Cooperative optimization operation method of multi-microgrid distribution system based on equivalent projection theory

Abstract

The invention discloses a method for collaborative optimization operation of a multi-microgrid distribution system based on equivalent projection theory, which belongs to the technical field of power system. The method analyzes the flexibility resources of distribution network and microgrid, and constructs a two-layer optimization operation model of a multi-microgrid distribution system considering network reconstruction and intelligent soft switch. Then, based on the improved Fourier-Motzkin elimination method and Gaussian elimination method, the low-dimensional projection feasible domain of the lower microgrid is calculated, and the information privacy of the lower microgrid is ensured by the equivalent projection of the feasible domain. Finally, a non-iterative solution method based on equivalent projection theory is proposed to solve the two-layer optimization operation model of the multi-microgrid distribution system, and the efficient solution of the network reconstruction problem is realized under the premise of ensuring the information privacy of each subject. The results show that after introducing network reconstruction and intelligent soft switch, the collaborative optimization problem of the multi-microgrid distribution system is studied, which can significantly reduce the operation cost of the system.

Description

Technical Field

[0001] The present invention belongs to the technical field of power systems, and in particular relates to a collaborative optimization method for a multi-microgrid distribution system based on equivalent projection theory. Background Art

[0002] With the rapid development of economy and society, energy crisis and environmental crisis are becoming increasingly prominent. The vigorous development of new energy has become the main trend of future energy system development. With the large-scale access of distributed renewable energy such as wind turbine (WT) and photovoltaic (PV) to the distribution network (DN), the initiative of DN has been enhanced. Microgrid (MG) is an effective management form of distributed generation (DG). Through the coordinated optimization between MG and DN, it can not only reduce the impact of renewable energy on DN, but also provide flexibility support for DN and reduce the operating cost of the distribution system.

[0003] Distribution network reconfiguration (DNR) and intelligent soft open point (SOP) are the main network-side flexibility resources of DN. Although it is difficult to directly supply flexible power to DN, they can provide channels for the transmission of other flexible resources and reduce the operating cost of the distribution system. DNR refers to improving the power flow distribution of the system by changing the topology of the network without adding additional equipment; SOP is a power electronic device installed on the tie line branch to replace the traditional tie switch, which can realize continuous reactive power compensation and active power precise control functions.

[0004] When solving the problem of collaborative optimization operation of multi-microgrid distribution systems with multiple network-side flexibility, the collaborative optimization between DN and MG is usually expressed as a two-layer model, which can be solved by centralized and distributed methods. The centralized method transforms the two-layer model into a single-layer model through the KKT condition and the Big M method, which introduces many binary variables and increases the computational burden of the model; the distributed method can protect the privacy of each subject and share the computational burden among all subjects, but its convergence can only be strictly guaranteed when the mathematical model of each subject is convex. Therefore, it is necessary to propose a non-iterative method based on equivalent projection theory to solve the limitations of centralized and distributed methods. The non-iterative method based on equivalent projection theory can not only overcome the problem of increased computational burden when converting the two-layer model in a centralized way, but also has the characteristics of protecting the privacy of each subject in the distributed method, so it is of great value to study this field. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a collaborative optimization operation method for a multi-microgrid distribution system based on equivalent projection theory, obtain the low-dimensional projection feasible region (projection feasible region, PFR) of MG through equivalent projection theory, upload it to the DN optimization operation model, complete the transformation and efficient solution of the two-layer optimization operation model, and realize the flexible and efficient operation of the multi-microgrid distribution system.

[0006] The method for collaborative optimization operation of a multi-microgrid distribution system based on equivalent projection theory includes the following steps:

[0007] Step 1: According to the resource flexibility of the network side and node measurement of the multi-microgrid distribution system, the multi-microgrid distribution system is layered, with the distribution network as the upper system and the microgrid as the lower system. The coupling characteristics between the distribution network and the microgrid are analyzed, and a two-layer optimization operation framework of the multi-microgrid distribution system based on the equivalent projection theory is established;

[0008] Step 2: Based on the two-layer optimization operation framework, a two-layer optimization operation model of a multi-microgrid distribution system considering network reconstruction and intelligent soft switching is constructed;

[0009] Step 3: Apply the non-iterative solution method based on equivalent projection theory to transform the double-layer optimization operation model into a single-layer model for unified solution of the double-layer problem.

[0010] The objective functions of the two-layer optimization operation model of the multi-microgrid distribution system are to minimize the distribution network operation cost and the microgrid operation cost.

[0011] The distribution network operation cost is minimized by adjusting the output of distributed power sources DG, changing the network topology, adjusting the transmission power of the intelligent soft switch SOP, and adjusting the active power interaction scheme according to the low-dimensional projection feasible domain PFR of the lower microgrid to improve the flexible operation level of the distribution network.

[0012] The microgrid of the lower system includes new energy units, energy storage equipment and conventional loads.

