A multi-modal single-objective hierarchical distributed integrated energy system economic dispatch method

Abstract

The present invention proposes a multimodal single-objective hierarchical distributed integrated energy system economic dispatch method, which combines the multimodal single-objective method and the hierarchical distributed consistency method for the economic dispatch of the integrated energy system. First, the multimodal single-objective method in the method is used to obtain a variety of multimodal economic dispatch single-objective schemes. Second, the hierarchical distributed consistency method in the method uses hierarchical operations to quickly obtain accurate economic dispatch decision schemes. The multimodal single-objective hierarchical distributed integrated energy system economic dispatch method can solve the multimodal problem of economic dispatch of large-scale integrated energy systems, diversify the decision schemes of economic dispatch, optimize the stability of economic dispatch, and improve the calculation speed and accuracy.

Description

Technical Field

[0001] The present invention belongs to the field of integrated energy of power systems, relates to a multi-modal single-objective hierarchical distributed method, and is suitable for economic dispatching of integrated energy systems. Background Art

[0002] The continuous expansion of the electricity market and the gradual increase in the scale of integrated energy systems will lead to slower computational speeds and reduced information privacy in the economic dispatch of integrated energy systems. Current research on single-objective economic dispatch of integrated energy systems fails to consider multimodal issues, resulting in a lack of diversity and comprehensiveness in economic dispatch solutions.

[0003] Therefore, a multimodal single-objective hierarchical distributed economic dispatch method for integrated energy systems is proposed. This method can retain the multimodal characteristics. In the multimodal single-objective hierarchical distributed economic dispatch problem of integrated energy systems, it can not only speed up the system's problem processing speed and ensure the privacy and robustness of the system, but also provide multimodal options for the dispatch of integrated energy systems, ensuring the stability and security of economic dispatch. Summary of the Invention

[0004] The present invention proposes a multimodal single-objective hierarchical distributed integrated energy system economic dispatch method, which combines the multimodal single-objective method with the hierarchical distributed consistency method for the economic dispatch of the integrated energy system, thereby increasing the diversity and stability of economic dispatch, improving the calculation speed and accuracy, and providing decision makers with a rich set of alternative decision-making solutions. The steps in the use process are as follows:

[0005] Step (1): The economic dispatch model of the integrated energy system is constructed, which includes thermal energy, hydropower, wind energy, solar energy and geothermal energy, and takes cost expenditure as the target, and follows the equality and inequality constraints of economic dispatch;

[0006] The objective function of the total power generation cost is:

[0007]

[0008] Where f(x) is the total power generation cost; represents the cost of the i-th generator set at time t; represents the cost of the jth thermal power generating unit at time t; represents the cost of the kth hydroelectric generator unit at time t; represents the cost of the lth wind turbine generator set at time t; represents the cost of the rth photovoltaic generator set at time t; represents the cost of the sth geothermal generator unit at time t; represents the cost of the u-th clean energy generator at time t; N is the total number of generators; N TG is the number of thermal power generator units; N HG is the number of hydropower generator sets; N WG is the number of wind turbine generator sets; N PG is the number of photovoltaic generator sets; N OG is the number of geothermal generator units; N CleanG is the number of clean energy generator sets; T is the statistical time of the objective function; is the power generation of the i-th generator set at time t; is the power generation of the jth thermal power generating unit at time t; is the power generation of the kth hydropower generator unit at time t; is the power generation of the lth wind turbine generator set at time t; is the power generation of the rth photovoltaic generator set at time t; is the power generation of the sth geothermal generator unit at time t; is the power generation of the uth clean energy generator at time t; a j 、b j and c j is the cost coefficient of the jth thermal power generating unit; b k is the cost coefficient of the kth hydropower generating unit; e l is the cost coefficient of the first wind turbine generator set; g r is the cost coefficient of the rth photovoltaic generator set; m s is the cost coefficient of the sth geothermal power generation unit; β u is the cost coefficient of the u-th clean energy generator set; the cost of the clean energy generator set is Replacement hydropower costs Wind power costs Light Energy and geothermal energy costs

