Game-based optimization scheduling method for active distribution network considering multi-microgrid energy storage sharing

Abstract

Considering the active distribution network game optimization scheduling method of multi-microgrid energy storage sharing, an active distribution network-multi-microgrid alliance-microgrid collaborative optimization scheduling model based on double game is established; the model parameters are obtained, and the initial value of the algorithm is given; based on the double-layer optimization theory and the definition of Nash equilibrium, the solution process of the lower optimization model is nested into the optimization problem of the upper model, the upper optimization model is solved by the PSO particle swarm algorithm, and the lower optimization model is solved by the Yalmip / Cplex toolbox; the Nash equilibrium state is reached through multiple rounds of iterations, and the final Nash equilibrium solution set is output; further based on the cooperative game theory, the Shapley value method is used to allocate the cooperative surplus of the multi-microgrid alliance. The present invention optimizes the time-of-use electricity price on the distribution network side, taking into account the interest balance between the distribution network and the multi-microgrids; on the multi-microgrid alliance side, the internal energy mutual assistance of the multi-microgrid is realized, and the local consumption capacity of new energy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network optimization and scheduling, and in particular to an active distribution network game optimization and scheduling method considering multi-microgrid energy storage sharing. Background Art

[0002] In order to promote the local consumption of renewable energy, wind and solar renewable energy is generally connected to the active distribution network (ADN) in the form of microgrids (MG), forming an active distribution network system containing multiple microgrids. Multiple microgrids jointly participate in the optimized operation and scheduling of ADN, which can improve the overall reliability and economy of the system, as recorded in the literature [1]: Chen Weinan, Hu Zhijian, Yue Jingpeng, et al. Two-layer energy management of active distribution network considering microgrid access [J]. China Southern Power Grid Technology, 2020, 14(7): 30-38.

[0003] When multiple microgrids are connected to the same distribution network and operate independently, the energy storage of each microgrid is only used to balance the power difference between its own renewable energy and its own load, and the excess or shortage of power continues to be balanced through interaction with the upper-level ADN. Due to the different characteristics of energy storage charging and discharging and energy production and consumption of each microgrid, in certain scheduling periods, there may be a phenomenon of excess energy in some microgrids and power shortage in adjacent microgrids. The energy storage of microgrids in independent operation mode only serves its own load and cannot coordinate the energy storage behavior of adjacent microgrids, which makes the energy storage charging and discharging behavior of multiple microgrids disordered from an overall perspective, resulting in a waste of energy storage resources.

[0004] The multi-microgrid grid-connected system has two complementary features: 1) the charging and discharging behaviors of the multi-microgrid energy storage applications are different and complementary; 2) the energy production and consumption characteristics of the multi-microgrids are different and complementary. Using these two features to form a multi-microgrid alliance (Multi Microgrid Coalition, MMG) is conducive to the interconnection and mutual assistance between microgrids, giving play to the complementary effect of the resources of each microgrid, and promoting the improvement of the operation efficiency of the active distribution network. For example, the document [2]: X. Feng, J. Guan and X. Guan. Optimal allocation of hybrid energy storage for microgrids based on multi-attribute utility theory [J]. Journal of Modern Power Systems and Clean Energy, 2018, 6 (1): 107-117.

[0005] Reference [3]: He Lijun, Cheng Shan, Chen Ziming. Two-layer optimal dispatch of multi-microgrids considering interactive power control and bilateral bidding transactions [J]. Power System Protection and Control, 2020, 48(11): 10-17. An optimal dispatching method considering the interactive power control of multi-microgrids and distribution networks is proposed, which suppresses the interactive power fluctuations between multi-microgrids and distribution networks and can reduce the overall cost of the multi-microgrid system.

[0006] Reference [4]: ​​Huang Zhanghao, Zhang Yachao, Zheng Feng, et al. Day-ahead and real-time energy management method for active distribution networks based on coordinated optimization of different stakeholders [J]. Power System Technology, 2021, 45(6): 2299-2308. A two-layer optimization scheduling model was established for distribution networks and microgrids with different stakeholders, achieving a balanced profit distribution of distribution networks and microgrid groups.

[0007] Reference [5]: Chen Bo, Sun Min, Zeng Wei, He Wei. Day-ahead optimal dispatching of active distribution network considering micro-energy grid access [J]. Journal of Electric Power System and Automation, 2020, 32(11): 102-108. The day-ahead dispatching strategy after multiple microgrids are connected to the active distribution network is studied, which effectively reduces the overall system loss.

[0008] Reference [6]: Y.Fu, Z.Zhang, Z.Li, et al. Energy management for hybrid AC / DC distribution system with microgrid clusters using master-slave game theory and robust optimization [J]. IEEE Transactions on Smart Grid, 2020, 11(2): 1510-1525. An energy management framework for hybrid AC / DC distribution system with microgrid clusters was established, and the interests of various entities were balanced through a master-slave game model.

[0009] Reference [7]: Marzband M, Javadi M, Pourmousavi SA, et al. An advanced retailelectricity market for active distribution systems and home microgrid interoperability based on game theory [J]. Electric Power Systems Research, 2018, 157: 187-199. A home microgrid optimization operation and retail market transaction model is proposed to improve the profits of all participants.

[0010] Reference [8]: Y. Zhang, X. Ai, J. Wen, J. Fang and H. He. Data-adaptive robust optimization method for the economic dispatch of active distribution networks [J]. IEEE Transactions on Smart Grid, 2019, 10(4): 3791-3800. An adaptive robust optimization strategy is proposed for ADN containing renewable energy, which improves the economic efficiency of system operation.

[0011] Reference [9]: H. Sheng, C. Wang, B. Li, J. Liang, et al. Multi-timescale active distribution network scheduling considering demand response and user comprehensive satisfaction [J]. IEEE Transactions on Industry Applications, 2021, 57(3): 1995-2005. A scheduling method considering user demand response and electricity satisfaction is proposed to smooth load fluctuations and promote the consumption of new energy.

[0012] Reference

[10] : Wei C, Fadlullah ZM, Kato N, et al. GT-CFS: A game theoretic coalition formulation strategy for reducing power loss in micro grids [J]. IEEE Transactions on Parallel & Distributed Systems, 2014, 25(9): 2307-2317. A multi-microgrid alliance scheduling strategy based on game theory was proposed to effectively reduce power loss.

