Data-driven feedback optimization control method and device considering distributed resource access

Through the data-driven feedback optimization control method, the cost-oriented renewable energy prediction model is constructed using historical power data, and the power system optimization scheduling model is embedded, which solves the problem that the traditional power system regulation architecture cannot adapt to dynamic changes in real time, and realizes the power resources Optimized distribution and stable system improvement.

CN119093325APending Publication Date: 2024-12-06ELECTRIC POWER RESEARCH INSTITUTE OF STATE GRID JIBEI ELECTRIC POWER CO LTD +2
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411024283.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-29
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The traditional power system regulation architecture is difficult to adapt to the dynamic changes in the power grid state and external environment in real time, resulting in the inability to accurately predict load demand and supply conditions, resulting in poor resource allocation, intermittent renewable energy, resulting in overpower or insufficient power, and increasing the risk of system instability.

Method used

Using the data-driven feedback optimization control method, the initial power system optimization scheduling model is established, historical power data is obtained, cost-oriented renewable energy prediction model is built, and it is embedded in the power system optimization scheduling model to achieve optimal resource allocation.

Benefits of technology

It improves the accuracy of renewable energy prediction, improves the economy and reliability of the system, optimizes the allocation of electricity resources, reduces operating costs, reduces dependence on fossil fuels, and supports environmental sustainability goals.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119093325A_ABST
    Figure CN119093325A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of power system operation, in particular to a data-driven feedback optimization control method and device considering distributed resource access, and the method comprises the steps: building an initial power system optimization scheduling model of a target distributed resource; acquiring historical power data of the target distributed resource, and selecting a plurality of feature types and representative historical scenes from the historical power data; a cost-considered empirical risk minimization problem model is constructed, a plurality of feature types and representative historical scenes are utilized to train the model to obtain a cost-oriented renewable energy prediction model, and the cost-oriented renewable energy prediction model is put into an initial power system optimization scheduling model to obtain a data-driven feedback optimization control model. And the target distributed resources are optimally allocated. Therefore, the problems that an electric power system can not realize optimal allocation of resources possibly, the intermittency of target distributed resources can cause surplus or insufficient electric power, and the risk of instability of the system is increased are solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power system operation, and in particular to a data-driven feedback optimization control method and device considering distributed resource access. Background Art

[0002] In the context of new power systems, the widespread use of renewable energy has led to a transformation of power systems from traditional, centralized energy production to more efficient, distributed and sustainable energy systems. This transformation not only includes the large-scale integration of renewable energy such as wind and solar energy, but also includes coping with the challenges of intermittency and unpredictability associated with these energy sources. In such a development context, traditional control architectures cannot fully utilize the large amount of data from smart grids, including real-time load data, weather forecasts, renewable energy production and consumer behavior patterns, and cannot effectively integrate intermittent renewable energy. At the same time, they cannot achieve the best in maintaining the stability of the power grid and optimizing power distribution to meet real-time needs, resulting in a certain waste of resources.

[0003] In addition, with the global commitment to reducing greenhouse gas emissions and achieving carbon neutrality, the green transformation of the power system has become particularly important. Data-driven feedback optimization control architecture helps optimize electricity production and consumption, reduce dependence on fossil fuels, and support broader environmental sustainability goals. At the same time, the economic benefits of this architecture cannot be ignored. It can reduce operating costs by improving energy efficiency while creating value for power companies and consumers.

[0004] The current power system control architecture is generally based on historical data or static assumptions. This method is difficult to adapt to the dynamic changes of the power grid status and external environment in real time. Due to the lack of real-time feedback, it may not be possible to accurately predict future load demand and supply conditions in the face of highly uncertain and rapidly changing market environments. As a result, the system may not be able to achieve the optimal allocation of resources, and the intermittent nature of renewable energy may lead to excess or insufficient electricity, while increasing the risk of system instability.

[0005] Therefore, in the context of today's power system developing towards higher efficiency, intelligence and sustainability, there is an urgent need for a data-driven power system feedback optimization control architecture to meet contemporary energy challenges and achieve future energy goals. Summary of the invention

[0006] The present invention provides a data-driven feedback optimization control method and device taking into account the access of distributed resources, so as to solve the problems that the power system may not be able to achieve the optimal allocation of resources, the intermittent nature of the target distributed resources may lead to excess or insufficient power, and at the same time increase the risk of system instability.

