A fully distributed power system dispatching method for arbitrary directed communication topology

Abstract

The present invention discloses a fully distributed power system scheduling method applicable to arbitrary directed communication topology, including using quadratic functions to calculate the operation and maintenance costs and carbon trading costs of a subsystem; based on the operation and maintenance costs and the carbon trading costs, introducing Lagrange multipliers, converting inequality constraints into equality constraints, and analyzing in combination with KKT conditions to obtain the optimal active power of the subsystem; and scheduling optimization based on consistency to optimize the optimal active power of the subsystem, including adding estimated variables of supply and demand deviations to each subsystem and introducing surplus variables; combining the estimated variables of supply and demand deviations and the surplus variables to respectively perform consistency iteration of states, calculate active power and estimate power inconsistency; each node makes decisions based on local information and simple communication with neighboring nodes, which greatly improves the robustness and flexibility of algorithms applicable to large-scale and complex power networks, and achieves an optimal balance between cost-effectiveness and environmental benefits.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching, and in particular to a fully distributed power system dispatching method applicable to any directional communication topology. Background Art

[0002] As the problem of climate change intensifies, the world has attached great importance to strategies to achieve carbon neutrality and the use of renewable energy. In this context, Smart Grid Economic Dispatch (SGED) came into being, aiming to optimize the comprehensive use of energy, promote energy efficiency and reduce carbon emissions, in order to support the development goals of a low-carbon economy. By incorporating innovative optimization algorithms, Smart Grid Economic Dispatch not only ensures the stability and security of energy supply, but also achieves the optimal balance between economic and environmental benefits, providing a technical path and practical solution for achieving the goal of carbon neutrality.

[0003] Traditional economic dispatch of power systems usually adopts centralized optimization methods to ensure the economic and reliable supply of energy. However, it is difficult to adapt to very large systems and is also susceptible to single point failures. As the system size grows, the communication and computing overheads will increase significantly. The existing leader-follower model has attracted widespread attention from academia and industry as an important optimization framework. This strategy relies on one or more leaders to formulate overall dispatch strategies, and many followers adjust their actions according to these strategies. Although this model is efficient and easy to implement in some cases, its reliance on centralized decision points and possible information bottlenecks limits its application potential in highly autonomous and dynamic environments.

[0004] Fully distributed economic dispatch methods have become a hot topic of research due to their significant advantages in autonomy, resilience, privacy protection and scalability. However, most of the existing fully distributed methods are based on undirected communication topologies, which may encounter limitations in practical applications. ChenW and LiT proposed a fully distributed economic dispatch method based on multi-agent consensus control and suitable for energy Internet. An important point is that these methods are designed to be applicable to undirected communication topologies. This means that in such a system, all communication links are bidirectional, and each node can communicate directly with any other node connected to it. It should be pointed out that due to the limitations of communication power, communication distance, etc., in actual networks, the communication path may also be directed, which means that the information flow can only propagate in one direction, and a dispatching method for directed communication topologies needs to be studied. In view of this, the present invention constructs a fully distributed power system dispatching method applicable to any directed communication topology to achieve the optimal balance between cost-effectiveness and environmental benefits. Summary of the invention

[0005] The purpose of the present invention is to provide a fully distributed power system scheduling method applicable to any directional communication topology, which can effectively solve the problems existing in the above-mentioned prior art.

[0006] To solve the above technical problems, the present invention adopts the following technical solution: a fully distributed power system scheduling method applicable to any directed communication topology, the power system includes several subsystems, each subsystem is embedded in a node, and the scheduling method includes the following steps:

[0007] The operation and maintenance costs and carbon trading costs of the subsystem are calculated using quadratic functions;

[0008] Based on the operation and maintenance costs and carbon trading costs, Lagrange multipliers are introduced to transform inequality constraints into equality constraints, and the optimal active power of the subsystem is obtained by combining the KKT conditions; and

[0009] Consistency-based dispatch optimization subsystem optimal active power, including:

[0010] Add estimation variables of supply-demand deviation and introduce surplus variables for each subsystem;

[0011] The estimated variables of supply-demand deviation and the surplus variables are combined to perform consistency iteration of the state, calculation of active power and estimation of power inconsistency respectively.