[0013] In the two-layer optimization operation framework of the multi-microgrid distribution system, there are independent energy and information interaction channels between the microgrid and the distribution network, which are used for the interaction of active power between the microgrid and the distribution network, the uploading of PFR information of the lower microgrid, and the downward transmission of the interconnection line power calculated after centralized optimization of the distribution network layer; if there is no energy and information interaction channel between the microgrids, indirect energy and information interaction is carried out through the distribution network.

[0014] The non-iterative solution method based on equivalent projection theory is:

[0015] Input: Constraints and objective functions of distribution network and microgrid;

[0016] Output: The optimal operating cost of the distribution network and microgrid and the power generated or consumed by the internal equipment; Step 1: Calculation of the equivalent projection of the lower microgrid

[0017] Generate a complete binary tree based on the requirements of the elements of each node in the complete binary tree;

[0018] Calculate u according to formula (42): l ;

[0019]

[0020] Among them, v l (θ) is the value of the θth node on the complete binary tree path l (excluding the root node);

[0021] According to equations (43)-(44) and Φ r Computation of high-dimensional feasible domain of microgrid min,l and max,l ;

[0022]

[0023] Among them, s l and r l are the indices of the first and last non-zero elements on path l, s′ l and r′ l Equal to Ts l +1 and Tr l +1, N stands for MG r The collection of devices in i,min (t) and e i,max (t) represent the lower and upper bounds of the energy of device i at time t, 1 is a T-dimensional unit column vector, and p i,max and p i,min are the upper and lower bound vector matrices of the active power of device i,<a,b> [c,d] Represents the inner product of the cth element to the dth element in vector a and vector b;

[0024] Calculate the projection feasible region Φ′ according to Gaussian elimination method and constraints r ;

[0025] Step 2: Coordinated optimization of multi-microgrid distribution system operation Receive the projected feasible domain Φ′ of each underlying microgrid r ;

[0026] According to formula (46), the optimal cost of the multi-microgrid distribution system is calculated;

[0027]

[0028] in, is the value of the tie line power calculated after unified solution at the DN layer;

[0029] The optimal cost must satisfy the upper and lower limit constraints of the active power and reactive power of the distribution network, the climbing constraint of the active power of the distribution network, the balance constraint of the active and reactive power of the node, the line flow constraint of the reconstructed distribution network, the upper and lower limit constraints of the node voltage, the radial constraint of the distribution network, the connectivity constraint of the distribution network, the number of tie-line switch actions, the active power balance constraint of the SOP, the active power loss of the converter in the SOP, and the transmission capacity constraint of the converter in the SOP;

[0030] According to the results of collaborative optimization operation, the coordination variables of each microgrid are obtained Step 3: Microgrid Optimization Scheduling Based on Given Coordination Variables Receive the coordination variables in each microgrid obtained in step 2

[0031]

[0032] According to formula (47), the output of each device in the microgrid is calculated;

[0033] Formula (47) satisfies the following constraints:

[0034] a. Microgrid power balance constraints

[0035]

[0036] Among them, P pv,r,j,t MG at time t r The active power output of photovoltaic power; P wt,r,j,t is MG at time t r The active power output of wind power; P ess,r,j,t MG at time t r The charging and discharging power of the energy storage device, charging is positive and discharging is negative; P d,r,j,t is MG at time t r of load power.

[0037] b. Photovoltaic and wind power output constraints

[0038]

[0039] Among them, equations (34) and (35) are the upper and lower limit constraints of photovoltaic and wind power output respectively; is MG at time t r The maximum active output of photovoltaic power; is MG at time t r The maximum active output of wind power.

[0040] c. Energy storage equipment operation constraints

[0041] In order to improve the solution efficiency of multi-microgrid distribution system and its application in the non-iterative solution method of equivalent projection theory, the constraints of traditional ESS are transformed into energy storage constraints based on power and energy boundaries (36)-(38).

[0042]

[0043] Among them, formula (36) is the maximum charge and discharge power constraint of ESS, and formulas (37) and (38) are the equality constraints of the upper and lower limits of ESS energy respectively; For MG r The upper limit of the charging and discharging power of ESS; η ch,r,j , η dis,r,j For MG r Medium ESS charging and discharging efficiency; MG r The upper and lower limits of the energy of ESS at time t are, For MG r The upper and lower limits of the energy of ESS, For MG r Initial energy state of ESS.

[0044] Through the above design scheme, the present invention can bring the following beneficial effects:

[0045] (1) A two-layer optimization operation framework for multi-microgrid distribution systems based on equivalent projection theory is proposed. Based on this framework, a two-layer optimization operation model for multi-microgrid distribution systems considering DNR and SOP is constructed. By adjusting the network topology and the transmission power of SOP, the flexibility level of the network side is improved, providing a channel for the transmission of flexibility resources on the node side of the multi-microgrid distribution system, thereby improving the flexibility level of the system.