[0009] The equality constraint for power balance is:

[0010]

[0011] in, is the total load demand value at time t;

[0012] The upper and lower limits of power are:

[0013]

[0014] in, is the lower limit of the power generation of the j-th thermal power generating unit; is the upper limit of the power generation capacity of the j-th thermal power generating unit; is the lower limit of the power generation of the kth hydropower generator unit; is the upper limit of the power generation of the kth hydropower generator unit; is the lower limit of the power generation of the first wind turbine generator set; is the upper limit of the power generation of the first wind turbine generator set; is the lower limit of the power generation of the rth photovoltaic generator set; is the upper limit of the power generation of the rth photovoltaic generator set; is the lower limit of the power generation capacity of the sth geothermal generator unit; is the upper limit of the power generation capacity of the sth geothermal power generating unit;

[0015] The ramp rate constraint of thermal power units is:

[0016]

[0017] in, is the power generation of the jth thermal power generating unit at time t-1; is the downward climbing rate of the jth thermal power unit; is the upward climbing rate of the jth thermal power unit; T 60 60 minutes;

[0018] The time constraint of the photoelectric motor group is:

[0019]

[0020] Step (2): Set the initial iteration time t=1;

[0021] Step (3): Input the predicted load value and set the initial iteration number k = 0;

[0022] Step (4): Divide the integrated energy system into four layers: the first layer is divided into region A, region B and region C; the second layer is divided into region Aa, region Ab, region Ac, region Ba, region Bb, region Bc, region Ca and region Cb; the third layer is divided into region Aa1, region Aa2, region Ab1, region Ab2, region Ac1, region Ac2, region Ba1, region Ba2, region Ba3, region Bb1, region Bb2, region Bc1, region Bc2, region Ca1, region Ca2, region Cb1, region Cb2 and region Cb3; the fourth layer is the third layer of each region Generator group; Step (5): Select the leader and follower agents of each layer, the first layer area A is the leader, and area B and area C are followers; the second layer area Aa, area Ba and area Ca are leaders, and the other second layer areas are followers; the third layer area Aa1, area Ab1, area Ac1, area Ba1, area Bb1, area Bc1, area Ca1 and area Cb1 are leaders, and the other third layer areas are followers; Step (6): Update the consistency variables of the first layer agents in combination with the topological structure diagram of the agents and formulas (12) and (13);

[0023] The directed graph G=(V,E) of the distributed network topology and the adjacency matrix A=(a ij ) constitutes a multi-agent network topology; the Laplace matrix L of the directed graph G = [l ij ]for:

[0024]

[0025] Among them, l ii and l ij is an element in the Laplace matrix L; a ij is an element in the adjacency matrix A;

[0026] The agent's row random matrix H ij for:

[0027]

[0028] The consistency variable λ is expressed as:

[0029] λ=[λ T ,λ Clean ] (9)

[0030]

[0031]

[0032] Where λ represents the consistency variable of the integrated energy system; λ Trepresents the cost increment of thermal power generating units; Clean represents the consistency variable of the clean energy generator set;

[0033] The consistency variable update formulas for followers and leaders are:

[0034]

[0035]

[0036] Among them, N IB represents the number of agents; k represents the kth iteration; represents the consistency variable of the nth agent at the kth iteration at time t; represents the consistency variable of the mth agent at the k+1th iteration at time t; τ represents the power balance adjustment factor of the distributed consensus method; ΔP t (k) represents the power deviation of the kth iteration at time t; H mn (k) represents the row random matrix of the agent;

[0037] Step (7): Calculate the active power output of each agent in the first layer according to formula (14);

[0038] The active power formula of each unit is:

[0039]