[0012] Preferably, the quadratic function for operation and maintenance cost is calculated as:

[0013]

[0014] Among them, P i Represents the active power output by the ith node; α i ,β i ,γ i They represent the cost coefficient of the i-th node respectively.

[0015] Preferably, the quadratic function for carbon trading cost is calculated as:

[0016]

[0017] Among them, ε represents the unit carbon price; E d,i and E q,i They represent the total carbon emissions and carbon emission quota of node i respectively; μ∈(0,1) represents the growth rate of the price corresponding to the increase of carbon emissions, r is an integer greater than or equal to 1, and E0 represents the carbon emissions corresponding to each carbon emission interval.

[0018] Preferably, in the calculation of carbon trading costs: define the value of m as: Where m is the parameter of the quadratic function f2 of carbon trading costs, reflecting the step range of current carbon emissions.

[0019] Preferably, the constraints in the power system include:

[0020] First,

[0021] Among them, Δ i is the power supply and demand deviation of each subsystem i, P D,i is the power demand of the ith subsystem; L loss,i is the amount of power transmission loss during the operation of the i-th subsystem, and its magnitude is L loss,i =σ i P i 2 , σ i represents the power transmission loss coefficient during the operation of the i-th subsystem;

[0022] Second, P i min ≤P i ≤P i max , where P i max ,P i min are the upper and lower limits of the output active power of subsystem i respectively.

[0023] Preferably, the optimal active power of the subsystem is:

[0024]

[0025] Among them, λ * The incremental cost when it is the optimal solution.

[0026] Preferably, the incremental cost λ of power subsystem i * The specific calculation is:

[0027] Create the Lagrangian function:

[0028]

[0029] Among them, ω1 and ω2 are weighting coefficients, whose values ​​are between 0 and 1 and satisfy ω1+ω2=1. represents the supply and demand deviation of the power system; λ is the Lagrange multiplier under the supply and demand balance of the power system; and u i ≥0 are the Lagrange multipliers of the inequality constraints of subsystem i;

[0030] Assume Pi * (i=1,2,...,N) is the local minimum point of the Lagrangian function, then there is a unique λ * , u i * (i=1,2,...,N) makes the KKT condition true:

[0031]

[0032] Definition of Ω P ={i|P i =P i min ,P i =P i max},Ω max ={i|P i =P i max} and Ω min ={i|P i =P i min According to the KKT condition of the optimal system, the incremental cost λ of power subsystem i can be obtained: * :

[0033]

[0034] Among them, θ i represents the carbon emission quota coefficient of node i, ε i is the carbon emission quota of subsystem i.

[0035] Preferably, an estimated variable y of the supply-demand deviation is added to each node of the subsystem i (k), introduce two surplus variables for each subsystem i and Among them, let y i (0) = Δ i (0) and Then there is and

[0036] Preferably, the scheduling based on consistency is specifically:

[0037] First, node i (i=1,2,...,N) follows:

[0038]

[0039] The incremental cost variable λ i (k) and surplus variables Iteration of λj (k) represents the information state of node j at the kth iteration; σ(k) represents the convergence coefficient, whose value will affect the convergence speed, and σ(k) = ρ k And ρ∈(0,1); parameter v>0,∈>0 is the scheduling gain to be designed, and the convergence speed of the scheduling algorithm can be adjusted by adjusting the gain parameter;

[0040] Second, node i (i=1,2,...,N) follows:

[0041]

[0042] Iterate the active power output, where And according to Δ i (k+1)=σ i (P i (k+1)) 2 +P D,i (k+1)-P i (k+1) Calculate the supply and demand deviation variable Δ i (k+1);

[0043] Third, node i (i = 1, 2, ..., N) estimates the estimated variable y of the supply and demand deviation i (k+1) and surplus variables

[0044]

[0045] When |λ i (k+1)-λ i When (k)|<ξ, the consistency iteration of node i (i=1,2,...,N) ends, and the node is recorded as λ * =λ i (k+1), and calculate the optimal active power, where ξ>0 is the given dispatching accuracy requirement value, which is used to characterize the accuracy of power dispatching.

