Flexible interconnection micro-grid group uncertainty absorption and load balancing method

Abstract

The invention discloses a flexible interconnected micro-grid group uncertainty absorption and load balancing method, which comprises the steps of S1, quantifying a micro-grid group uncertainty source, and determining an energy storage two-stage regulation and control strategy; s2, a distributed robust opportunity constraint-based microgrid group optimization regulation and control model is constructed by taking the minimum operation cost as a target, and the model comprises a line transmission power opportunity constraint, a node voltage opportunity constraint, an energy storage charging and discharging power opportunity constraint and an energy storage capacity safety opportunity constraint; the method further comprises energy storage capacity constraint, network power flow constraint, VSC capacity constraint and alternating current area power balance constraint. S3, performing linear reconstruction on the uncertainty penalty cost by adopting a Wasserstein Metric definition and a dual theorem, and converting an opportunity constraint into a deterministic constraint; s4, carrying out linear reconstruction on other nonlinear parts of the model objective function and the constraint condition; and S5, solving the micro-grid group optimization regulation and control model, and optimizing, regulating and controlling the flexible interconnection micro-grid group.

Description

Technical Field

[0001] The present invention relates to the technical field of microgrid control, and particularly to a method for flexible interconnection of microgrid groups to eliminate uncertainty and balance load. Background Art

[0002] With the advancement of the green transformation of the power industry, the distribution network is undergoing profound changes: the large-scale access of distributed new energy and the continuous growth of DC load demand have brought unprecedented challenges. Specifically, the unbalanced layout of new energy leads to unbalanced load distribution, which in turn causes voltage over-limit problems; at the same time, the uncertainty caused by the prediction error of new energy output exacerbates voltage fluctuations, seriously threatening the safe and stable operation of the distribution network.

[0003] However, with the rapid development of power electronics technology, multiple low-voltage distribution network areas are flexibly interconnected through voltage source converters (VSCs), and the formed flexible interconnection microgrid architecture based on chain segment topology has advantages such as spatio-temporal complementarity of sources and loads, complementarity between uncertainty generation sources and energy storage, and coordinated scheduling of multiple resources, providing an effective way to solve the above problems.

[0004] The emergence of low-voltage AC / DC flexible interconnection microgrid groups has promoted the evolution of the distribution network towards a multi-level architecture. Through autonomous optimization, the lower-level microgrid groups can effectively reduce the regulation scale and coordination complexity of the upper-level distribution network. However, with a large number of wind power, photovoltaic, and energy storage connected to the DC bus of the flexible interconnection microgrid group and energy interaction with the superior distribution network through VSCs, the operation and regulation of the system face higher uncertainty and complexity. Traditional system regulation methods do not consider source-load uncertainty. Although an "optimal" regulation solution can be obtained, due to ignoring the actual fluctuations of energy storage output and the purchase and sale electricity of each level of the distribution area, the feasibility of the solution is greatly reduced, increasing the system regulation difficulty and making the "optimal" result often difficult to be actually implemented. Currently, the methods for dealing with multi-source uncertainty in the system mainly include stochastic optimization and robust optimization. The former has a clear model structure and simple derivation, but it requires the probability distribution of uncertain quantities to be known, and in practice, the distribution information is often difficult to accurately obtain, and the solution efficiency also decreases significantly when the number of scenarios increases; the latter is modeled based on the worst-case scenario, which can ensure the safe and stable operation of the system under any scenario, but this method is too conservative, resulting in low system operation economy. Therefore, it is urgent to study an autonomous regulation method for flexible interconnection microgrid groups considering uncertainty to achieve the coordinated optimization of microgrid group resources and ensure the economic, safe, and stable operation of the microgrid group. Summary of the Invention

[0005] To balance the economy and regulation reliability of the system, the present invention provides a method for uncertainty accommodation and load balancing in a flexible interconnected microgrid group. This method introduces a distributionally robust chance-constrained optimization method driven by historical data, which can effectively reduce the over-conservatism of robust optimization while improving the solution efficiency, and achieve improved operating performance under multiple scenarios.

[0006] To solve the above technical problems, the present invention adopts the following technical method: A method for uncertainty accommodation and load balancing in a flexible interconnected microgrid group, including: S1. Quantify the uncertainties of wind power, photovoltaic power, and DC load in the microgrid group, and determine a two-stage regulation strategy for energy storage day-ahead scheduling and real-time adjustment; S2. With the goal of minimizing the sum of the superior power purchase and sale cost, energy storage operation cost, load balancing penalty cost, and uncertainty penalty cost, construct an optimal regulation model for the microgrid group based on distributionally robust chance constraints. The constraints of this model include chance constraints and other constraints. The chance constraints include line transmission power chance constraints, node voltage chance constraints, energy storage charge and discharge power chance constraints, and energy storage capacity safety chance constraints. The other constraints include energy storage capacity constraints, network power flow constraints, VSC capacity constraints, and AC area power balance constraints; S3. Use the Wasserstein Metric definition and the duality theorem to linearly reconstruct the uncertainty penalty cost in the model objective function, and transform the chance constraints of the model into deterministic constraints; S4. Linearly reconstruct other non-linear parts of the model objective function and constraint conditions; S5. Obtain the network topology, equipment, and distribution network parameter information of the microgrid group, collect several groups of historical data sample points of wind power and photovoltaic historical output fluctuations at each moment, organize the equivalently linearized optimal regulation model of the microgrid group and solve it to obtain regulation parameters, and optimize and regulate the flexible interconnected microgrid group.

[0007] Further, in S3, when using the Wasserstein Metric definition and the duality theorem to linearly reconstruct the uncertainty penalty function, first collect a series of historical sample data sets of random variables as the historical uncertainty fluctuation data set of the uncertainty source. Then, before obtaining the distribution fuzzy set of the random variable, determine the Wasserstein ball radius according to the fuzzy set radius size formula. Finally, based on the Wasserstein Metric definition, transform the uncertainty penalty function into a dual problem and simplify it by introducing a supremum approximation.

[0008] Further, in S3, when converting the chance constraints of the microgrid group optimal regulation model into deterministic constraints using the Wasserstein Metric definition and the duality theorem, first calculate the mean and variance of the historical sample data sets of the random variables at each moment, and normalize all the sample points into the standard coordinate system. Then, solve the problem of minimizing the side length of the hypercube to obtain the side length of the hypercube of the sample points at a certain confidence level in the standard coordinate system, and determine the corresponding hypercube. Next, anti-normalize the hypercube into the original coordinate system to obtain the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding box uncertainty set. Finally, use the duality theorem to transform the original chance constraint problem into a robust problem under this box uncertainty set, and then: t Substitute the side length of the hypercube of the sample points at each moment to obtain the side length of the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding parallelepiped. Then, anti-normalize the parallelepiped into the original coordinate system to obtain the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding box uncertainty set. Finally, use the duality theorem to transform the original chance constraint problem into a robust problem under this box uncertainty set, and then: t Substitute the side length of the hypercube of the sample points at each moment to obtain the side length of the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding parallelepiped. Then, anti-normalize the parallelepiped into the original coordinate system to obtain the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding box uncertainty set. Finally, use the duality theorem to transform the original chance constraint problem into a robust problem under this box uncertainty set, and then: Substitute the network power flow constraint into the line transmission power chance constraint, and equivalently transform it to obtain the line transmission power deterministic constraint; Substitute the network power flow constraint into the node voltage chance constraint, and equivalently transform it to obtain the node voltage deterministic constraint; Substitute the energy storage output into the energy storage charge and discharge power chance constraint, and equivalently transform it to obtain the energy storage output deterministic constraint; Substitute the energy storage capacity constraint into the energy storage capacity safety chance constraint, and equivalently transform it to obtain the energy storage capacity deterministic constraint.

