Micro-grid two-stage optimization method and system considering energy storage networking capability

Abstract

The invention discloses a micro-grid two-stage optimization method and system considering energy storage networking capability, and relates to the technical field of micro-grid optimization, and the method comprises the steps: obtaining a wind and light data set based on the wind power and photovoltaic power of a target micro-grid; performing clustering based on the wind-light data set to obtain scene parameters; obtaining a wind and light uncertainty set based on the scene parameters; energy storage rated power and energy storage rated capacity are obtained based on the target micro-grid to serve as first-stage optimization variables; acquiring system load power as a second-stage optimization variable based on the target microgrid; obtaining a coefficient column vector when the daily comprehensive operation cost is lowest based on the second-stage optimization variable; establishing a two-stage robust optimization model of the target microgrid based on the wind and light uncertainty set, the first-stage optimization variable, the second-stage optimization variable, the coefficient column vector and the constraint condition; based on the two-stage robust optimization model, a column constraint generation algorithm is adopted for solving, and an optimal scheduling scheme is obtained. And the robustness of micro-grid optimization is improved.

Description

Technical Field

[0001] This invention relates to the field of microgrid optimization technology, and more specifically to a two-stage optimization method and system for microgrids that takes into account the energy storage network construction capability. Background Technology

[0002] Currently, a microgrid is a small-scale power generation and distribution system composed of distributed power sources, energy storage, energy conversion devices, loads, and related monitoring devices. It can flexibly schedule distributed units based on actual operating environment and economic requirements, guided by energy optimization, to achieve integrated operation of "generation, grid, load, and storage".

[0003] Wind and solar power often utilize grid-following (GFL) control strategies to connect to microgrids, injecting power into the system. Energy storage provides flexible power reserves and dispatch capabilities to the microgrid when there is an imbalance between renewable energy output and load demand. Proper configuration and dispatch of energy storage can help improve renewable energy utilization and the overall operating efficiency of the microgrid. However, as the penetration rate of grid-following renewable energy increases, the proportion of synchronous generators decreases, microgrid inertia decreases, and active power disturbances can lead to drastic frequency changes, resulting in insufficient microgrid stability. Existing energy storage configuration and dispatch schemes do not fully consider the volatility of renewable energy and loads, are limited to typical scenarios and fixed prediction errors, and lack robustness.

[0004] Therefore, how to fully consider the volatility of new energy sources and loads, and not be limited to typical scenarios and fixed prediction errors, so as to improve the robustness of microgrid optimization, is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a two-stage optimization method and system for microgrids that takes into account the energy storage network construction capability. It fully considers the volatility of new energy sources and loads, and is not limited to typical scenarios and fixed prediction errors, thereby improving the robustness of microgrid optimization.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A two-stage optimization method for microgrids considering energy storage grid integration capabilities includes:

[0008] A wind and solar power dataset is obtained based on the wind power and solar power of the target microgrid;

[0009] Scene parameters are obtained by clustering the aforementioned wind and light dataset;

[0010] The uncertain set of wind and light is obtained based on the scene parameters;

[0011] Based on the target microgrid, the rated power and rated capacity of energy storage are obtained as the first-stage optimization variables.

[0012] The system load power obtained based on the target microgrid is used as the optimization variable for the second stage.

[0013] Based on the second-stage optimization variables, obtain the coefficient column vector when the daily comprehensive operating cost is at its lowest;

[0014] A two-stage robust optimization model for the target microgrid is established based on the aforementioned wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints.

[0015] The optimal scheduling scheme is obtained by using a column constraint generation algorithm based on the two-stage robust optimization model.

[0016] Preferably, the method for obtaining the landscape dataset is as follows:

[0017] Based on the wind power and the photovoltaic power, establish the edge probability density function of wind power output and the edge probability density function of photovoltaic power output;

[0018] Based on the integration of the wind power output edge probability density function and the photovoltaic output edge probability density function respectively, the wind power edge distribution function and the photovoltaic edge distribution function are obtained respectively;

[0019] Based on the Frank Copula function, the wind power edge distribution function and the photovoltaic edge distribution function are combined to obtain the wind-solar joint probability distribution function;

[0020] Based on the joint probability distribution function of wind and solar energy, sampling and inverse transformation are performed to generate multiple wind and solar energy datasets.

[0021] Preferably, the method for obtaining the scene parameters is as follows:

[0022] Clustering is performed on the aforementioned landscape dataset, and similar scenes are merged to obtain two typical landscape scenes;

[0023] Based on the two typical wind and solar scenarios, the wind power of the first scenario, the photovoltaic power of the first scenario, the wind power of the second scenario, and the photovoltaic power of the second scenario are obtained.

[0024] The wind power benchmark value is obtained by averaging the wind power in the first scenario and the wind power in the second scenario.

[0025] A photovoltaic baseline value is obtained by averaging the photovoltaic power of the first scenario and the photovoltaic power of the second scenario.

[0026] The wind power deviation is obtained by subtracting the wind power power in the first scenario from the wind power benchmark value.

[0027] The photovoltaic deviation is obtained by subtracting the photovoltaic power in the first scenario from the photovoltaic reference value.

[0028] The predicted wind power value is obtained based on the wind power benchmark value and the wind power deviation;

[0029] The photovoltaic power prediction value is obtained based on the photovoltaic baseline value and the photovoltaic deviation;

[0030] The predicted wind power and the predicted photovoltaic power are used as the scenario parameters.

[0031] Preferably, the method for obtaining the wind and solar uncertainty set is as follows:

[0032] Obtain the load uncertainty, system load, wind power uncertainty, and photovoltaic power uncertainty of the target microgrid;

[0033] The wind and solar uncertainty set is composed of the load uncertainty variable, the system load, the wind power uncertainty variable, the photovoltaic power uncertainty variable, the wind power prediction value, and the photovoltaic power prediction value.

[0034] Preferably, the system load power includes: wind power, photovoltaic power, grid-type energy storage power, generator power, and grid interconnection power;

[0035] The grid-type energy storage power includes energy storage charging power and energy storage discharging power.

[0036] Preferably, the method for obtaining the coefficient column vector is as follows:

[0037] The wind power operating cost is obtained based on the wind power output and the unit price of wind turbine power generation.

[0038] The cost of photovoltaic power generation is obtained based on the photovoltaic power and the unit price of photovoltaic power generation.

[0039] The operating cost of grid-type energy storage is obtained based on the energy storage charging power, the energy storage discharging power, the energy storage unit power charging cost, the energy storage unit power discharging cost, and the energy storage charging and discharging efficiency.