[0046] Beneficial effects: In the present invention, a quadratic function is used to calculate the operation and maintenance cost and carbon trading cost of the subsystem, and the Lagrange multiplier method is introduced to solve the optimization problem; secondly, the optimal active power of the scheduling optimization subsystem is based on consistency, which is suitable for large-scale complex power networks. Based on the consistency theory, the incremental cost is regarded as a consistency variable, and distributed scheduling is realized through iterative calculation. Each node makes decisions based on local information and simple communication with neighboring nodes, which greatly improves the robustness and flexibility of the algorithm suitable for large-scale and complex power networks; the entire system achieves net zero emissions, and can achieve the optimal balance between cost-effectiveness and environmental benefits while ensuring the stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0048] In the attached picture:

[0049] Figure 1 It is a framework diagram of the power system of the present invention;

[0050] Figure 2 is a diagram of the IEEE-14 system structure of the present invention;

[0051] Figure 3 It is the IEEE-14 directed communication topology diagram of the present invention;

[0052] Figure 4 It is an iterative curve diagram of the electricity incremental cost of each node of the present invention;

[0053] Figure 5 It is an iterative curve diagram of the power supply of each node of the present invention;

[0054] Figure 6 is an iterative curve diagram of supply and demand deviation of the power system of the present invention;

[0055] Figure 7 It is the power supply incremental cost (ρ=0.99) diagram of each node of the present invention;

[0056] Figure 8 It is the power supply incremental cost (ρ=0.95) diagram of each node of the present invention;

[0057] Fig. 9 It is a graph of the power supply incremental cost (ρ=0.90) of each node of the present invention. DETAILED DESCRIPTION

[0058] The embodiments of the present invention are described below in conjunction with the drawings in the embodiments of the present invention. The terms used in the implementation mode of the present invention are only used to explain the specific embodiments of the present invention, and are not intended to limit the present invention. The embodiments of the present application are described below in conjunction with the drawings.

[0059] For power system: Assume that the power system under study has N subsystems, and a local intelligent agent (also called node) is embedded in each subsystem, which is denoted as node i (i = 1, 2, ..., N). Use a directed graph G = (V, E, A, B) to represent the communication relationship between N nodes in the system, where V = {1, 2, ..., N} represents the node set, Represents the edge set, A,B∈R N×Ndenote the input and output weight matrices respectively. A directed edge (j,i)∈E indicates that node i receives information from node j, that is, information flows from node j to node i. For any node i∈V, Input the neighbor node set, Output the neighbor node set for it. In general, there is And there is and That is, there is no self-loop at node i. For the input weight matrix A, if Then A ij ∈(0,1); otherwise Then A ij =0. For the output weight matrix B, if Then B ij ∈(0,1); otherwise Then B ij = 0. The in-degree of node i is out degree In-degree, the in-degree matrix and out-degree matrix of the system are D = diag (d1, d2, ..., d N )and Let L = DA and Then L1 N =0 N and That is, the matrix L and The row sum and column sum are zero respectively;

[0060] A fully distributed power system dispatching method for arbitrary directed communication topology, see Figure 1 As shown in the figure, the control end at the top layer of the power system framework focuses on the monitoring and optimization of the power system to meet zero-carbon emissions and power load requirements. On the power generation side, the diagram shows a variety of clean energy generation methods, including thermal energy, hydropower, wind power and solar power generation. These power generation methods are not only diverse, but also renewable, which helps to achieve the goal of zero carbon emissions. On the power transmission side, electricity is transmitted from the power station through the transmission network, and after being adjusted by the transformer, it is delivered to power consumers and energy storage equipment. The power storage and consumer layer mainly includes electric vehicles, smart homes, smart cities, and industrial plants. Electricity can also be stored in the form of batteries for use in powering electric vehicles, including:

[0061] The operation and maintenance costs and carbon trading costs of the subsystem are calculated using quadratic functions;

[0062] For subsystem i, the following quadratic function is used:

[0063] Operation and maintenance costs:

[0064] Among them, P i Represents the active power output by the ith node; α i ,β i ,γ i They represent the cost coefficient of the i-th node respectively.