[0009] Furthermore, the regulation parameters include the day-ahead output reference and real-time adjustment factor of each energy storage in the microgrid group, and the upstream power purchase and sale volume of each substation area.

[0010] Furthermore, in S1, quantify the uncertainties of the wind power, photovoltaic power, and DC load in the microgrid group as follows: (1) (2) (3) In the formula, represents the column vector of the actual output of each uncertainty source at time ; represents the uncertainty source-node incidence matrix, including the photovoltaic-node incidence matrix , the wind power-node incidence matrix , and the DC load-node incidence matrix ; represents the predicted value of the uncertainty source at time , including the predicted value of photovoltaic power at time , the predicted value of wind power , and the predicted value of DC load ; is the prediction error column vector of each uncertainty source, representing the uncertain quantity; In S1, the energy storage output of the microgrid cluster is divided into day-ahead scheduling and real-time adjustment. The output of the energy storage at each node is as follows: (4) In the formula, represents the real-time output of each energy storage at time represents the energy storage-node incidence matrix, is the benchmark of the day-ahead output of the energy storage; is the real-time adjustment factor matrix, and its specific definition is: (5) In the formula, the th row and th column element of represents the proportion of the mitigation of the uncertain quantity of the th uncertainty source by the th energy storage device, is the number of energy storage devices, (6) (7) Formula (6) means that the proportion of each energy storage in mitigating each uncertain quantity is less than 100%; Formula (7) means that at time the total compensation of the th uncertain quantity by all energy storages is less than or equal to the uncertain quantity itself.

[0011] Furthermore, in S2, the optimized regulation model of the microgrid cluster based on distributionally robust chance-constrained is: (8) In the formula, is the optimization variable; are the superior power purchase and sale cost, energy storage operation cost, load sharing penalty cost, and uncertainty penalty cost respectively; (9) In the formula, is the set of all scheduling times, is the set of the numbers of all nodes in the microgrid; and are the power purchase cost and the power sale unit price respectively; is the The power interaction volume between a node and the superior distribution network; operator Satisfy ; (10) In the formula, Is the set of numbers of all energy storages in the microgrid; Is the operating cost per unit output of the energy storage; Is the active power output of each energy storage; (11) In the formula, Is the equal load penalty coefficient; , Are respectively the 、 Transformer capacities of the Is the Power interaction volume between the Is the number of nodes in the microgrid group; (12) (13) (14) In the formula, Is the uncertainty penalty coefficient; Represents the set of numbers of all uncertainty sources; Represents the Uncertainty distribution fuzzy set of the Uncertainty source at time Represents the maximum expected value under all possible probabilities; Is the column vector The Element in, , Whose physical meaning represents the proportion of the uncertainty eliminated by all energy storages in the microgrid group for the Uncertainty generated by the Represents a column vector of all 1s with a dimension of n×1; The chance constraints of the microgrid group optimization and control model constructed in the above S2 include: 1) Line transmission power chance constraint (15) In the formula, Represents the set of all probability distributions that the uncertain quantity may satisfy; Represents the minimum probability for the formula in the parentheses to hold; Represents at DC line transmission power vector of the microgrid cluster at a moment, and are the minimum and maximum transmission power vectors of the DC line respectively; represents the confidence level; 2) Node voltage chance constraint (16) In the formula, represents the DC bus voltage vector of the microgrid cluster at moment, and are the maximum and minimum values of the DC bus voltage respectively; 3) Energy storage charge and discharge power chance constraint (17) In the formula, represents the output column vector of the energy storage at moment, and it is stipulated that the energy storage discharge is the positive direction; and are the maximum and minimum values of the energy storage charge and discharge power respectively; 4) Energy storage capacity safety chance constraint (18) In the formula, represents the energy storage capacity vectors of the microgrid cluster at moment; and are the lower and upper limits of the energy storage capacity respectively.

[0012] For the above-mentioned S2, other constraints of the microgrid cluster optimization control model constructed based on the distributionally robust chance constraint include: 1) Energy storage capacity constraint (19) In the formula, is the energy storage charging efficiency; is the energy storage self-discharge coefficient; represents the daily scheduling interval, taking 1; 2) Network power flow constraint (20) (21) (22) In the formula, represents the node injection power; represents the active power injected by the VSC, is the augmented node-line incidence matrix, obtained from the microgrid topological structure diagram, i.e., the connection relationship of the distribution transformers; is the augmented line resistance matrix; represents the column vector of the DC bus voltage of the microgrid group; is the voltage reference; 3) VSC capacity constraint (23) In the formula, and respectively represent the active power injected by the VSC and the reactive power output by the VSC at the th moment at the th node; is the capacity of the VSC at the th node; 4) AC area power balance constraint (24) In the formula, represents the column vector of the power purchased and sold by the superior, is the column vector of the AC load of the AC node.