[0040] The construction cost of grid-type energy storage is obtained based on the unit cost of energy storage power capacity, the unit cost of energy capacity, the benchmark discount rate, and the energy storage operating life.

[0041] The generator operating cost is obtained based on the generator power and operating cost parameters.

[0042] The grid purchase cost is obtained based on the grid interaction power and the grid-side peak-valley electricity price.

[0043] The daily comprehensive operating cost of the target microgrid is obtained based on the wind power operating cost, the photovoltaic power generation cost, the grid-connected energy storage operating cost, the grid-connected energy storage construction cost, the generator operating cost, and the grid power purchase cost.

[0044] Based on the lowest daily comprehensive operating cost, an optimization function is constructed and the first coefficient column vector and the second coefficient column vector are extracted.

[0045] The first coefficient column vector and the second coefficient column vector together form the coefficient column vector.

[0046] Preferably, the constraints include:

[0047] Power balance constraints, wind power constraints, photovoltaic power constraints, generator power constraints, generator ramping power constraints, energy storage rated power constraints, energy storage discharge power constraints, energy storage charging power constraints, energy storage capacity constraints, energy storage charge and discharge balance constraints, grid interaction power constraints, microgrid system inertia constraints, and system reserve power constraints.

[0048] Preferably, the method for constructing the two-stage robust optimization model is as follows:

[0049] The first coefficient matrix is ​​obtained based on the power balance constraint and the energy storage charge-discharge balance constraint;

[0050] The second coefficient matrix is ​​obtained based on the wind power constraint, the photovoltaic power constraint, the generator power constraint, the generator ramping power constraint, and the grid interaction power constraint.

[0051] The third coefficient matrix is ​​obtained based on the energy storage rated power constraint, the energy storage discharge power constraint, the energy storage charging power constraint, the energy storage capacity constraint, the microgrid system inertia constraint, and the system reserve power constraint;

[0052] The wind and solar coefficient matrix is ​​obtained based on the aforementioned wind and solar uncertainty set;

[0053] Obtain the uncertainties of wind power and photovoltaic power, and combine them with the uncertainties of the load to obtain a column vector of uncertainties;

[0054] The two-stage robust optimization model is constructed based on the first coefficient matrix, the second coefficient matrix, the third coefficient matrix, the wind and light coefficient matrix, the column vector, the first-stage optimization variable, the second-stage optimization variable, and the coefficient column vector.

[0055] The preferred approach yields the optimal scheduling scheme, which specifically includes:

[0056] S1 decomposes the main problem and sub-problems based on the two-stage robust optimization model;

[0057] S2 sets a set of uncertain variables as the initial worst-case scenario and sets the upper and lower bounds of the initial operating cost;

[0058] S3 solves the main problem based on the initial worst-case scenario and obtains the objective function value of the main problem as the current master optimal solution;

[0059] S4 uses the current master optimal solution as a new lower bound for operating costs;

[0060] S5 substitutes the current master optimal solution into the subproblem to obtain the objective function value of the subproblem as the current suboptimal solution;

[0061] S6 updates the new upper bound of operating cost based on the current primary optimal solution, the current suboptimal solution, and the initial upper bound of operating cost;

[0062] S7 calculates the cost difference by subtracting the new upper bound of operating cost from the new lower bound of operating cost;

[0063] S8 determines whether the cost difference is greater than the convergence threshold;

[0064] If S9 is true, it indicates non-convergence; determine whether a solution exists.

[0065] If S10 exists, then add the first constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output;

[0066] If S11 does not exist, then add a second constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output;

[0067] If S12 is not true, it indicates convergence, the iteration stops, and the current principal optimal solution and the current suboptimal solution are taken as the final principal optimal solution and the final suboptimal solution.

[0068] S13 obtains the optimal scheduling scheme based on the final master optimal solution and the final suboptimal solution.

[0069] A two-stage optimization system for microgrids that considers energy storage network construction capabilities includes: a dataset acquisition module, a first variable acquisition module, a second variable acquisition module, a column vector acquisition module, an optimization model construction module, and an optimal solution output module;

[0070] The dataset acquisition module is used to obtain a wind and solar power dataset based on the wind power and photovoltaic power of the target microgrid; to obtain scene parameters by clustering the wind and solar power dataset; and to obtain a wind and solar power uncertainty set based on the scene parameters.

[0071] The first variable acquisition module is used to acquire the rated power and rated capacity of energy storage based on the target microgrid as the first-stage optimization variables;

[0072] The second variable acquisition module is used to acquire the system load power based on the target microgrid as the second-stage optimization variable;

[0073] The column vector acquisition module is used to obtain the coefficient column vector when the daily comprehensive operating cost is the lowest based on the second-stage optimization variables.

[0074] The optimization model construction module is used to establish a two-stage robust optimization model of the target microgrid based on the wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints.

[0075] The optimal solution output module is used to solve the optimal scheduling scheme by using a column constraint generation algorithm based on the two-stage robust optimization model.

[0076] As can be seen from the above technical solutions, compared with the prior art, this invention discloses a two-stage optimization method and system for microgrids that considers energy storage network construction capabilities. This invention takes into account energy storage network construction capabilities and aims to optimize the overall cost of the microgrid under the worst-case scenario. It establishes an objective function for the microgrid optimization model, considering that the system frequency changes meet load requirements under anticipated accidents, and establishes inertia constraints and reserve power constraints. Considering the correlation between wind and solar power output in the same region, a joint wind-solar probability distribution model is established based on kernel density estimation and the Frank Copula function. Based on k-means clustering, the wind-solar fluctuation boundaries and their corresponding probabilities are determined. Considering the volatility of wind and solar loads, uncertainty adjustment parameters are introduced to form an uncertain set of wind and solar loads. Based on this, this invention proposes a two-stage robust optimization strategy for microgrids, which can optimize the overall system operating cost under relatively severe scenarios, reduce the conservatism of traditional robust models, reduce energy storage configuration costs, and ensure that the system retains sufficient inertia and reserve power during the scheduling cycle under actual operating conditions. Attached Figure Description

[0077] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0078] Figure 1 The flowchart of a two-stage optimization method for microgrids that considers energy storage network construction capabilities is provided by the present invention.

[0079] Figure 2 This is a schematic diagram of the upper and lower bounds of the robust optimization strategy provided by the present invention.

[0080] Figure 3 This is a schematic diagram illustrating the comparison results of wind power optimization scenarios provided by the present invention.