[0040] in, represents the active power output of the i-th agent at time t in the k+1 iteration; represents the active power output of the jth thermal power generating unit at time t in the k+1 iteration; represents the active power output of the u-th clean energy generator at time t in the k+1 iteration; It is expressed as the upper limit of active power of the μth area in the ωth layer, which is the sum of the upper limits of active power of the lower sub-areas or units divided by this area; represents the consistency variable of the i-th agent at time t in the k+1 iteration; represents the cost increment of the j-th thermal power generating unit at time t in the k+1 iteration; represents the consistency variable of the u-th clean energy unit at time t in the k+1 iteration; represents the number of thermal power units in the μth region of the ωth layer; represents the number of clean energy generating units in the μth region of the ωth layer; U represents the total number of regions in the layer; W represents the total number of layers;

[0041] Step (8): Correct the first layer active power output obtained in step (7) according to formula (15);

[0042] The correction formula for active power is:

[0043]

[0044] in, represents the lower limit of active power output of the i-th power generation agent; represents the upper limit of active power output of the i-th power generation agent;

[0045] Step (9): Calculate the active power deviation value of the first layer according to formula (16);

[0046] Active power deviation ΔP t (k+1) is:

[0047]

[0048] Step (10): Determine whether the multi-modal active power deviation value of the first layer obtained in step (9) meets the accuracy requirements

[0049]

[0050] Where δ represents the maximum allowable power deviation; Indicates the allowable value of multi-modal power deviation;

[0051] if Then set the number of iteration steps k = k + 1, and return to step (6) to continue the iterative calculation of the first layer;

[0052] if The obtained active output of a group of agents is stored, and then the number of iteration steps k = k + 1 is set, and the process returns to step (6) to continue the iterative calculation of the first layer;

[0053] if The multimodal active power deviation value of the first layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent agents is obtained;

[0054] Step (11): Perform multi-modal judgment and decision congestion distance screening on the active power outputs of the multiple groups of intelligent generators obtained in step (10) to obtain multi-modal active power outputs that meet the requirements of the decision maker;

[0055] Multimodal judgment is:

[0056]

[0057] Among them, x1 and x2 are two sets of multimodal solutions; ε represents the critical threshold of multimodality;

[0058] The decision space crowding distance screening is:

[0059]

[0060] in, represents the crowding distance of the jth decision variable of the i-th solution; x i+1,j represents the jth decision variable of the i+1th solution; x i,j represents the jth decision variable of the i-th solution; The jth decision variable representing the maximum objective value; The jth decision variable representing the minimum objective value; x i-1,j represents the j-th decision variable of the i-1-th group of solutions; represents the crowding distance of the i-th group of solutions; D represents the number of decision variables; ξ represents the minimum decision crowding distance allowed by multimodality;

[0061] Step (12): Input the multiple groups of multi-modal power generation of the first layer obtained in step (11) as load values ​​into the second layer, and set the initial iteration step number k = 0;

[0062] Step (13): Update the consistency variables of the second-layer agents based on the topological structure diagram of the agents and formulas (12) and (13);

[0063] Step (14): Calculate the active power output of each intelligent unit in the second layer according to formula (14);

[0064] Step (15): Correct the active power obtained in step (14) according to formula (15);

[0065] Step (16): Calculate the second layer active power deviation value according to formula (16);

[0066] Step (17): determining whether the multi-modal active power deviation value obtained in step (16) meets the accuracy requirement;

[0067] if Then set the number of iteration steps k = k + 1, and return to step (13) to continue the iterative calculation of the second layer;

[0068] if The obtained active output of a group of agents is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (13) to continue the iterative calculation of the second layer;

[0069] if The multi-modal active power deviation value of the second layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent generator sets is obtained;

[0070] Step (18): Perform multimodal judgment and decision crowding distance screening on the active power output of multiple groups of intelligent agents in the second layer obtained in step (17) to obtain multimodal active power output that meets the requirements of the decision maker;

[0071] Step (19): Input the multiple groups of multi-modal power generation of the second layer obtained in step (18) as load values ​​into the third layer, and set the initial iteration step number k=0;