[0013] At present, there have been a lot of research results on the optimal scheduling of active distribution networks containing multiple microgrids, but few involve shared energy storage among multiple microgrids. Summary of the invention

[0014] Aiming at the problems of disordered charging and discharging and waste of resources generated after multiple microgrids are connected to the active distribution network, the present invention provides an active distribution network game optimization scheduling method considering the energy storage sharing of multiple microgrids, and establishes an active distribution network (ADN)-multi-microgrid alliance (MMG)-microgrid (MG) collaborative optimization scheduling model based on double game; the active distribution network ADN and the multi-microgrid alliance MMG form a master-slave game relationship, and the active distribution network ADN encourages the multi-microgrid alliance MMG to purchase electricity by changing the purchase and sale price to maximize the profit of electricity sales; the multi-microgrid alliance MMG responds to the quotation of the active distribution network ADN to formulate a power purchase strategy to minimize the total operating cost; further based on the cooperative game theory, the Shapley value method (SVM) is used to allocate the cooperative surplus within the multi-microgrid alliance MMG. The double game model is solved by nesting the Yalmip / Cplex solver with the particle swarm algorithm, the upper ADN is solved by the particle swarm algorithm, and the optimization problem of the lower MMG is solved by Yalmip / Cplex. This method optimizes the time-of-use electricity price of ADN, takes into account the balance of interests between the distribution network and multiple microgrids, and realizes power mutual assistance within multiple microgrids, further improving the local consumption capacity of new energy.

[0015] The technical solution adopted by the present invention is:

[0016] The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing includes the following steps:

[0017] Step 1: According to the game relationship between the active distribution network ADN and the multi-microgrid alliance MMG, a master-slave game optimization model is established;

[0018] Step 2: Construct the constraints of the master-slave game optimization model;

[0019] Step 3: Obtain the parameters of the active distribution network ADN and the multi-microgrid alliance MMG, as well as the particle swarm algorithm parameters used to solve the master-slave game optimization model;

[0020] Step 4: Set the initial value of the Stackelberg-Nash equilibrium solution of the master-slave game optimization model And calculate the response set of the lower MMG And substitute it into the upper layer ADN for a new round of optimization;

[0021] Step 5: In the kth round of cycles, the upper ADN uses the equilibrium solution of the k-1th round of MMG As the input strategy, find the optimal strategy set of ADN in the current kth round And the corresponding response set of the lower-level MMG

[0022] Step 6: Determine whether the solution of step 5 is a Stackelberg-Nash equilibrium solution; if the conditions are met, output the game equilibrium solution and go to step 7; otherwise, execute k=k+1 and return to step 5;

[0023] Step 7: Allocate the cooperation surplus of the multi-microgrid alliance MMG according to the SVM method and calculate the benefits shared by each microgrid member.

[0024] In step 1, the master-slave game optimization model is as follows:

[0025] As the upper-level leader, ADN sets the electricity sales price and electricity purchase price based on the total system operation cost; as the lower-level follower, MMG adjusts the transaction purchase and sales electricity volume according to the electricity sales price set by ADN to achieve the lowest operating cost.

[0026] ADN uses the electricity purchase and sales prices it sets to guide MMG to dynamically adjust its electricity purchase and sales strategies to ensure the balance of power within ADN and its own optimal economic benefits. MMG adjusts its electricity strategy in response to the electricity purchase and sales prices. ADN also dynamically adjusts the electricity price based on the updated alliance strategy. When MMG's electricity trading strategy and ADN's electricity price strategy remain stable, this optimization model reaches the game Nash equilibrium state.

[0027] The mathematical expression of the master-slave game model G can be described as:

[0028] G = {{ADN,MMG}; {ρ s ,ρ b}; {P PCC,MMG,b ,P PCC,MMG,s}; {J ADN}; {J MMG}} (1);

[0029] Where: {ADN,MMG} represents the two parties in the game, namely the active distribution network and the multi-microgrid alliance; {ρ b ,ρ s} represents the ADN electricity purchase and sale price strategy set; ρ b represents the ADN repurchase price; ρ s represents the electricity price sold by ADN; {P PCC,MMG,b ,P PCC,MMG,s} represents the MMG power purchase and sales strategy set, P PCC,MMG,b , P PCC,MMG,s Respectively represent the amount of electricity purchased and sold by MMG.

[0030] When the operating cost of ADN is ADN and MMG's operating costs MMG When the optimal solution is reached, the game model G has a unique Nash equilibrium solution and satisfies the following conditions:

[0031]

[0032] In the formula, the equilibrium solution of the game model is They respectively represent: ADN electricity sales price, ADN electricity purchase price, MMG electricity purchase amount, and MMG electricity sales amount.

[0033] The upper-level ADN objective function is the total operating cost of the system, including the power generation cost and operation and maintenance cost of the unit, the power interaction cost between ADN and the large power grid, and the power interaction cost between ADN and MMG. Its objective function is shown in formula (3).

[0034]

[0035] In the formula, C ADN represents the total operating cost of ADN; C PCC,MMG represents the power interaction cost between ADN and MMG; C g It represents the power interaction cost between ADN and the upper grid; is the power generation cost of the controllable units in ADN; is the operation and maintenance cost of the generator sets in ADN.

[0036] The objective function of the lower-level MMG is shown in formula (4).

[0037] J MMG =minC MMG =min(C MTΣ +C PCC,MMG +C OMΣ ) (4);

[0038] Where: C MMG represents the total cost of MMG operation; C MTΣ represents the total cost of gas turbine power generation in MMG; C OMΣ Represents the total operation and maintenance cost of the generator sets within the MMG.

[0039] In step 2, the constraints of the master-slave game optimization model include:

[0040] The constraints of the upper-layer ADN and the lower-layer MMG. The constraints of energy mutual assistance among multiple microgrids are included in the lower-layer MMG model.

[0041] The constraints of the upper layer ADN are as follows:

[0042] 1) Distribution network flow constraints:

[0043]

[0044] In formula (5), n1 is the number of ADN nodes; Pi , Q i are the injected active and reactive power of node i respectively; U i is the voltage amplitude of node i; G ij , B ij , δ ij are the conductance, susceptance, and voltage phase difference between node i and node j, respectively.

[0045] 2) ADN line power constraints:

[0046]

[0047] In formula (6), is the active power on the lth branch of the distribution network during period t, is the maximum transmission power of the lth branch.

[0048] 3) ADN node voltage constraints:

[0049]

[0050] In the formula, are the lower and upper limits of the voltage at node i respectively.

[0051] 4) ADN unit output constraints:

[0052]

[0053] In the formula, are the lower and upper limits of the output of unit j respectively; Represents the output of unit j at time t.