[0081] Figure 4 This is a schematic diagram illustrating the comparison results of photovoltaic optimization scenarios provided by the present invention.

[0082] Figure 5 This is a schematic diagram illustrating the comparison results of load optimization scenarios provided by the present invention.

[0083] Figure 6 A schematic diagram of the scheduling scheme corresponding to the robust optimization method of the present invention.

[0084] Figure 7 This is a schematic diagram of the scheduling scheme corresponding to the traditional robust optimization method provided by the present invention.

[0085] Figure 8 This is a schematic diagram of the scheduling scheme corresponding to the deterministic optimization method provided by the present invention.

[0086] Figure 9 This diagram illustrates the minimum rotational kinetic energy under the three optimization methods provided by this invention.

[0087] Figure 10 The diagram illustrates the backup power under the three optimization methods provided by this invention. Detailed Implementation

[0088] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0089] Example 1

[0090] like Figure 1 As shown, this invention discloses a two-stage optimization method for microgrids that considers energy storage network construction capabilities, including:

[0091] A wind and solar power dataset is obtained based on the wind power and solar power of the target microgrid;

[0092] Scene parameters are obtained by clustering the wind and light dataset;

[0093] The uncertainty set of the landscape is obtained based on scene parameters;

[0094] The rated power and rated capacity of energy storage are obtained from the target microgrid as the first-stage optimization variables.

[0095] The system load power obtained from the target microgrid is used as the second-stage optimization variable;

[0096] The coefficient column vector at which the daily comprehensive operating cost is lowest is obtained based on the second-stage optimization variables;

[0097] A two-stage robust optimization model for the target microgrid is established based on the wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints.

[0098] The optimal scheduling scheme is obtained by using a column constraint generation algorithm based on a two-stage robust optimization model.

[0099] Example 2

[0100] This invention discloses a two-stage optimization method for microgrids that considers energy storage network construction capabilities, including:

[0101] A wind and solar power dataset is obtained based on the wind power and solar power of the target microgrid.

[0102] The preferred method for obtaining the landscape dataset is as follows:

[0103] Based on the corresponding wind power and photovoltaic power, establish the edge probability density function of wind power output and the edge probability density function of photovoltaic power output;

[0104] By integrating the edge probability density function of wind power output and the edge probability density function of photovoltaic power output respectively, the edge distribution function of wind power and the edge distribution function of photovoltaic power are obtained respectively;

[0105] Based on the Frank Copula function combined with the wind power edge distribution function and the photovoltaic edge distribution function, the wind-solar joint probability distribution function is obtained;

[0106] Multiple wind and solar datasets are generated by sampling and inverse transformation based on the joint probability distribution function of wind and solar energy.

[0107] Preferably, in this embodiment, the wind power includes the wind power x during time period t. t Wind power X during time period t on day i i,t The specific marginal probability density function of wind power output is as follows:

[0108]

[0109] Where n represents the number of samples, h x This represents the wind power output window width, and K represents the Gaussian kernel function, ensuring the continuity of wind and solar power output changes.

[0110] Preferably, the Gaussian kernel function K is as follows:

[0111]

[0112] Preferably, in this embodiment, the photovoltaic power includes the photovoltaic power y during time period t. t Photovoltaic power Y during time period t on day i i,t The specific edge probability density function of photovoltaic power output is:

[0113]

[0114] Among them, h y This indicates the width of the wind power output window.

[0115] Preferably, the wind-solar joint probability distribution function is as follows:

[0116] F(x t ,y t )=C(F(x t ),F(y t ));

[0117] Where F(x) t F(y) represents the edge distribution function of wind power. t ) represents the photovoltaic edge distribution function, and C represents the FrankCopula function.

[0118] Preferably, this embodiment uses kernel density estimation for nonparametric estimation, directly mining data distribution characteristics from the sample to calculate the marginal distribution of photovoltaic and wind power output.

[0119] Scene parameters are obtained by clustering the wind and light dataset.

[0120] Preferably, the method for obtaining scene parameters is as follows:

[0121] Clustering was performed on the landscape dataset, and similar scenes were merged to obtain two typical landscape scenes;

[0122] Based on two typical wind and solar power scenarios, the wind power in the first scenario, the photovoltaic power in the first scenario, the wind power in the second scenario, and the photovoltaic power in the second scenario are obtained.

[0123] The wind power baseline value is obtained by averaging the wind power in the first scenario and the wind power in the second scenario.

[0124] The photovoltaic baseline value is obtained by averaging the photovoltaic power of the first scenario and the photovoltaic power of the second scenario.

[0125] The wind power deviation is obtained by subtracting the wind power output from the wind power benchmark value in the first scenario.

[0126] The photovoltaic deviation is obtained by subtracting the photovoltaic power from the photovoltaic reference value in the first scenario.

[0127] The predicted wind power value is obtained based on the wind power benchmark value and the wind power deviation;

[0128] The photovoltaic power prediction value is obtained based on the photovoltaic baseline value and the photovoltaic deviation;

[0129] The predicted values ​​of wind power and photovoltaic power are used as scenario parameters.

[0130] Preferably, considering the large number of generated wind and light datasets and the high similarity between different scenes, k-means++ clustering is used to effectively merge similar scenes in order to obtain the upper and lower boundaries of wind and light power, thereby obtaining two typical wind and light scenes.

[0131] The preferred selection yields two typical landscape scenes, specifically including:

[0132] Set two center points;

[0133] Calculate the Euclidean distance from the data points in the landscape dataset to the two center points;

[0134] Cluster all data based on the Euclidean distance to obtain multi-class data;

[0135] The midpoint of the same type of data is used as the new center point of the class.

[0136] Repeat this process until the sum of the distances between each data point and the center point remains constant, ultimately yielding two typical landscape scenes.

[0137] Preferably, the wind power forecast value is as follows:

[0138]

[0139] in, Indicates the benchmark value for wind power. Indicates wind power deviation, This represents the total fluctuation of wind power in the t-th time period under the k-th scenario.

[0140] Preferably, the photovoltaic power prediction value is as follows:

[0141]

[0142] in, Indicates the photovoltaic benchmark value. Indicates photovoltaic deviation, This represents the total fluctuation of photovoltaic power in the k-th scenario during the t-th time period.

[0143] Preferably, the total fluctuation of wind power in the t-th time period under the k-th scenario. Specifically:

[0144]

[0145] Where p1 represents the probability corresponding to the first scenario. p1 represents the wind power uncertainty in the t-th time period under the first scenario, and p2 represents the probability corresponding to the second scenario. This represents the uncertainty of wind power in the t-th time period under the second scenario.