[0072] Step (20): Update the consistency variables of the third-layer agent based on the agent's topological structure diagram and formulas (12) and (13);

[0073] Step (21): Calculate the active power output of each intelligent unit in the third layer according to formula (14);

[0074] Step (22): Correct the active power output obtained in step (21) according to formula (15);

[0075] Step (23): Calculate the third layer active power deviation value according to formula (16);

[0076] Step (24): determining whether the multi-modal active power deviation value obtained in step (23) meets the accuracy requirement;

[0077] if Then set the number of iteration steps k = k + 1, and return to step (20) to continue the iterative calculation of the third layer;

[0078] if The obtained active output of a group of agents is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (20) to continue the iterative calculation of the third layer;

[0079] if The multi-modal active power deviation value of the third layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent generators is obtained;

[0080] Step (25): Perform multimodal judgment and decision crowding distance screening on the active power output of multiple groups of intelligent agents in the third layer obtained in step (24) to obtain multimodal active power output that meets the requirements of the decision maker;

[0081] Step (26): input the multiple groups of multi-modal power generation of the third layer obtained in step (25) as load values ​​into the fourth layer, and set the initial iteration step number k=0;

[0082] Step (27): Update the consistency variables of each generator set in the fourth layer by combining the topological structure diagram of the intelligent agent and formulas (12) and (13); Step (28): Calculate the active power output of each generator set in the fourth layer according to formula (14);

[0083] Step (29): Correct the active power output obtained in step (28) according to formula (15);

[0084] Step (30): Calculate the fourth layer active power deviation value according to formula (16);

[0085] Step (31): determining whether the multi-modal active power deviation value obtained in step (30) meets the accuracy requirement;

[0086] if Then set the number of iteration steps k = k + 1, and return to step (27) to continue the iterative calculation of the fourth layer;

[0087] if The obtained active output of a group of generator sets is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (27) to continue the iterative calculation of the fourth layer;

[0088] if The multi-modal active power deviation value of the fourth layer meets the requirements, the iteration ends, and the active power output of multiple generator sets is obtained;

[0089] Step (32): Perform multi-modal judgment and decision congestion distance screening on the active output of the plurality of generator sets in the fourth layer obtained in step (31), and obtain a multi-modal active output decision scheme that meets the requirements of the decision maker;

[0090] Step (33): Determine whether t≤T is satisfied. If so, the iteration ends; if not, proceed to step (3).

[0091] The present invention has the following advantages and effects compared to the prior art:

[0092] (1) An analytical framework for a multimodal, single-objective, hierarchical distributed approach is established for the economic dispatch problem of integrated energy systems.

[0093] (2) The constructed multi-modal single-objective hierarchical distributed method adopts hierarchical scheduling and only exchanges consistent variables in adjacent areas. It does not need to exchange detailed operating information of each unit, which speeds up the calculation speed and ensures information privacy. It also solves the multi-modal characteristics problem in the scheduling of integrated energy systems and increases the diversity and comprehensiveness of economic scheduling solutions.

[0094] (3) The economic dispatch schemes of the integrated energy system obtained contain multiple interchangeable economic dispatch schemes. When faced with some sudden accidents or serious wind, solar, and water abandonment in some areas, decision makers have more options to choose from, which increases the stability and security of economic dispatch. BRIEF DESCRIPTION OF THE DRAWINGS

[0095] Figure 1 It is a hierarchical distributed schematic diagram of the method of the present invention.

[0096] Figure 2 It is a multi-modal single-objective distributed consistency flow chart of the method of the present invention.