[0065] The quadratic function of carbon trading cost is represented by f2. At present, the research on carbon trading mechanism is mainly divided into traditional and step-by-step. The traditional carbon trading mechanism is mainly related to the output power of the system and will not have a significant effect on carbon emission reduction. For node i, its traditional carbon trading mechanism function is:

[0066] Where ε represents the unit carbon price; E d,i and E q,i They represent the total carbon emissions and carbon emission quota of node i respectively; and θ i They represent the carbon emission intensity coefficient and carbon emission quota coefficient of node i respectively.

[0067] Different from the traditional carbon trading mechanism, the tiered carbon trading mechanism divides the carbon emission intensity into multiple intervals and designs different unit carbon prices ε for different intervals. Its main purpose is to increase the carbon emission cost through the increase of ε as the carbon emission increases, so as to limit the carbon emission of the system. Basis:

[0068] E q,i +(r-1)E0<E d,i ≤E q,i +rE0 (3);

[0069] To determine the carbon emission interval r, according to Construct a penalty coefficient that increases with the increase of r, and establish a step-by-step carbon trading function:

[0070]

[0071] Where r is an integer greater than or equal to 1; E0 represents the carbon emissions corresponding to each carbon emission interval, which is represented as a fixed value here; μ∈(0,1) represents the growth rate corresponding to the price as the carbon emissions increase. In formula (4), for the sake of simplicity, according to:

[0072]

[0073] Define the value of variable m.

[0074] For the constraints in the power system, according to: Δ i =L loss,i +P D,i -P i (6);

[0075] Define the power supply and demand deviation Δ of each subsystem i i , where P D,i is the power demand of the ith subsystem; L loss,i is the amount of power transmission loss during the operation of the i-th subsystem, and its magnitude is L loss,i =σ i P i 2 , and σ i represents the power transmission loss coefficient in the operation of the i-th subsystem. Then the supply and demand balance constraint of the entire power system is:

[0076]

[0077] In order to ensure the stable operation of the system and meet the demand, the power output of the system must meet the upper and lower limit constraints during operation: P i min ≤P i ≤P i max (8);

[0078] Where: P i max ,P i min are the upper and lower limits of the output active power of subsystem i respectively.

[0079] The low-carbon economic dispatching goal of the power system is to optimize the output of each unit, thereby minimizing the system's operating cost and carbon trading cost; establish the optimization function: minF = ω1f1 + ω2f2 (9);

[0080] Wherein ω1 and ω2 are weighting coefficients, whose values ​​are between 0 and 1 and satisfy the requirement of ω1+ω2=1.

[0081] The Lagrange multiplier method is used to solve the optimization problem with equality and inequality constraints. By introducing the Lagrange multiplier, the inequality constraint is transformed into an equality constraint, and the KKT (Karush-Kuhn-Tucker) condition is combined for analysis. The Lagrange function is established as follows:

[0082]

[0083] Where: represents the supply and demand deviation of the power system; λ is the Lagrange multiplier under the supply and demand balance of the power system; and u i ≥0 are the Lagrange multipliers of the inequality constraints of subsystem i; assuming that P i *(i=1,2,...,N) is the local minimum point of equation (10), then there is a unique λ * , u i * (i=1,2,...,N) so that the KKT condition:

[0084]

[0085] Establish; define Ω P ={i|P i =P i min ,P i =P i max},Ω max ={i|P i =P i max} and Ω min ={i|P i =P i min According to the KKT condition of the optimal system, the incremental cost λ of power subsystem i can be obtained: * :

[0086]

[0087] At the same time, if i∈Ω p , there is P i * =P i max or P i * =P i min . According to the system function balance constraint, we have:

[0088]

[0089] Established.