[0054] 5) Time-of-use electricity price constraints:

[0055] The basic model of time-of-use electricity price is:

[0056]

[0057]

[0058] Formulas (9) and (10) represent the time-of-use electricity sales price and electricity purchase price constraints set by ADN respectively; T g ,T p ,T f Represent valley, flat and peak time periods respectively; ρ sg ,ρ sp ,ρ sf Represent the electricity prices during off-peak, flat and peak periods respectively; ρ bg ,ρ bp ,ρ bf Represent the repurchase electricity prices during valley, flat and peak periods respectively; the time-of-use electricity price needs to satisfy the inequality relationship ρsf ≥ρ sp ≥ρ sg , bf ≥ρ bp ≥ρ bg , and the electricity sales price in the same period is greater than the electricity purchase price

[0059] The constraints of the lower MMG are as follows:

[0060] 1) Power balance constraints of each microgrid:

[0061]

[0062] In the formula, is the load of microgrid i in period t; For the new energy output of microgrid i in period t, is the power purchased by microgrid i from the distribution grid, is the power purchased by the distribution network from microgrid i, is the output of the gas turbine in microgrid i; represents the mutual aid power of microgrid i in the multi-microgrid alliance, where n represents the number of alliance members; D ji and D ij Both are binary state variables, indicating the state of power mutual assistance between microgrids. When 0 is taken, it means no power is transferred. ji =1 means that microgrid j transfers power to microgrid i, D ij =1 means that microgrid i transfers power to microgrid j; represents the power value transmitted from microgrid j to microgrid i at time t; represents the power value transmitted from microgrid i to microgrid j at time t; and They represent the energy storage charging and discharging power of microgrid i at time t respectively.

[0063] 2) Constraints on energy mutual assistance within MMG:

[0064]

[0065] In the formula, represents the remaining power of microgrid i in period t, It represents the residual load of microgrid i in period t.

[0066] 3) Tie-line constraints between each microgrid and ADN:

[0067]

[0068] In the formula, and They represent the lower and upper limits of the power of the tie line between microgrid i and ADN respectively.

[0069] 4) Upper and lower limits of gas turbine output for each microgrid:

[0070]

[0071] In the formula, and They represent the lower and upper limits of the gas turbine power of microgrid i, respectively; represents the output of the gas turbine of microgrid i at time t.

[0072] 5) Constraints on energy storage charging and discharging of each microgrid:

[0073]

[0074] Where P i Cap is the energy storage power capacity of microgrid i.

[0075] 6) Upper and lower bounds of energy storage capacity of each microgrid:

[0076]

[0077] In the formula, is the minimum state of charge of the energy storage of microgrid i; is the remaining energy storage capacity of microgrid i at time t; is the energy storage capacity of microgrid i.

[0078] 7) The relationship between the power consumption of each microgrid energy storage in two adjacent time periods:

[0079]

[0080] In the formula, is the remaining energy storage capacity of microgrid i at time t-1; are the charging efficiency and discharging efficiency of the energy storage of microgrid i respectively; △t represents the scheduling time interval.

[0081] In step 3, the initial parameters of the active distribution network ADN are obtained, including the distribution network grid structure, node data, branch data, capacity and operating parameters of new energy units and controllable units, tie line power limit, initial electricity sales price, initial repurchase price, etc.

[0082] Obtain the operating parameters of the multi-microgrid alliance MMG, including the capacity and operating parameters of each microgrid gas turbine, new energy unit, and energy storage system.

[0083] Obtain the particle swarm algorithm parameters for solving the master-slave game optimization model, including: population size, maximum number of iterations, particle flight speed range, learning factor, initial inertia weight, etc.;

[0084] In step 4, the follower MMG responds to the initial purchase and sale electricity price of the leader ADN, with the minimum total operating cost of MMG as the optimization goal, as shown in formula (4), satisfying the constraints: formula (11) to formula (17), and obtaining the response set of MMG under the first round of ADN electricity price incentives Substitute the response set of MMG into the upper ADN model, use formula (3) as the objective function, consider the constraints (5) to (10), and solve the new electricity price strategy set for the next round of ADN: MMG continues to respond to ADN's electricity price strategy and gives a response set In this way, the solution is solved by repeated iterations to obtain the optimal strategy set of the game participants in each round.

[0085] in: They represent the initial values ​​of ADN electricity purchase and sale price strategies respectively; They represent the initial values ​​of MMG's electricity purchase and sales strategies respectively; They represent ADN’s electricity purchase and sale price strategies in the first cycle respectively; They respectively represent MMG’s electricity purchase and sales strategies in the first cycle.

[0086] In step 5, the strategies of each game participant obtained in the kth cycle are as follows:

[0087]

[0088] Where: γ 1,k is the strategy set of the leader ADN in round k; γ 2,k is the response set of follower MMG in round k; represents the electricity purchase and sale price strategy of ADN in the kth round; represents MMG's electricity purchase and sales strategy in round k; represents the strategy of MMG in round k-1 As input, find the optimal strategy of the current k-th round leader ADN; Represents the strategy of the leader ADN in round k As an incentive signal, find the optimal response strategy of the follower MMG in the current kth round; arg represents the independent variable, argminJ ADN Represents the objective function J ADN The optimal strategy of ADN corresponding to the minimum value

[0089] In step 6, the method for determining whether a Stackelberg-Nash equilibrium solution is found is as follows:

[0090] If the optimization results of each game participant in the kth round are consistent with those in the k-1th round, that is:

[0091]

[0092] This means that the game model finds the Stackelberg-Nash equilibrium solution Otherwise, update the iteration cycle round k=k+1 and return to step 5 to continue solving until the Stackelberg-Nash equilibrium solution is found.

[0093] In formula (19): is the Nash equilibrium strategy set of ADN in the k-round cycle; is the Nash equilibrium strategy set of MMG in the k-round cycle; Represents the ADN electricity purchase and sales price and MMG electricity purchase and sales strategy corresponding to the Nash equilibrium solution; represents the optimal electricity price of ADN in the k-1th round; represents the optimal repurchase price of ADN in the k-1th round; represents the optimal electricity purchase of MMG in the k-1th round; It represents the optimal electricity sales of MMG in the k-1th round.

[0094] In step 7, the cooperative surplus allocation model is specifically as follows:

[0095] The cooperative surplus generated by the alliance members through cooperation is used to obtain the benefits allocated to each member using the SVM method, which is calculated as shown in formula (20).

[0096]

[0097] In formula (20), s is the alliance set composed of all microgrids, v(i) is the income allocated to microgrid i, |s| is the number of subsets, ω(|s|) is the weight factor, v(s) is the cooperation surplus of sub-alliance s, v(s / i) is the cooperation surplus of sub-alliance excluding member i, n is the number of microgrids participating in the alliance; (n-|s|)! represents the factorial of the number of MMG members n minus the number of members in subset s; (|s|-1)! represents the factorial of the number of members in subset s minus 1; n! represents the factorial of the number of MMG members n.

[0098] The present invention is an active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing, and its technical effects are as follows:

[0099] 1) The present invention makes full use of the game relationship of competition and cooperation among the stakeholders involved in the scheduling, establishes an active distribution network-multi-microgrid alliance-microgrid collaborative optimization scheduling model based on multiple games, and combines the Nash equilibrium definition with the particle swarm optimization algorithm to solve the model, and obtains the optimal scheduling scheme for each system in the active distribution network.

