Microgrid scheduling method and device, and storage medium

Abstract

Provided are a microgrid scheduling method and device, and a storage medium. The method comprises: first, determining a distributionally robust optimization model on the basis of power grid data of a microgrid, wherein the model can be used for generating, on the basis of given data, an effective scheduling strategy for the microgrid; then performing model conversion on the distributionally robust optimization model to obtain a target convex optimization model; solving the target convex optimization model to obtain a specific target scheduling strategy for the microgrid; and finally, performing a corresponding scheduling operation on the microgrid according to this strategy.

Description

Microgrid dispatching methods, equipment and storage media

[0001] Cross-references to related applications

[0002] This application is based on and claims priority to Chinese Patent Application No. 2025100817129, filed on January 16, 2025, the entire contents of which are incorporated herein by reference. Technical Field

[0003] The embodiments of this application relate to, but are not limited to, the field of microgrid dispatching, and particularly to a microgrid dispatching method, device, and storage medium. Background Technology

[0004] Microgrids, as an effective carrier for distributed energy management and optimized dispatch, integrate various distributed energy sources (such as solar and wind power). However, due to the intermittency and uncertainty of renewable energy generation, microgrid dispatch faces significant challenges, necessitating the development of efficient and reliable dispatch strategies to ensure stable grid operation and economic benefits.

[0005] Traditional microgrid dispatching methods primarily rely on predictive models and deterministic optimization algorithms. However, due to prediction errors and uncertainties, these methods often fail to achieve ideal dispatching results in practical applications. To address this challenge, Distributed Robust Optimization (DRO) emerges as an optimization method to handle uncertainty. DRO constructs an optimization model capable of withstanding worst-case uncertainties, thereby maximizing economic benefits while ensuring the robustness of the dispatching strategy. In constructing the DRO model, the accuracy of microgrid power flow calculations is crucial to ensuring model effectiveness. Traditional linear approximation methods, such as the DC power flow method, while computationally simple, suffer from significant errors when handling complex grid structures and nonlinear constraints. Traditional methods such as convex cone relaxation and the Newton-Raphson method have limitations in terms of computational speed and feasibility, making them unsuitable for real-time dispatching. Furthermore, traditional optimization algorithms are difficult to directly apply to microgrid dispatching problems with complex constraints and uncertainties when solving DRO models. Summary of the Invention

[0006] This application provides a microgrid scheduling method, device, and storage medium.

[0007] On one hand, embodiments of this application provide a microgrid scheduling method, comprising: determining a sub-Bruker optimization model based on grid data of the microgrid, wherein the sub-Bruker optimization model is used to obtain a scheduling strategy for the microgrid; performing model transformation on the sub-Bruker optimization model to obtain a target convex optimization model; solving the target convex optimization model to obtain a target scheduling strategy for the microgrid; and scheduling the microgrid according to the target scheduling strategy.

[0008] On the other hand, embodiments of this application also provide a microgrid scheduling device, including: at least one processor; at least one memory for storing at least one program; at least one program is run by at least one processor to perform the microgrid scheduling method as described above.

[0009] On the other hand, embodiments of this application also provide a computer-readable storage medium storing computer-executable instructions for executing the microgrid scheduling method described above.

[0010] On the other hand, embodiments of this application also provide a computer program product, including a computer program or computer instructions, wherein the computer program or computer instructions are stored in a computer-readable storage medium, the processor of the microgrid dispatching device reads the computer program or computer instructions from the computer-readable storage medium, and the processor executes the computer program or computer instructions, causing the microgrid dispatching device to perform the microgrid dispatching method as described above. Attached Figure Description

[0011] Figure 1 is a flowchart of a microgrid scheduling method provided in an embodiment of this application;

[0012] Figure 2 is a detailed flowchart of step S110 in Figure 1 provided in an embodiment of this application;

[0013] Figure 3 is a detailed flowchart of step S230 in Figure 2 provided in an embodiment of this application;

[0014] Figure 4 is a flowchart of a specific embodiment of the present application for determining a linear approximate power flow model;

[0015] Figure 5 is a detailed flowchart of step S320 in Figure 3 provided in an embodiment of this application;

[0016] Figure 6 is a detailed flowchart of step S130 in Figure 1 provided in an embodiment of this application;

[0017] Figure 7 is a detailed flowchart of step S610 in Figure 6 provided in an embodiment of this application;

[0018] Figure 8 is a diagram showing the relationship of a microgrid scheduling method provided in a specific example of this application.

[0019] Figure 9 is a flowchart of the steps of a microgrid scheduling method provided in a specific example of this application. Detailed Implementation

[0020] To make the objectives, technical methods, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] It should be noted that although the flowchart shows a logical order, in some cases, the steps shown or described may be performed in a different order than that shown in the flowchart. In the description of the specification, claims, and the foregoing drawings, "multiple" means two or more; "greater than," "less than," and "exceeding" are understood to exclude the stated number; "above," "below," and "within" are understood to include the stated number. The use of terms such as "first" and "second" is merely for distinguishing technical features and should not be construed as indicating or implying relative importance, or implicitly specifying the number of indicated technical features or their sequential relationship.

[0022] With the transformation of the global energy structure and the rapid development of smart grid technology, microgrids, as an effective carrier for distributed energy management and optimized dispatch, are increasingly becoming a research and practice hotspot. Microgrids integrate various distributed energy sources (such as solar and wind power) and achieve local energy production, storage, and efficient utilization through advanced power electronic equipment and intelligent control systems. However, due to the intermittency and uncertainty of renewable energy generation, microgrid dispatch faces significant challenges, necessitating the development of efficient and reliable dispatch strategies to ensure stable grid operation and economic benefits.

[0023] It is worth noting that traditional microgrid dispatching methods mainly rely on prediction models and deterministic optimization algorithms. However, due to the influence of prediction errors and uncertainties, these methods often fail to achieve ideal dispatching results in practical applications. To address this challenge, Distributed Robust Optimization (DRO), as an optimization method for handling uncertainty, constructs an optimization model capable of resisting uncertainty in the worst-case scenario, thereby maximizing economic benefits while ensuring the robustness of the dispatching strategy. Specifically, its basic idea is to construct a fuzzy set of the probability distribution of random variables and solve for the optimal decision under the worst-case distribution of the random variables in the fuzzy set. This ensures the feasibility of the system under the worst-case scenario and also prevents the dispatching results for economic objectives from being overly conservative, combining the advantages of robust optimization and stochastic optimization. The types of DRO models can be classified according to the different ways of constructing the fuzzy set of probability distribution. Compared with traditional fuzzy sets based on moment information, fuzzy sets based on Wasserstein distance can make fuller use of historical data to obtain an uncertainty set that is closer to the actual probability distribution, effectively reducing the conservatism of the DRO method.