[0007] The first aspect of the embodiment of the present invention provides a data-driven feedback optimization control method considering the access of distributed resources, comprising the following steps: establishing an initial power system optimization scheduling model for the target distributed resources; obtaining the historical power data of the target distributed resources, and selecting multiple feature types and multiple representative historical scenarios from the historical power data; constructing an empirical risk minimization problem model considering cost, and using the multiple feature types and the multiple representative historical scenarios to train the empirical risk minimization problem model considering cost to obtain a cost-oriented renewable energy prediction model; embedding the cost-oriented renewable energy prediction model into the initial power system optimization scheduling model to obtain a data-driven feedback optimization control model; and using the data-driven feedback optimization control model to optimally allocate the target distributed resources.

[0008] Optionally, the initial power system optimization scheduling model includes an objective function and constraints, wherein the objective function includes minimizing the start-up and shutdown costs and minimizing the operating costs of the target distributed resources, and the constraints include the flow constraints, safe and stable operation constraints, distributed power supply constraints and energy storage constraints of the target distributed resources.

[0009] Optionally, the acquiring historical power data of the target distributed resource and selecting a plurality of feature types and a plurality of representative historical scenarios from the historical power data includes:

[0010] Obtain historical power data of the target distributed resource to construct a linear multivariable function prediction model based on feature vectors of the historical power data; predict future power data of the target distributed resource based on the linear multivariable function prediction model, and calculate an average standard regression coefficient; quantify the importance of each feature vector in the linear multivariable function prediction model by the average standard regression coefficient to select the multiple feature types; obtain actual power data of the target distributed resource, and use Wasserstein distance to compare the difference between the actual power data and the future power data to identify the multiple representative historical scenarios with median Wasserstein distance values.

[0011] Optionally, the empirical risk minimization problem model considering cost is:

[0012]

[0013] in, is the SPO loss, and is the predicted power and actual power under a given scenario s, and They are and The optimal target value under .

[0014] Optionally, the cost-oriented renewable energy prediction model is:

[0015]

[0016]

[0017] Where H is a set of decision variables, x1,...x |H| , y1, ..., y |H| is the decision variable in each scenario, S is the set of all scenarios s, is the SPO loss, and is the predicted power and actual power under a given scenario s, λ is a hyperparameter, ‖H‖1 is the 1-norm of matrix H, c T is the integer variable cost coefficient matrix, x s is an integer variable, d T is the continuous variable cost coefficient matrix, y s is a continuous variable, for The optimal objective value under , A and B are the coefficient matrices of the variables in the inequality constraints containing integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of the variables in the equality constraints and inequality constraints of continuous variables, D and F are constant matrices, and G is the coefficient matrix considering random variables.

[0018] Optionally, the data-driven feedback optimization control model is:

[0019] z(f)=min x,y [c T x+d T y]

[0020] stAx+By≤L

[0021] Cy=D

[0022] Ey≤F

[0023] Gy≤H * f

[0024] Hx≤1

[0025] Among them, z(f) is the objective function value, f is the eigenvalue, c T is the integer variable cost coefficient matrix, x is the integer decision variable, d Tis the coefficient matrix of continuous variables, y is a continuous decision variable, A and B are the coefficient matrices of variables in inequality constraints involving integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of variables in equality constraints and inequality constraints of continuous variables, D and F are constant matrices, G is the coefficient matrix of the linear combination of y subject to the upper bounds of the external parameters f after linear transformation through the coefficient matrix H, H * is the prediction vector considering the cost.

[0026] The second aspect of the present invention provides a data-driven feedback optimization control device that takes into account the access of distributed resources, including: a construction module, which is used to establish an initial power system optimization scheduling model for the target distributed resources; a selection module, which is used to obtain the historical power data of the target distributed resources, and select feature types and representative historical scenarios from the historical power data; a training module, which is used to construct an empirical risk minimization problem model that takes cost into consideration, and use the feature types and the representative historical scenarios to train the empirical risk minimization problem model that takes cost into consideration to obtain a cost-oriented renewable energy prediction model; an embedding module, which is used to embed the cost-oriented renewable energy prediction model into the initial power system optimization scheduling model to obtain a data-driven feedback optimization control model; and an allocation module, which is used to use the data-driven feedback optimization control model to optimally allocate the target distributed resources.

[0027] An embodiment of the third aspect of the present invention provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the data-driven feedback optimization control method considering distributed resource access as described in the above embodiment.

[0028] A fourth aspect of the present invention provides a computer program product, which, when executed by a processor, implements the above-mentioned data-driven feedback optimization control method taking into account distributed resource access.