[0100] 2) The present invention utilizes the master-slave game competition between ADN and MMG to optimize the electricity price and MMG energy use strategy when ADN scores, and takes into account the balance of interests between ADN and multiple microgrids.

[0101] 3) The present invention makes full use of the complementary characteristics of the charging and discharging behaviors and energy production and consumption behaviors of each microgrid energy storage system. By forming an alliance among multiple microgrids and distributing the cooperative surplus after the alliance according to cooperative game theory, the power mutual assistance within the multiple microgrids is maximized, further improving the local consumption capacity of new energy.

[0102] 4) The present invention optimizes the time-of-use electricity price on the ADN side, taking into account the balance of interests between ADN and MMG; on the MMG side, it realizes energy mutual assistance within multiple microgrids and improves the local consumption capacity of new energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] Figure 1 This is a two-layer game framework diagram of the active distribution network (ADN) containing multiple microgrid alliances (MMGs).

[0104] Figure 2 Flowchart for solving the ADN-MMG master-slave game model.

[0105] Figure 3 This is the system diagram of the ADN test example.

[0106] Figure 4(a) is the load and new energy output curve of MG1;

[0107] Figure 4(b) is the load and new energy output curve of MG2;

[0108] Figure 4(c) is the load and new energy output curve of MG3;

[0109] Figure 4(d) shows the load and renewable energy output curve of ADN.

[0110] Figure 5(a) is a schematic diagram of the output of each device in MG1;

[0111] Figure 5(b) is a schematic diagram of the output of each device in MG2;

[0112] Figure 5(c) is a schematic diagram of the output of each device in MG3.

[0113] Figure 6 This is the energy mutual assistance power diagram within the MMG. DETAILED DESCRIPTION

[0114] Considering the active distribution network game optimization scheduling method of multi-microgrid energy storage sharing, an active distribution network-microgrid alliance-microgrid collaborative optimization scheduling model based on double game is established, which is implemented in the following steps:

[0115] Step 1:

[0116] S1.1: Building the master-slave game relationship between active distribution network (ADN) and multi-microgrid alliance (MMG):

[0117] ADN and MMG belong to different stakeholders and have their own operating goals. In joint dispatch, based on game theory, it is advisable to adopt the master-slave game dispatch between ADN and MMG; while the members of MMG have the same operating goals, the cooperative game dispatch between MMG members can be adopted. Thus, a two-layer game dispatch framework between active distribution network-multi-microgrid alliance-microgrid is formed, such as Figure 1 shown.

[0118] The two-layer game process is as follows:

[0119] 1) As the leader, ADN sets the initial electricity purchase and sales prices based on the peak and valley periods of the net load curve.

[0120] 2) According to the electricity price set by ADN, multiple microgrids cooperate to become MMGs, respond to electricity price strategies as an alliance, formulate corresponding alliance electricity purchase and sales (power) strategies, and report the strategies to ADN.

[0121] 3) ADN updates the electricity price according to the electricity purchase and sales strategy reported by MMG, and guides MMG to dynamically adjust the electricity purchase and sales strategy to ensure the balance of power within ADN and the optimization of its own economic benefits.

[0122] 4) MMG adjusts its electricity strategy again according to the updated electricity price, and ADN also dynamically adjusts the electricity price according to the updated alliance strategy until the MMG electricity strategy and ADN electricity price strategy remain stable and the Nash equilibrium solution of the game is reached.

[0123] 5) MMG members cooperate to distribute the remaining profits according to their own contribution, ensuring the improvement of energy utilization and the optimal distribution of alliance power among MG units.

[0124] The mathematical expression of the master-slave game model G can be described as:

[0125] G = {{ADN,MMG}; {ρ s ,ρ b}; {P PCC,MMG,b ,P PCC,MMG,s}; {J ADN}; {J MMG}} (1)

[0126] Where: {ADN,MMG} represents the two parties in the game, namely the active distribution network and the multi-microgrid alliance; {ρ b ,ρ s} represents the ADN electricity purchase and sale price strategy set; {P PCC,MMG,b ,P PCC,MMG,s} represents the MMG electricity purchase and sales strategy set.

[0127] When the operating cost of ADN is ADN and MMG's operating costs MMG When the optimal solution is achieved together, the master-slave game optimization model G has a unique Nash equilibrium solution, as follows:

[0128]

[0129] In the formula, the equilibrium solution of the master-slave game optimization model is They respectively represent: ADN electricity sales price, ADN electricity purchase price, MMG electricity purchase amount and electricity sales amount.

[0130] S1.2: Construct the objective function of the upper leader (ADN);

[0131] The objective function of the upper-level ADN is the total operating cost of the system, which includes the power generation cost and operation and maintenance cost of the unit, the power interaction cost with the large power grid, and the power interaction cost between the multi-microgrid alliance system. Its objective function is shown in formula (3).

[0132]

[0133] In the formula, C ADN represents the total operating cost of ADN; C PCC,MMG represents the power interaction cost between ADN and MMG; C g It represents the power interaction cost between ADN and the upper grid; is the power generation cost of the controllable units in ADN; is the operation and maintenance cost of the generator sets in ADN.

[0134] 1) Interaction cost between ADN and large power grid:

[0135]

[0136] In the formula, are the electricity sales price and electricity purchase price of the large power grid in period t, They respectively represent the power purchased and sold by ADN to the large power grid during period t.

[0137] 2) Interaction cost between ADN and MMG:

[0138]

[0139] In the formula, are the total electricity sold and purchased by MMG in period t respectively.

[0140] 3) ADN unit power generation cost:

[0141]

[0142] In the formula, a mj 、b mj 、c mj is the power generation cost coefficient of unit j, k is the number of units, is the output of the jth unit in period t.

[0143] 4) ADN unit operation and maintenance costs:

[0144]

[0145] In the formula, k mj is the unit power operation and maintenance cost coefficient of unit j.

[0146] S1.3: Construct the objective function of the lower follower (MMG):

[0147] The objective function of the lower MMG includes the total power generation cost C of the gas turbine in the microgrid MTΣ , the total interaction cost C of the MMG and ADN tie line PCC,MMG , MMG total operation and maintenance cost C OMΣ , and its objective function is shown in formula (4).

[0148] J MMG =minC MMG =min(C MTΣ +C PCC,MMG +C OMΣ ) (4);

[0149] 1) Total cost of power generation from gas turbines in MMG:

[0150] C MTΣ =C MT,Σfuel +C MT,Σen ;

[0151]

[0152] In the formula, C MT,Σfuel is the total cost of gas turbine power generation in MMG; C MT,Σen is the pollution control cost of the gas turbine in the MMG. M,i、b M,i 、c M,i are the gas turbine power generation cost coefficients of microgrid i, is the output power of the gas turbine of microgrid i in period t; m is the number of pollutant types; p is the different types of pollutants; α ep is the emission cost of pollutants; ep For emissions.