[0024] In constructing a distributed bar optimization model, the accuracy of microgrid power flow calculation is crucial to ensuring the model's effectiveness. Traditional linear approximation methods, such as the DC power flow method, while computationally simple, suffer from significant errors when handling complex grid structures and nonlinear constraints. While traditional methods like convex cone relaxation and the Newton-Raphson method offer high accuracy, they have limitations in computational speed and feasibility, making them unsuitable for real-time scheduling requirements.

[0025] Furthermore, solving the subbulk bar optimization model also faces many challenges. Traditional optimization algorithms are often difficult to apply directly to microgrid scheduling problems with complex constraints and uncertainties. The primal-dual interior-point method, as an efficient and stable optimization algorithm, has significant advantages in solving convex optimization problems. However, considering the characteristics of the subbulk bar scheduling problem in microgrids, the traditional primal-dual interior-point method requires targeted improvements in initial point selection and algorithm design.

[0026] To improve the robustness and solution efficiency of microgrid scheduling strategies, this application provides a microgrid scheduling method, microgrid scheduling equipment, computer-readable storage medium, and computer program product. First, a sub-Bruker optimization model is determined based on the microgrid's grid data. This model can generate an effective scheduling strategy for the microgrid based on given data. Then, the sub-Bruker optimization model is transformed to obtain an easily solvable objective convex optimization model. By solving this objective convex optimization model, a specific target scheduling strategy for the microgrid can be derived. Finally, based on this strategy, actual scheduling operations can be performed on the microgrid. In this application, the optimization model constructed using the sub-Bruker optimization method can withstand worst-case uncertainties, thus ensuring the stability of the scheduling strategy and the feasibility of the system even when facing prediction errors and various uncertainties. In the process of constructing the sub-Bruker optimization model, historical grid data of the microgrid is fully utilized, resulting in an uncertainty set that more closely approximates the actual probability distribution. This approach significantly reduces the conservatism of the sub-Bruker optimization method, making the scheduling strategy more closely aligned with actual operational needs, thereby improving the overall system performance and economic benefits. Furthermore, addressing the limitations of traditional linear approximation methods in handling complex power grid structures and nonlinear constraints, the adopted sub-Bruker optimization model significantly improves computational efficiency while maintaining computational accuracy. Moreover, by appropriately transforming and solving the optimization model, the problem of obtaining reliable initial points in traditional methods can be resolved, thus effectively overcoming the limitations of traditional optimization algorithms in solving the sub-Bruker scheduling problem in microgrids. In summary, the embodiments of this application significantly improve the robustness, solution efficiency, and economy of microgrid scheduling strategies.

[0027] Based on the above analysis, the embodiments of this application will be further described below with reference to the accompanying drawings.

[0028] Referring to Figure 1, which is a flowchart of a microgrid scheduling method provided in an embodiment of this application, the method includes, but is not limited to, steps S110 to S140.

[0029] Step S110: Based on the microgrid grid data, determine the sub-Blu-rod optimization model, which can be used to obtain the scheduling strategy for the microgrid.

[0030] Step S120: Transform the bipolar bar optimization model to obtain the target convex optimization model.

[0031] Step S130: Solve the target convex optimization model to obtain the target scheduling strategy for the microgrid.

[0032] Step S140: Dispatch the microgrid according to the target scheduling strategy.

[0033] As we understand it, a microgrid is a small-scale power system composed of distributed generation, loads, energy storage, power distribution, and control systems, capable of self-control, protection, and management. Microgrid grid data forms the basis for constructing a distributed bar optimization model to formulate microgrid dispatch strategies. This data typically includes distributed generation data, load data, energy storage system data, and electricity price data. Distributed generation data includes photovoltaic power generation data, wind power generation data, and other distributed generation data, such as the power generation, fuel consumption, and emissions data of micro gas turbines and fuel cells. Load data reflects the electricity demand of users in the microgrid, such as load demand at each time period, load type (e.g., residential load, commercial load, industrial load), and load fluctuations. Energy storage system data includes battery capacity, charging and discharging power, and remaining capacity. Electricity price data (e.g., time-of-use pricing data) reflects the electricity market price situation at different times. After obtaining this grid data, a distributed bar optimization model can be constructed. This model can consider data uncertainties and formulate dispatch strategies that maintain good performance under various possible conditions.

[0034] In some examples, as shown in Figure 2, the specific process of determining the sub-bulb optimization model based on the microgrid's grid data in step S110 may include, but is not limited to, steps S210 to S230.

[0035] Step S210: Based on the microgrid's grid data, determine the decision variables and random variables used to construct the sub-Bruker optimization model.

[0036] Step S220: Obtain historical data of random variables of the microgrid, and determine the fuzzy set used to construct the sub-Bruker optimization model based on the random variables and the historical data of random variables.

[0037] Step S230: Determine the sub-Blu-ray optimization model based on the decision variables and fuzzy sets.

[0038] In some examples, when determining the decision variables and random variables used to construct the sub-Bruker optimization model based on microgrid grid data, the analysis can start from the microgrid's operating characteristics and uncertainties. It should be noted that within the sub-Bruker optimization model of a microgrid, decision variables typically encompass a series of controllable parameters. These parameters include, but are not limited to, generator dispatch commands (e.g., output power settings), energy storage system operating variables (e.g., switching of charge / discharge states and power levels), and load management strategies (which may involve load shedding, shifting, or dispatching). The optimal configuration of these decision variables directly determines the overall operating state of the microgrid and its energy utilization efficiency. On the other hand, random variables represent the uncertainties brought about by external factors in the microgrid that are difficult to predict accurately or control effectively. Among these factors, the prediction error of renewable energy (such as solar and wind power) output power is particularly noteworthy. The existence of these uncertainties may affect the microgrid's operating costs (e.g., additional dispatch costs due to energy supply and demand mismatch), the stability of energy supply (e.g., increased risk of power outages or voltage fluctuations), and overall economic performance (e.g., the possibility of reduced revenue or operating losses).

[0039] In some examples, in power system optimization, the input data often exhibits uncertainty due to factors such as load forecasting uncertainty and the volatility of renewable energy generation. This embodiment defines fuzzy sets to describe the range of uncertainty in the input data and incorporates it into the optimization model, which helps the model better adapt to actual conditions.

[0040] In some examples, fuzzy sets can represent not only the uncertainty of the probability distribution of a random variable, but also the uncertainty of its range of values. In constructing fuzzy sets, the Wasserstein distance can be used to measure the distance between the true distribution and the empirical distribution. Specifically, with the empirical distribution as the center, all distributions whose distance from the empirical distribution does not exceed a certain Wasserstein distance are included in the fuzzy set. Specifically, let μ and ν be two probability distributions, and W_p(μ,ν) represent the Wasserstein distance between them, where p≥1 is a constant. Then, the Wasserstein fuzzy set centered at μ and with radius r can be defined as: the set of all probability distributions whose Wasserstein distance from μ does not exceed r. This set forms a sphere in the probability distribution space, and μ is the center of this sphere.