[0013] In S3, the uncertainty penalty function is linearly reconstructed by using the Wasserstein Metric definition and the duality theorem as: 1) Obtain the Wasserstein ball radius Collect a series of random variables historical sample data set, that is, the historical uncertainty fluctuation data set of the uncertainty source, and the sample size is M ; before obtaining the distribution fuzzy set of the random variable , the size of the fuzzy set radius can be determined by the following formula: (35) In the formula, represents the diameter size of the data support set; is the confidence level including all distributions of the random variable ; 2) Problem transformation Based on the Wasserstein Metric definition, the original objective function is transformed into a dual problem: (36) In the formula, is the dual multiplier; is the auxiliary variable, is the set of sample numbers; represents the feasible set of the uncertain quantity ; The constraint conditions contained in Equation (36) face the problem of scalability, that is, as the sample data volume gradually increases, the solution scale gradually expands; therefore, the supremum approximation is introduced, and Equation (36) is further simplified to: (37) (25) where 、 represent the maximum and minimum values of the historical data of the -th uncertainty source at the -th moment; is the radius size of the fuzzy set; represents the number of samples; represents the historical sample data set of the -th component in the uncertain quantity at each moment; In the above S3, the chance constraint of the microgrid group optimal regulation model is transformed into a deterministic constraint by using the Wasserstein Metric definition and the duality theorem; 1) Normalize all historical sample points to the standard coordinate system Calculate the mean and variance of the historical sample data set of the random variable at each moment, and normalize all sample points to the standard coordinate system: (38) where represents the coordinates of the sample point in the standard coordinate system; 2) Obtain the side length of the hypercube in the standard coordinate system, solve the problem of minimizing the side length of the hypercube, and obtain the hypercube enclosing the sample points at a certain confidence level at the -th moment : (39) where is the dual variable, is the radius size of the fuzzy set obtained according to Equation (35) for the sample point in the standard coordinate system; is the set containing the coordinates of each vertex of the hypercube, which is further described as: (40) where is a column vector of all 1s with a dimension of ; 3) Inverse normalization of the hypercube to the original coordinate system Inverse normalize the obtained hypercube to the original coordinate system to obtain a parallelotope of sample points at a certain confidence level in the original coordinate system Lower envelope at a certain moment , and its vertices can be obtained by substituting Equation (40) into the following formula: (41) is the set of all vertex coordinates of the parallelotope, which is further described as the following box uncertainty set: (42) This formula (42) describes the internal space of the parallelotope , is a constant matrix, and the element in its th row and th column is 1, the element in its th row and th column is -1, and the rest of the elements are all 0; is a constant column vector, and the th row and the th row are both the absolute value of the coordinates of the th vertex in ; 4) Proposition proof Let the optimization variable be , and all chance-constrained problems in the model can be described in a unified form: (43) The above formula always holds, that is, the maximum value on the left side of the inequality is less than or equal to 0: (44) For the problem of finding the maximum value of the left side of the inequality in Equation (44), using the duality theorem, Equation (44) is further transformed into: (45) Therefore, Equation (43) can be equivalently transformed into the following proposition: (46) 5) Transformation of chance constraints into deterministic constraints Substitute Equation (20) and Equation (21) into the chance constraint Equation (15), and the following robust problem can be obtained: (47) Based on the above-proved proposition, Equation (47) can be equivalently transformed into the following deterministic constraint of line transmission capacity: (26) In the formula, and are dual variables related to power capacity constraints; Similarly, substituting Equation (20) and Equation (22) into Equation (16), based on the above-proven proposition, the equivalent transformation yields the deterministic constraint of the node voltage as follows: (27) In the formula, and are dual variables related to voltage constraints; Substituting Equation (4) into Equation (17), based on the above-proven proposition, the equivalent transformation yields the deterministic constraint of the energy storage output as follows: (28) In the formula, and are dual variables related to energy storage power constraints; Substituting Equation (19) into Equation (18), based on the above-proven proposition, the equivalent transformation yields the deterministic constraint of the energy storage capacity as follows: (29) In the formula, is the self-discharge coefficient of the energy storage; is the initial capacity of the energy storage; is the dual variable related to energy storage capacity constraints.

[0014] Preferably, the confidence level of the chance constraint is 95%.

[0015] Furthermore, in S4, auxiliary variables are introduced to linearly reconstruct the upper-level power purchase and sale cost function and the energy storage operation cost function; 1) Introduce auxiliary variable The upper-level power purchase and sale cost function is linearly reconstructed as: (30) 2) Introduce auxiliary variable The energy storage operation cost function is linearly reconstructed as: (31).

[0016] Furthermore, in S4, the constraint of the linearized converter capacity constraint containing quadratic terms is approximated and simplified by approximating the circular feasible interval of the VSC capacity constraint to an inscribed regular dodecagon feasible interval, expressed in the standard straight-line equation form of each side length, as follows: (32) In the formula, , and is the coefficient of the straight line equation of the th side in a regular dodecagon.

[0017] Furthermore, in the above S5, the obtained equivalent linearized optimal control model of the microgrid group is: (33) (34) In the formula, is a set of optimization variables, including: optimization variable , as well as auxiliary variables and dual multipliers introduced in the equivalent transformation process.

[0018] Compared with the traditional control method, the flexible interconnected microgrid group uncertainty accommodation and load balancing method provided by the present invention introduces a distributionally robust chance-constrained optimization method based on historical data-driven, which improves the solution efficiency while effectively reducing the overly conservative problem of robust optimization, realizes the equal load sharing of the load in the microgrid group, the efficient control of energy storage and VSC, the complete accommodation of wind power and photovoltaic power, and also reduces the adverse effects caused by the fluctuations of new energy power generation on the microgrid group and the superior distribution network, realizing the economic, safe and stable operation of the flexible interconnected microgrid group. Moreover, the present invention also optimizes the distributionally robust chance-constrained method of the main grid by using the duality principle, so that the uncertain sources in the system of the present invention can be of any dimension, which greatly simplifies the solution difficulty of the system model. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flowchart of the flexible interconnected microgrid group uncertainty accommodation and load balancing method involved in the present invention; Figure 2 is a basic topological structure diagram of the flexible interconnected microgrid group in the embodiment of the present invention; Figure 3 is a normalized output curve diagram of wind power, photovoltaic power, AC and DC loads in the embodiment of the present invention; Figure 4 is a schematic diagram of the remaining total uncertainty in the microgrid group system in the embodiment of the present invention; Figure 5 is a schematic diagram of the uncertainty mitigation ratio of each uncertainty source at different times based on the RO method in the embodiment of the present invention; Figure 6 is a schematic diagram of the uncertainty mitigation ratio of each uncertainty source at different times based on the DRCO method in the embodiment of the present invention; Figure 7 is a superior purchase and sale fluctuation curve diagram of substation area 2 in the microgrid group system in the embodiment of the present invention; Figure 8 Schematic diagram of the DC bus voltage fluctuation of substation area 2 in the WU method in the embodiment of the present invention Figure 9 Schematic diagram of the DC bus voltage fluctuation of substation area 2 in the RO method in the embodiment of the present invention Figure 10 Schematic diagram of the DC bus voltage fluctuation of substation area 2 in the DRCO method in the embodiment of the present invention Figure 11 Comparison diagram of the total operating cost distribution under three methods in the embodiment of the present invention Figure 12 Schematic diagram of the output of the energy storage in substation area 3 under the RO and DRCO methods in the embodiment of the present invention Figure 13 Schematic diagram of the load conditions of each substation area in the microgrid group under different scenarios in the embodiment of the present invention Detailed implementation manners

[0020] For the convenience of understanding by those skilled in the art, the present invention will be further described below in conjunction with embodiments and the accompanying drawings. The content mentioned in the embodiments does not limit the present invention

[0021] The flow of a method for flexible interconnection of a microgrid group to eliminate uncertainty and balance load is as Figure 1 shown, and the specific steps are as follows S1. Propose a two-stage real-time adjustment strategy for energy storage

[0022] In a low-voltage flexible interconnected AC-DC microgrid, wind power, photovoltaic power, DC loads, etc. are centrally connected to the DC bus of each substation area. The uncertainty brought by the prediction error of their output will have an adverse impact on the distribution network