[0153] 2) Total interaction cost between MMG and ADN interconnection lines:

[0154]

[0155] In the formula, and They represent the amount of electricity purchased and sold by microgrid i in period t respectively.

[0156] 3) MMG total operation and maintenance cost

[0157]

[0158] In the formula, is the maintenance cost coefficient of the jth micro-source in microgrid i; is the power generation of the jth micro-source in microgrid i during period t.

[0159] Step 2: Construct the constraints of the master-slave game optimization model:

[0160] S2.1: The ADN constraints of the upper leader are as follows:

[0161] 1) Distribution network flow constraints:

[0162]

[0163] In formula (5), n1 is the number of ADN nodes; P i , Q i are the injected active and reactive power of node i respectively; U i is the voltage amplitude of node i; G ij , B ij , δ ij are the conductance, susceptance, and voltage phase difference between node i and node j, respectively.

[0164] 2) ADN line power constraints:

[0165] 0≤P l t ≤P l max (6);

[0166] In formula (6), P lt is the active power on the lth branch of the distribution network during period t, P l max is the maximum transmission power of the lth branch.

[0167] 3) ADN node voltage constraints:

[0168]

[0169] In the formula, are the lower and upper limits of the voltage at node i respectively.

[0170] 4) ADN unit output constraints:

[0171]

[0172] In the formula, are the lower and upper limits of the output of unit j respectively; Represents the output of unit j at time t.

[0173] 5) Time-of-use electricity price constraints:

[0174] The basic model of time-of-use electricity price is:

[0175]

[0176]

[0177] Formulas (9) and (10) represent the time-of-use electricity sales price and electricity purchase price constraints set by ADN respectively; T g ,T p ,T f Represent valley, flat and peak time periods respectively; ρ sg ,ρ sp ,ρ sf Represent the electricity prices during off-peak, flat and peak periods respectively; ρ bg ,ρ bp ,ρ bf Represent the repurchase electricity prices during valley, flat and peak periods respectively; the time-of-use electricity price needs to satisfy the inequality relationship ρ sf ≥ρ sp ≥ρ sg , bf ≥ρ bp ≥ρ bg , and the electricity sales price in the same period is greater than the electricity purchase price S2.2: Constraints of the lower MMG:

[0178] 1) Power balance constraints of each microgrid:

[0179]

[0180] In the formula, is the load of microgrid i in period t; For the new energy output of microgrid i in period t, is the power purchased by microgrid i from ADN, is the power repurchased by ADN from microgrid i, is the output of the gas turbine in microgrid i; represents the mutual aid power of microgrid i in MMG, where n represents the number of alliance members; D ji and D ij Both are binary state variables, indicating the state of power mutual assistance between microgrids. When 0 is taken, it means no power is transferred. ji =1 means that microgrid j transfers power to microgrid i, D ij =1 means that microgrid i transfers power to microgrid j. represents the power value transmitted from microgrid j to microgrid i at time t; Represents the power value transmitted from microgrid i to microgrid j at time t. and They represent the energy storage charging and discharging power of microgrid i at time t respectively.

[0181] 2) Constraints on energy mutual assistance within MMG:

[0182]

[0183] In the formula, represents the remaining power of microgrid i in period t, It represents the residual load of microgrid i in period t.

[0184] 3) Constraints on the interconnection lines between each microgrid and ADN:

[0185]

[0186] In the formula, and They represent the lower and upper limits of the power of the tie line between microgrid i and ADN respectively.

[0187] 4) Upper and lower limits of gas turbine output for each microgrid:

[0188]

[0189] In the formula, and They represent the lower and upper limits of the gas turbine power of microgrid i, respectively; represents the output of the gas turbine of microgrid i at time t.

[0190] 5) Constraints on energy storage charging and discharging of each microgrid:

[0191]

[0192] Where P i Cap is the energy storage power capacity of microgrid i.

[0193] 6) Upper and lower bounds of energy storage capacity of each microgrid:

[0194]

[0195] In the formula, is the minimum state of charge of the energy storage of microgrid i; is the remaining energy storage capacity of microgrid i at time t; is the energy storage capacity of microgrid i.

[0196] 7) The relationship between the power consumption of each microgrid energy storage in two adjacent time periods:

[0197]

[0198] In the formula, is the remaining energy storage capacity of microgrid i at time t-1; are the charging efficiency and discharging efficiency of the energy storage of microgrid i respectively; △t represents the scheduling time interval.

[0199] Step 3: Obtain the parameters of ADN and MMG, as well as the particle swarm algorithm parameters used to solve the model;

[0200] The parameters of the active distribution network and its connected multi-microgrid system include: system grid structure, line parameters, load distribution and its forecast data, distributed power output limit, etc.; the particle swarm algorithm parameters used for model solution include: population size, maximum number of iterations, particle flight speed range, learning factor, initial inertia weight, etc. In order to enhance the local convergence ability of the particle swarm algorithm, the inertia weight will decrease with the increase of the number of iterations. The adjustment formula of the inertia weight is as follows:

[0201] w=w max -(w max -w min )·Iter / D

[0202] Where w is the inertia weight, w max ,w min are the maximum and minimum inertia weights respectively, Iter is the current iteration round, and D is the maximum number of iterations.

[0203] This two-layer optimization model uses a nested particle swarm algorithm to find the Nash equilibrium solution of the master-slave game. The outer optimization corresponds to the leader game model and is solved by the particle swarm algorithm; the inner optimization corresponds to the follower game model and is solved by the Yalmip / Cplex toolbox. The solution process is as follows: Figure 2 The main solution steps are shown in steps 4 to 6.

[0204] Step 4: Set the initial value of the Stackelberg-Nash equilibrium solution of the master-slave model Calculate the MMG response set And substitute it into the upper-level ADN model for a new round of optimization.

[0205] The optimization process is as follows: the follower MMG responds to the initial purchase and sale electricity price of the leader ADN, with the minimum total operating cost of MMG as the optimization goal (Formula (4)), and the constraints (Formula (11) to (Formula (17)) are satisfied to obtain the response set under the first round of ADN electricity price incentives Substitute the MMG response set into the ADN optimization model, use formula (3) as the objective function, consider the constraints (5) to (10), and solve the new electricity price strategy set for the next round of ADN: MMG continues to respond to ADN's electricity price strategy and gives a response set The solution is solved by repeated iterations in this way to obtain the optimal strategy set of the game participants in each round.

[0206] in: They represent the initial values ​​of ADN electricity purchase and sale price strategies respectively; They represent the initial values ​​of MMG's electricity purchase and sales strategies respectively; They represent ADN’s electricity purchase and sale price strategies in the first cycle respectively; They respectively represent MMG’s electricity purchase and sales strategies in the first cycle.

[0207] Step 5: In the kth round of cycles, the upper ADN uses the equilibrium solution of the k-1th round of MMG As input, find the optimal strategy set of ADN and the corresponding lower-level response set.