[0146] Preferably, represents the total photovoltaic fluctuation in the t-th time period under the k-th scenario. Specifically:

[0147]

[0148] in, This represents the photovoltaic uncertainty in the t-th time period under the first scenario. This represents the photovoltaic uncertainty in the t-th time period under the second scenario.

[0149] Preferably, the probability p corresponding to the k-th scene k for:

[0150] p k =n k / n D ;

[0151] Where, n k Let n represent the number of landscape datasets corresponding to the k-th typical scene in {1,2}. D This indicates the total amount of data.

[0152] The uncertainty set of the landscape is obtained based on the scene parameters.

[0153] The preferred method for obtaining the uncertain set of wind and light is as follows:

[0154] Obtain the load uncertainty, system load, wind power uncertainty, and photovoltaic power uncertainty of the target microgrid;

[0155] The wind and solar uncertainty set is composed of load uncertainty, system load uncertainty, wind power uncertainty, photovoltaic power uncertainty, wind power prediction value, and photovoltaic power prediction value.

[0156] Preferably, due to the seasonality of wind and solar power and prediction errors, microgrids face numerous random factors in actual operation. Deterministic optimization models rely on typical input datasets, resulting in inadequate energy storage configurations and operation schemes. Therefore, it is necessary to account for the impact of uncertainty in the model. This involves considering both the fluctuations in renewable energy output and the fact that load power fluctuations are assumed to be 10% of the predicted value. Let g be defined. l,tLet be the load uncertainty variable, which is a real number between -1 and 1.

[0157] Preferably, the uncertain set U of wind and light is as follows:

[0158]

[0159] Among them, P Load,t This represents the load power of the microgrid during time period t. Indicates the load power reference value, g l,t This represents an uncertain variable in the load. This represents the lower limit of the adjustment parameter for the total uncertainty of load power. Γ represents the upper limit of the total uncertainty adjustment parameter for load power. low Γ represents the lower limit of the adjustment parameter for the total uncertainty of wind and solar power. up g represents the upper limit of the total uncertainty adjustment parameter for wind and solar power. wt,t G represents the uncertain variable in wind power output. pv,t This represents the uncertain variable in photovoltaic power.

[0160] Preferably, the total uncertainty adjustment parameter takes the value of an integer within a certain range. It can be used to adjust the conservatism of the optimal solution. The larger the value, the more conservative the solution, and vice versa.

[0161] The rated power and rated capacity of energy storage are obtained from the target microgrid as the first-stage optimization variables.

[0162] Preferably, the first-stage optimization variable x is as follows:

[0163]

[0164] Among them, P e E represents the rated power of energy storage. e Indicates the rated capacity of energy storage. This indicates the charging and discharging operation status of the grid-type energy storage at time t. A value of 1 indicates discharging, and a value of 0 indicates charging.

[0165] The system load power obtained from the target microgrid is used as the second-stage optimization variable.

[0166] Preferably, the system load power includes: wind power P WT,t Photovoltaic power P PV,t Grid-type energy storage power P Gfm,t Generator power P Gen,t Power interaction P with the power grid Grid,t .

[0167] Grid-type energy storage power P Gfm,t Including energy storage charging power and energy storage discharge power

[0168] Preferably, the second-stage optimization variable y is as follows:

[0169]

[0170] The coefficient column vector at which the daily comprehensive operating cost is lowest is obtained based on the second-stage optimization variables.

[0171] Preferably, the method for obtaining the coefficient column vector is as follows:

[0172] The operating cost of wind power is obtained based on wind power output and the unit price of wind turbine power generation.

[0173] The cost of photovoltaic power generation is obtained based on photovoltaic power output and photovoltaic power generation unit price.

[0174] The operating cost of grid-type energy storage is obtained based on energy storage charging power, energy storage discharging power, energy storage unit power charging cost, energy storage unit power discharging cost, and energy storage charging and discharging efficiency.

[0175] The construction cost of grid-type energy storage is obtained based on the unit cost of energy storage power capacity, the unit cost of energy capacity, the benchmark discount rate, and the energy storage operating life.

[0176] The generator operating cost is obtained based on generator power and operating cost parameters;

[0177] The grid purchase cost is obtained based on grid interaction power and grid-side peak-valley electricity price.

[0178] The daily comprehensive operating cost of the target microgrid is obtained based on the operating cost of wind power, the cost of photovoltaic power generation, the operating cost of grid-connected energy storage, the construction cost of grid-connected energy storage, the operating cost of generators, and the cost of purchasing electricity from the grid.

[0179] An optimization function is constructed based on the lowest daily comprehensive operating cost, and the first and second coefficient column vectors are extracted.

[0180] The first coefficient column vector and the second coefficient column vector together form the coefficient column vector.

[0181] Preferably, wind power operating costs

[0182]

[0183] Among them, c WT This indicates the unit price of wind turbine power generation.

[0184] Preferably, the cost of photovoltaic power generation

[0185]

[0186] Among them, c PV This indicates the unit price of wind turbine power generation.

[0187] Preferably, the operating cost of grid-based energy storage for:

[0188]

[0189] Among them, K s1 and K s2 These represent the unit power discharge cost and unit power charging cost of grid-based energy storage, respectively. Indicates the energy storage charging power. η represents the energy storage discharge power, and η represents the energy storage charge and discharge efficiency.

[0190] Preferably, the cost of grid-based energy storage mainly includes the construction cost and the operation cost, where the construction cost C... In for:

[0191]

[0192] Among them, K P and K E Let n represent the unit cost of energy storage power capacity and the unit cost of energy capacity, respectively; r represents the benchmark discount rate; and n represents the unit cost of energy storage power capacity and the unit cost of energy capacity. e This indicates the operational lifespan of the energy storage system.

[0193] Preferably, generator operating cost for:

[0194]

[0195] Where, α Gen and β Gen All of these represent operating cost parameters.

[0196] Preferably, the cost of purchasing electricity from the grid. for:

[0197]

[0198] in, This represents the electricity purchase cost from the power grid during the t-th time period.

[0199] The preferred optimization function is as follows:

[0200] Goal:

[0201] Preferably, an optimization function is constructed based on the lowest daily comprehensive operating cost, and the first coefficient column vector c1 and the second coefficient column vector c2 are extracted.

[0202] A two-stage robust optimization model for the target microgrid is established based on the wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints.