[0090] The optimal incremental cost λ can be solved by formula (11): * The incremental cost λ when combined with the optimal solution * , from formula (13) we can get Then the optimal active power output of subsystem i is:

[0091]

[0092] It should be noted that, when solving λ from equation (13) * The process is a centralized calculation process, which requires all subsystems to set their own parameters, including α i ,β i,θ i ,ε i ,P i max ,P i min Etc., and then transmit it to a centralized computing center. If the computing center fails or is untrustworthy, it will cause scheduling failure or data privacy leakage.

[0093] Consistency-based dispatch optimization subsystem optimal active power:

[0094] To avoid the incremental cost λ * The shortcomings of centralized computing, for the directed communication topology G, with an incremental cost λ i (k) is the consistency variable, denoted by Δ i (k), λ i (k) is the supply-demand deviation and incremental cost of node i at the kth iteration; an estimated variable y of the supply-demand deviation is added to each node i (k), considering that the communication network topology of the power system is an arbitrary directed graph, two surplus variables are introduced for each subsystem i and Let y i (0) = Δ i (0) and So there and

[0095] The consistency-based scheduling process consists of three steps: consistency iteration of states, calculation of active power, and estimation of power inconsistency.

[0096] First, node i (i=1,2,...,N) follows:

[0097]

[0098] The incremental cost variable λ i (k) and surplus variables Iteration of λ j (k) represents the information state of node j at the kth iteration; σ(k) represents the convergence coefficient, whose value will affect the convergence speed, and σ(k) = ρ k And ρ∈(0,1); parameter v>0, ∈>0 is the scheduling gain to be designed, and the convergence speed of the scheduling algorithm can be adjusted by adjusting the gain parameter.

[0099] Next, node i (i = 1, 2, ..., N) is

[0100]

[0101] Iterate the active power output, where And according to Δ i (k+1)=σ i (P i (k+1)) 2 +P D,i (k+1)-P i (k+1) Calculate the supply and demand deviation variable Δ i (k+1);

[0102] Furthermore, node i (i = 1, 2, ..., N) estimates the estimated variable y of the supply-demand deviation i (k+1) and surplus variables

[0103]

[0104] Finally, when |λ i (k+1)-λ i When (k)|<ξ, the consistency iteration of node i (i=1,2,...,N) ends, and the node is recorded as λ * =λ i (k+1), and the final optimized output P is obtained according to formula (12) i * , where ξ>0 is the given dispatching accuracy requirement value, which is used to characterize the accuracy of power dispatching.

[0105] For the convergence of scheduling algorithms:

[0106] Condition 1 The communication topology G of the power system is a strongly connected directed graph, that is, there is a communication path between any two nodes in the network G;

[0107] Condition 2: Input adjacency weight A between nodes i and j ij and output neighbor weight B ij according to:

[0108]

[0109] Lemma 1 The distributed consensus algorithm (1) achieves average consistency if and only if 1 is a single-valued eigenvalue of the matrix M and the moduli of all other eigenvalues ​​are less than 1.

[0110] When conditions 1 and 2 are met, if the scheduling gain further satisfies: 0<v<1, 0<∈<1(19);

[0111] Then the consistency scheduling algorithm (15);

[0112]

[0113] Convergence, or the consistency of the incremental costs of the power system, is:

[0114]

[0115] And there is and

[0116] Based on the above: Through numerical simulation experiments and IEEE-14 system simulation experiments; In order to verify the effectiveness of the proposed algorithm, the IEEE-14 node system is used to verify the effectiveness of the algorithm. Figure 2 This is the IEEE-14 system structure diagram. Figure 3 is the IEEE-14 directed communication topology diagram; based on the IEEE-14 system simulation, nodes 1, 2, 3, 6, and 8 represent generators, and the remaining nodes represent loads. Considering that in the power system, nodes 4, 5, 7, 10, 11, 12, 13, and 14 do not have any power generation capacity, and are just pure load nodes. Load nodes are consumption points rather than supply points, and will not provide or absorb any power to the grid, that is, they will not participate in the power supply or frequency regulation of the grid. Therefore, P is set i,min =P i,max = 0. The initial power and detailed parameter settings are shown in the following table:

[0117]

[0118] This example verifies the effectiveness of the fully distributed low-carbon economic dispatch algorithm for power systems with arbitrary communication topology in the IEEE-14 node system. In the IEEE-14 node simulation system, the feedback gain g is set to 0.8 between (0, 1), and the other parameters are set to: ρ = 0.85, and the algorithm is iterated 100 times. The simulation results are shown in Figure 4-Figure 6 As shown. Figure 4 It represents the iterative curve of the electricity incremental cost of each node in the IEEE-14 system, Figure 5 Describe the power supply iteration curve of each node in the IEEE-14 system, Figure 6 represents the supply and demand deviation iteration curve of the IEEE-14 system power system; Figure 4-Figure 6 It can be seen that in the iterative process of the algorithm, after the 27th iteration, the algorithm has reached the optimal state. Figure 4 It can be seen that when the iteration converges, the incremental power supply cost of all nodes reaches the same value, and the incremental cost λ1(k)=λ2(k)=λ3(k)=...=λ * =128.102 yuan / MW, at this time all nodes will output at the same incremental rate principle, and the IEEE-14 node system is in the optimal economic operation state. Figure 5It can be seen that when each node runs for the 27th time, its power supply will no longer change. At this time, P i (λ i (k)) = P i (λ * ) is the optimal value. Figure 6 It can be seen that the economic dispatch problem solved by the fully distributed system satisfies the supply and demand balance, and the entire system operates in a power balance state. After 27 iterations, the power mismatch eventually approaches 0, which means that the difference between the power generation and the load demand in the system gradually decreases, and finally reaches a relatively balanced state. The effectiveness of the fully distributed economic dispatch with a directed communication topology is verified.

[0119] By changing the value of the convergence factor ρ, the convergence speed of the gain cost of each node is observed, so as to obtain the influence of the convergence factor on the convergence speed.

[0120] When ρ=0.99, ρ=0.95, ρ=0.9, the incremental cost iterations are as follows: Figure 7-Figure 9 shown.

[0121] Depend on Figure 7-9 It can be seen that 1) when ρ = 0.99, after 44 iterations, the incremental power supply cost of all nodes converges to a consistent value of 93.6825 yuan / MW. 2) when ρ = 0.95, after 40 iterations, the incremental power supply cost of all nodes converges to a consistent value of 93.8734 yuan / MW. 3) when ρ = 0.90, the incremental power supply cost is as follows Fig. 9 As shown in Figure 2, after 37 iterations, the incremental power supply cost of all nodes converges to a consistent value of 103.436 yuan / MW. In summary, by increasing the convergence factor ρ to 0.99, the Figure 7-Figure 9 It can be seen that the convergence speed of the incremental power supply cost slows down, that is, as the convergence factor ρ increases, the convergence speed of the scheduling algorithm becomes slower.

[0122] The embodiments of the present invention are described in detail above in conjunction with the accompanying drawings, but the present invention is not limited to the above embodiments. For ordinary technicians in this technical field, after knowing the contents recorded in the present invention, they can make several equivalent changes and substitutions without departing from the principle of the present invention. These equivalent changes and substitutions should also be regarded as belonging to the protection scope of the present invention.