[0097] Figure 3 It is an overall flow chart of the method of the present invention. DETAILED DESCRIPTION

[0098] The present invention proposes a multi-modal single-objective hierarchical distributed integrated energy system economic dispatch method, which is described in detail with reference to the accompanying drawings as follows:

[0099] Figure 1 This is a schematic diagram of the hierarchical distributed system of the present invention. The integrated energy system is divided into four layers: the first layer is divided into regions A, B, and C; the second layer is divided into regions Aa, Ab, Ac, Ba, Bb, Bc, Ca, and Cb; the third layer is divided into regions Aa1, Aa2, Ab1, Ab2, Ac1, Ac2, Ba1, Ba2, Ba3, Bb1, Bb2, Bc1, Bc2, Ca1, Ca2, Cb1, Cb2, and Cb3; the fourth layer comprises the generators within each region of the third layer. Each layer selects a leader and followers.

[0100] Figure 2 This is a flow chart of the multimodal single-objective distributed consistency method of the present invention. The specific steps are:

[0101] Step 1: Input the predicted load value and set the initial iteration number k = 0;

[0102] Step 2: Determine the leader and follower of the agent, and update the consistency variables of the agent based on the topological structure diagram of the agent and formulas (12) and (13);

[0103] Step 3: Calculate the active power output of each intelligent unit according to formula (14);

[0104] Step 4: Correct the active power output obtained in step 3 according to formula (15);

[0105] Step 5: Calculate the active power deviation value according to formula (16);

[0106] Step 6: Determine whether the multi-mode active power deviation value obtained in step 5 meets the accuracy requirements. If Then set the number of iterations k = k + 1 and return to step 2; if The obtained active output of a group of agents is stored, and then the number of iteration steps k=k+1 is set, and return to step 2; if The multi-modal active power deviation value meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent generator sets is obtained;

[0107] Step 7: Perform multimodal judgment and decision congestion distance screening on the active power output of multiple groups of intelligent generators obtained in step 6 to obtain a multimodal active power dispatching plan that meets the requirements of the decision maker.

[0108] Figure 3 It is the overall flow chart of the method of the present invention. Its specific steps are:

[0109] Step 1: Set the initial iteration time t=1;

[0110] Step 2: Input the predicted load value and set the initial iteration number k = 0;

[0111] Step 3: Divide the system into four layers. The first layer is divided into area A, area B, and area C; the second layer is divided into area Aa, area Ab, area Ac, area Ba, area Bb, area BC, area Ca, and area Cb; the third layer is divided into area Aa1, area Aa2, area Ab1, area Ab2, area Ac1, area Ac2, area Ba1, area Ba2, area Ba3, area Bb1, area Bb2, area Bc1, area Bc2, area Ca1, area Ca2, area Cb1, area Cb2, and area Cb3; the fourth layer is the generator sets in each area of ​​the third layer;

[0112] Step 4: Select the leader and follower agents in each layer. Region A in the first layer is the leader, and regions B and C are followers. Regions Aa, Ba, and Ca in the second layer are leaders, and the other second-layer regions are followers. Regions Aa1, Ab1, Ac1, Ba1, Bb1, Bc1, Ca1, and Cb1 in the third layer are leaders, and the other third-layer regions are followers.

[0113] Step 5: Exploitation Figure 2 The multimodal single-objective distributed consensus method in

[15] is used to optimize the first layer of the system.

[0114] Step 6: Input the multiple groups of multi-modal power generation of the three regions in the first layer obtained in step 5 as load values ​​into the second layer;

[0115] Step 7: Exploitation Figure 2 The multimodal single-objective distributed consensus method in

[15] is used to optimize the second layer of the system.

[0116] Step 8: Input the multiple groups of multi-modal power generation of the eight regions in the second layer obtained in step 5 as load values ​​into the third layer;

[0117] Step 9: Exploitation Figure 2 The multimodal single-objective distributed consensus method in

[15] is used to optimize the third layer of the system.

[0118] Step 10: Input the multi-modal power generation of the 18 regions in the third layer obtained in step 5 as load values ​​into the fourth layer. Treat each generator group in the fourth layer as an agent, and select the leader and follower agents.