[0023] 1. Define the output mode of uncertainty sources To quantify the uncertainty of wind power, photovoltaic power, and DC loads in the microgrid group and formulate strategies to reduce uncertainty, the output of the above uncertainty sources at each node is specified as follows (1) (2) (3) In the formula, represents the actual output column vector of each uncertainty source at time represents the uncertainty source-node incidence matrix, including the photovoltaic-node incidence matrix , the wind power-node incidence matrix , and the DC load-node incidence matrix , these matrices are obtained from the parameter information of the node positions where wind power, photovoltaic power, and loads are connected to the microgrid; denote the predicted values of the uncertainty sources at time including the predicted photovoltaic power at time , the predicted wind power , and the predicted DC load power . These values are obtained from the output powers and normalized output curves of wind power, photovoltaic power, and DC loads; is the column vector of the prediction errors of each uncertainty source, representing the uncertain quantity; 2. Energy storage control strategy The energy storage output of the microgrid group is divided into day-ahead scheduling and real-time adjustment: among them, day-ahead scheduling is to ensure the safe, stable, and economic operation of the microgrid group system, and real-time adjustment is to suppress the adverse effects brought by uncertainties. The energy storage output at each node can be defined as: (4) In the formula, denote the real-time output of each energy storage at time denote the energy storage-node association matrix, which is obtained from the parameter information of the node positions where the energy storage is connected to the microgrid; is the day-ahead output reference of the energy storage; is the real-time adjustment factor matrix, and its specific definition is: (5) In the formula, the element in the th row and th column of , denote the proportion of the mitigation amount of the uncertain quantity of the th uncertainty source by the th energy storage device, is the number of energy storage devices, is the number of uncertainty sources, (6) (7) Equation (6) means that the proportion of each energy storage mitigating each uncertain quantity is less than 100%, that is, overcompensation of uncertainties is not allowed; Equation (7) means that at time the total compensation amount of the th uncertain quantity by all energy storage devices is less than or equal to the uncertain quantity itself.

[0024] S2. Construct an optimal control model for the microgrid group based on distributionally robust chance constraints.

[0025] To achieve the economic, safe and stable operation of the microgrid cluster, balance the loads of each substation area, and reduce the adverse effects of uncertainties on the microgrid cluster, an optimal regulation model of the microgrid cluster based on the distributionally robust chance-constrained method is constructed.

[0026] 1. Objective function The objective function of the optimal regulation model of the microgrid cluster is: (8) In the formula, is the optimization variable; are the upstream power purchase and sale cost, energy storage operation cost, load balancing penalty cost, and uncertainty penalty cost respectively; 1) Upstream power purchase and sale cost When the microgrid cluster has excess energy, it can be sold to the distribution network at a lower price; when the energy is insufficient, it can purchase power from the upstream distribution network; therefore, the power purchase and sale cost is: (9) In the formula, is the set of all scheduling times, is the set of the numbers of all nodes in the microgrid; and are the power purchase cost (i.e., time-of-use electricity price) and the power sale unit price (i.e., the power sale price of the microgrid cluster to the distribution network) respectively; is the power interaction volume between the -th node and the upstream distribution network; the operator satisfies ; 2) Energy storage operation cost In the microgrid cluster, the energy storage operation needs to bear a certain cost : (10) In the formula, is the set of the numbers of all energy storages in the microgrid; is the unit output operation cost of the energy storage; is the active power output of each energy storage; 3) Load balancing penalty cost To relieve the operation pressure of the transformers in each substation area of the microgrid cluster, the load balancing penalty cost of each node in the microgrid cluster is set as: (11) In the formula, is the load balancing penalty coefficient, The value of is set to obtain a good load balancing effect during the debugging and simulation process. After comprehensive debugging, is taken as 1. If a better load balancing effect is required, can be gradually increased. Get larger; , Are the transformer capacities of the 、 th node / substation area in the microgrid; Is the th node's power interaction volume with the superior distribution network; Is the number of nodes / substation areas in the microgrid group; 4) Uncertainty penalty cost To mitigate the adverse effects of uncertainty on each substation area, the uncertainty penalty cost Is defined as: (12) (13) (14) In the formula, Is the uncertainty penalty coefficient, Is a very large number. To reduce the remaining uncertainty in the system, a relatively good optimization result is obtained during the debugging and simulation process, and the value is 10^6; Represents the set of numbers of all uncertainty sources; Represents the th uncertainty source's uncertainty distribution fuzzy set at moment; Represents the maximum expected value under all possible probabilities; Is a column vector The th element in, , whose physical meaning represents the proportion of the uncertainty eliminated by all energy storages in the microgrid group for the th uncertainty source; Represents a column vector of all 1s with a dimension size of n×1; 2. Construct chance constraints The low-voltage AC-DC distribution network formed by the interconnection of multiple substation areas through VSCs is the flexible interconnected microgrid involved in the present invention. New energy devices such as wind power, photovoltaic power, and energy storage are directly connected to the DC bus. Therefore, the prediction errors of uncertainty sources such as wind power, photovoltaic power, and DC loads will cause fluctuations in the DC bus voltage, the transmission power of the DC line, and the output of the energy storage. It is necessary to limit the risk of their fluctuations exceeding the limit. Therefore, chance constraints for the network security power flow of the microgrid group and chance constraints for the safe operation of the energy storage are constructed.

[0027] 1) Chance constraint for line transmission power To make full use of energy storage resources and limit the line transmission power within a safe range, the line transmission power at each moment must satisfy the safe transmission constraints with a certain confidence level, that is: (15) In the formula, represents the set of all probability distributions that the uncertain quantity may satisfy; represents the minimum probability for the formula in the brackets to hold; represents at the DC line transmission power vector of the microgrid cluster at a certain moment, 、 are the minimum and maximum transmission power vectors of the DC line respectively; represents the confidence level; 2) Node voltage chance constraint The voltage of each node at each moment must satisfy the safety constraint with a certain confidence level, that is: (16) In the formula, represents at the DC bus voltage vector of the microgrid cluster at a certain moment, 、 are the maximum and minimum values of the DC bus voltage respectively; 3) Energy storage charge and discharge power chance constraint With a certain confidence level, the charge and discharge power of each energy storage at each moment must satisfy the safe operation constraints of the energy storage, that is: (17) In the formula, represents at the output column vector of the energy storage at a certain moment, and it is stipulated that the energy storage discharge is in the positive direction; 、 are the maximum and minimum values of the charge and discharge power of the energy storage respectively; 4) Energy storage capacity safety chance constraint With a certain confidence level, the capacity of each energy storage at each moment must satisfy the safe operation constraints of the energy storage, that is: (18) In the formula, represents at the capacity vector of each energy storage in the microgrid cluster at a certain moment; 、 are the lower and upper limits of the energy storage capacity respectively.

[0028] 3. Construct other constraints 1) Energy storage capacity constraint The capacity safety constraint of the energy storage capacity at each moment is: (19) Wherein, is the energy storage charging efficiency; is the self-discharge coefficient of the energy storage; represents the daily scheduling interval, taking 1; 2) Network power flow constraint The relationship between the node injection power, the network line transmission power, and the network node voltage is expressed as follows: (20) (21) (22) Wherein, represents the node injection power; represents the active power injected by the VSC; is the augmented node-line incidence matrix, obtained from the microgrid topology structure diagram, i.e., the connection relationship of the distribution transformer area; is the augmented line resistance matrix, obtained by multiplying the unit resistance of the DC line by the distance of the distribution transformer area line; represents the DC bus voltage column vector of the microgrid group; is the voltage reference; 3) VSC capacity constraint The operation of the VSC satisfies the capacity safety constraint: (23) Wherein, , respectively represent the active power injected by the VSC and the reactive power output at the th moment at the th node; is the capacity of the VSC at the th node; 4) AC area power balance constraint The AC area node is the power interaction node between the microgrid and the distribution network, satisfying the power balance: (24) Wherein, represents the superior power purchase and sale power column vector, is the AC load column vector of the AC node.