[0208] The strategies of each game participant obtained in the kth cycle are as follows:

[0209]

[0210] Where: γ 1,k is the strategy set of the leader ADN in round k; γ 2,k is the response set of follower MMG in round k; represents the electricity purchase and sale price strategy of ADN in the kth round; Represents MMG's electricity purchase and sales strategy in round k.

[0211] Step 6: The method to determine whether the Stackelberg-Nash equilibrium solution of the model is found is:

[0212] If the optimization results of each game participant in the kth round are consistent with those in the k-1th round, that is:

[0213]

[0214] This means that the game model finds the Stackelberg-Nash equilibrium solution Otherwise, update the iteration cycle round k=k+1 and return to step 5 to continue solving until the Stackelberg-Nash equilibrium solution is found.

[0215] In formula (19): is the Nash equilibrium strategy set of ADN in the k-round cycle; is the Nash equilibrium strategy set of MMG in the k-round cycle; Represents the ADN electricity purchase and sales price and MMG electricity purchase and sales strategy corresponding to the Nash equilibrium solution.

[0216] Step 7: Use the SVM method to allocate the cooperative surplus of MMG and calculate the cost saved by each microgrid member.

[0217] The SVM value can be interpreted as the average marginal contribution of all possible alliance situations of participant i. The more contributions a member makes, the more benefits he or she gets. The calculation expression for allocating the cooperative surplus using the SVM method is shown in formula (20).

[0218]

[0219] In the formula, s is the alliance set composed of all microgrids, v(i) is the income allocated to microgrid i, |s| is the number of subsets, ω(|s|) is the weight factor, v(s) is the cooperation surplus of sub-alliance s, v(s / i) is the cooperation surplus of sub-alliance excluding member i, and n is the number of microgrids participating in the alliance. i Represents the marginal contribution of participant i.

[0220] The cooperative surplus is defined as the difference between the total revenue generated by the cooperation of all members in the alliance and the sum of the revenue of each member acting independently. Combined with the master-slave game model mentioned above, the difference in revenue between multi-microgrid alliance and non-alignment mainly comes from the energy mutual assistance within the alliance. In other words, if the members consider power mutual assistance, it means that the members are in an alliance, otherwise they are non-aligned. The difference in revenue before and after power mutual assistance can be understood as cooperative surplus.

[0221] Specifically, assume that MMG contains three members, namely MG1, MG2 and MG3. Let N = {MG1, MG2, MG3}. There are seven forms of coalitions consisting of three participants, namely Φ = {{MG1}, {MG2}, {MG3}, {MG1, MG2}, {MG1, MG3}, {MG2, MG3}, {MG1, MG2, MG3}}.

[0222] Taking microgrid MG1 as an example, the benefit of each member is obtained by allocating the cooperative surplus through the SVM method:

[0223]

[0224] Where: x(MG1) represents the profit of microgrid MG1 when the cooperative surplus is allocated using the SVM method. MC1({MG1}) represents the marginal contribution of MG1 when the alliance form is {MG1}. Correspondingly, MC1({MG1,MG2}) represents the marginal contribution of MG1 when the alliance form is {MG1,MG2}; MC1({MG1,MG2,MG3}) represents the marginal contribution of MG1 when the alliance form is {MG1,MG2,MG3}.

[0225] Similarly, the benefits allocated to microgrids MG2 and MG3 can also be calculated using the SVM method.

[0226] The SVM method can be found in the literature

[11] : Liu Nian, Zhao Jing, Wang Jie, Lei Jinyong. Photovoltaic microgrid group transaction model based on cooperative game theory [J]. Transactions of the Chinese Society of Electrotechnical Engineering, 2018, 33(8): 1903-1910.

[0227] Verification example:

[0228] The example is based on the IEEE 14-node distribution network system and performs a simulation calculation on the day-ahead dispatch. The network topology is as follows: Figure 3 As shown in the figure. In the system, three microgrids are set up at nodes 6, 11, and 13, and the three microgrids form a multi-microgrid alliance. Node 1 is the public connection point between the ADN and the upper high-voltage power grid, and nodes 2, 3, and 8 are respectively equipped with photovoltaic units, wind turbine units, and controllable units of the distribution network. The new energy output and load curves of the distribution network and each microgrid on a typical day are shown in the figure. Figure 4(a) to Figure 4(d) shown.

[0229] The maximum transmission power of the interconnection lines between microgrids 1 to 3 and the active distribution network is 750kW. The operation and maintenance coefficients of photovoltaic power generation and wind power generation are 0.0096 yuan / kW and 0.0296 yuan / kW respectively. The electricity price of the high-voltage main grid is 0.55 yuan / kW. The gas turbine parameters in each microgrid are shown in Table 1, the pollution emission cost of the gas turbine is shown in Table 2, and the energy storage parameters of each microgrid are shown in Table 3. The charging and discharging efficiency of each microgrid energy storage system is 0.95.

[0230] The particle swarm algorithm parameters used for model solution include: population size is set to 500, maximum number of iterations is set to 2000, particle flight speed range is [-1, 1], learning factor is 1.5, and initial inertia weight is set to 0.9. The test calculation of the present invention uses MatlabR2018b software combined with Yalmip plug-in to call Cplex solver for solution. The computer configuration is Intel Core i7 processor, main frequency 1.8GHz, memory 16GB.

[0231] Table 1 Gas turbine related parameters in each microgrid

[0232]

[0233] Table 2 Gas turbine pollutant emission costs

[0234]

[0235] Table 3 Parameters of each microgrid energy storage system

[0236]

[0237] Analysis of case test results:

[0238] (1) Analysis of the optimization results of the master-slave game between ADN and MMG:

[0239] The optimized time-of-use electricity price is shown in Table 4, and the optimized output dispatch results of each device in microgrid MG1~MG3 are shown in Figure 5.

[0240] Table 4 ADN time-of-use electricity price at game equilibrium (yuan / kWh)

[0241]

[0242] Combination Figure 5(a) to Figure 5(c) , taking microgrid 1 as an example to analyze the dispatching strategy of the present invention:

[0243] 1) ADN valley electricity price period:

[0244] During the period of 0:00-6:00, the valley electricity price period of ADN, the photovoltaic output in microgrid 1 is 0, which is insufficient to meet the power demand of the load. Since the electricity price is low at this time, the gas turbine is not used for power supply, and most of the power is obtained by transmission from other microgrids. During the period of 6:00-8:00, the photovoltaic output gradually increases to meet the power supply demand of the load. The "valley charging" strategy is adopted to charge the energy storage battery of the microgrid, reducing the power sold to ADN.