[0203] Preferably, the constraints include:

[0204] Power balance constraints, wind power constraints, photovoltaic power constraints, generator power constraints, generator ramping power constraints, energy storage rated power constraints, energy storage discharge power constraints, energy storage charging power constraints, energy storage capacity constraints, energy storage charge and discharge balance constraints, grid interaction power constraints, microgrid system inertia constraints, and system reserve power constraints.

[0205] Preferably, the power balance constraint is:

[0206] Preferably, the wind power constraint is:

[0207]

[0208] in, This indicates the available wind power capacity for each time period.

[0209] Preferably, the photovoltaic power constraint is:

[0210]

[0211] in, This indicates the available photovoltaic power within each time period.

[0212] Preferably, the generator power constraint is:

[0213]

[0214] in, and These represent the upper and lower limits of the generator's output power, respectively.

[0215] Preferably, the generator ramping power constraint is:

[0216]

[0217] Among them, P Gen,t-1 This represents the generator power at time t-1. This indicates the generator's maximum ramping power.

[0218] Preferably, the rated power constraint for energy storage is:

[0219] P e ≤k e E e ;

[0220] Where, k e This represents the energy storage coefficient, used to meet the requirements for energy storage operation for 1-2 hours.

[0221] Preferably, the energy storage charging power constraint is:

[0222]

[0223] in, Indicates the energy storage charging power. This represents the charging and discharging operation state of the grid-type energy storage at time t. A value of 1 indicates discharging, and a value of 0 indicates charging. k represents the upper limit of the charging and discharging power of energy storage. g This indicates the overload capacity of energy storage.

[0224] Preferably, the energy storage discharge power constraint is:

[0225]

[0226] in, This indicates the energy storage discharge power. The energy storage charging power constraint and the energy storage discharge power constraint ensure that both the energy storage charging power and the energy storage discharge power are within the power limit range.

[0227] Preferably, the energy storage capacity constraint is:

[0228]

[0229] Among them, E n,t This represents the energy storage capacity at time t. and These represent the lower and upper limits of the energy storage capacity, respectively. This constraint prevents the energy storage unit from being overcharged or over-discharged.

[0230]

[0231] Among them, E Gfm (0) represents the initial energy of the stored energy.

[0232] Preferably, the energy storage charge-discharge balance constraint is:

[0233]

[0234] Where, N T This represents the entire scheduling time period. This constraint is beneficial for the cyclic scheduling of energy storage, ensuring that the energy storage capacity is equal at the beginning and end of the scheduling process.

[0235] Preferably, the power grid interaction constraint is:

[0236]

[0237] in, This represents the maximum power transmitted through the tie line between the microgrid and the upper-level grid. It takes into account that power is not transmitted in reverse during periods of high renewable energy generation. The renewable energy is consumed locally through coordinated dispatch. Power is purchased from the upper-level grid only when the power generation units within the microgrid cannot meet the load demand.

[0238] Preferably, the inertia constraint of the microgrid system is:

[0239]

[0240] in, P represents the inertia of a microgrid system. CS This represents the active power disturbance under the extreme anticipated fault condition, where f0 represents the rated frequency, and RoCoF max This represents the maximum rate of frequency change. This constraint ensures the frequency stability of the microgrid, so that the rate of frequency change of the system after a disturbance should be limited to a controllable range.

[0241] Preferably, the inertia of the microgrid system is as follows:

[0242]

[0243] Among them, H G Represents the inertial time constant. This indicates the generator's maximum power. This represents the equivalent inertial time constant of energy storage during time period t. This represents the maximum output power of grid-type energy storage.

[0244] Preferably, the equivalent inertial time constant of energy storage during time period t. Specifically:

[0245]

[0246] Among them, H max and H min These represent the upper and lower limits of the energy storage inertial time constant, respectively. This represents the maximum power transmitted via the tie line between the microgrid and the grid-connected energy storage.

[0247] Preferably, the release of the inertia support capacity of grid-type energy storage is essentially due to the power response capability of the energy storage, with the energy storage inertia time constant ranging from 4s to 12s. Considering that in actual operation, the transmission power of the tie line is relatively large during peak load periods, and the power deficit caused by the disconnection of the tie line has a significant impact on the system frequency stability, the equivalent inertia time constant of the energy storage during time period t is determined. The inertial time constant of the energy storage is set to be proportional to the tie-line power. When the inertial time constant is large, the energy released by the energy storage is large, the frequency change rate decreases, which is beneficial to maintaining the power balance of the system. When the inertial time constant is small, the energy released by the energy storage is small, the frequency change rate is large, which may lead to frequency stability problems in the system.

[0248] The preferred method for constructing a two-stage robust optimization model is as follows:

[0249] The first coefficient matrix is ​​obtained based on power balance constraints and energy storage charge-discharge balance constraints.

[0250] The second coefficient matrix is ​​obtained based on wind power constraints, photovoltaic power constraints, generator power constraints, generator ramping power constraints, and grid interaction power constraints.

[0251] The third coefficient matrix is ​​obtained based on the constraints of rated power of energy storage, discharge power of energy storage, charging power of energy storage, capacity of energy storage, inertia of microgrid system and reserve power of system.

[0252] Obtain the wind and solar coefficient matrix based on the wind and solar uncertainty set;

[0253] Obtain the uncertainties of wind power and photovoltaic power, and combine them with the uncertainties of load to obtain a column vector of uncertainties;

[0254] A two-stage robust optimization model is constructed by combining the first coefficient matrix, the second coefficient matrix, the third coefficient matrix, the wind and light coefficient matrix, the column vector, the first-stage optimization variables, the second-stage optimization variables, and the coefficient column vector.

[0255] Preferably, the column vector g of the uncertainties is specifically:

[0256]

[0257] in, This represents the uncertainty of wind power in the t-th time period under the first scenario. This represents the uncertainty of wind power in the t-th time period under the second scenario. This represents the uncertainty of photovoltaic power in the t-th time period under the first scenario. Let g represent the uncertainty of photovoltaic power in the t-th time period under the second scenario. To ensure the robustness of the energy storage configuration and scheduling scheme, and to satisfy the constraints under various "wind-solar-load" scenarios, an optimal solution exists. That is, to find the economically optimal configuration and scheduling scheme when g changes towards the worst-case scenario within the uncertainty set U.

[0258] The preferred two-stage robust optimization model is as follows:

[0259]

[0260] Where c1 and c2 represent the first and second coefficient column vectors respectively, K represents the first coefficient matrix, D represents the second coefficient matrix, F and G both represent the third coefficient matrix, d, h, u1 and u2 all represent constant column vectors, and I represents the wind and light coefficient matrix.