[0029] S3, the uncertainty penalty cost in the model objective function is linearly reconstructed using the Wasserstein Metric definition and the duality theorem, and the chance constraints of the model are transformed into deterministic constraints.

[0030] 1. Reconstruction of the uncertainty penalty function based on the distributionally robust method.

[0031] In the distributionally robust problem, Wasserstein Metrics has received extensive attention due to its simplicity and intuitiveness. Starting from its basic definition, this invention uses the duality theorem to gradually simplify the high-dimensional and difficult-to-solve equation (12) into a linearly solvable problem.

[0032] 1) Obtain the Wasserstein ball radius Collect a series of random variables The historical sample data set, that is, the historical uncertainty fluctuation data set of the uncertainty source, with a sample size of M ; Before obtaining the distribution fuzzy set of the random variable The size of the fuzzy set radius Can be determined by the following formula: (35) In the formula, Represents the diameter size of the data support set; Is the confidence level containing all distributions of the random variable ; 2) Problem transformation Based on the definition of Wasserstein Metric, transform the original objective function Into a dual problem: (36) In the formula, Is the dual multiplier; Is an auxiliary variable, Is the set of sample numbers; Represents the feasible set of the uncertain quantity ; The constraint conditions contained in formula (36) face the problem of scalability, that is, as the sample data volume gradually increases, the solution scale gradually expands; therefore, a supremum approximation is introduced, and formula (36) is further simplified to: (37) (25) In the formula, , Represents the maximum and minimum values of the historical data of the -th uncertainty source at the -th moment; Is the size of the fuzzy set radius, determined by formula (35); Represents the number of samples; Represents the historical sample data set of the -th component in the uncertain quantity at each moment, which is generated based on the Gaussian distribution function; 2. Reconstruct the probability constraint based on the distributionally robust chance constraint.

[0033] The implementation idea of transforming the chance constraint in the present invention is as follows: Since there are a small number of extreme scenarios or error noises in the data-driven historical dataset, a distributionally robust method based on the Wasserstein Metric is adopted to obtain a box uncertainty set that envelopes the sample points under a certain confidence level, and the original chance constraint problem is transformed into a robust problem under this set.

[0034] 1) Normalize all historical sample points to the standard coordinate system Calculate the random variables at each moment Historical sample dataset The mean value Variance , and all sample points Are normalized to the standard coordinate system: (38) In the formula, Represents the coordinates of the sample point in the standard coordinate system; 2) Obtain the side length of the hypercube in the standard coordinate system Solve the problem of minimizing the side length of the hypercube to obtain the hypercube That envelopes the sample points at the moment In the standard coordinate system with a certain confidence level Side length : (39) In the formula, Is the dual variable, Is the sample point in the standard coordinate system The radius size of the fuzzy set obtained according to formula (35); Is the set containing the vertex coordinates of the hypercube, which is further described as: (40) In the formula, Is a column vector of all 1s with a dimension of ; 3) Inverse normalize the hypercube to the original coordinate system Inverse normalize the obtained hypercube to the original coordinate system to obtain a parallelepiped That envelopes the sample points at the moment In the original coordinate system with a certain confidence level , and its vertices can be obtained by substituting formula (40) into the following formula: (41) is a set containing the vertex coordinates of a parallelepiped, which is further described as the following box uncertainty set: (42) Equation (42) describes the parallelepiped internal space, is a constant matrix, and its row and column element is 1, the row and column element is -1, and the remaining elements are all 0; is a constant column vector, and the row and the row are both the absolute values of the coordinates of the th vertex in ; 4) Proposition proof Let the variable , and all chance-constrained problems in the model can be described in a unified form: (43) The above equation always holds, that is, the maximum value of the left side of the inequality is less than or equal to 0: (44) For the problem of finding the maximum value of the left side of the inequality in Equation (44), using the duality theorem, Equation (44) is further transformed into: (45) Therefore, Equation (43) can be equivalently transformed into the following proposition: (46) 5) Chance constraint is transformed into a deterministic constraint Substituting Equation (20) and Equation (21) into the chance constraint Equation (15), the following robust problem can be obtained: (47) Based on the above-proved proposition, Equation (47) can be equivalently transformed into the following deterministic constraint for the line transmission capacity: (26) In the formula, 、 are the dual variables related to the power capacity constraint; Similarly, substituting Equation (20) and Equation (22) into Equation (16), based on the above-proved proposition, the deterministic constraint of the node voltage is equivalently transformed into: (27) In the formula, , is the dual variable related to the voltage constraint; Substituting Equation (4) into Equation (17), based on the above-proven proposition, the deterministic constraint of the energy storage output is equivalently transformed into: (28) wherein, , are the dual variables related to the energy storage power constraint; Substituting Equation (19) into Equation (18), based on the above-proven proposition, the deterministic constraint of the energy storage capacity is equivalently transformed into: (29) wherein, is the self-discharge coefficient of the energy storage; is the initial capacity of the energy storage; is the dual variable related to the energy storage capacity constraint.

[0035] Here, it is worth mentioning that traditional regulation methods can only transform the uncertain sources of the system into one-dimensional or two-dimensional ones, while the present invention optimizes the distributionally robust chance-constrained method of the main grid through the above dual principle, enabling the uncertain sources in the system of the present invention to be of any dimension.

[0036] Preferably, the confidence level of the foregoing chance constraint is 95%.

[0037] S4. Linearly reconstruct the other non-linear parts of the model objective function and constraint conditions.

[0038] 1. Reconstruction of the total power purchase and sale cost objective function 1) Reconstruction of the superior power purchase and sale cost The original power purchase and sale cost is non-linear. By introducing an auxiliary variable , it can be re-transformed into: (30) 2) Reconstruction of the energy storage operation cost The original energy storage operation cost is non-linear. By introducing an auxiliary variable , it can be re-transformed into: (31).

[0039] 2. Linearize the constraint with quadratic terms The circular feasible interval of the VSC capacity constraint is approximately simplified to the inscribed regular dodecagon feasible interval, that is, Constraint (23) can be simplified and expressed in the standard straight-line equation form of each side length: (32) In the formula, , and are the coefficients of the straight-line equation of the th side in a regular dodecagon. The currently publicly available coefficients are shown in Table 1 below: Table 1 Coefficients of the straight-line equation ; S5, organize and solve the equivalent linearized optimal regulation model of the microgrid group.

[0040] Based on the above derivation and transformation, the optimal model of the flexible interconnected microgrid group based on distributionally robust chance constraints is: (33) (34) In the formula, is the set of optimization variables, including: the optimization variable and the auxiliary variables and dual multipliers introduced in the equivalent transformation process.