[0029] A fifth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned data-driven feedback optimization control method considering distributed resource access.

[0030] The data-driven feedback optimization control method and device considering the access of distributed resources proposed in the embodiment of the present invention optimizes the original feedback-free control architecture based on the data-driven feedback optimization control architecture. By effectively utilizing the existing feature data and training the renewable energy prediction model considering the cost, and finally embedding it into the original model to realize the feedback optimization control model, the feedback optimization control architecture can improve the prediction accuracy of renewable energy, thereby improving the economy and reliability of the entire system.

[0031] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0033] Figure 1 A flow chart of a data-driven feedback optimization control method considering distributed resource access provided by an embodiment of the present invention;

[0034] Figure 2 A flow chart of the sub-gradient iterative algorithm provided by an embodiment of the present invention;

[0035] Figure 3 A block diagram of a data-driven feedback optimization control device considering distributed resource access provided by an embodiment of the present invention;

[0036] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0038] The following describes a data-driven feedback optimization control method and device considering distributed resource access according to an embodiment of the present invention with reference to the accompanying drawings.

[0039] Figure 1 A flow chart of a data-driven feedback optimization control method considering distributed resource access provided by an embodiment of the present invention.

[0040] like Figure 1 As shown, the data-driven feedback optimization control method considering distributed resource access includes the following steps:

[0041] In step S101, an initial power system optimization dispatch model of target distributed resources is established.

[0042] In some embodiments, the initial power system optimization scheduling model includes an objective function and constraints, wherein the objective function includes minimizing the start-up and shutdown costs and minimizing the operating costs of the target distributed resources, and the constraints include the flow constraints, safe and stable operation constraints, distributed power supply constraints, and energy storage constraints of the target distributed resources.

[0043] In the actual execution process, the start-stop cost and operation cost of the target distributed resources are minimized as the objective function, and the expression is as follows:

[0044]

[0045] Among them, c T and d T are cost coefficient matrices, x is an integer variable, such as the start / stop plan of the controllable generator set, the charging / discharging state of the energy storage, and the power purchase and sale state between the distribution microgrid, and y is a continuous variable, such as the output of the controllable generator set, the charging / discharging power of the energy storage, and the power purchase and sale power between the distribution microgrid. T x is the start-up and shutdown cost of the controllable unit, d T y represents the system operating cost.

[0046] The construction includes inequality constraints and equality constraints such as power flow constraints, safe and stable operation constraints, distributed power constraints, and energy storage constraints, which can be specifically expressed as follows:

[0047]

[0048] Among them, A and B are currently the coefficient matrices of variables x and y in the inequality constraints containing integers and continuous variables, and L is a constant matrix; C and E are currently the coefficient matrices of variable y in the equality constraints and inequality constraints of continuous variables, and D and F are constant matrices; the other two constraints: one is the renewable energy processing constraint, ξ is the predicted value of renewable energy, and the other is a 0 / 1 variable constraint, which is the energy storage charging and discharging state constraint.

[0049] In step S102, historical power data of the target distributed resource is obtained, and a plurality of feature types and a plurality of representative historical scenarios are selected from the historical power data.

[0050] In some embodiments, historical power data of a target distributed resource is obtained, and multiple feature types and multiple representative historical scenarios are selected from the historical power data, including:

[0051] Obtain historical power data of target distributed resources to construct a linear multivariable function prediction model based on feature vectors of the historical power data;

[0052] Predicting the future power data of the target distributed resources based on the linear multivariable function prediction model and calculating the average standardized regression coefficient;

[0053] The importance of each feature vector in the linear multivariate function prediction model is quantified by the average standardized regression coefficient to select multiple feature types;

[0054] The actual power data of the target distributed resources is obtained, and the difference between the actual power data and the future power data is compared using the Wasserstein distance to identify multiple representative historical scenarios with median Wasserstein distance values.

[0055] During the actual execution process, the most useful feature types and representative historical scenarios for the prediction model will be selected from the historical power data of the target distributed resources, thereby ensuring that the model can learn the most critical information for the future operation of the power system.

[0056] When selecting feature types to train a prediction model, too few feature types may perform poorly during training and actual condition processing (i.e., underfitting), while using too many feature types for training may result in the model having many parameters and becoming too complex, thereby failing to generalize well during actual condition processing (i.e., overfitting). Therefore, embodiments of the present invention utilize feature selection to determine an appropriate set of relevant feature types for a prediction model, thereby deriving an accurate and generalizable prediction model while avoiding potential underfitting and overfitting problems.