[0261] The optimal scheduling scheme is obtained by using a column constraint generation algorithm based on a two-stage robust optimization model.

[0262] The preferred approach yields the optimal scheduling scheme, which specifically includes:

[0263] S1 decomposes the main problem and sub-problems based on a two-stage robust optimization model;

[0264] S2 sets a set of uncertain variables as the initial worst-case scenario and sets the upper and lower bounds of the initial operating cost;

[0265] S3 solves the main problem based on the initial worst-case scenario and obtains the objective function value of the main problem as the current master optimal solution;

[0266] S4 uses the current master optimal solution as the new lower bound for operating cost;

[0267] S5 substitutes the current master optimal solution into the subproblem to obtain the objective function value of the subproblem as the current suboptimal solution;

[0268] S6 updates the new upper bound of operating cost based on the current primary optimal solution, the current suboptimal solution, and the initial upper bound of operating cost;

[0269] S7 calculates the cost difference by subtracting the new upper bound and the new lower bound of operating costs.

[0270] S8 determines whether the cost difference is greater than the convergence threshold;

[0271] If S9 is true, it indicates non-convergence; determine whether a solution exists.

[0272] If S10 exists, then add the first constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output;

[0273] If S11 does not exist, then add a second constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output;

[0274] If S12 is not true, it indicates convergence, the iteration stops, and the current principal optimal solution and the current suboptimal solution are taken as the final principal optimal solution and the final suboptimal solution.

[0275] S13 obtains the optimal scheduling scheme based on the final master optimal solution and the final suboptimal solution.

[0276] Preferably, based on a two-stage robust optimization model decomposition, k is defined as the iteration number, and the value of the uncertain variable under the worst-case scenario obtained after the l-th iteration is... δ is an auxiliary variable, and the main problem can be expressed as:

[0277]

[0278] Among them, y l Let I represent the column vector of decision variables in the l-th iteration. l Let g represent the wind and solar coefficient matrix in the l-th iteration. l This represents the column vector of uncertainties in the l-th iteration.

[0279] Preferably, given and , the inner optimization problem is linear, and the Gurobi solver can be used to obtain the KKT conditions. This transforms the inner problem into a max form, which is then merged with the outer max problem to obtain the subproblem:

[0280]

[0281] Where x* represents the column vector of decision variables obtained in the first stage of optimization. After the above derivation and transformation, the two-stage robust model can be decoupled into a main problem and subproblems with mixed integer linear form, which can then be solved using the C&CG algorithm.

[0282] Preferably, the new upper bound of operating costs UB n for

[0283]

[0284] Where UB represents the upper bound of the initial operating cost. This represents the current principal optimal solution. This represents the current suboptimal solution.

[0285] Preferably, the first constraint is: Where y k+1 This indicates the addition of a variable.

[0286] Preferably, the second constraint is:

[0287]

[0288] Solvability is ensured by the introduced second constraint.

[0289] Example 3

[0290] Verification of the method of this invention:

[0291] Based on the actual conditions in a certain area, the peak-hour electricity price is 0.55 yuan / kWh, the off-peak electricity price is 0.35 yuan / kWh, and the grid connection price for wind and solar power is 0.33 yuan / kWh; the upper limit for purchasing electricity from the grid is 200kW; the unit power cost required for configuring energy storage is 35 yuan, and the unit capacity cost is 144 yuan; the charging cost of energy storage is 0.23 yuan / kWh, and the discharging cost is 0.38 yuan / kWh; the initial SOC of energy storage is 0.5, the lower limit of SOC is 0.2, the upper limit of SOC is 0.8, and the charging and discharging efficiency is taken as 0.95; the overload capacity of grid-type energy storage is 1.5; the maximum power of the synchronous generator is 150kW, the ramping power is 20kW / h, the operating cost coefficient is 0.67, and the inertia time constant is 8s; the anticipated active power disturbance is 60kW, the rated frequency is 50Hz, and the maximum frequency change rate is 1 (Hz / s).

[0292] Considering load fluctuations between 0 and 20, and total fluctuations of wind and solar power between 6 and 12; during the optimization process, such as Figure 2 As shown, the upper and lower bounds of the robust optimization strategy proposed in this invention are represented by the dashed lines in the figure. After three iterations, the objective function converges to 3830. The required energy storage rated power is approximately 85.6kW, and the rated capacity is approximately 124.2kWh, corresponding to a daily investment cost of approximately 9.2 yuan. If the probability of the wind-solar boundary event is not considered, the result of the traditional robust optimization method is approximately 4042, yielding an energy storage rated power of approximately 104.3kW and a rated capacity of approximately 198.6kWh, corresponding to a daily investment cost of approximately 14.2 yuan.

[0293] The microgrid operating conditions obtained by the two robust optimization strategies are as follows: Figures 3-5 As shown. On the one hand, compared to the proposed strategy, when the probability of wind and solar load scenarios is not considered, the wind and solar power tends to reach the lower limit of the uncertainty range. On the other hand, the load scenarios obtained by the two robust methods are relatively similar, and the overall load power is higher than the benchmark value, tending to approach the upper limit of the uncertainty range. The traditional robust optimization strategy considers more severe microgrid conditions and is more conservative. When the uncertainty of "wind and solar load" is ignored, under the benchmark condition, the optimization result of the deterministic method is approximately 3237, resulting in a rated energy storage power of 57.6kW, a rated capacity of 303.3kWh, and a daily investment construction cost of 20.2 yuan for energy storage. Compared to the deterministic optimization model, the energy storage rated power configured by the strategy of this invention is increased by 48.6%, the rated capacity is reduced by 59%, and the configuration cost is reduced by 54%. The traditional robust optimization strategy is more conservative, with a configured energy storage rated power increased by 81.1%, a rated capacity reduced by 34.5%, and a configuration cost reduced by 29%.

[0294] In the corresponding scenario, the proposed robust optimization strategy, the traditional robust optimization strategy, and the day-ahead scheduling scheme of the deterministic optimization strategy are as follows: Figures 6-8 As shown, all three strategies can meet the system power balance requirements. Meanwhile, throughout the entire scheduling cycle, the State of Charge (SOC) of grid-based energy storage can be maintained between 20% and 80%.

[0295] In the diagram, PV represents photovoltaic power, Wind represents wind power, Generator represents motor power, Grid represents grid interaction power, Load represents load power, GFM discharge represents grid-connected energy storage discharge power, GFM charge represents grid-connected energy storage charging power, and SOC represents state of charge.