[0245] 2) ADN flat electricity price period:

[0246] At 8:00-10:00 and 14:00-18:00, the photovoltaic output is large, and the output of new energy is generally greater than the load. The power is smoothed by charging the energy storage battery, and the excess new energy power is also transmitted to other microgrids that are short of power. At 23:00-24:00, the output of new energy is insufficient, and the energy storage battery is used to discharge, while the electricity transmitted by other microgrids is accepted to cooperate with the gas turbine output to make up for the power shortage.

[0247] 3) ADN peak electricity price period:

[0248] During the period from 10:00 to 14:00, the photovoltaic output is relatively large. At this time, charging of the energy storage battery is not considered. Instead, the excess new energy is transferred to other microgrids. Since the electricity price is relatively high at this time, the gas turbine is used to generate more power, thereby increasing the electricity sales during this period. During the period from 20:00 to 23:00, the new energy output is insufficient, and the energy storage battery is used first to discharge, while the gas turbine is used at the same time, which can reduce the purchase of electricity from ADN.

[0249] The comparison of economic benefits and peak-to-valley differences between the two parties in the game is shown in Table 5:

[0250] Table 5 Comparison of economic benefits of ADN and MMG before and after the game

[0251]

[0252] As can be seen from Table 5, ADN adopts time-of-use electricity prices to encourage MMG to adopt rational charging and discharging strategies, and encourages MMG to output more controllable units during peak hours, so as to meet ADN's peak-shaving and valley-filling scheduling needs. Under the condition that the revenue from electricity sales remains unchanged, the incentive effect of time-of-use electricity prices can improve the load curve of ADN, reduce the peak-valley difference on a typical day, and reduce the load standard deviation, that is, the overall trend of the load curve becomes smoother. At the same time, the microgrid alliance adopts a rational charging and discharging strategy under the time-of-use electricity price, which can reduce operating costs.

[0253] (2) Analysis of the mutual benefit of multiple microgrids and the benefits of each member

[0254] Table 6 shows the cost savings of various forms of cooperative alliances. Table 7 is a comparison table of the operating costs of various microgrids.

[0255] Table 6 Cost savings of various forms of cooperation (yuan)

[0256]

[0257] Table 7 Comparison of costs before and after the formation of microgrids (yuan)

[0258]

[0259] From Table 6 and Table 7, we can see that the total operating cost of the multi-microgrid alliance is 4802.21 yuan, which is less than the sum of the costs of the three MGs operating independently, which is 5161.54 yuan. In addition, the costs shared by the three MGs after forming an alliance are all less than the costs of their independent operations, indicating that the conditions for cooperation can be met. Figure 6 The power mutual assistance relationship between microgrids shows that the new energy of MG1 and MG3 is photovoltaic, and the output is zero in the early morning and at night, while the new energy of MG2 is wind power, with a large output at night and a low output during the day. Therefore, MG2 performs power mutual assistance with MG1 and MG3 respectively, and MG1 and MG3 are both photovoltaic microgrids with homogeneity, and there is only power interaction at 18:00. Therefore, when distributing benefits according to the alliance contribution, the cost savings allocated to MG1 and MG3 are relatively small, while the cost savings allocated to MG2 are large, as shown in Table 7.

Claims

1. Active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing, characterized by The following steps are involved: Step 1: According to the game relationship between the active distribution network ADN and the multi-microgrid alliance MMG, a master-slave game optimization model is established; Step 2: Construct the constraints of the master-slave game optimization model; Step 3: Obtain the parameters of the active distribution network ADN and the multi-microgrid alliance MMG, as well as the particle swarm algorithm parameters used to solve the master-slave game optimization model; Step 4: Set the initial value of the Stackelberg-Nash equilibrium solution of the master-slave game optimization model And calculate the response set of the lower MMG And substitute it into the upper layer ADN for a new round of optimization; Step 5: In the kth round of cycles, the upper ADN uses the equilibrium solution of the k-1th round of MMG As the input strategy, find the optimal strategy set of ADN in the current kth round And the corresponding response set of the lower-level MMG Step 6: Determine whether the solution of step 5 is a Stackelberg-Nash equilibrium solution; if the conditions are met, output the game equilibrium solution and go to step 7; otherwise, execute k=k+1 and return to step 5; Step 7: Allocate the cooperation surplus of the multi-microgrid alliance MMG according to the SVM method and calculate the benefits shared by each microgrid member; In step 1, the master-slave game optimization model is as follows: The mathematical expression of the master-slave game model G can be described as: G={{ADN,MMG};{ρ s ,r b };{P PCC,MMG,b ,P PCC,MMG,s };{J ADN };{J MMG }} (1); Where: {ADN,MMG} represents the two parties in the game, namely the active distribution network and the multi-microgrid alliance; {ρ b ,ρ s } represents the ADN electricity purchase and sale price strategy set; ρ b represents the ADN repurchase price; ρ s represents the electricity price sold by ADN; {P PCC,MMG,b ,P PCC,MMG,s } represents the MMG power purchase and sales strategy set, P PCC,MMG,b , P PCC,MMG,s They represent the electricity purchased and sold by MMG respectively; When the operating cost of ADN is ADN and MMG's operating costs MMG When the optimal solution is reached, the game model G has a unique Nash equilibrium solution and satisfies the following conditions: In the formula, the equilibrium solution of the game model is They represent: ADN electricity sales price, ADN electricity purchase price, MMG electricity purchase price, MMG electricity sales price; The upper-level ADN objective function is the total operating cost of the system, including the power generation cost and operation and maintenance cost of the unit, the power interaction cost between ADN and the large power grid, and the power interaction cost between ADN and MMG. Its objective function is shown in formula (3): In the formula, C ADN represents the total operating cost of ADN; C PCC,MMG represents the power interaction cost between ADN and MMG; C g It represents the power interaction cost between ADN and the upper grid; is the power generation cost of the controllable units in ADN; is the operation and maintenance cost of the generator sets in ADN; The objective function of the lower-level MMG is shown in formula (4): J MMG =minC MMG =min(C MTΣ +C PCC,MMG +C OMΣ ) (4); Where: C MMG represents the total operating cost of MMG; C MTΣ represents the total cost of gas turbine power generation in MMG; C OMΣ Represents the total operation and maintenance cost of the generator sets within the MMG.

2. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 1 is characterized by: In step 2, the constraints of the master-slave game optimization model include: the constraints of the upper layer ADN are as follows: 1) Distribution network flow constraints: In formula (5), n1 is the number of ADN nodes; P i , Q i are the injected active and reactive power of node i respectively; U i is the voltage amplitude of node i; G ij , B ij , δ ij are the conductance value, susceptance value, and voltage phase angle difference between node i and node j respectively; 2) ADN line power constraints: 0≤P l t ≤P l max (6); In formula (6), P l t is the active power on the lth branch of the distribution network during period t, P l max is the maximum transmission power of the lth branch; 3) ADN node voltage constraints: In the formula, are the lower and upper limits of the voltage at node i respectively; 4) ADN unit output constraints: In the formula, are the lower and upper limits of the output of unit j respectively; represents the output of unit j at time t; 5) Time-of-use electricity price constraints: The basic model of time-of-use electricity price is: Formulas (9) and (10) represent the time-of-use electricity sales price and electricity purchase price constraints set by ADN respectively; T g ,T p ,T f Represent valley, flat and peak time periods respectively; ρ sg ,ρ sp ,ρ sf Represent the electricity prices during off-peak, flat and peak periods respectively; ρ bg ,ρ bp ,ρ bf Represent the repurchase electricity prices during valley, flat and peak periods respectively; the time-of-use electricity price needs to satisfy the inequality relationship ρ sf ≥ρ sp ≥ρ sg , bf ≥ρ bp ≥ρ bg , and the electricity sales price in the same period is greater than the electricity purchase price The constraints of the lower MMG are as follows: 1) Power balance constraints of each microgrid: In formula (11): is the load of microgrid i in period t; For the new energy output of microgrid i in period t, is the power purchased by microgrid i from the distribution grid, is the power purchased by the distribution network from microgrid i, is the output of the gas turbine in microgrid i; In formula (11): represents the mutual aid power of microgrid i in the multi-microgrid alliance, where n represents the number of alliance members; D ji and D ij Both are binary state variables, indicating the state of power mutual assistance between microgrids. When 0 is taken, it means no power is transferred. ji =1 means that microgrid j transfers power to microgrid i, D ij =1 means that microgrid i transfers power to microgrid j; In formula (11): represents the power value transmitted from microgrid j to microgrid i at time t; represents the power value transmitted from microgrid i to microgrid j at time t; and They represent the energy storage charging and discharging power of microgrid i at time t respectively; 2) Constraints on energy mutual assistance within MMG: In the formula, represents the remaining power of microgrid i in period t, represents the residual load of microgrid i in period t; 3) Tie-line constraints between each microgrid and ADN: In the formula, and They represent the lower and upper limits of the power of the tie line between microgrid i and ADN respectively; 4) Upper and lower limits of gas turbine output for each microgrid: In the formula, and They represent the lower and upper limits of the gas turbine power of microgrid i, respectively; represents the output of the gas turbine of microgrid i at time t; 5) Constraints on energy storage charging and discharging of each microgrid: Where P i Cap is the energy storage power capacity of microgrid i; 6) Upper and lower bounds of energy storage capacity of each microgrid: In the formula, is the minimum state of charge of the energy storage of microgrid i; is the remaining energy storage capacity of microgrid i at time t; is the energy storage capacity of microgrid i; 7) The relationship between the power consumption of each microgrid energy storage in two adjacent time periods: In the formula, is the remaining energy storage capacity of microgrid i at time t-1; are the charging efficiency and discharging efficiency of the energy storage of microgrid i respectively; △t represents the scheduling time interval.

3. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 1 is characterized by: In the step 3, the initial parameters of the active distribution network ADN are obtained, including the distribution network grid structure, node data, branch data, capacity and operating parameters of new energy units and controllable units, tie line power limit, initial electricity sales price, and initial repurchase price; Obtain the operating parameters of the multi-microgrid alliance MMG, including the capacity and operating parameters of each microgrid gas turbine, new energy unit, and energy storage system; Obtain the particle swarm algorithm parameters for solving the master-slave game optimization model, including: population size, maximum number of iterations, particle flight speed range, learning factor, and initial inertia weight.

4. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 2 is characterized by: In step 4: The follower MMG responds to the initial purchase and sale electricity price of the leader ADN, with the minimum total operating cost of MMG as the optimization goal, as shown in formula (4), satisfying the constraints: formula (11) to formula (17), and obtaining the response set of MMG under the first round of ADN electricity price incentives Substitute the response set of MMG into the upper ADN model, use formula (3) as the objective function, consider the constraints (5) to (10), and solve the new electricity price strategy set for the next round of ADN: MMG continues to respond to ADN's electricity price strategy and gives a response set In this way, the solution is repeatedly solved in alternating iterations to obtain the optimal strategy set of the game participants in each round; in: They represent the initial values ​​of ADN electricity purchase and sale price strategies respectively; They represent the initial values ​​of MMG's electricity purchase and sales strategies respectively; They represent ADN’s electricity purchase and sale price strategies in the first cycle respectively; They respectively represent MMG’s electricity purchase and sales strategies in the first cycle.

5. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 1 is characterized by: In step 5, the strategies of each game participant obtained in the kth cycle are as follows: Where: γ 1,k is the strategy set of the leader ADN in round k; γ 2,k is the response set of follower MMG in round k; represents the electricity purchase and sale price strategy of ADN in the kth round; represents the electricity purchase and sales strategy of MMG in round k; represents the strategy of MMG in round k-1 As input, find the optimal strategy of the current k-th round leader ADN; the character arg represents the independent variable, argminJ ADN Representative J ADN The optimal ADN strategy corresponding to the minimum value Represents the strategy of the leader ADN in round k As an incentive signal, find the optimal response strategy of the follower MMG in the current kth round.

6. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 1 is characterized by: In step 6, the method for determining whether a Stackelberg-Nash equilibrium solution is found is: If the optimization results of each game participant in the kth round are consistent with those in the k-1th round, that is: This means that the game model finds the Stackelberg-Nash equilibrium solution Otherwise, update the iteration cycle round k=k+1 and return to step 5 to continue solving until the Stackelberg-Nash equilibrium solution is found; In formula (19): is the Nash equilibrium strategy set of ADN in the k-round cycle; is the Nash equilibrium strategy set of MMG in the k-round cycle; Represents the ADN electricity purchase and sales price and MMG electricity purchase and sales strategy corresponding to the Nash equilibrium solution; represents the optimal electricity price of ADN in the k-1th round; represents the optimal repurchase price of ADN in the k-1th round; represents the optimal electricity purchase of MMG in the k-1th round; It represents the optimal electricity sales of MMG in the k-1th round.

7. The active distribution network game optimization scheduling method considering multi-microgrid energy storage sharing according to claim 1 is characterized by: In step 7, the cooperative surplus allocation model is specifically as follows: The cooperative surplus generated by the alliance members through cooperation is used to obtain the benefits allocated to each member using the SVM method, which is calculated as shown in formula (20); In formula (20), s is the alliance set composed of all microgrids, v(i) is the income allocated to microgrid i, |s| is the number of subsets, ω(s) is the weight factor, v(s) is the cooperation surplus of sub-alliance s, v(si) is the cooperation surplus of sub-alliance excluding member i, n is the number of microgrids participating in the alliance; (n-|s)! represents the factorial of the number of MMG members n minus the number of members in subset s; (s|-1)! represents the factorial of the number of members in subset s minus 1; n! represents the factorial of the number of MMG members n.

Images