[0119] Step 11: Exploitation Figure 2 The multimodal single-objective distributed consensus method in

[15] is used to optimize and solve the fourth layer of the system.

[0120] Step 12: Determine whether t≤T is satisfied. If so, the iteration ends; if not, go to step 2.

[0121] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A multi-modal single-objective hierarchical distributed integrated energy system economic dispatch method, characterized in that: The multimodal single-objective method and the hierarchical distributed consensus method are combined for the economic dispatch of integrated energy systems. This increases the diversity and stability of economic dispatch, improves the calculation speed and accuracy, and provides decision makers with a rich set of alternative decision-making solutions. The steps in the application process are as follows: Step (1): The economic dispatch model of the integrated energy system is constructed, which includes thermal energy, hydropower, wind energy, solar energy and geothermal energy, and takes cost expenditure as the target, and follows the equality and inequality constraints of economic dispatch; The objective function of the total power generation cost is: Where f(x) is the total power generation cost; represents the cost of the i-th generator set at time t; represents the cost of the jth thermal power generating unit at time t; represents the cost of the kth hydroelectric generator unit at time t; represents the cost of the lth wind turbine generator set at time t; represents the cost of the rth photovoltaic generator set at time t; represents the cost of the sth geothermal generator unit at time t; represents the cost of the u-th clean energy generator at time t; N is the total number of generators; N TG is the number of thermal power generator units; N HG is the number of hydropower generator sets; N WG is the number of wind turbine generator sets; N PG is the number of photovoltaic generator sets; N OG is the number of geothermal generator units; N CleanG is the number of clean energy generator sets; T is the statistical time of the objective function; is the power generation of the i-th generator set at time t; is the power generation of the jth thermal power generating unit at time t; is the power generation of the kth hydropower generator unit at time t; is the power generation of the lth wind turbine generator set at time t; is the power generation of the rth photovoltaic generator set at time t; is the power generation of the sth geothermal generator unit at time t; is the power generation of the uth clean energy generator at time t; a j 、b j and c j is the cost coefficient of the jth thermal power generating unit; d k is the cost coefficient of the kth hydropower generating unit; e l is the cost coefficient of the first wind turbine generator set; g r is the cost coefficient of the rth photovoltaic generator set; m s is the cost coefficient of the sth geothermal power generation unit; β u is the cost coefficient of the u-th clean energy generator set; the cost of the clean energy generator set is Replacement hydropower costs Wind power costs Light Energy and geothermal energy costs The equality constraint for power balance is: in, is the total load demand value at time t; The upper and lower limits of power are: in, is the lower limit of the power generation of the j-th thermal power generating unit; is the upper limit of the power generation capacity of the j-th thermal power generating unit; is the lower limit of the power generation of the kth hydropower generator unit; is the upper limit of the power generation of the kth hydropower generator unit; is the lower limit of the power generation of the first wind turbine generator set; is the upper limit of the power generation of the first wind turbine generator set; is the lower limit of the power generation of the rth photovoltaic generator set; is the upper limit of the power generation of the rth photovoltaic generator set; is the lower limit of the power generation capacity of the sth geothermal generator unit; is the upper limit of the power generation capacity of the sth geothermal power generating unit; The ramp rate constraint of thermal power units is: in, is the power generation of the jth thermal power generating unit at time t-1; is the downward climbing rate of the jth thermal power unit; is the upward climbing rate of the jth thermal power unit; T 60 60 minutes; The time constraint of the photoelectric motor group is: Step (2): Set the initial iteration time t=1; Step (3): Input the predicted load value and set the initial iteration number k = 0; Step (4): Divide the integrated energy system into four layers: the first layer is divided into region A, region B and region C; the second layer is divided into region Aa, region Ab, region Ac, region Ba, region Bb, region Bc, region Ca and region Cb; the third layer is divided into region Aa1, region Aa2, region Ab1, region Ab2, region Ac1, region Ac2, region Ba1, region Ba2, region Ba3, region Bb1, region Bb2, region Bc1, region Bc2, region