[0041] Obtain the microgrid group network topology, equipment, and distribution network parameter information, and collect 100 sets of historical data sample points of wind power and photovoltaic historical output fluctuations at each moment. As Figure 2 shown, the basic topology of the flexible interconnected microgrid group is a chain-type DC bus structure, which is divided into a low-voltage AC area and a low-voltage DC area. The former realizes energy interaction with the upper-level distribution network, and the latter accesses a large number of distributed new energy sources to ensure the safe, economic, and stable operation inside the microgrid group. There are 3 energy storage devices connected in the microgrid group. The maximum discharge capacity of each energy storage is 50 kW, the maximum charge capacity is -50 kW, the maximum capacity is 200 kWh, the upper and lower limits of the capacity are set to 0.2 and 0.9 of the maximum capacity respectively, the initial capacity is 0.4 of the maximum capacity, the self-discharge coefficient is 0.998, and the charge efficiency is 100%; 40 kW of DC loads are connected to the DC buses of all nodes. Figure 3The normalized output curves of wind power, photovoltaic power, and AC / DC loads are shown. A 90kW photovoltaic device and a 60kW wind power device are connected to the DC bus of Node 1; a 300kW photovoltaic is connected to Node 2; a 150kW photovoltaic and a 150kW wind power are connected to Node 3; a 120kW photovoltaic is connected to Node 4. The unit resistance of the DC line is set to 0.0754Ω / km, and the distances between Substations 1-4 are 2.5km, 1.2km, and 1.6km respectively; the VSC capacities connected to Substations 1-4 are 100kVA, 500kVA, 500kVA, and 100kVA respectively. The capacities of the distribution transformers in each substation are the same as the VSC capacities, and the connected AC loads are 90kW, 420kW, 420kW, and 60kW. The maximum transmission capacity of the line is set to 300kW. The upper and lower limits of the node voltage are set to 1.05 and 0.95 respectively; the reference node voltage is set to 1. The time-of-use electricity price is set as follows: the peak electricity price is 0.14$ / kWh (18:00-22:00), the valley electricity price is 0.06$ / kWh (11:00-15:00), and the normal electricity price is set to 0.10$ / kWh; the selling price of the microgrid cluster to the distribution network is 0.04$ / kWh, and the operating cost of the energy storage is set to 0.015$ / kWh.

[0042] To verify the effectiveness of the method proposed in the present invention, 1000 sets of scenario data are generated based on the Gaussian distribution function with a mean of 0 and a variance of 10% of the output of the uncertainty source. In the MATLAB environment on the PC side, the equivalent linearized microgrid cluster optimization and control model obtained is solved by calling the CPLEX solver based on YALMIP to obtain the control parameters of the flexible interconnected microgrid system, mainly including the day-ahead output reference and real-time adjustment factors of each energy storage in the microgrid cluster, as well as the upper-level power purchase and sales volumes of each substation, that is, the optimization variables , , . According to these key control parameters, the flexible interconnected microgrid cluster is optimized and controlled, so as to achieve the load sharing within the microgrid cluster, the efficient control of the energy storage and VSC, and the complete consumption of wind power and photovoltaic power.

[0043] To verify the superiority of the method proposed in the present invention (hereinafter referred to as DRCO), the simulation results of this method are compared with those of two other traditional benchmark methods, namely: the robust optimization method (hereinafter referred to as RO), and the optimization method without considering uncertainty (hereinafter referred to as WU). Figure 4 Shows the total remaining uncertainty in the microgrid cluster system. WU does not consider the accommodation of uncertainty, so Figure 4The remaining uncertainty shown by the WU method is the original total uncertainty generated by each uncertainty source in the system. Based on limited energy storage resources, both the RO method and the method proposed in the present invention can absorb part of the uncertainty, but the method proposed in the present invention has a better effect on absorbing uncertainty than the RO method. This is because the operation of the energy storage in the RO method is mainly to ensure the safe and stable operation of the microgrid cluster system. For example, maintaining the DC bus within a safe range leaves little remaining capacity to absorb uncertainty. In contrast, the method proposed in the present invention relaxes the constraints on parameters such as voltage, enabling the BESs to absorb more uncertainty.

[0044] Figure 5 and Figure 6 It further illustrates the uncertainty mitigation of each uncertainty source at each moment under the RO method and the DRCO method. As time goes by, the ability of the RO method to absorb uncertainty weakens because the energy storage capacity has been exhausted. In contrast, DRCO maintains a high uncertainty absorption performance during periods of high uncertainty. For example, when the wind power (at night) and photovoltaic (at noon) generation are high, this ensures that the remaining uncertainty in the microgrid cluster system always remains at a relatively low level. Figure 7 It further illustrates the fluctuation of the energy interaction between the second distribution area and the upper-level distribution network. Obviously, compared with other methods, the DRCO method provides the best performance in suppressing the fluctuation of power purchase / sale.

[0045] Figure 8 、 Figure 9 、 Figure 10 It respectively gives the DC bus voltage fluctuations of the second distribution area under the WU, RO, and DRCO methods. Under the WU method, a high proportion of the voltage exceeds the upper limit; RO can ensure that the voltage remains within the safe range in all cases, but the voltage distribution is mainly around 1 p.u., making the voltage optimization effect too robust and conservative, limiting the reuse potential of the energy storage. In contrast, DRCO only exceeds the upper limit in very few cases, while improving the operation flexibility of the energy storage, it highly probably ensures the voltage safety of the system.

[0046] Table 2 shows the average total cost and its components of the microgrid cluster system under the WU, RO, and DRCO methods. Affected by the internal uncertainty of the system, the distribution of the purchase and sale cost reflects the characteristics of the remaining uncertainty, and the real-time output of the energy storage is also closely related to the uncertainty. Therefore, in the present invention, the 95% confidence level risk value (VaR) of the real-time output of the energy storage under all scenarios is used as a metric for the real-time regulation cost of the energy storage. Since the remaining uncertainty within the system may have an adverse impact on the upper-level distribution network, the system needs to absorb this uncertainty as much as possible internally. However, in the case of limited internal energy storage resources, all three methods need to pay a certain penalty cost to the distribution network. As can be seen from Table 2, although the average purchase and sale cost of the WU method is lower than that of the RO and DRCO methods, due to its higher day-ahead output cost of the energy storage and uncertainty penalty cost, its total cost is ultimately the highest. In addition, except for the real-time energy storage regulation cost, the cost of the RO method is generally higher than that of the DRCO method. This is because the RO method is optimized based on the worst-case scenario to ensure the absolute safety and stability of the system. Therefore, under the RO method, the operation actions of the energy storage mainly prioritize ensuring the safety of the system, sacrificing to a certain extent the economic benefits of the microgrid cluster.

[0047] It is worth mentioning that Table 2 shows the actual operating costs of the system after optimizing and regulating the system using the DRCO, WU, and RO methods. This actual operating cost is not exactly the same as the cost item in the objective function of the model proposed in the present invention, mainly because they serve different purposes respectively. The cost in the objective function is for optimizing decisions and is not equivalent to the actual operating cost. Specifically, the present invention introduces energy storage cost, power purchase and sale cost, and load balancing penalty cost, etc. in the objective function, mainly to guide the model to make more reasonable scheduling decisions in the face of uncertainties. For example, in real life, the power grid does not charge users the load balancing penalty cost. Adding this item in the objective function is to guide each substation area to perform power mutual assistance and achieve capacity sharing. In the transformed equivalent model, the uncertain quantities have been eliminated. However, in the simulation, in order to simulate the uncertain quantities that will occur during actual operation, 1000 uncertain quantities are randomly generated using the Gaussian function. Therefore, the "mean value" (a statistic) is added to the power purchase cost. Thus, these cost items themselves are not used to directly restore the economic expenditures in the actual operation process, but serve the model solving process. However, in the simulation stage of this embodiment, actual uncertainty samples are introduced to evaluate the performance of the scheduling strategy. Although the present invention equivalently processes uncertainties through the distributionally robust chance-constrained method during model solving, in order to test the effect of this scheduling strategy under real uncertain conditions, 1000 sets of sample scenarios are generated using the Gaussian distribution in this embodiment to simulate the operation process. Finally, based on the energy storage day-ahead output benchmark, real-time adjustment factor, and substation area power purchase quantity obtained by model optimization, according to the 1000 generated uncertain quantities, the data shown in Table 2 are calculated according to the actual operating cost of the distribution network: Table 2 Total Operating Cost within the Microgrid Cluster ; Figure 11 It further shows the overall operating cost distribution of the microgrid cluster system under the three methods. From Figure 11 it can be observed that the expected cost of the DRCO method is significantly lower than that of the RO and WU methods. In addition, since the DRCO method minimizes the remaining uncertainties within the system, its total cost distribution is also more concentrated.