[0041] In some examples, when constructing the fuzzy set for a split-Bruker optimization model, the empirical probability distribution of the random variable can be determined first based on historical data. This empirical probability distribution can be viewed as an approximation of the random variable data. If the data points are very dense or it is assumed that each data point is an independent, equally likely observation, then this empirical distribution might be constructed based on some form of the Dirac distribution (i.e., each data point is a discrete distribution with a probability of 1 / n, where n is the number of data points). Next, based on the random variable and the empirical probability distribution, the fuzzy set used to construct the split-Bruker optimization model is determined. For example, to construct the fuzzy set for the split-Bruker optimization model, a metric based on Wasserstein distance can be used. Wasserstein distance measures the distance between two probability distributions, taking into account not only the quality differences between the distributions but also the positional differences of these qualities in the distribution space. Therefore, a fuzzy set based on Wasserstein distance can be defined as the set of all probability distributions whose Wasserstein distance to the empirical probability distribution does not exceed a certain preset threshold.

[0042] In some examples, when there are multiple random variables, when determining the fuzzy set used to construct the subbulbar optimization model based on the random variables and empirical probability distributions, we can first determine the distance value between each random variable and the empirical probability distribution. Then, among the multiple random variables, those with distance values ​​less than or equal to a preset distance threshold are selected as the fuzzy set used to construct the subbulbar optimization model. Specifically, when faced with multiple random variables, in order to determine the fuzzy set used to construct the subbulbar optimization model based on these random variables and empirical probability distributions, we can first calculate the distance value between the empirical probability distribution of each random variable and a certain reference distribution (or ideal distribution). This distance value can measure the degree of difference between the actual distribution and the ideal distribution of the random variable. Next, among the multiple random variables, those random variables with distance values ​​less than or equal to the preset distance threshold are selected. These random variables have smaller differences from the ideal distribution and are therefore more likely to represent typical or main distribution characteristics in the actual situation.

[0043] In some examples, after determining the fuzzy set, a sub-Bruker optimization model can be further determined based on the decision variables and the fuzzy set. The core of this model lies in considering the uncertainty of random variables and striving to find the optimal scheduling strategy under uncertain conditions. Specifically, a sub-Bruker optimization model typically includes an objective function (e.g., aiming to minimize operating costs or maximize profits) and a series of constraints (such as power balance constraints, equipment capacity constraints, etc.). These constraints ensure that the formulated decision strategy is practically feasible in actual operation.

[0044] In some examples, during the process of determining the sub-bar optimization model based on decision variables and fuzzy sets, the objective function and corresponding constraints can be constructed based on the decision variables. At the same time, fuzzy sets can be constructed based on random variables and incorporated into the optimization model to comprehensively consider the impact of uncertainties on the decision strategy.

[0045] Referring to Figure 3, which is a flowchart of step S230 provided in an embodiment of this application, the step includes but is not limited to steps S310 and S320.

[0046] Step S310: Based on the decision variables, determine the linear approximate power flow model and power generation equipment constraints used to construct the sub-Blu-ray bar optimization model.

[0047] Step S320: Determine the sub-Blu-ray optimization model based on the fuzzy set, the linear approximate power flow model, and the constraints of the power generation equipment.

[0048] Understandably, power flow calculations are fundamental to power system analysis, used to determine voltage magnitudes and phase angles at nodes and power flow in branches. However, traditional power flow equations are nonlinear, increasing the complexity of the optimization problem. Therefore, a linear approximation of the power flow model can be used to simplify the problem when constructing a subbulb bar optimization model. This approximation method linearizes the power flow equations, making the optimization problem easier to solve while maintaining a certain level of accuracy. Furthermore, generating equipment may be constrained by various physical and economic factors during operation, such as generator output limitations, ramp rate limits, and minimum start-up and shutdown times. These constraints are taken into account when constructing the subbulb bar optimization model to ensure the feasibility of the obtained solution in practical operation.

[0049] In some examples, a subbulb bar optimization model can be constructed by considering a linear approximate power flow model and generator constraints, combined with uncertainties described by fuzzy sets. This model aims to find a decision scheme whose performance remains acceptable even in the worst case. Compared to traditional stochastic optimization models, the subbulb bar optimization model in this embodiment does not require precise knowledge of the probability distribution of uncertainties; instead, it optimizes based on the set of uncertainties (i.e., fuzzy sets). This makes the model more flexible and practical, especially when sufficient data is lacking to accurately estimate the probability distribution.

[0050] In some examples, a linear approximate power flow model can be obtained by the following steps: first, determining the node voltage constraints and power flow constraints of the microgrid based on the decision variables; then, determining the linear approximate power flow model based on the node voltage constraints and power flow constraints. The node voltage constraints are used to ensure that the voltage of each node in the grid is within a safe range. In a microgrid, node voltages may change due to the integration of distributed generation and load fluctuations. Therefore, setting upper and lower voltage limits ensures the stable operation of the grid. Power flow constraints describe the flow of power in the grid. In a microgrid, power flow constraints include branch power limits and power balance conditions. These constraints ensure that the power flow in the grid does not exceed the carrying capacity of the equipment while meeting load demands. It is understood that traditional power flow equations are nonlinear, involving complex calculations such as the product of voltage magnitude and phase angle. To simplify the problem, linearization methods can be used to approximate the power flow equations as linear equations. For example, linearization can be achieved by ignoring higher-order terms or using Taylor series expansion. Based on the linearization, and combining the node voltage constraints and power flow constraints, a linear approximate power flow model can be constructed. The model approximates the relationship between decision variables and grid conditions (such as node voltage and branch power) as a linear relationship.

[0051] Referring to Figure 4, which is a flowchart of a specific process for determining a linear approximate power flow model according to an embodiment of this application, the specific process for determining the linear approximate power flow model based on node voltage constraints and power flow constraints may include, but is not limited to, steps 410 to 430.

[0052] Step 410: Perform a first linear approximation on the power flow constraints to obtain the first linear approximation result.

[0053] Step 420: Perform a second linear approximation on the node voltage constraint based on the first linear approximation result to obtain the second linear approximation result.

[0054] Step 430: Determine the linear approximation power flow model based on the first linear approximation result and the second linear approximation result.

[0055] In some examples, the primary objective in step 410 is to linearize the power flow constraints describing the power flow in the power grid. These constraints typically encompass the balance of active and reactive power, as well as limitations on line power transmission. To achieve this, fixed-point linearization (FPL) can be used when performing a first linear approximation of the power flow constraints. FPL linearizes the original nonlinear equations based on an initially guessed solution (i.e., a fixed point). Specifically, the nonlinear equations are Taylor-expanded at the fixed point, and higher-order terms are ignored, resulting in a simplified linear equation.