[0296] Under the three optimization methods, the system inertia and reserve power are as follows: Figures 9-10 As shown. Overall, the system inertia corresponding to the deterministic optimization strategy is smaller than that of the robust optimization strategy. At 19h, the energy storage transmission power is relatively large, and the minimum rotational kinetic energy of the system increases. The system inertia corresponding to the robust optimization strategy, the proposed robust optimization strategy, and the traditional deterministic optimization strategy are 2430kW·s, 2117kW·s, and 1546kW·s, respectively. At the same time, all three optimization strategies can provide sufficient reserved power for the system. In the early morning (0-5h) and evening (20-24h), the reserved power of the three strategies is relatively high. The reserved power corresponding to the traditional robust optimization strategy, the proposed robust optimization strategy, and the deterministic optimization strategy are 295kW, 265kW, and 226kW, respectively. At noon (10-15h) and 19h, due to the increased load and the decrease in the output level of new energy sources, the reserved power of all three strategies decreases. The minimum reserved power of the two robust optimization strategies is 60kW, and the minimum reserved power of the traditional deterministic optimization strategy is 81kW.

[0297] This invention clarifies the operational constraints of each distributed generation unit in a microgrid, analyzes the frequency characteristics of the microgrid, and establishes inertia constraints and reserve power constraints to ensure that system frequency changes meet load requirements under anticipated contingencies. Then, considering the correlation between wind and solar power output in the same region, a joint wind-solar probability distribution model is constructed based on kernel density estimation and the Frank Copula function, and k-means clustering is used to determine the wind-solar fluctuation boundaries and their corresponding probabilities. Finally, considering load forecasting errors, uncertainty adjustment parameters are introduced to form an uncertain set of "wind-solar-load". Based on column constraint generation algorithms and KKT conditions, a two-stage robust optimization configuration and scheduling strategy for the microgrid is formed. Results show that, on the one hand, compared with deterministic optimization strategies, the proposed strategy increases the rated power of energy storage by 48.6% and reduces capacity by 59%, exhibiting better robustness while retaining sufficient inertia and reserve power. On the other hand, the proposed strategy reduces the conservatism of traditional robust optimization strategies and lowers energy storage configuration costs by approximately 54%.

[0298] Example 4

[0299] A two-stage optimization system for microgrids that considers energy storage network construction capabilities includes: a dataset acquisition module, a first variable acquisition module, a second variable acquisition module, a column vector acquisition module, an optimization model construction module, and an optimal solution output module;

[0300] The dataset acquisition module is used to obtain wind and solar power datasets based on the wind power and solar power of the target microgrid; to obtain scene parameters by clustering the wind and solar power datasets; and to obtain the wind and solar uncertainty set based on the scene parameters.

[0301] The first variable acquisition module is used to obtain the rated power and rated capacity of energy storage based on the target microgrid as the first-stage optimization variables.

[0302] The second variable acquisition module is used to obtain the system load power based on the target microgrid as the second-stage optimization variable;

[0303] The column vector acquisition module is used to obtain the coefficient column vector when the daily comprehensive operating cost is the lowest based on the second-stage optimization variables.

[0304] The optimization model building module is used to establish a two-stage robust optimization model of the target microgrid based on the wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints.

[0305] The optimal solution output module is used to solve the optimal scheduling scheme based on the two-stage robust optimization model using a column constraint generation algorithm.

[0306] Preferably, in this embodiment, the functional implementation methods of each functional module correspond one-to-one with the methods described above, and will not be described in detail here.

[0307] Example 5

[0308] Based on the same inventive concept, the present invention also provides a computer device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0309] Memory, used to store computer programs;

[0310] When the processor executes a program stored in the memory, it is able to implement a two-stage optimization method for microgrids that takes into account the energy storage network capability, as in Embodiment 1 or 2.

[0311] The electronic device may include a processor, a communications interface, a memory, and a communication bus, wherein the processor, communications interface, and memory communicate with each other via the communication bus. The processor can call logical instructions in the memory to execute a two-stage optimization method for microgrids considering energy storage network capabilities, as described in Embodiment 1 or 2.

[0312] Furthermore, when the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0313] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0314] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A two-stage optimization method for microgrids considering energy storage grid construction capabilities, characterized in that, include: A wind and solar power dataset is obtained based on the wind power and solar power of the target microgrid; Scene parameters are obtained by clustering the aforementioned wind and light dataset; The uncertain set of wind and light is obtained based on the scene parameters; Based on the target microgrid, the rated power and rated capacity of energy storage are obtained as the first-stage optimization variables. The system load power obtained based on the target microgrid is used as the optimization variable for the second stage. Based on the second-stage optimization variables, obtain the coefficient column vector when the daily comprehensive operating cost is at its lowest; A two-stage robust optimization model for the target microgrid is established based on the aforementioned wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints. The optimal scheduling scheme is obtained by using a column constraint generation algorithm based on the two-stage robust optimization model.

2. The two-stage optimization method for microgrids considering energy storage network construction capability according to claim 1, characterized in that, The method for obtaining the landscape dataset is as follows: Based on the wind power and the photovoltaic power, establish the edge probability density function of wind power output and the edge probability density function of photovoltaic power output; Based on the integration of the wind power output edge probability density function and the photovoltaic output edge probability density function respectively, the wind power edge distribution function and the photovoltaic edge distribution function are obtained respectively; Based on the Frank Copula function, the wind power edge distribution function and the photovoltaic edge distribution function are combined to obtain the wind-solar joint probability distribution function; Based on the joint probability distribution function of wind and solar energy, sampling and inverse transformation are performed to generate multiple wind and solar energy datasets.

3. The two-stage optimization method for microgrids considering energy storage grid construction capability according to claim 1, characterized in that, The method for obtaining the scene parameters is as follows: Clustering is performed on the aforementioned landscape dataset, and similar scenes are merged to obtain two typical landscape scenes; Based on the two typical wind and solar scenarios, the wind power of the first scenario, the photovoltaic power of the first scenario, the wind power of the second scenario, and the photovoltaic power of the second scenario are obtained. The wind power benchmark value is obtained by averaging the wind power in the first scenario and the wind power in the second scenario. A photovoltaic baseline value is obtained by averaging the photovoltaic power of the first scenario and the photovoltaic power of the second scenario. The wind power deviation is obtained by subtracting the wind power power in the first scenario from the wind power benchmark value. The photovoltaic deviation is obtained by subtracting the photovoltaic power in the first scenario from the photovoltaic reference value. The predicted wind power value is obtained based on the wind power benchmark value and the wind power deviation; The photovoltaic power prediction value is obtained based on the photovoltaic baseline value and the photovoltaic deviation; The predicted wind power and the predicted photovoltaic power are used as the scenario parameters.