[0046] (2) A non-iterative solution method based on equivalent projection theory is proposed. The proposed method has the advantage of protecting the privacy of each subject in mainstream distributed algorithms, while overcoming the problem that distributed algorithms need to be iterated repeatedly. In particular, it can significantly improve the solution rate for mixed integer linear programming problems containing a large number of integer variables. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present invention is further described below with reference to the accompanying drawings and specific embodiments:

[0048] Figure 1 It is a schematic diagram of the double-layer optimization operation framework of the multi-microgrid distribution system based on the equivalent projection theory of the present invention.

[0049] Figure 2 It is a schematic diagram of the energy boundary of the energy storage device according to a specific embodiment of the present invention.

[0050] Figure 3 It is a schematic diagram of the specific form of a complete binary tree according to a specific embodiment of the present invention.

[0051] Figure 4 The topological structure diagram of the IEEE 33-node distribution network system improved by the specific implementation method of the present invention

[0052] Figure 5 It is a schematic diagram of the second-order approximate projection feasible region of the microgrid 1 according to a specific embodiment of the present invention.

[0053] Figure 6 It is a schematic diagram of the optimization scheduling results of the microgrid of the third-order approximate equivalent projection model according to the specific implementation method of the present invention.

[0054] Figure 7 The figure is a schematic diagram of active power transmission of the intelligent soft switch under different cases according to the specific implementation mode of the present invention.

[0055] Figure 8 The figure is a schematic diagram of reactive power transmission of the intelligent soft switch under different cases according to the specific implementation mode of the present invention.

[0056] Fig. 9 It is a schematic flow chart of the collaborative optimization operation method of a multi-microgrid distribution system based on equivalent projection theory of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention are described in further detail below.

[0058] A collaborative optimization operation method for multi-microgrid distribution system based on equivalent projection theory, such as Fig. 9 As shown, the calculation method includes the following steps:

[0059] 101: First, the multi-microgrid distribution system is hierarchically processed, and the flexibility resources existing in DN and MG are introduced respectively. DN is defined as the upper system and MG is defined as the lower system. A two-layer optimization operation framework of the multi-microgrid distribution system based on equivalent projection theory is established.

[0060] 102: Then, according to the two-layer optimization operation framework of multi-microgrid distribution system, the upper-layer DN optimization operation model and the lower-layer microgrid optimization operation model are established respectively, and the objective functions are the minimum DN operation cost and the minimum MG operation cost respectively, and the influence of network-side flexibility resources represented by SOP and DNR on the coordinated optimization of the two is explored;

[0061] 103: Finally, based on step 101 and step 102, the double-layer optimization operation problem is transformed and solved by a non-iterative solution method based on equivalent projection theory to obtain the optimal operation state of the multi-microgrid distribution system, and at the same time verify the influence of SOP and DNR on the coordinated optimization operation of the multi-microgrid distribution system.

[0062] In summary, the embodiment of the present invention can quickly and accurately solve the optimal operating state of the system after the collaborative optimization of the multi-microgrid power distribution system through the above steps 101 to 103.

[0063] The scheme of the present invention is described in detail below with reference to specific calculation formulas and examples:

[0064] 201: Establish a two-layer optimization operation framework for multi-microgrid distribution system based on equivalent projection theory. The framework structure is as follows Figure 1 As shown. In the two-layer optimization operation framework constructed in this section, the upper system is DN, the lower system is MG, the multi-microgrid distribution system contains N MGs, and the two-layer optimization problem is converted into a single-layer problem based on the equivalent projection theory for unified solution. The internal structure and dispatchable scheme of the multi-microgrid distribution system are as follows.

[0065] As an important carrier of electric energy transmission, the upper-layer system DN contains a large number of node-side and network-side flexibility resources. In the optimization operation framework studied in this section, DN can improve the flexible operation level of DN and reduce the operation cost of the distribution system by adjusting the output of DG, changing the network topology, adjusting the transmission power of SOP, and adjusting the active power interaction scheme according to the PFR of the lower-layer MG.

[0066] The lower-level system MG includes new energy units such as WT and PV, energy storage systems (ESS) and conventional loads. It can obtain certain flexibility by flexibly adjusting the output of the units and the charging and discharging behavior of the energy storage equipment, and transmit the flexibility that MG can provide to the upper-level DN in a more intuitive way through the form of PFR, thereby improving the flexibility level of DN. At the same time, MG will perform power decomposition based on the interconnection line power obtained by solving the centralized optimization operation problem of the upper-level DN, and then obtain the output of the equipment in MG.

[0067] exist Figure 1 In the two-layer optimization operation framework shown, there is an independent energy and information interaction channel between each MG and DN, which is used for the interaction of active power between MG and DN, the uploading of PFR information of the lower-layer MG, and the downward transmission of the tie line power calculated after centralized optimization of the DN layer. It is assumed that there is no energy and information interaction channel between the MGs, and only indirect energy and information interaction can be carried out through the DN.