[0056] In some examples, after obtaining a linear approximation of the power flow constraints (i.e., the first linear approximation), the nodal voltage constraints can be further linearized. Nodal voltage constraints typically involve limitations on voltage magnitude and phase angle. Therefore, a first-order Taylor (FOT) method can be used when performing a second linear approximation of the nodal voltage constraints. The FOT method performs a Taylor expansion of the nonlinear equations based on a given point (usually the operating point or equilibrium point) and retains only the first-order terms, thus obtaining a linearized voltage constraint equation. It is worth noting that this process depends to some extent on the linear approximation obtained in the first step, because there is a close relationship between nodal voltage and power flow state.

[0057] In some examples, in step 430, the linear approximation results obtained in the previous two steps are combined to construct a complete linear approximation power flow model. This model can be used for subsequent power grid analysis, optimization, or control tasks. Specifically, by using the FPL linear approximation of power flow constraints and the FOT method to linearly approximate node voltage constraints, the originally nonlinear constraints can be approximated as affine functions of decision variables and random variables. This linearization facilitates subsequent optimization and analysis.

[0058] In some examples, after constructing the linear approximate power flow model, verification and validation can be performed according to actual needs. The model's accuracy and applicability are checked by comparing it with actual power grid data. If the model has significant errors or does not conform to reality, adjustments and optimizations are required.

[0059] In some examples, after determining the linear approximate power flow model, a sub-Bruker optimization model can be further constructed by combining fuzzy sets, the linear model, and the actual operating constraints of the generating equipment. This model aims to address uncertainties in the power system and ensure the stability and economy of the power grid under various possible conditions. As shown in Figure 5, step S320 determines the specific process of the sub-Bruker optimization model based on fuzzy sets, the linear approximate power flow model, and the constraints of the generating equipment, including but not limited to steps S510 and S530.

[0060] Step S510: Determine the objective function of the fuzzy bar based on the fuzzy set.

[0061] Step S520: Determine the sub-Bluer constraint conditions based on the fuzzy set, the linear approximate power flow model, and the constraints of the power generation equipment.

[0062] Step S530: Based on the objective function and constraints of the sub-Bruker bar, obtain the sub-Bruker bar optimization model.

[0063] In some examples, the significance of the fuzzy bar objective function lies in minimizing the expected value of the objective function under the most unfavorable probability distribution in the fuzzy set.

[0064] In some examples, step S510 aims to define a sub-Bruker objective function based on a fuzzy set. The fuzzy set is used here to describe uncertainties that may exist in grid operation, such as fluctuations in load demand and the intermittency of renewable energy generation. The sub-Bruker objective function aims to minimize the expected mathematical cost under the most unfavorable probability distribution. Here, the most unfavorable probability distribution refers to the distribution that maximizes the objective function value within the range of uncertainties described by the fuzzy set. By optimizing this objective function, it can be ensured that the overall performance of the system remains within an acceptable range when facing uncertainties.

[0065] In some examples, when determining the objective function of the fuzzy set, the specific form of the fuzzy set, including its membership function and parameters, can be determined first. Then, the objective function is derived based on these definitions, which typically involves multiple aspects such as generation costs, load shedding costs, and grid losses. Finally, using the concepts of mathematical expectation and probability distribution, the objective function is expressed as the expected value under the most unfavorable probability distribution.

[0066] In some examples, in step S520, constraints for the subbulk optimization model can be defined by combining fuzzy sets, a linear approximate power flow model, and actual operating constraints of generating equipment. The linear approximate power flow model provides a simplified description of the grid state, making the optimization problem easier to solve. In this model, the power flow equations are linearized, significantly reducing computational complexity. The inclusion of constraints on generating equipment ensures the feasibility of the solution in practice. Since fuzzy sets describe uncertainties, the impact of these uncertainties on grid operation can be considered in the constraints. For example, the range of constraint variations under different load demands or renewable energy generation levels can be defined.

[0067] In some examples, in step S530, the objective function and constraints obtained in the previous two steps are combined to construct a complete sub-Bru bar optimization model.

[0068] In some examples, after obtaining the subbulb bar optimization model, to further simplify and solve the model, it can be further transformed into a deterministic, finite-dimensional convex optimization problem. This transformation not only makes the model easier to handle but also allows for efficient finding of the optimal solution using existing convex optimization theories and algorithms. In this embodiment, the specific process of transforming the subbulb bar optimization model to obtain the target convex optimization model in step S120 includes: first, transforming the subbulb bar optimization model to obtain a transformed convex optimization model; then, adding auxiliary constraints and auxiliary variables to the transformed convex optimization model to obtain the target convex optimization model. Specifically, in the stage of obtaining the transformed convex optimization model, the uncertainty in the subbulb bar optimization model is mainly eliminated or transformed in some way to obtain a convex optimization model that is simpler and easier to handle. The uncertainty in the subbulb bar optimization model can usually be described by fuzzy sets. To transform it into a convex optimization model, some traditional mathematical methods, such as worst-case analysis and duality theory, can be used to transform uncertainty into deterministic constraints or objective functions. During the transformation process, it is necessary to ensure that the resulting model remains convex, meaning that the objective function and constraints must maintain convexity when dealing with uncertainties. For example, if the objective function or constraints in the original model contain non-convex terms, they can be transformed into a convex form by introducing auxiliary variables, relaxation conditions, or using convex approximation. After obtaining the convex optimization model through the initial transformation, further auxiliary constraints and variables need to be added to refine the model. Auxiliary constraints can be used to ensure that the transformed model still satisfies certain key characteristics of the original problem, such as the feasibility and optimality of the solution. For example, if relaxation variables are introduced during the transformation process, additional constraints may need to be added to ensure that the values ​​of these variables are within a reasonable range. The introduction of auxiliary variables can often be used to simplify the model structure or improve solution efficiency. For example, new variables can be introduced to represent certain complex expressions in the original problem, thereby simplifying the form of the objective function or constraints.

[0069] In some examples, the target convex optimization model includes inequality constraints. As shown in Figure 6, the specific process of solving the target convex optimization model in step S130 to obtain the target scheduling strategy for the microgrid includes, but is not limited to, steps S610 and S620.

[0070] Step S610: Determine the target parameters that satisfy the inequality constraints.

[0071] Step S620: Solve the target convex optimization model based on the target parameters to obtain the target scheduling strategy for the microgrid.

[0072] In some examples, in objective convex optimization models, inequality constraints define the feasible region of decision variables. These constraints ensure that the solutions obtained from the optimization process are meaningful and feasible in practical applications. For microgrid dispatching problems, inequality constraints may include generator output power limits, energy storage device charging and discharging limits, and line transmission capacity limits, etc.

[0073] In some examples, the objective parameters can be considered as initial values ​​that strictly satisfy inequality constraints. Methods for determining initial values ​​include: Heuristic methods: Selecting a reasonable set of initial values ​​based on historical data, expert experience, or system characteristics. These values ​​should strictly satisfy all inequality constraints. Relaxation methods: If directly finding initial values ​​that satisfy all constraints is difficult, some constraints can be ignored first (usually relaxing the less stringent constraints), and after finding a set of solutions, they can be gradually adjusted to satisfy all constraints. Numerical methods: Using numerical optimization algorithms (such as interior point methods, projected gradient methods, etc.) to search for feasible initial points on the constraint boundaries.