4. The two-stage optimization method for microgrids considering energy storage network construction capability according to claim 3, characterized in that, The method for obtaining the uncertain set of wind and light is as follows: Obtain the load uncertainty, system load, wind power uncertainty, and photovoltaic power uncertainty of the target microgrid; The wind and solar uncertainty set is composed of the load uncertainty variable, the system load, the wind power uncertainty variable, the photovoltaic power uncertainty variable, the wind power prediction value, and the photovoltaic power prediction value.

5. A two-stage optimization method for microgrids considering energy storage network construction capability according to claim 4, characterized in that, The system load power includes: wind power, photovoltaic power, grid-type energy storage power, generator power, and grid interconnection power; The grid-type energy storage power includes energy storage charging power and energy storage discharging power.

6. A two-stage optimization method for microgrids considering energy storage network construction capability according to claim 5, characterized in that, The method for obtaining the coefficient column vector is as follows: The wind power operating cost is obtained based on the wind power output and the unit price of wind turbine power generation. The cost of photovoltaic power generation is obtained based on the photovoltaic power and the unit price of photovoltaic power generation. The operating cost of grid-type energy storage is obtained based on the energy storage charging power, the energy storage discharging power, the energy storage unit power charging cost, the energy storage unit power discharging cost, and the energy storage charging and discharging efficiency. The construction cost of grid-type energy storage is obtained based on the unit cost of energy storage power capacity, the unit cost of energy capacity, the benchmark discount rate, and the energy storage operating life. The generator operating cost is obtained based on the generator power and operating cost parameters. The grid purchase cost is obtained based on the grid interaction power and the grid-side peak-valley electricity price. The daily comprehensive operating cost of the target microgrid is obtained based on the wind power operating cost, the photovoltaic power generation cost, the grid-connected energy storage operating cost, the grid-connected energy storage construction cost, the generator operating cost, and the grid power purchase cost. Based on the lowest daily comprehensive operating cost, an optimization function is constructed and the first coefficient column vector and the second coefficient column vector are extracted. The first coefficient column vector and the second coefficient column vector together form the coefficient column vector.

7. A two-stage optimization method for microgrids considering energy storage grid construction capability according to claim 6, characterized in that, The constraints include: Power balance constraints, wind power constraints, photovoltaic power constraints, generator power constraints, generator ramping power constraints, energy storage rated power constraints, energy storage discharge power constraints, energy storage charging power constraints, energy storage capacity constraints, energy storage charge and discharge balance constraints, grid interaction power constraints, microgrid system inertia constraints, and system reserve power constraints.

8. A two-stage optimization method for microgrids considering energy storage network construction capability according to claim 7, characterized in that, The method for constructing the two-stage robust optimization model is as follows: The first coefficient matrix is ​​obtained based on the power balance constraint and the energy storage charge-discharge balance constraint; The second coefficient matrix is ​​obtained based on the wind power constraint, the photovoltaic power constraint, the generator power constraint, the generator ramping power constraint, and the grid interaction power constraint. The third coefficient matrix is ​​obtained based on the energy storage rated power constraint, the energy storage discharge power constraint, the energy storage charging power constraint, the energy storage capacity constraint, the microgrid system inertia constraint, and the system reserve power constraint; The wind and solar coefficient matrix is ​​obtained based on the aforementioned wind and solar uncertainty set; Obtain the uncertainties of wind power and photovoltaic power, and combine them with the uncertainties of the load to obtain a column vector of uncertainties; The two-stage robust optimization model is constructed based on the first coefficient matrix, the second coefficient matrix, the third coefficient matrix, the wind and light coefficient matrix, the column vector, the first-stage optimization variable, the second-stage optimization variable, and the coefficient column vector.

9. A two-stage optimization method for microgrids considering energy storage grid construction capability according to claim 1, characterized in that, Obtaining the optimal scheduling scheme specifically includes: S1 decomposes the main problem and sub-problems based on the two-stage robust optimization model; S2 sets a set of uncertain variables as the initial worst-case scenario and sets the upper and lower bounds of the initial operating cost; S3 solves the main problem based on the initial worst-case scenario and obtains the objective function value of the main problem as the current master optimal solution; S4 uses the current master optimal solution as a new lower bound for operating costs; S5 substitutes the current master optimal solution into the subproblem to obtain the objective function value of the subproblem as the current suboptimal solution; S6 updates the new upper bound of operating cost based on the current primary optimal solution, the current suboptimal solution, and the initial upper bound of operating cost; S7 calculates the cost difference by subtracting the new upper bound of operating cost from the new lower bound of operating cost; S8 determines whether the cost difference is greater than the convergence threshold; If S9 is true, it indicates non-convergence; determine whether a solution exists. If S10 exists, then add the first constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output; If S11 does not exist, then add a second constraint, return to S3 iteration, and continue until the final principal optimal solution and the final suboptimal solution are output; If S12 is not true, it indicates convergence, the iteration stops, and the current principal optimal solution and the current suboptimal solution are taken as the final principal optimal solution and the final suboptimal solution. S13 obtains the optimal scheduling scheme based on the final master optimal solution and the final suboptimal solution.

10. A two-stage optimization system for microgrids considering energy storage grid construction capabilities, used to execute a two-stage optimization method for microgrids considering energy storage grid construction capabilities as described in any one of claims 1-9, characterized in that, include: The system includes a dataset acquisition module, a first variable acquisition module, a second variable acquisition module, a column vector acquisition module, an optimization model construction module, and an optimal solution output module. The dataset acquisition module is used to obtain a wind and solar power dataset based on the wind power and photovoltaic power of the target microgrid; to obtain scene parameters by clustering the wind and solar power dataset; and to obtain a wind and solar power uncertainty set based on the scene parameters. The first variable acquisition module is used to acquire the rated power and rated capacity of energy storage based on the target microgrid as the first-stage optimization variables; The second variable acquisition module is used to acquire the system load power based on the target microgrid as the second-stage optimization variable; The column vector acquisition module is used to obtain the coefficient column vector when the daily comprehensive operating cost is the lowest based on the second-stage optimization variables. The optimization model construction module is used to establish a two-stage robust optimization model of the target microgrid based on the wind and solar uncertainty set, the first-stage optimization variables, the second-stage optimization variables, the coefficient column vector, and the constraints. The optimal solution output module is used to solve the optimal scheduling scheme by using a column constraint generation algorithm based on the two-stage robust optimization model.

Images