[0068] Specifically, 1. Distribution network optimization operation model

[0069] 1) Objective function

[0070] The objective function of the DN optimization operation model is to minimize the operation cost F of DN in the entire scheduling period. DN The operating cost of DN includes the cost of purchasing electricity from the transmission grid. TN 、Distributed generation power generation cost f DG , network loss cost f LOSS and the revenue from selling electricity to MG MGs , as shown in formulas (1)-(5):

[0071] f DN =f TN +f DG +f LOSS -f MGs (1)

[0072]

[0073] Among them, λ TN,t , g , LOSS , DN,t They represent the unit cost of electricity purchased from the transmission grid, DG power generation, network loss, and the time-of-use electricity price of the DN node; P TN,j,t P is the reactive power injected into node j by the transmission network at time t; g,j,t is the active power generated by the DG located at node j at time t; is MG at time t r Active power injected into node j; Ω DG is the set of DG nodes connected to DN; Ω MGs is the set of MG nodes connected to DN; binary variable α ij,t Indicates the connectivity state of branch ij at time t. When α ij,t When it is 1, the branch is conducting, α ij,t When it is 0, the branch is disconnected; I ij,t is the square of the current flowing through branch ij at time t; r ij is the resistance value of branch ij at time t; Ω l It is the set of all branches of DN.

[0074] 2) Constraints

[0075] a. Output and ramp constraints of DG in distribution network

[0076] P g,j,min ≤P g,j,t ≤P g,j,max (6)

[0077] Q g,j,min ≤Q g,j,t ≤Q g,j,max (7)

[0078]

[0079] Among them, equations (6) and (7) are the upper and lower limit constraints of DG active power and reactive power, respectively, and equations (8) and (9) are the ramp constraints of DG active power. g,j,t is the reactive power injected into node j by DG at time t; P g,j,max , P g,j,min is the upper and lower limits of the DG active power at node j; Q g,j,max , Q g,j,min are the upper and lower limits of the reactive power of the DG at node j; are the upper and lower limits of the active power ramp of the DG at node j.

[0080] b. Node active and reactive power balance constraints

[0081]

[0082] Among them, equations (10) and (11) are the active and reactive power balance constraints of the node, respectively. 1 (j), v 2 (j) are the sets of branches flowing into and out of node j respectively; P ij,t , Q ij,t are the active and reactive power flowing through branch ij at time t respectively; P D,j,t , Q D,j,t are the active and reactive power corresponding to the load of node j at time t respectively; are the active and reactive power injected into node j by SOP at time t; Q TN,j,t is the reactive power injected into node j by the transmission grid at time t.

[0083] c. Reconstructed line power flow and node voltage constraints

[0084] In order to facilitate the solution of the proposed model, the linear power flow constraint with second-order cone relaxation is adopted as shown in equation (12). At the same time, since some branches of DN will be disconnected during the reconstruction process, the large M method is used to relax the voltage drop constraint and obtain equation (12-17).

[0085] ||2P ij,t 2Q ij,t I ij,t -u i,t || 2 ≤I ij,t +u i,t(12)

[0086]

[0087] -M(1-α ij,t )≤u i,t -u j,t -2(r ij P ij,t +x ij Q ij,t )-((r ij ) 2 +(x ij ) 2 )I ij,t ≤M(1-α ij,t )(16)

[0088] (u i,min ) 2 ≤u i,t ≤(u i,max ) 2 (17)

[0089] Among them, equations (12)-(16) are the DN line power flow constraints, equation (17) is the upper and lower limit constraints of node i voltage; u i,max 、u i,min are the upper and lower limits of the voltage at node i respectively; are the upper limits of active and reactive power flowing through branch ij respectively; It is the upper limit of the square of the current flowing through branch ij.

[0090] d. Distribution network radiality and connectivity constraints

[0091] Since DN usually adopts a radial operation mode with closed-loop design and open-loop operation, it is necessary to ensure that the reconstructed grid structure meets the radial and connectivity constraints when performing DNR.

[0092]

[0093] β ij,t +β ji,t =α ij,t (19)

[0094]

[0095] β 1j,t =0 (21)

[0096]

[0097] Among them, formula (18) is the DN radial constraint, formulas (19)-(21) are DN connectivity constraints, and formula (22) is the tie line switch operation times constraint; n b is the number of nodes in DN; binary variable β ij,t represents the flow direction of branch ij at time t, β ij,t =1 means that node j is the parent node of node i, that is, the flow direction in branch ij is from node j to node i; The upper limit of the number of switch actions allowed for DN.

[0098] e.SOP power constraint

[0099] In the optimized operation of DN, SOP can not only improve the branch power flow, but also play a role in supporting the node voltage as a reactive power compensation device. In the operation process of SOP, while considering the active power balance, it is also necessary to consider the active power loss of the converter and the transmission capacity constraint, as shown in equations (23)-(27).

[0100]

[0101] Among them, formula (23) is the active power balance constraint of SOP, formula (24) and (25) are the active power loss of the converter in SOP, and formula (26) and (27) are the transmission capacity constraints of the converter in SOP; are the active power losses caused by SOP at both sides of node i and node j at time t; η sop is the loss coefficient of the converter on both sides of the SOP; are the transmission capacities of the converters on both sides of SOP at node i and node j respectively.