[0074] Understandably, a good initial value can accelerate the convergence process and reduce computation time. An unreasonable initial value may cause the algorithm to converge to a local optimum, or even fail to converge.

[0075] In some examples, as shown in Figure 7, the specific process of determining the target parameter that satisfies the inequality constraint in step S610 may include, but is not limited to, steps S710 to S730.

[0076] Step S710: Determine the objective function used to obtain the initial values ​​of the primal-dual interior point method based on the objective convex optimization model.

[0077] Step S720: For any given first parameter, obtain the corresponding second parameter according to the objective function.

[0078] Step S730: Obtain the target parameters based on the first and second parameters.

[0079] Understandably, the primal-dual interior-point method is an optimization algorithm that simultaneously improves both the primal and dual variables. It combines Newton's method, the logarithmic barrier function method, and the Lagrangian function, finding the common optimal solution for both the primal and dual problems by iteratively approximating the central path (a series of intermediate solutions guided by the barrier term). In the primal-dual interior-point method, the barrier function ensures that inequality constraints are satisfied, while the Lagrangian function transforms the primal problem into an unconstrained optimization problem. Through iterative computation, the algorithm gradually adjusts the values ​​of the decision and dual variables until it finds the optimal solution that satisfies all constraints.

[0080] In some examples, the objective convex optimization model obtained by transforming the bilabial bar optimization model typically includes an objective function (e.g., minimizing generation costs or maximizing energy efficiency) and a series of inequalities and equality constraints (e.g., power supply and demand balance, equipment capacity limitations, etc.). For the primal-dual interior-point method, a suitable objective function can be chosen, and its initial value (or initial point) can be determined. This initial value will serve as the starting point for the iterative process of the primal-dual interior-point method.

[0081] In some examples, in step S720, for each given first parameter (such as a decision variable value), the corresponding Lagrange multiplier (or other form of dual variable, referred to here as the second parameter) can be calculated based on the objective function and constraints. Here, the Lagrange multipliers are variables associated with the constraints, reflecting their impact on the objective function. In the primal-dual interior-point method, a barrier function is typically introduced to ensure that inequality constraints are satisfied, while the Lagrange multipliers are used to adjust the weights of the barrier function, thereby guiding the algorithm to gradually approach the optimal solution.

[0082] In some examples, step S730 updates the target parameters using the previously calculated first and second parameters. In the primal-dual interior-point method, this typically means performing a series of iterative calculations, updating the values ​​of the decision variable and dual variable based on the current values ​​in each iteration. The iteration process continues until a certain convergence condition is met. This convergence condition might include the change in the objective function value being less than a certain threshold, the changes in the decision variable and dual variable being less than a certain threshold, or the number of iterations reaching a preset upper limit. When the convergence condition is met, the target parameters satisfying the inequality constraints can be considered to have been found.

[0083] CSI report submitted

[0084] CSI report submitted

[0085] In some examples, after determining the target parameters that satisfy the inequality constraints, the target convex optimization model can be solved based on these parameters to obtain the target scheduling strategy for the microgrid. In this process, the target parameters and the target convex optimization model can be used as input parameters for the primal-dual interior-point method. By executing the primal-dual interior-point method, the optimal solution satisfying all constraints can be found, and these solutions can constitute the target scheduling strategy for the microgrid.

[0086] In some examples, after obtaining the target scheduling strategy, the microgrid can be scheduled according to that strategy. This includes adjusting generator output, energy storage device charging and discharging schedules, and load allocation to ensure the microgrid operates according to predetermined goals. By implementing effective scheduling strategies, the energy efficiency of the microgrid can be improved, operating costs reduced, and its stable operation ensured.

[0087] The microgrid scheduling method provided in this embodiment will be illustrated below with a specific example.

[0088] Referring to Figure 8, which is a diagram illustrating a microgrid scheduling method provided in a specific example of this application, after obtaining historical data of random variables, microgrid structure information, and equipment information in the microgrid scheduling problem, a power flow model can be established based on this data and information, and then linearly approximated. Next, based on this linearly approximated power flow model, a subbulb optimization model is constructed and transformed into a convex optimization problem (i.e., the objective convex optimization model). Then, the applicable primal-dual interior-point method is used to solve this transformed convex optimization problem. Finally, the objective scheduling strategy can be obtained, the results can be derived, and applied.

[0089] Specifically, the flowchart of this example is shown in Figure 9.

[0090] Step 1: Based on the microgrid's grid data, determine the decision variables and random variables used to construct the distributed bar optimization model. This step typically involves an in-depth analysis of the microgrid topology to identify the equipment and related variables at each node. Specifically, decision variables mainly cover variables related to equipment that needs to be scheduled, such as generator output and the charging / discharging state of energy storage systems. These variables need to be controlled during the optimization process to achieve the predetermined energy management objectives. Meanwhile, random variables are used to characterize unpredictable or difficult-to-predict factors, such as the prediction error of the output power of renewable energy sources (e.g., solar, wind power).

[0091] Assuming the bus number connecting the microgrid to the upper-level grid is 0, the voltage of this node within the microgrid will be considered constant, and its phase will always remain 0 within the microgrid. Furthermore, this node is equipped by default with a special generator model whose generation cost is equal to the electricity price, and whose generation capacity is unlimited. The practical significance of this generator model lies in its representation of the energy transmission channel between the microgrid and the upper-level grid, allowing the microgrid to obtain or transmit electrical energy from the upper-level grid when needed. This configuration helps to treat the upper-level grid as a reliable and cost-effective energy source or storage point when optimizing the microgrid's energy management strategy, thereby enabling more flexible adjustments to energy allocation and scheduling within the microgrid.

[0092] In this example, the decision variables encompass the generator output power and the charging and discharging power of the energy storage system over the next 24 hours. However, it should be noted that the electricity purchased by bus node 0 connected to the upstream grid is not included in these decision variables. The scheduling period (e.g., 12 hours, 24 hours) and decision interval (e.g., 1 hour, 0.5 hours, 0.25 hours) must be determined based on external input. For ease of subsequent processing, the decision variable x can be constructed as a column vector, in which the variables are arranged in order of time and number.

[0093] In equation (1), P G,i,t With Q G,i,t The active and reactive power output of the generator at time t of the i-th node; P ES+,i,t P ES-,i,t With S OC,i,t Let x represent the energy storage charging power, discharging power, and state of charge of the i-th node at time t, respectively; x is the value divided by P. G,0,t Q G,0,t The column vector consisting of the above variables, excluding the decision variables, is represented by the subscripts. This represents the largest index in the set of nodes containing generators. This represents the largest number in the set of nodes containing energy storage; This represents the maximum time period contained in a scheduling cycle.