Ca1, region Ca2, region Cb1, region Cb2 and region Cb3; the fourth layer is the third layer of each region Generator group; Step (5): Select the leader and follower agents of each layer, the first layer area A is the leader, and area B and area C are followers; the second layer area Aa, area Ba and area Ca are leaders, and the other second layer areas are followers; the third layer area Aa1, area Ab1, area Ac1, area Ba1, area Bb1, area Bc1, area Ca1 and area Cb1 are leaders, and the other third layer areas are followers; Step (6): Update the consistency variables of the first layer agents in combination with the topological structure diagram of the agents and formulas (12) and (13); The directed graph G=(V,E) of the distributed network topology and the adjacency matrix A=(a ij ) constitutes a multi-agent network topology; the Laplace matrix L of the directed graph G = [l ij ]for: Among them, l ii and l ij is an element in the Laplace matrix L; a ij is an element in the adjacency matrix A; The agent's row random matrix H ij for: The consistency variable λ is expressed as: λ=[λ T ,l Clean ] (9) Where λ represents the consistency variable of the integrated energy system; λ T represents the cost increment of thermal power generating units; Clean represents the consistency variable of the clean energy generator set; The consistency variable update formulas for followers and leaders are: Among them, N IB represents the number of agents; k represents the kth iteration; represents the consistency variable of the nth agent at the kth iteration at time t; represents the consensus variable of the mth agent at the k+1th iteration at time t; τ represents the power balance adjustment factor of the distributed consensus method; ΔP t (k) represents the power deviation of the kth iteration at time t; H mn (k) represents the row random matrix of the agent; Step (7): Calculate the active power output of each agent in the first layer according to formula (14); The active power formula of each unit is: in, represents the active power output of the i-th agent at time t in the k+1 iteration; represents the active power output of the jth thermal power generating unit at time t in the k+1 iteration; represents the active power output of the u-th clean energy generator at time t in the k+1 iteration; It is expressed as the upper limit of active power of the μth area in the ωth layer, which is the sum of the upper limits of active power of the lower sub-areas or units divided by this area; represents the consistency variable of the i-th agent at time t in the k+1 iteration; represents the cost increment of the j-th thermal power generating unit at time t in the k+1 iteration; represents the consistency variable of the u-th clean energy unit at time t in the k+1 iteration; represents the number of thermal power units in the μth region of the ωth layer; represents the number of clean energy generating units in the μth region of the ωth layer; U represents the total number of regions in the layer; W represents the total number of layers; Step (8): Correct the first layer active power output obtained in step (7) according to formula (15); The correction formula for active power is: in, represents the lower limit of active power output of the i-th power generation agent; represents the upper limit of active power output of the i-th power generation agent; Step (9): Calculate the active power deviation value of the first layer according to formula (16); Active power deviation ΔP t (k+1) is: Step (10): Determine whether the multi-modal active power deviation value of the first layer obtained in step (9) meets the accuracy requirements Where δ represents the maximum allowable power deviation; Indicates the allowable value of multi-modal power deviation; if Then set the number of iteration steps k = k + 1, and return to step (6) to continue the iterative calculation of the first layer; if The obtained active output of a group of agents is stored, and then the number of iteration steps k = k + 1 is set, and the process returns to step (6) to continue the iterative calculation of the first layer; if The multimodal active power deviation value of the first layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent agents is obtained; Step (11): Perform multi-modal judgment and decision congestion distance screening on the active power outputs of the multiple groups of intelligent generators obtained in step (10) to obtain multi-modal active power outputs that meet the requirements of the decision maker; Multimodal judgment is: |f(x1)-f(x2)|<ε (18) where x1 and x2 are two sets of multimodal solutions; ε represents the critical threshold of the multimodal solution. The decision space crowding distance screening is: in, represents the crowding distance of the jth decision variable of the i-th solution; x i+1,j represents the jth decision variable of the i+1th solution; x i,j represents the jth decision variable of the i-th solution; The jth decision variable representing the maximum objective value; The jth decision variable representing the minimum