[0102] Since there are nonlinear terms in equations (23)-(27), the model is difficult to solve, so it is converted into a second-order cone form, that is:

[0103]

[0104]

[0105] 2. Microgrid Optimization Operation Model

[0106] 1) Objective function

[0107] The objective function of the MG optimization operation model is to minimize the electricity purchase cost F of MG from the upper DN. MG,r .

[0108]

[0109] 2) Constraints

[0110] a. Microgrid power balance constraints

[0111]

[0112] Among them, P pv,r,j,t is MG at time t r The active power output of photovoltaic power; P wt,r,j,t is MG at time t r The active power output of wind power; P ess,r,j,t is MG at time t r The charging and discharging power of the energy storage device, charging is positive and discharging is negative; P d,r,j,t is MG at time t r of load power.

[0113] b. Photovoltaic and wind power output constraints

[0114]

[0115] Among them, equations (34) and (35) are the upper and lower limit constraints of photovoltaic and wind power output respectively; is MG at time t r The maximum active output of photovoltaic power; is MG at time t r The maximum active output of wind power.

[0116] c. Energy storage equipment operation constraints

[0117] In order to improve the solution efficiency of the multi-microgrid distribution system and facilitate the application of the non-iterative solution method based on equivalent projection theory in step 203, the constraints of the traditional ESS are transformed into energy storage constraints (36)-(38) based on power and energy boundaries. The schematic diagram of the ESS energy boundary is shown in Figure 2 shown.

[0118]

[0119]

[0120] Among them, formula (36) is the maximum charge and discharge power constraint of ESS, and formulas (37) and (38) are the equality constraints of the upper and lower limits of ESS energy respectively; For MG r The upper limit of the charging and discharging power of ESS; η ch,r,j , η dis,r,j For MG r Medium ESS charging and discharging efficiency; MG r The upper and lower limits of the energy of ESS at time t are, For MG r The upper and lower limits of the energy of ESS, For MG rInitial energy state of ESS.

[0121] 203: Non-iterative solution method based on equivalent projection theory

[0122] Since there are a large number of binary variables in the multi-microgrid distribution system optimization operation model considering DNR, the solution time of the model increases significantly. When a distributed algorithm that requires repeated iterations is adopted, the solution time will increase dramatically, which is difficult to meet actual needs. In view of the above problems, the present invention proposes a non-iterative solution method based on equivalent projection theory, which has the advantages of distributed algorithm privacy protection and the characteristics of fast solution speed of centralized method.

[0123] 1. Equivalent Projection Theory

[0124] Equivalent projection refers to eliminating internal variables from the high-dimensional feasible domain consisting of coordination variables and internal variable constraints without changing the optimality of the system, thereby obtaining a low-dimensional PFR containing only coordination variables. r As a lower-level subsystem that needs projection, MG r The high-dimensional operation feasible domain of

[0125]

[0126] Among them, Φ r For MG r The high-dimensional feasible domain of r With y r denote coordination variables and internal variables respectively, N x With N y Represents the variable x r With y r At the same time, because MG r The variables in are all bounded and their constraints are all linear constraints, that is, MG r The feasible domain is Convex polyhedron in dimensional space.

[0127] After eliminating internal variables by equivalent projection method, MG r The low-dimensional PFR based on equivalent projection is

[0128]

[0129] Among them, Φ′ r For MG r The low-dimensional PFR, formula (40) indicates that no matter the coordination variable x r Whatever the value, there must be an internal variable y that satisfies equation (39) r .

[0130] 2. Feasible domain projection method based on improved Fourier-Motzkin elimination method

[0131] Fourier–Motzkin elimination (FME) is a mathematical method used to eliminate variables in a set of linear inequalities. When FME is applied to the variable elimination problem of a multi-period coupled model with inequality constraints, the idea is to eliminate the variables in the inequality constraints from the last period forward, and then obtain a set of inequalities without the eliminated variables. However, when FME is used for variable elimination, only one variable can be eliminated at a time and multiple inequality transformations are required, which will generate a large number of redundant constraints. Therefore, it is not suitable for the feasible domain projection problem considering multi-period coupling. Therefore, it is necessary to improve the traditional FME method to obtain the improved Fourier–Motzkin elimination (IFME) without redundant constraints.

[0132] According to the algorithm principle of IFME method and the related constraints of lower-level MG, MG r The aggregate feasible domain of all devices in can be expressed as

[0133]

[0134] Among them, min,l ,Ψ max,l are the lower and upper limits of the energy corresponding to path l, respectively. P is a T-dimensional column vector consisting of P(T), P(T-1)…P(1), u l is the T-dimensional column vector of the coefficients corresponding to each element in the vector P, l is the path of the complete binary tree selected in the process of variable elimination, C represents the set of all paths of the complete binary tree with a height of T+1, and the specific form of the complete binary tree where the path l is located is as follows Figure 3 As shown. At the same time, the requirements for the elements of each node in the complete binary tree are as follows:

[0135] 1) The root node data is 0;

[0136] 2) The left child node data of each node is 0, and the right child node data is 1 or -1;

[0137] 3) If the data on the parent node and all ancestor nodes of the child node are 0, the data of the node is 1;

[0138] 4) If the data on the parent node and all ancestor nodes of the child node are not all zero, the data of the node has the opposite sign to that of its nearest non-zero ancestor node.