[0094] Furthermore, the photovoltaic output power prediction value is a given constant, while the photovoltaic output power prediction error can be regarded as a random variable ξ. For ease of subsequent processing, the random variable ξ can be constructed as a column vector, in which the variables are arranged in chronological and numerical order.

[0095] In equation (2), P PV,i,t Let be the photovoltaic output active power at node i at time t. For its predicted value, Its prediction error; ξ is for all The column vector is composed of subscripts. This represents the largest number in the set of nodes containing photovoltaic elements.

[0096] Step 2: An empirical distribution based on the Dirac distribution can be constructed based on historical data of photovoltaic prediction errors in the microgrid scheduling problem, and a fuzzy set based on the Wasserstein distance can be constructed accordingly.

[0097] First, construct an empirical probability distribution based on existing historical data. In an empirical probability distribution, all M historical data samples are possible values ​​with equal probability (1 / M). It's worth noting that the more samples there are, the closer the empirical probability distribution closely approximates the actual probability distribution of the prediction error.

[0098] In equation (3), for The sample corresponds to the Dirac distribution; M is the sample size. This is the constructed empirical distribution.

[0099] It should be noted that the Wasserstein distance can be used to represent the magnitude of the difference between two probability distributions, and the constructed fuzzy set is based on the empirical probability distribution. Let be a set of probability distributions centered at δ, whose Wasserstein distances from the empirical probability distribution are all less than a given δ. Generally, the more samples there are, the more likely the actual probability distributions are to lie within the fuzzy set.

[0100] In equation (4), EMD is the Wasserstein distance; P M This is the constructed uncertain set (i.e., fuzzy set).

[0101] Step 3: Construct linear approximation constraints using the FPL and FOT methods. Specifically, generator constraints do not require linear approximation, power flow constraints can use the FPL linear approximation, while voltage magnitude constraints require both the FPL and FOT linear approximations. The resulting constraints can be stored as matrices, which will be used in subsequent steps to construct the distributed bar optimization model.

[0102] Specifically, for generator constraints, there is usually an upper and lower limit to the output power at each time period.

[0103] P Gmin,i,t ≤P G,i,t ≤P Gmax,i,t i≠0

[0104] Q Gmin,i,t ≤Q G,i,t ≤Q Gmax,i,t i≠0 (5) ,

[0105] In equation (5), P Gmin,i,t With P Gmax,i,t P G,i,t At the corresponding upper and lower limits, Q Gmin,i,t With Q Gmax,i,t Similarly.

[0106] In addition, there are also limitations on the energy storage charging and discharging power and state of charge constraints.

[0107] In equation (6), P ES+,i,t P represents the energy storage charging power at node i at time t. ES+max,i,t With P ES+min,i,t These are its upper and lower limits, respectively; P ES-,i,t P represents the energy storage and discharge power at node i at time t. ES-max,i,t With P ES-min,i,t These are its upper and lower limits, respectively; η ES+,i For charging efficiency, η ES-,i For discharge efficiency, Δt is a given decision interval, E ES,i S represents the rated energy storage capacity at node i; OC,i,t This represents the energy storage charge state of node i at time t. With S OC,i,t These are its upper and lower limits, respectively.

[0108] It should be noted that equations (5) and (6) are both linear constraints.

[0109] For power flow constraints and their linear approximations, the physical meaning of power flow constraints lies in establishing the relationship between node injected power and node voltage. Since microgrid transmission lines are mostly short-distance, line-to-ground admittance is usually not considered. Based on the microgrid topology and branch admittance data, the relationship between node voltage and branch current can be established.

[0110] In equation (7), I i,j,t Let S be the current from node i to node j at time t. i,j,t Let G be the complex power from node i to node j at time t. i,j +jB i,j V is the branch admittance from node i to node j. i,t Let be the voltage at node i at time t.

[0111] Furthermore, by transforming it into the relationship between node injected power and node voltage, we can obtain...

[0112] In equation (8), I i,t Inject current S at node i at time t. i,t Inject power into the corresponding node; N i Let i be the set of all nodes connected to node i.

[0113] Furthermore, by transforming equation (8) into matrix form, we can obtain equation (9).

[0114] In equation (9), I bus,tV bus,t With S bus,t Let I be a column vector, consisting of the injected current I at each node at time t. i,t Node voltage V i,t Injection power S i,t Composition. Y bus Let be the admittance matrix, which is the matrix representation of admittance in equation (8). The diagonal elements are the sum of the branch admittances between the nodes connected to the corresponding node in the row, and the remaining elements are the negatives of the branch admittances between the corresponding node in the row and the corresponding node in the column. It should be noted that equation (9) is equivalent to equation (8).

[0115] Furthermore, by dividing the variables in equation (9) into blocks according to the node 0 (referring to the connection point between the microgrid and the upper-level power grid), equation (9) can be equivalently transformed into equation (10).

[0116] In equation (10), Y 00 Y 0L Y L0 and Y LL The introduced submatrix constant is Y. bus Submatrices; S 0,t S represents the injected power of the balancing node at time t. L,t Let S represent the injected power at the non-equilibrium node at time t. bus,t submatrices; V 0,t V represents the node voltage at the equilibrium node at time t. L,t Let V represent the node voltage of the non-equilibrium node at time t. bus,t The submatrix.

[0117] Furthermore, the FPL method is employed, based on the injection power at all given times. and node voltage By linearly approximating the above equation, we can obtain equation (11).

[0118] Transforming equation (11) into a linear constraint in the real number field, we can obtain equation (12).

[0119] In equation (12), V L,real With V L,imag These represent the real and imaginary parts of the column vector, respectively.

[0120] Furthermore, by introducing the relationship between injected power and the power balance between node load and power generation equipment, we can obtain equation (13).

[0121] P i,t =P G,i,t +P PV,i,t +PD,i,t Q i,t =Q G,i,t +Q D,i,t (13),

[0122] In equation (13), P i,t With Q i,t P represents the active and reactive power injected into node i at time t, respectively. D,i,t With Q D,i,t Let represent the active power and reactive power of the load at time t at node i, respectively.

[0123] After substituting equation (13) into the power flow constraint (i.e. equation (12)), the real and imaginary components of the power injected at node 0 and the voltage at other nodes can be linearly represented by decision variables and random variables. Furthermore, the constraint equations (12) and (13) obtained after substitution are both linear constraints.

[0124] Additionally, the magnitude of the node voltage is limited to...

[0125] V i,t,min ≤|V i,t |≤V i,t,max (14),

[0126] Among them, V i,t,min With V i,t,max These represent the lower and upper limits of the voltage at node i, respectively.

[0127] Furthermore, by applying the FOT method to linearly approximate equation (14), we can obtain equation (15).

[0128] Furthermore, by using traditional mathematical methods (such as substitution), by substituting equations (12) and (13) into equation (15), the constraint conditions for linear approximation can be obtained, denoted as equation (16).

