flexible load power output, c FLnis the secondary cost of flexible load power output.

[0132] When line loss is considered, the optimal power output expression of each agent is no longer equation (7), but equation (8).

[0133]

[0134] Where, L t is the loss coefficient of the generator when considering the line loss model, L n It is the loss coefficient of the flexible load when considering the line loss model.

[0135] The present invention can effectively handle changes in communication topology, realize economic optimization scheduling of dispatchable resources in a microgrid, and has good "plug and play" capability; in this way, the system's ability to adaptively schedule resources in the face of dynamic changes ensures the operating efficiency and stability of the microgrid.

[0136] The present invention has the following advantages:

[0137] (1) It realizes fully distributed control of microgrid economic dispatch, significantly reducing the communication load and the system's sensitivity to single-point failures;

[0138] (2) The improved precise diffusion algorithm accelerates the system convergence speed and improves the execution efficiency of the scheduling scheme;

[0139] (3) Enhanced system stability and scalability, making it easier to adapt to changes in microgrid scale and the access to new energy.

[0140] The embodiment of the present invention uses the improved IEEE 33 node as an example to verify the feasibility of the proposed method. The lines in the figure represent different agents in the microgrid. Through different agents, the trend is observed. The following scenario is set to the grid-connected mode, without considering the upper and lower limits of the generator, that is, removing the constraint term in formula (5) to verify the effectiveness of the strategy in distributed economic dispatch. When connected to the grid, the renewable energy unit works in the MPPT mode and uses the consensus algorithm for calculation. The convergence process of the consensus variable x is as follows: Figure 2 As shown. Figure 2 It can be seen that all consensus variables in the microgrid are converged and converted into electricity prices, λ=ep=0.1710.

[0141] The calculation results of the strategy and algorithm proposed by the present invention are as follows Figure 3-4 As shown. Figure 3 It can be seen that by improving the calculation of the exact diffusion algorithm, x can be obtained with Figure 2The same results are obtained. However, comparing the iteration time and the convergence process of consensus variables, the consensus algorithm requires nearly 300 iterations to obtain accurate results, while the modified EDA only requires 35 iterations, which greatly improves the calculation speed.

[0142] According to the values ​​of the above consensus variables, the output of each device can be calculated: P DG1 =147.50kW, P DG2 =117.00kW, P DG3 =96.67kW, P DG4 =82.14kW, P DG5 =71.25kW, P DG6 =62.78kW, P FL1 =71.92kW, P FL2 =96.82kW, P B =-33.21kW, P MG =287.13kW. Total cost: C MG =462.73$. The multi-agent microgrid system achieves supply and demand balance, and the microgrid is economically optimal at this time.

[0143] The present invention provides a fully distributed economic dispatch method that can achieve fast, accurate and stable global optimization, which is of great significance for improving the operating efficiency and reliability of large-scale, dynamic microgrids.

[0144] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-agent fully distributed microgrid economic dispatch method based on an accurate diffusion strategy, characterized in that: The following steps are involved: Step S1, constructing a multi-agent microgrid system; Step S2, based on a multi-agent microgrid system, an economic dispatch optimization model is established with the goal of minimizing the operation cost of the microgrid; Step S3, using an improved exact diffusion algorithm to solve the economic dispatch optimization model to obtain a dispatch plan.

2. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 1 is characterized in that: In step S1, the structure of the multi-agent microgrid system includes: The upper-level microgrid control center agent; The bottom layer is controlled locally by device agents, including battery agents, flexible load agents, renewable energy generation device agents, and traditional distributed generation device agents. Among them, renewable energy power generation equipment agents include photovoltaic agents and wind turbine agents, and traditional distributed power generation equipment agents include micro-generator agents.

3. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 2 is characterized in that: The cost function of the battery agent is as follows: In the formula, C Bj (P Bj ) is the power cost of battery j, P Bj is the power of battery j, C Bj is the power cost, a Bj 、b Bj 、c Bj is the function cost coefficient, SOC represents the current state of charge of the battery; The cost function of the traditional distributed generation equipment (DGA) agent is as follows: In the formula, C DGt (P DGt ) is the power cost of conventional generator t, P DGt is the power of the conventional generator t, C DGt is the power cost, a DGt 、b DGt 、c DGt is the function cost coefficient, The cost function of the flexible load agent is: In the formula, C FLn (P FLn ) is the power cost of flexible load n, P FLn is the power of the flexible load n, C FLn is the power cost, a n , b n 、c n is the function cost coefficient, The cost function of the microgrid control center agent is as follows: C MG (P MG )=C DGt (P DGt )+C Bj (P Bj )-C FLn (P FLn )+ep×P MG (4) In the formula, C MG (P MG ) is the cost of exchanging operation between the microgrid and the public grid, P MG is the amount of electricity exchanged between the microgrid and the public grid, C MG is the operating cost of the microgrid, and ep is the electricity price.

4. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 3 is characterized in that: In step S2, the objective function and constraints of the economic dispatch optimization model are: In the formula, C MG is the microgrid operating cost, q i is the weight coefficient of each agent cost function, and q i >0,P i is the power of the agent, C i is the agent at different power outputs P i The operating cost is P L is the normal load demand, P REk is the power of renewable energy, S Bj is the state variable of the battery, P FLn,min , P FLn,max is the lower and upper power limits of the flexible load n, P DGt,min , P DGt,max is the lower and upper power limit of generator t; P REk,min , P REk,max The lower and upper limits for renewable energy power; S Bj,min , S Bj,max Charge the battery to the lower and upper limits.

5. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 1 is characterized in that: In step S3, the specific steps of solving are as follows: Step S31, information sharing and initialization, each agent shares state information with neighboring agents through a sparse communication network and sets initial consensus variables; Step S32, each agent calculates its own local optimization problem based on the current information and state, while taking into account the constraints of the system; Step S33, adopting the precise diffusion strategy, each agent updates its own state variable by adding a correction term after calculating the local optimization result; Step S34, the agent updates its own consensus variables and optimization results according to the corrected information and the status of neighboring agents, until the consensus variables of all agents converge to the same value and the system reaches the overall optimal economic dispatch solution; In step S35, the system determines whether convergence is achieved according to the preset convergence conditions; if the convergence conditions are met, the entire optimization process ends, and the agent determines the final scheduling decision.

6. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 1 is characterized in that: The method also includes step S4, a microgrid economic dispatch process based on a precise diffusion strategy, including: initialization, calculation of a marginal cost function, optimization solution based on a modified EDA, power allocation considering a line loss model, and verification and adjustment of satisfying various agent constraints.

7. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 6 is characterized in that: In step S4, the microgrid economic dispatch process based on the precise diffusion strategy specifically includes: Step S41, compile the adjacency matrix according to the communication topology, and update the element a in the coefficient matrix A ij , a ij represents the weight of information received by agent i from agent j; Step S42, randomly select an initial value x i,0 ,make Initial data processing, calculation step size μ i ; Step S43, determining whether the microgrid is connected to the public grid; If connected to the grid, P MG If it is not equal to 0, the microgrid and the public grid exchange power; otherwise, P MG is 0, distributed generation and load need to maintain power balance; Step S44, calculating the marginal cost function of each agent; Step S45, obtaining the economic optimal solution of agent i at time k according to the modified EDA; Step S46, calculating the active power P of agent i at time k according to whether the line loss model is considered i,k ; Step S47, determine whether the power meets the agent's constraints: if not, take the upper or lower limit of the constraint and update x i,k , time k is increased by 1, and the process returns to step 45; otherwise, it is determined whether the absolute value of the net load is less than an infinitesimal random number: if not, time k is increased by 1, and the process returns to step 45; if yes, the economic optimization process ends.

8. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 7 is characterized in that: In step S44, when calculating the marginal cost function of each agent, the marginal cost calculation and equilibrium conditions in the economic dispatch of the microgrid are considered; Set the consensus variable x to the marginal cost as follows: In the formula, C t is the cost function of the power generation equipment, C1…C n is the cost function of each scheduling agent, P1…P n is the power output of each agent, λ is the optimal value of the consensus variable, b t is the linear operating cost of the generator; Using the Lagrange multiplier method, it is concluded that when the marginal cost of each dispatchable agent converges to the same value, that is, when the consensus variable x converges to a certain λ value, the microgrid has the greatest economic benefit, and the determined value is the electricity price when connected to the grid.

9. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 7 is characterized in that: In step S46, if the line loss model is not considered, the active power P of agent i at time k is calculated based on the optimized power output of each agent. i,k ; The optimized power of each agent is: Where b FLn is the linear cost of flexible load power output, c FLn is the secondary cost of flexible load power output.

10. The multi-agent fully distributed microgrid economic dispatch method based on precise diffusion strategy according to claim 7, characterized in that: In step S46, if the line loss model is considered, the active power P of agent i at time k is calculated based on the optimized power output of each agent after considering the line loss. i,k ; The optimized power output of each agent after considering line loss is: Where, L t is the loss coefficient of the generator when considering the line loss model, L n It is the loss coefficient of the flexible load when considering the line loss model.

Images