[0139]

[0140] Among them, v l (θ) is the value of the θth node on the complete binary tree path l (excluding the root node).

[0141] The parameter Ψ in formula (41) min,l , max,l The calculation formula is shown in equations (43)-(44):

[0142]

[0143] Among them, s l and r l are the indices of the first and last non-zero elements on path l, s′ l and r′ l Equal to Ts l +1 and Tr l +1, N stands for MG r The collection of devices in i,min (t) and e i,max (t) represent the lower and upper bounds of the energy of device i at time t, 1 is a T-dimensional unit column vector, and p i,max and p i,min are the upper and lower bound vector matrices of the active power of device i,<a,b> [c,d] Represents the inner product of the cth element to the dth element in vector a and vector b.

[0144] Although the IFME method has greatly reduced the redundancy of calculation compared with the traditional FME method, the calculation complexity of IFME is still too complicated to consider the day-ahead optimization problem of network reconstruction. Therefore, it is necessary to approximate the exact projection method and select a suitable approximate model.

[0145] The method for selecting the approximate model is: select branches whose total number of non-zero elements is less than or equal to a given value K from the branches of the complete binary tree, and then obtain a K-order approximate model. The constraint conditions are shown in formula (45):

[0146]

[0147] 3. Non-iterative solution calculation process based on equivalent projection theory

[0148] After the PFR calculation of the lower MG is completed by applying the non-iterative solution method based on equivalent projection theory, it is necessary to uniformly solve the optimization problem in the upper DN and decompose the tie line power. Therefore, it is necessary to reconstruct the objective function and constraint conditions of the optimization operation model of DN and MG proposed in step 202, as shown in equations (46)-(47):

[0149]

[0150] Among them, formula (46) is the objective function and constraint after the upper layer DN is reconstructed, and formula (47) is the objective function and constraint satisfied when the lower layer MG performs power decomposition. is the value of the tie line power calculated after a unified solution at the DN layer.

[0151] Based on the equivalent projection theory, the non-iterative solution process of the optimal operation problem of multi-microgrid distribution system can be divided into three steps, as shown in Algorithm 1.

[0152]

[0153]

[0154] In summary, the embodiment of the present invention can quickly and accurately solve the collaborative optimization problem of a multi-microgrid distribution system considering DNR and SOP through the above steps 201 to 203.

[0155] Specific application examples of the solution of the present invention:

[0156] 1. Description of the multi-microgrid distribution system example

[0157] The improved IEEE 33-node DN system topology is as follows: Figure 4 As shown in the figure, it contains 32 traditional branches and 5 tie line branches. The SOP is connected to the tie line branch of node 21-8. The four MGs are connected to the DN through nodes 12, 19, 5, and 30, respectively, and are named MG 1 MG 2 MG 3 MG 4 ,The specific parameters of the energy storage equipment are shown in Table 1.

[0158] Table 1 Parameters of energy storage equipment in each microgrid

[0159]

[0160] 2. Calculation of microgrid projection feasible domain based on equivalent projection theory

[0161] In order to ensure the privacy of the lower-level MGs and clarify the dispatchable range of the MGs, the microgrid PFR based on the equivalent projection theory is calculated by using equations (41)-(45). Since the internal structures of the MGs are the same and only the equipment parameters are different, only the second-order low-dimensional PFR of MG1 based on equivalent projection is given, as follows: Figure 5 shown.

[0162] Figure 5 For MG 1 In the above second-order approximate PFR, the gap between Ψmax and Ψmin reflects the MG1 The active power flexibility that can be provided to DN. 2 As MG increases, the gap between the upper and lower bounds increases. 1 The flexibility of the DG in the middle will gradually increase with time, and the gap corresponding to a few time periods decreases because the ESS is not only limited by the output power, but also by the initial charge state of the ESS and the upper and lower energy boundaries.

[0163] 3. Feasibility verification of non-iterative solution method based on equivalent projection theory

[0164] When the accurate PFR calculation method based on equivalent projection theory is used in the day-ahead optimization operation, the number of parameters and constraints that need to be solved are 2(2 24 -1) and 2 24 -1, the computational complexity of this calculation model obviously does not meet the requirements of actual engineering, so the approximate feasible domain calculation method based on equivalent projection theory is adopted. Therefore, it is necessary to consider both the approximate accuracy and the solution efficiency to select the appropriate approximate feasible domain order.

[0165] Table 2 Comparison of different approximate model orders and number of variables in microgrid

[0166]

[0167] Table 2 compares the order and number of variables of different approximate models of MG. As can be seen from Table 2, as the order of the approximate model gradually increases, the number of paths selected in the complete binary tree gradually increases, and the number of parameters in the approximate model also gradually increases, which will gradually increase the accuracy of the approximate model, but the computational complexity of the approximate model will also increase exponentially.