[0129] A cons,2-1 x+A cons,2-2 ξ≤a cons,2 (16),

[0130] Among them, A cons,2-1 A cons,2-2 a cons,2 The introduced matrix constants are used to represent the coefficients of the decision variables, the coefficients of the random variables, and the remaining constants in equation (15). Expressing linear constraints in matrix form is a common mathematical method and engineering practice. The elements of each row correspond to the coefficients and constants of an inequality constraint in equation (15). Specifically, after substituting equations (12) and (13) into equation (15) using substitution, the coefficients of all elements of x constitute A.cons,2-1 A cons,2-1 The element in the i-th row and j-th column represents the coefficient of the j-th element of x in the i-th constraint; the coefficients of all elements of ξ constitute A. cons,2-2 A cons,2-2 The element in the i-th row and j-th column represents the coefficient of the j-th element of ξ in the i-th constraint; the remaining constants constitute a. cons,2 a cons,2 The elements in the i-th row represent all constants of the i-th constraint shifted to the right of the less than or equal to sign.

[0131] Step 4: Construct the sub-Bruker optimization model, including the objective function based on mathematical expectation and joint chance constraints. Both can ultimately have their coefficients converted into matrix coefficients, which will be used in subsequent steps to transform the sub-Bruker optimization model.

[0132] Understandably, the significance of the fuzzy objective function lies in minimizing the expected value of the objective function under the most unfavorable probability distribution within the fuzzy set. Taking linear electricity purchase cost and generation cost as an example, its form can be...

[0133] Among them, E P c represents the expected value under probability P. G,i,t Let be the generator power generation cost at node i at time t.

[0134] By substituting equations (12) and (13) into equation (17) using the substitution method, and substituting the power flow constraints and their linear approximations from step three, equation (17) can be transformed into the form shown in equation (18) through a series of traditional mathematical transformation methods. Equation (18) is the objective function of the sub-Bruker optimization model.

[0135] In equation (18), c obj,1-1 c obj,1-2 c obg,1-3 The introduced matrix constant is used to represent the coefficients of the decision variables, the coefficients of the random variables, and the other constants in equation (17). Depending on the specific model, if a coefficient does not exist, the corresponding element is set to zero. Specifically, by substituting equations (12) and (13) into equation (17), the coefficients of all elements in equation (17) constitute c. obj,1-1 The coefficients of all the elements of ξ constitute c. obj,1-2 The summation of all remaining constants yields c. obg,1-3 .

[0136] It is worth noting that all constraints that do not involve random variables remain unchanged. Specifically, these include equations (5) and (6), which can be represented by equation (19). Using matrices is a common mathematical method and engineering convention when representing these linear constraints.

[0137] In equation (19), A cons,1 a cons,1 B cons,1 b cons,1 As introduced matrix constants, each row's elements correspond to a coefficient and constant in one of the inequalities or equality constraints in equations (5) and (6). Specifically, after expressing equations (5) and (6) in matrix form, the coefficients of all x elements constitute A. cons,1 and B cons,1 A cons,1 The element in the i-th row and j-th column represents the coefficient of the j-th element of x in the i-th inequality constraint, B cons,1 The element in the i-th row and j-th column represents the coefficient of the j-th element of x in the i-th equality constraint; the remaining constants constitute a. cons,1 and b cons,1 a cons,1 The elements in the i-th row represent all the constants of the i-th inequality constraint shifted to the right of the less than or equal to sign, b cons,1 The elements in the i-th row represent the values ​​of all constants in the i-th equality constraint moved to the right of the equals sign.

[0138] Furthermore, all constraints involving random variables are treated as chance constraints. The significance of the pluralistic bar constraint lies in ensuring that, under the most unfavorable probability distribution of the fuzzy set, the probability that all chance constraints are true is greater than a preset threshold.

[0139] Where 1-ε is the given probability magnitude.

[0140] Step 5: Transform the bibliometric optimization model into a deterministic, finite-dimensional convex optimization problem.

[0141] Define a j (x) and b j (x) is as follows:

[0142] Where a is the horizontal dimension, b is the element, and C represents matrix A. cons,2-2 number of rows.

[0143] The resulting finite-dimensional convex optimization problem is:

[0144] In equation (22), λ obj τobj,i β cons , λ cons τ cons,i The auxiliary variable introduced here has no physical meaning on its own. It should be noted that C here uses the definition from equation (21), which is A. cons,2-2 The number of rows; here M is defined using equation (3), which is the empirical distribution. The number of samples used; the dual norm operator here. This refers to the dual norm of the norm used in the Wasserstein distance in step two.

[0145] Step Six: Solve the transformed convex optimization problem using the primal-dual interior-point method. This problem includes the objective function and constraints obtained in Step Five.

[0146] The convex optimization problem to be solved (i.e., the convex optimization model) can be denoted as follows:

[0147] Where f0(y) is a convex function, corresponding to the objective function; f i (y) is a convex function, corresponding to the left side of the inequality constraint; A and b are usually used to represent the coefficient matrix and constant vector in the equality constraint, but in practical applications, the equality constraint often does not exist, so A is usually regarded as an empty matrix and b is regarded as an empty vector.

[0148] Subsequently, to obtain initial values ​​that strictly satisfy the inequality constraints, constraints and auxiliary variables are added, and equation (23) can be transformed into the following form:

[0149] Here, s is an auxiliary variable introduced.

[0150] For any given y0, the initial value [y0] that strictly satisfies the inequality constraint can be calculated according to equation (25). T ,s0] T .

[0151] s0=max{f i (y0)}+ω (25),

[0152] Where ω is any positive real number.

[0153] Subsequently, by executing the traditional primal-dual interior-point method algorithm, the convex optimization problem can be solved, thereby obtaining the target scheduling strategy of the microgrid. It is worth noting that in the first iteration of the primal-dual interior-point method, the initial point is x0 = [y0...]. T ,s0] T .

[0154] Step 7: Execute the obtained target scheduling strategy in the actual microgrid.

[0155] In addition, one embodiment of this application discloses a microgrid scheduling device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the microgrid scheduling method as described in any of the preceding embodiments.

[0156] In addition, one embodiment of this application discloses a computer-readable storage medium storing computer-executable instructions for performing the microgrid scheduling method as described in any of the preceding embodiments.

[0157] Furthermore, one embodiment of this application discloses a computer program product, including a computer program or computer instructions, which are stored in a computer-readable storage medium. The processor of the device reads the computer program or computer instructions from the computer-readable storage medium and executes the computer program or computer instructions, causing the device to perform the microgrid scheduling method as described in any of the preceding embodiments.