[0048] To further reveal the difference in the output cost of the energy storage system under the RO and DRCO methods, Figure 12 the day-ahead and real-time output adjustment curves of the energy storage connected to the DC area of Substation Area 3 are given. From Figure 12It can be clearly seen that under the RO method, in order to ensure the security and stability of the system (especially voltage security), the charge and discharge scheduling of energy storage mainly focuses on off-peak and non-valley electricity price periods. However, due to the large occupation of energy storage capacity in the day-ahead stage, its ability to regulate the uncertainty of real-time renewable energy output is greatly reduced. Therefore, the in-day regulation range of energy storage under the RO method is generally lower than that of the DRCO method. In contrast, by appropriately relaxing the constraints, the DRCO method not only effectively alleviates the impact of uncertainty, but also adopts the strategy of charging during low electricity price periods (11:00) and discharging during high electricity price periods (21:00) for arbitrage, thereby reducing the purchase and sale electricity costs from the upper-level distribution network and optimizing the overall revenue of the system. Therefore, while allowing a slight deviation in the DC bus voltage within the microgrid cluster system, the DRCO method fully exploits the potential of multifunctional reuse of the energy storage system, ensuring the economy and stability of the microgrid cluster. Generally speaking, the DRCO method significantly reduces the conservativeness (RO method) and instability (WU method) of the system, and effectively inhibits the transfer of uncertainty to the upstream distribution network.

[0049] Figure 13 shows the load rate distribution of each node in the microgrid cluster system when considering (Scenario A, i.e., the method of the present invention) and not considering (Scenario B) the balanced load target. As Figure 13 shown, the method of the present invention effectively achieves power balance and capacity sharing between different substations.

[0050] The above embodiments are preferred implementation schemes of the present invention. In addition, the present invention can also be implemented in other ways. Any obvious replacement without departing from the concept of the technical solution of the present invention is within the protection scope of the present invention.

[0051] In order to make it more convenient for those of ordinary skill in the art to understand the improvements of the present invention over the prior art, some of the drawings and descriptions of the present invention have been simplified, and for the sake of clarity, some other elements have also been omitted from this application document. Those of ordinary skill in the art should be aware that these omitted elements may also constitute the content of the present invention.

Claims

1. A method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group, characterized in that Including: S1. Quantify the uncertainties of wind power, photovoltaic power, and DC load in the microgrid cluster, and determine the two-stage control strategy for energy storage day-ahead scheduling and real-time adjustment; S2. With the goal of minimizing the sum of the superior power purchase and sale cost, energy storage operation cost, load sharing penalty cost, and uncertainty penalty cost, construct an optimal control model for the microgrid cluster based on distributionally robust chance constraints. The constraints of this model include chance constraints and other constraints. The chance constraints include line transmission power chance constraints, node voltage chance constraints, energy storage charge and discharge power chance constraints, and energy storage capacity safety chance constraints. The other constraints include energy storage capacity constraints, network power flow constraints, VSC capacity constraints, and AC area power balance constraints; S3. Use the Wasserstein Metric definition and duality theorem to linearly reconstruct the uncertainty penalty cost in the model objective function, and transform the chance constraints of the model into deterministic constraints; S4. Linearly reconstruct the other non-linear parts of the model objective function and constraint conditions; S5. Obtain the microgrid cluster network topology, equipment, and distribution network parameter information, collect several groups of historical data sample points of wind power and photovoltaic historical output fluctuations at each moment, organize the equivalently linearized optimal control model of the microgrid cluster and solve it to obtain the control parameters, and optimize and control the flexible interconnected microgrid cluster.

2. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 1, wherein: In S3, when using the Wasserstein Metric definition and duality theorem to linearly reconstruct the uncertainty penalty function, first collect a series of historical sample data sets of random variables as the historical uncertainty fluctuation data set of the uncertainty source. Then, before obtaining the distribution fuzzy set of the random variable, determine the Wasserstein ball radius according to the fuzzy set radius size formula. Finally, based on the Wasserstein Metric definition, transform the uncertainty penalty function into a dual problem and simplify it by introducing a supremum approximation.

3. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 2, wherein: In S3, when converting the chance constraints of the microgrid group optimal regulation model into deterministic constraints by using the Wasserstein Metric definition and the duality theorem, first calculate the mean and variance of the historical sample data sets of the random variables at each moment, and normalize all the sample points into the standard coordinate system. Then, solve the problem of minimizing the side length of the hypercube to obtain the side length of the hypercube of the sample points at a certain confidence level in the standard coordinate system, and determine the corresponding hypercube. Next, anti-normalize the hypercube into the original coordinate system to obtain the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding box uncertainty set. Finally, use the duality theorem to transform the original chance constraint problem into a robust problem under this box uncertainty set, and then: t The side length of the hypercube of the sample points at a certain moment, and determine the corresponding hypercube. Then, anti-normalize the hypercube into the original coordinate system to obtain the parallelepiped of the sample points at a certain confidence level in the original coordinate system, and determine the corresponding box uncertainty set. t Finally, use the duality theorem to transform the original chance constraint problem into a robust problem under this box uncertainty set, and then: Substitute the network power flow constraint into the line transmission power chance constraint, and equivalently transform it to obtain the line transmission power deterministic constraint; Substitute the network power flow constraint into the node voltage chance constraint, and equivalently transform it to obtain the node voltage deterministic constraint; Substitute the energy storage output into the energy storage charge and discharge power chance constraint, and equivalently transform it to obtain the energy storage output deterministic constraint; Substitute the energy storage capacity constraint into the energy storage capacity safety chance constraint, and equivalently transform it to obtain the energy storage capacity deterministic constraint.

4. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 3, characterized in that: The control parameters include the day-ahead output reference and real-time adjustment factor of each energy storage in the microgrid cluster, and the superior power purchase and sale volume of each substation area.