[0168] Table 3 Comparison of the impact of different approximate model orders on the dispatch results of multi-microgrid distribution system

[0169]

[0170] Table 3 compares the impact of different approximate model orders on the optimization operation results of multi-microgrid distribution system. all Represents the total operating cost of the multi-microgrid distribution system during the entire scheduling cycle. As can be seen from Table 3, the total operating cost of the system gradually increases with the increase of the order of the approximate model, the decomposition error gradually decreases from 13069.25kW of the first-order approximate model to 0kW, and the solution time of the entire model increases from 91.12s to 170.60s. The reason for the above phenomenon is that in the process of increasing the order of the approximate model, the PFR accuracy of the lower-level MG gradually increases, and the searchable range of the optimal solution gradually decreases, which increases the operating cost and reduces the decomposition error. The reason for the increase in solution time is that the number of variables and constraints that constitute the projected feasible domain increases with the increase of the order.

[0171] At the same time, it can be seen from Table 3 that when the order of the approximate model increases to the third order, the decomposition error has completely dropped to 0, and the solution time of the fourth-order approximate model is 64.10s longer than that of the third-order approximate model. Therefore, taking into account both the approximate accuracy and solution efficiency, the third-order approximate equivalent projection model is used in the subsequent research of this invention. The interactive power and equipment output results of each MG obtained under the third-order approximate equivalent model are shown in Figure 6 shown.

[0172] 4Scene Settings

[0173] In order to verify the effectiveness of the optimal operation method of the multi-microgrid distribution system considering DNR, this section sets up three typical scenarios. The settings of different scenarios are as follows:

[0174] Scenario 1: In a multi-microgrid distribution system, only the coordinated optimization scheduling between DN and MG is considered;

[0175] Scenario 2: Based on Scenario 1, further consider the role of SOP;

[0176] Scenario 3: Based on Scenario 1, the effects of DNR and SOP are considered simultaneously.

[0177] The comparison of the optimized operation cost of the multi-microgrid distribution system under different scenarios is shown in Table 4. Among them, the operation cost of the system includes the operation cost C of DN DN and each MG operating cost C MG1 , C MG2 , C MG3 , C MG4 .

[0178] Table 4 Comparison of operating costs of multi-microgrid distribution systems under different scenarios

[0179]

[0180] It can be seen from Table 4 that the operating cost of the entire distribution system gradually decreases under scenarios 1, 2, and 3. DN The cost of C is lower than that of C in scenario 1. DN The cost values ​​of the intelligent soft switch in different scenarios are reduced by 7.22% and 3.07% respectively. The active and reactive power transmission conditions of the intelligent soft switch in different scenarios and the switching action results of the DN dynamic reconstruction in different periods under scenario 3 are shown as follows: Figure 7 , Figure 8 And shown in Table 5.

[0181] Table 5 Switching action results of dynamic reconstruction of distribution network at different time periods

[0182]

[0183]

[0184] The reason why the cost of scenario 2 is lower than that of scenario 1 is that SOP is connected to the tie line branch of nodes 21-8 respectively, and active power and reactive power are injected into nodes 21 and 8 by adjusting SOP, so as to realize continuous real-time control of DN power, optimize the power flow distribution of DN, improve the network side flexibility level of DN, and improve the economic operation of multi-microgrid distribution system. Scenario 3 is based on adjusting SOP to inject power into nodes in scenario 2, and further controls the operation status of tie switches and section switches on branches in five time periods of 0:00-3:00, 4:00-13:00, 14:00-17:00, 18:00-20:00 and 21:00-24:00 according to Table 5, dynamically adjusts the network topology of DN, increases the power flow path available for DN to choose, further improves the network side flexibility level of DN, provides channels for the transmission of flexible resources such as DG and MG in DN, and further reduces the operation cost of the system.

[0185] In summary, the embodiments of the present invention study the collaborative optimization problem of multi-microgrid distribution systems by introducing network reconstruction and intelligent soft switching, which can significantly reduce the operating cost of the system.