objective value; x i-1,j represents the j-th decision variable of the i-1-th group of solutions; represents the crowding distance of the i-th group of solutions; D represents the number of decision variables; ξ represents the minimum decision crowding distance allowed by multimodality; Step (12): Input the multiple groups of multi-modal power generation of the first layer obtained in step (11) as load values ​​into the second layer, and set the initial iteration step number k = 0; Step (13): Update the consistency variables of the second-layer agents based on the topological structure diagram of the agents and formulas (12) and (13); Step (14): Calculate the active power output of each intelligent unit in the second layer according to formula (14); Step (15): Correct the active power obtained in step (14) according to formula (15); Step (16): Calculate the second layer active power deviation value according to formula (16); Step (17): determining whether the multi-modal active power deviation value obtained in step (16) meets the accuracy requirement; if Then set the number of iteration steps k = k + 1, and return to step (13) to continue the iterative calculation of the second layer; if The obtained active output of a group of agents is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (13) to continue the iterative calculation of the second layer; if The multi-modal active power deviation value of the second layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent generator sets is obtained; Step (18): Perform multimodal judgment and decision crowding distance screening on the active power output of multiple groups of intelligent agents in the second layer obtained in step (17) to obtain multimodal active power output that meets the requirements of the decision maker; Step (19): Input the multiple groups of multi-modal power generation of the second layer obtained in step (18) as load values ​​into the third layer, and set the initial iteration step number k=0; Step (20): Update the consistency variables of the third-layer agent based on the agent's topological structure diagram and formulas (12) and (13); Step (21): Calculate the active power output of each intelligent unit in the third layer according to formula (14); Step (22): Correct the active power output obtained in step (21) according to formula (15); Step (23): Calculate the third layer active power deviation value according to formula (16); Step (24): determining whether the multi-modal active power deviation value obtained in step (23) meets the accuracy requirement; if Then set the number of iteration steps k = k + 1, and return to step (20) to continue the iterative calculation of the third layer; if The obtained active output of a group of agents is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (20) to continue the iterative calculation of the third layer; if The multi-modal active power deviation value of the third layer meets the requirements, the iteration ends, and the active power output of multiple groups of intelligent generators is obtained; Step (25): Perform multimodal judgment and decision crowding distance screening on the active power output of multiple groups of intelligent agents in the third layer obtained in step (24) to obtain multimodal active power output that meets the requirements of the decision maker; Step (26): input the multiple groups of multi-modal power generation of the third layer obtained in step (25) as load values ​​into the fourth layer, and set the initial iteration step number k=0; Step (27): Update the consistency variables of each generator set in the fourth layer by combining the topological structure diagram of the intelligent agent and formulas (12) and (13); Step (28): Calculate the active power output of each generator set in the fourth layer according to formula (14); Step (29): Correct the active power output obtained in step (28) according to formula (15); Step (30): Calculate the fourth layer active power deviation value according to formula (16); Step (31): determining whether the multi-modal active power deviation value obtained in step (30) meets the accuracy requirement; if Then set the number of iteration steps k = k + 1, and return to step (27) to continue the iterative calculation of the fourth layer; if The obtained active output of a group of generator sets is stored, and then the number of iteration steps k=k+1 is set, and the process returns to step (27) to continue the iterative calculation of the fourth layer; if The multi-modal active power deviation value of the fourth layer meets the requirements, the iteration ends, and the active power output of multiple generator sets is obtained; Step (32): Perform multi-modal judgment and decision congestion distance screening on the active output of the plurality of generator sets in the fourth layer obtained in step (31), and obtain a multi-modal active output decision scheme that meets the requirements of the decision maker; Step (33): Determine whether t≤T is satisfied. If so, the iteration ends; if not, proceed to step (3).

Images