5. The method for uncertainty absorption and load balancing of a flexible interconnected microgrid group according to claim 4, wherein: In S1, the uncertainties of wind power, photovoltaic power, and DC load in the microgrid cluster are quantified as follows: （1） （2） （3） In the formula, represents the actual output column vector of each uncertainty source at the moment; represents the uncertainty source-node incidence matrix, including the photovoltaic-node incidence matrix , the wind power-node incidence matrix , and the DC load-node incidence matrix ; represents the predicted values of the uncertainty sources at the moment, including the predicted photovoltaic values at the moment , the predicted wind power values , and the predicted DC load values ; is the predicted error column vector of each uncertainty source, representing the uncertain quantity; In S1, the energy storage output of the microgrid cluster is divided into day-ahead scheduling and real-time adjustment, and the output of the energy storage at each node is as follows: （4） In the formula, represents the real-time output of each energy storage at a certain moment; represents the energy storage-node incidence matrix, is the reference of the energy storage's day-ahead output; is the real-time adjustment factor matrix, and its specific definition is: （5） In the formula, the row, column element of the th energy storage device to the uncertainty mitigation ratio of the th uncertainty source, is the number of energy storage devices, the elements in satisfy the following conditions: （6） （7） Equation (6) indicates that the proportion of each energy storage to mitigate each uncertainty is less than 100%; Equation (7) indicates At time the total compensation amount of all energy storages for the uncertainty itself is less than or equal to the uncertainty.

6. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 5, wherein: In S2, the optimal control model for the microgrid cluster based on distributionally robust chance constraints is: （8） In the formula, is the optimization variable; are the upper-level power purchase and sale cost, energy storage operation cost, load sharing penalty cost, and uncertainty penalty cost respectively; （9） In the formula, is the set of all scheduling times, is the set of numbers of all nodes in the microgrid; and are the power purchase cost and the power selling unit price respectively; is the power interaction volume between the -th node and the superior distribution network; the operator satisfies ; （10） In the formula, is the set of numbers of all energy storages in the microgrid; is the operating cost per unit output of the energy storage; is the active power output of each energy storage; （11） In the formula, is the load sharing penalty coefficient; , are the transformer capacities of the 、 th nodes in the microgrid respectively; is the power interaction quantity between the th node and the upstream distribution network; is the number of nodes in the microgrid cluster; （12） （13） （14） In the formula, is the uncertainty penalty coefficient; represents the set of numbers of all uncertainty sources; represents the th uncertainty source's uncertainty distribution fuzzy set at moment; represents the maximum expected value under all possible probabilities; is the th element in the column vector , , and its physical meaning represents the proportion of the uncertainty generated by the th uncertainty source absorbed by all energy storages in the microgrid cluster; represents a column vector of all 1s with a dimension of n×1; In S2, the chance constraints of the optimal control model for the microgrid cluster based on distributionally robust chance constraints include: 1) Line transmission power chance constraint （15） In the formula, represents the set of all probability distributions that the uncertainty may satisfy; represents the minimum probability for the expression in the parentheses to hold; represents at the transmission power vector of the DC line of the microgrid cluster at the moment, and are the minimum and maximum transmission power vectors of the DC line respectively; represents the confidence level; 2) Node voltage chance constraint （16） In the formula, represents the DC bus voltage vector of the microgrid cluster at moment, , are the maximum and minimum values of the DC bus voltage respectively; 3) Energy storage charge and discharge power chance constraint （17） In the formula, represents the output column vector of the energy storage at moment, and it is stipulated that the energy storage discharging is the positive direction; , are the maximum and minimum charge and discharge powers of the energy storage respectively; 4) Energy storage capacity safety chance constraint （18） In the formula, represents the vector of the energy storage capacities of each microgrid group at moment; , are respectively the lower limit and upper limit of the energy storage capacity. In step S2, other constraints of the microgrid group optimal regulation model constructed based on distributionally robust chance constraint include: 1) Energy storage capacity constraint （19） In the formula, is the energy storage charging efficiency; is the self-discharge coefficient of the energy storage; represents the daily scheduling interval and takes 1; 2) Network power flow constraint （20） （21） （22） In the formula, represents the node injection power; represents the active power injected by the VSC, is the augmented node-line incidence matrix, is the augmented line resistance matrix; represents the column vector of the DC bus voltage of the microgrid cluster; is the voltage reference; 3) VSC capacity constraint （23） In the formula, , respectively represent the active power injected by the VSC and the reactive power output at the th moment at the th node; is the capacity of the VSC at the th node; 4) AC area power balance constraint （24） wherein, represents the upper-level power purchase and sale electric power column vector, is the AC load column vector of the AC node; In step S3, the uncertainty penalty function is linearly reconstructed by using the Wasserstein Metric definition and the duality theorem as: （25） In the formula, is the dual multiplier; is the radius size of the fuzzy set; , denote the maximum and minimum values of the historical data of the -th uncertainty source at the -th moment; denotes the number of samples; denotes the historical sample data set of the -th component in the uncertain quantity at each moment; ​ In step S3, the chance constraint of the microgrid group optimal regulation model is transformed into a deterministic constraint by using the Wasserstein Metric definition and the duality theorem as follows: Substituting Equation (20) and Equation (21) into the chance constraint Equation (15), the deterministic constraint of the line transmission power is equivalently transformed as: （26） In the formula, , are dual variables related to the power capacity constraint; Substituting Equation (20) and Equation (22) into Equation (16), the deterministic constraint of the node voltage is equivalently transformed as: （27） wherein, and are dual variables related to voltage constraints; Substituting Equation (4) into Equation (17), the deterministic constraint of the energy storage output is equivalently transformed as: （28） wherein, and are dual variables related to the energy storage power constraint; Substituting Equation (19) into Equation (18), the deterministic constraint of the energy storage capacity is equivalently transformed as: （29） Wherein, is the self-discharge coefficient of the energy storage; is the initial capacity of the energy storage; is the dual variable related to the energy storage capacity constraint; is the constant matrix of the parallelepiped . The element in the -th row and the -th column is 1, and the element in the -th row and the -th column is -1, and the rest of the elements are all 0; is the constant column vector of the parallelepiped . The elements in the -th row and the -th row are both the absolute values of the coordinates of the -th vertex in , .

7. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 6, characterized in that: The confidence level of the chance constraint is 95%.

8. The method for uncertainty absorption and load balancing of a flexible interconnected microgrid group according to claim 7, wherein: In step S4, auxiliary variables are introduced to linearly reconstruct the upper-level power purchase and sale cost function and the energy storage operation cost function; 1) Introduce auxiliary variables Linearly reconstruct the upper-level purchase and sale electricity cost function as follows: （30） 2) Introduce auxiliary variables Linearly reconstruct the energy storage operation cost function as follows: （31）。 9. The method for flexible interconnected microgrid group uncertainty accommodation and load balancing according to claim 8, characterized in that: In step S4, the constraint with quadratic terms in the VSC capacity constraint is linearized, and the circular feasible interval of the VSC capacity constraint is approximately simplified to an inscribed regular dodecagon feasible interval, which is expressed in the standard straight-line equation form of each side length as follows: （32） In the formula, , and are the coefficients of the straight line equation of the th side in a regular dodecagon.

10. The method for uncertainty accommodation and load balancing of a flexible interconnected microgrid group according to claim 9, wherein: In step S5, the equivalently linearized microgrid group optimal regulation model obtained after sorting is: （33） （34） wherein is a set of optimization variables, including: optimization variable and auxiliary variables and dual multipliers introduced in the equivalent transformation process.

Images