Claims

1. A fully distributed power system dispatching method applicable to any directed communication topology, wherein the power system comprises several subsystems, each of which is embedded in a node, and is characterized in that: The scheduling method includes the following steps: The operation and maintenance costs and carbon trading costs of the subsystem are calculated using quadratic functions; Based on the operation and maintenance cost and carbon trading cost, Lagrange multipliers are introduced to transform the inequality constraints into equality constraints, and the optimal active power of the subsystem is obtained by combining the KKT conditions. as well as Consistency-based dispatch optimization subsystem optimal active power, including: Add estimation variables of supply-demand deviation and introduce surplus variables for each subsystem; Combine the estimated variables of supply and demand deviation and the surplus variables to perform consistency iteration of the state, calculation of active power and estimation of power inconsistency respectively; Add an estimated variable y of the supply and demand deviation to each node of the subsystem i (k), introduce two surplus variables for each subsystem i and Among them, let y i (0) = Δ i (0) and Then there is and Consistency-based scheduling is as follows: First, node i (i=1,2,...,N) follows: The incremental cost variable λ i (k) and surplus variables Iteration of λ j (k) represents the information state of node j at the kth iteration; σ(k) represents the convergence coefficient, whose value will affect the convergence speed, and σ(k) = ρ k And ρ∈(0,1); parameter ν>0, ∈>0 is the scheduling gain to be designed, and the convergence speed of the scheduling algorithm can be adjusted by adjusting the gain parameter; Second, node i (i=1,2,...,N) follows: Perform the iteration of active power output, where And according to Δ i (k+1)=σ i (P i (k+1)) 2 +P D,i (k+1)-P i (k+1) Calculate the supply and demand deviation variable Δ i (k+1), P i represents the active power output by the ith node, P D,i is the power demand of the ith subsystem; Third, node i (i = 1, 2, ..., N) estimates the estimated variable y of the supply and demand deviation i (k+1) and surplus variables When |λ i (k+1)-λ i When (k)|<ξ, the consistency iteration of node i (i=1,2,...,N) ends, and the node is recorded as λ * =λ i (k+1), and calculate the optimal active power, where ξ>0 is the given dispatching accuracy requirement value, which is used to characterize the accuracy of power dispatching; Among them, P i Represents the active power output by the ith node; α i ,β i They represent the cost coefficient of the i-th node respectively; m is the parameter of the carbon trading cost quadratic function f2, reflecting the step interval of the current carbon emission; P i max ,P i min are the upper and lower limits of the output active power of subsystem i respectively; * is the incremental cost of power subsystem i; ω1 and ω2 are weighting coefficients; θ i represents the carbon emission quota coefficient of node i, ε i is the carbon emission quota of subsystem i.

2. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 1, characterized in that: The quadratic function for operation and maintenance cost is calculated as: Among them, γ i represents the cost coefficient of the i-th node.

3. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 2, characterized in that: The quadratic function for carbon trading costs is calculated as: Among them, ε represents the unit carbon price; E d,i and E q,i They represent the total carbon emissions and carbon emission quota of node i respectively; μ∈(0,1) represents the growth rate of the price corresponding to the increase of carbon emissions, r is an integer greater than or equal to 1, and E0 represents the carbon emissions corresponding to each carbon emission interval.

4. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 3 is characterized in that: In the calculation of carbon trading costs: the value of m is defined as:

5. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 3, characterized in that: The constraints in the power system include: First, Among them, Δ i is the power supply and demand deviation of each subsystem i; L loss,i is the amount of power transmission loss during the operation of the i-th subsystem, and its magnitude is L loss,i =σ i P i 2 , σ i represents the power transmission loss coefficient during the operation of the i-th subsystem; Second, P i min ≤P i ≤P i max .

6. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 5, characterized in that: The optimal active power of the subsystem is: Among them, λ * The incremental cost when it is the optimal solution.

7. The fully distributed power system dispatching method applicable to any directed communication topology according to claim 6, characterized in that: λ * The specific calculation is: Create the Lagrangian function: Among them, ω1 and ω2 are between 0 and 1 and satisfy ω1+ω2=1. represents the supply and demand deviation of the power system; λ is the Lagrange multiplier under the supply and demand balance of the power system; and u i ≥0 are the Lagrange multipliers of the inequality constraints of subsystem i; Assume P i * (i=1,2,...,N) is the local minimum point of the Lagrangian function, then there is a unique λ * , u i * (i=1,2,...,N) makes the KKT condition true: Definition of Ω P ={i|P i =P i min ,P i =P i max },Ω max ={i|P i =P i max } and Ω min ={i|P i =P i min According to the KKT condition of the optimal system, the incremental cost λ of power subsystem i can be obtained: * :

Images