Claims

1. A method for collaborative optimization operation of a multi-microgrid distribution system based on equivalent projection theory, which is characterized by: The following steps are included: Step 1: According to the resource flexibility of the network side and node measurement of the multi-microgrid distribution system, the multi-microgrid distribution system is layered, with the distribution network as the upper system and the microgrid as the lower system. The coupling characteristics between the distribution network and the microgrid are analyzed, and a two-layer optimization operation framework of the multi-microgrid distribution system based on the equivalent projection theory is established; Step 2: Based on the two-layer optimization operation framework, a two-layer optimization operation model of a multi-microgrid distribution system considering network reconstruction and intelligent soft switching is constructed; Step 3: Apply a non-iterative solution method based on equivalent projection theory to transform the double-layer optimization operation model into a single-layer model for unified solution of the double-layer problem; The non-iterative solution method based on equivalent projection theory is: Input: Constraints and objective functions of distribution network and microgrid; Output: The optimal operating cost of the distribution network and microgrid and the power generated or consumed by the internal equipment; Step 1: Calculation of equivalent projection of lower-level microgrid Generate a complete binary tree based on the requirements of the elements of each node in the complete binary tree; Calculate u according to formula (1): l ; Among them, v l (θ) is the value of the θth node on the complete binary tree path l; According to formula (2)-(3) and Φ r Computation of high-dimensional feasible domain of microgrid min,l andΨ max,l ; Among them, s l and r l are the indices of the first and last non-zero elements on path l, s′ l and r′ l Equal to Ts l +1 and Tr l +1, N stands for MG r The collection of devices in i,min (t) and e i,max (t) represent the lower and upper bounds of the energy of device i at time t, 1 is a T-dimensional unit column vector, and p i,max and p i,min are the upper and lower bound vector matrices of the active power of device i,<a,b> [c,d] Represents the inner product of the cth element to the dth element in vector a and vector b; Calculate the projection feasible region Φ′ according to Gaussian elimination method and constraints r ; Step 2: Coordinated optimization of multi-microgrid distribution system operation Receive the projected feasible domain Φ′ of each lower microgrid r ; According to formula (4), the optimal cost of the multi-microgrid distribution system is calculated; in, is the value of the tie line power calculated after unified solution at the DN layer, Ω MGs is the set of MG nodes connected to DN, λ DN,t is the time-of-use electricity price of the DN node; The optimal cost must satisfy the upper and lower limit constraints of the active power and reactive power of the distribution network, the climbing constraint of the active power of the distribution network, the balance constraint of the active and reactive power of the node, the line flow constraint of the reconstructed distribution network, the upper and lower limit constraints of the node voltage, the radial constraint of the distribution network, the connectivity constraint of the distribution network, the number of tie-line switch actions, the active power balance constraint of the SOP, the active power loss of the converter in the SOP, and the transmission capacity constraint of the converter in the SOP; According to the results of collaborative optimization operation, the coordination variables of each microgrid are obtained. Step 3: Microgrid Optimal Dispatch Based on Given Coordination Variables Receive the coordinated variables in each microgrid obtained in step 2 According to formula (5), the output of each device in the microgrid is calculated; Formula (5) satisfies the following constraints: a. Microgrid power balance constraints Among them, P pv,r,j,t MG at time t r The active power output of photovoltaic power; P wt,r,j,t MG at time t r The active power output of wind power; P ess,r,j,t MG at time t r The charging and discharging power of the energy storage device, charging is positive and discharging is negative; P d,r,j,t is MG at time t r Load power; b. Photovoltaic and wind power output constraints Among them, equations (7) and (8) are the upper and lower limit constraints of photovoltaic and wind power output respectively; MG at time t r The maximum active output of photovoltaic power; MG at time t r The maximum active output of wind power; c. Energy storage equipment operation constraints In order to improve the solution efficiency of multi-microgrid distribution system and its application in the non-iterative solution method of equivalent projection theory, the constraints of traditional ESS are transformed into energy storage constraints based on power and energy boundaries (9)-(11). Among them, formula (9) is the maximum charge and discharge power constraint of ESS, and formulas (10) and (11) are the equality constraints of the upper and lower limits of ESS energy respectively; For MG r The upper limit of the charging and discharging power of ESS; η ch,r,j , η dis,r,j For MG r Medium ESS charging and discharging efficiency; MG r The upper and lower limits of the energy of ESS at time t are, For MG r The upper and lower limits of the energy of ESS, For MG r Initial energy state of ESS.

2. The method for collaborative optimization operation of a multi-microgrid power distribution system based on equivalent projection theory according to claim 1 is characterized in that: The objective functions of the two-layer optimization operation model of the multi-microgrid distribution system are to minimize the distribution network operation cost and the microgrid operation cost.

3. The method for collaborative optimization operation of a multi-microgrid power distribution system based on equivalent projection theory according to claim 2 is characterized in that: The distribution network operation cost is minimized by adjusting the output of distributed power sources DG, changing the network topology, adjusting the transmission power of the intelligent soft switch SOP, and adjusting the active power interaction scheme according to the low-dimensional projection feasible domain PFR of the lower microgrid to improve the flexible operation level of the distribution network.

4. The method for collaborative optimization operation of a multi-microgrid power distribution system based on equivalent projection theory according to claim 2 is characterized in that: The microgrid of the lower system includes new energy units, energy storage equipment and conventional loads.

5. The method for collaborative optimization operation of a multi-microgrid power distribution system based on equivalent projection theory according to claim 1 is characterized in that: In the two-layer optimization operation framework of the multi-microgrid distribution system, there are independent energy and information interaction channels between the microgrid and the distribution network, which are used for the interaction of active power between the microgrid and the distribution network, the uploading of PFR information of the lower microgrid, and the downward transmission of the tie line power calculated after centralized optimization of the distribution network layer; If there is no energy and information exchange channel between microgrids, indirect energy and information exchange is carried out through the distribution network.

Images