[0158] In this embodiment, a sub-Bluhl bar optimization model is first determined based on the microgrid's grid data. This model can generate an effective dispatch strategy for the microgrid based on given data. Subsequently, the sub-Bluhl bar optimization model is transformed to obtain a target convex optimization model. By solving the target convex optimization model, a specific target dispatch strategy for the microgrid can be derived. Finally, based on this strategy, corresponding dispatch operations can be performed on the microgrid. In this embodiment, the optimization model constructed using the sub-Bluhl bar optimization method can withstand the uncertainty under worst-case conditions, thus ensuring the stability of the dispatch strategy and the feasibility of the system even when facing prediction errors and various uncertainties. In the process of constructing the sub-Bluhl bar optimization model, historical grid data of the microgrid is fully utilized, thereby obtaining an uncertainty set that is closer to the actual probability distribution. This approach greatly reduces the conservatism of the sub-Bluhl bar optimization method, making the dispatch strategy closer to actual operational needs. In addition, addressing the limitations of traditional linear approximation methods in handling complex grid structures and nonlinear constraints, the sub-Bluhl bar optimization model significantly improves computational efficiency while ensuring computational accuracy. Furthermore, by appropriately transforming and solving the optimization model, the problem of obtaining reliable initial points using traditional methods can be solved, thereby effectively overcoming the limitations of traditional optimization algorithms in solving the microgrid distributed robust scheduling problem. In summary, the embodiments of this application can effectively improve the robustness, economy, and solution efficiency of microgrid scheduling strategies. Those skilled in the art will understand that all or some steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as integrated circuits, such as application-specific integrated circuits. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0159] The above describes several embodiments of this application in detail, but this application is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the scope of this application, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A microgrid dispatching method, comprising: Based on the grid data of the microgrid, a sub-Bluerg optimization model is determined, which is used to obtain the scheduling strategy for the microgrid. The sub-bar optimization model is transformed to obtain the target convex optimization model; Solving the target convex optimization model yields the target scheduling strategy for the microgrid. The microgrid is scheduled according to the target scheduling strategy.

2. The microgrid dispatching method according to claim 1, wherein, The process of transforming the sub-bar optimization model to obtain the target convex optimization model includes: The sub-Blubar optimization model is transformed to obtain a transformed convex optimization model; By adding auxiliary constraints and auxiliary variables to the transformed convex optimization model, the target convex optimization model is obtained.

3. The microgrid dispatching method according to claim 1, wherein, The target convex optimization model includes inequality constraints; solving the target convex optimization model to obtain the target scheduling strategy for the microgrid includes: Determine the target parameters that satisfy the inequality constraints; The target convex optimization model is solved based on the target parameters to obtain the target scheduling strategy for the microgrid.

4. The microgrid dispatching method according to claim 3, wherein, Determining the target parameter that satisfies the inequality constraint includes: Based on the target convex optimization model, determine the objective function used to obtain the initial values ​​of the primal-dual interior point method; For any given first parameter, the corresponding second parameter is obtained according to the objective function; The target parameters are obtained based on the first parameter and the second parameter.

5. The microgrid dispatching method according to claim 4, wherein, The step of solving the target convex optimization model based on the target parameters to obtain the target scheduling strategy for the microgrid includes: The target parameters and the target convex optimization model are used as input parameters of the original-dual interior-point method. The original-dual interior-point method is executed to obtain the target scheduling strategy for the microgrid.

6. The microgrid dispatching method according to claim 1, wherein, The step of determining the sub-bar optimization model based on the microgrid's grid data includes: Based on the microgrid's grid data, determine the decision variables and random variables used to construct the sub-Bruker optimization model; Obtain historical data of random variables of the microgrid, and determine the fuzzy set used to construct the sub-Bruker optimization model based on the random variables and the historical data of the random variables; The sub-Blubar optimization model is determined based on the decision variables and the fuzzy set.

7. The microgrid dispatching method according to claim 6, wherein, The step of determining the fuzzy set used to construct the sub-Bruker optimization model based on the random variable and its historical data includes: Based on the historical data of the random variables, an empirical probability distribution is obtained; Based on the random variables and the empirical probability distribution, a fuzzy set is determined for constructing the sub-Bruker optimization model.

8. The microgrid dispatching method according to claim 7, wherein, The number of random variables is multiple; determining the fuzzy set used to construct the sub-Bruker optimization model based on the random variables and the empirical probability distribution includes: Determine the distance value between each of the random variables and the empirical probability distribution; Among the multiple random variables, the random variables whose distance values ​​are less than or equal to a preset distance threshold are determined as the fuzzy set used to construct the sub-Bruker optimization model.

9. The microgrid dispatching method according to claim 6, wherein, The step of determining the sub-Bruker optimization model based on the decision variables and the fuzzy set includes: Based on the decision variables, determine the linear approximate power flow model and power generation equipment constraints used to construct the sub-Blu-ray bar optimization model; The sub-Blu-rod optimization model is determined based on the fuzzy set, the linear approximate power flow model, and the constraints of the power generation equipment.

10. The microgrid dispatching method according to claim 9, wherein, The linear approximate power flow model is obtained according to the following steps: Based on the decision variables, determine the node voltage constraints and power flow constraints of the microgrid; The linear approximate power flow model is determined based on the node voltage constraints and the power flow constraints.

11. The microgrid dispatching method according to claim 10, wherein, Determining the linear approximate power flow model based on the node voltage constraints and the power flow constraints includes: The power flow constraints are subjected to a first linear approximation to obtain a first linear approximation result; Based on the first linear approximation result, the node voltage constraint condition is subjected to a second linear approximation process to obtain a second linear approximation result; The linear approximation power flow model is determined based on the first linear approximation result and the second linear approximation result.

12. The microgrid dispatching method according to claim 11, wherein, The first linear approximation process includes fixed-point linearization, and the second linear approximation process includes first-order Taylor processing.

13. The microgrid dispatching method according to claim 9, wherein, The step of determining the sub-Blule optimization model based on the fuzzy set, the linear approximate power flow model, and the constraints of the power generation equipment includes: Determine the sub-Blule objective function based on the fuzzy set; Based on the fuzzy set, the linear approximate power flow model, and the constraints of the power generation equipment, determine the sub-Blule bar constraints; The sub-Bruker optimization model is obtained based on the sub-Bruker objective function and the sub-Bruker constraints.

14. A microgrid dispatching device, comprising: At least one processor; At least one memory for storing at least one program; wherein, At least one of the programs is executed by at least one of the processors to perform the microgrid scheduling method according to any one of claims 1 to 13.

15. A computer-readable storage medium storing computer-executable instructions, wherein, The computer-executable instructions are used to execute the microgrid scheduling method according to any one of claims 1 to 13.

16. A computer program product comprising a computer program or computer instructions, wherein, The computer program or the computer instructions are stored in a computer-readable storage medium. The processor of the microgrid dispatching device reads the computer program or the computer instructions from the computer-readable storage medium. The processor executes the computer program or the computer instructions, causing the microgrid dispatching device to perform the microgrid dispatching method according to any one of claims 1 to 13.

Images