[0057] For example, the importance of each characteristic type for renewable energy is quantified by the following steps:

[0058] For renewable energy i at time t, a linear multivariable function based on the feature vectors of historical power data is constructed to predict the output of future renewable energy:

[0059]

[0060] in, is the historical power of renewable energy, is the historical feature type, β t,j,|N| is the regression coefficient, H is the number of historical data, and N is the total number of feature types.

[0061] Furthermore, the average standardized regression coefficient is calculated to quantify the importance of each feature in the prediction model:

[0062]

[0063] in, is the average standardized regression coefficient, σ tj , σ tn are the standard deviations of renewable energy power and characteristics, is the total time, Assemble for renewable energy.

[0064] If a feature type n corresponds to A larger value indicates that the feature is more important for the prediction model. The relevance of each feature type to RES power output is quantified based on a large amount of historical scenario data to ensure that the identified feature types are both reasonable and explainable.

[0065] After determining the feature types most relevant to renewable energy output, NT representative historical scenarios need to be identified from the most recent NH-day historical data for effectively training the forecast model for the upcoming dispatch date.

[0066] To quantify the extremeness of the latest historical scenario, the Wasserstein distance W is used. h′ To compare the predicted power and actual power The difference between them identifies NT representative scenarios with median Wasserstein distance values ​​as training scenarios for future scheduling:

[0067]

[0068]

[0069] in, and The predicted power and actual power The empirical cumulative distribution function of .

[0070] In step S103, an empirical risk minimization problem model considering cost is constructed, and the empirical risk minimization problem model considering cost is trained using multiple feature types and multiple representative historical scenarios to obtain a cost-oriented renewable energy prediction model.

[0071] In the actual implementation process, by determining the appropriate feature types and representative scenarios, a cost-oriented empirical risk minimization problem was further established based on SPO (asmart predict-then-optimize) to train the cost-oriented renewable energy prediction model.

[0072] Among them, the empirical risk minimization problem considering the cost is modeled: First, define the SPO loss To quantify the system operating cost losses due to imperfect renewable energy forecast information:

[0073]

[0074] in, is the SPO loss, and is the predicted power and actual power under a given scenario s, and They are and The optimal target value under .

[0075] based on Combined with the initial power system optimization dispatch model, the following cost-oriented renewable energy forecasting model can be formed:

[0076]

[0077]

[0078] Where H is a set of decision variables, x1,...x |H| , y1, ..., y |H| is the decision variable in each scenario, S is the set of all scenarios s, is the SPO loss, and is the predicted power and actual power under a given scenario s, λ is a hyperparameter, ‖H‖1 is the 1-norm of matrix H, c T is the integer variable cost coefficient matrix, x s is an integer variable, d T is the continuous variable cost coefficient matrix, y s is a continuous variable, for The optimal target value under , A, B are currently the variables x in the inequality constraints containing integers and continuous variables s ,y s The coefficient matrix of L is the constant matrix, C and E are the variables y in the equality and inequality constraints of continuous variables. s The coefficient matrix of , D and F are constant matrices, and G is the coefficient matrix of the considered random variables.

[0079] Furthermore, the empirical risk minimization problem considering the cost is solved: Due to the coupling of H across all training scenarios, the original problem (7) may be difficult to handle. Therefore, a decomposition method based on Lagrangian relaxation is adopted to solve it.

[0080] LR decomposition can be used to dualize certain coupled constraints and decompose the original problem into |S| disjoint subproblems that can be solved in parallel:

[0081]

[0082] Each sub-problem corresponds to a training scenario s. In the end, instead of directly solving formula (7), we can solve formula (8) to process |S| single scenario problems in parallel, thus reducing the computational complexity.

[0083] In order to approach the global optimum of the original problem (7), μ should be determined s′,e The correct value of μ is updated by the subgradient algorithm. s′,e , until the optimality gap is small enough, such as Figure 3 As shown. In formula (9), α c is the scalar step size determined by equation (10), where ω is a constant scalar selected between 0 and 2.

[0084]

[0085]

[0086] Among them, UB * , LB c Find the upper and lower bounds of problem (7) obtained by section problem (8) respectively.

[0087] In step S104, the cost-oriented renewable energy prediction model is embedded in the initial power system optimization dispatch model to obtain a data-driven feedback optimization control model.

[0088] In some embodiments, a feedback optimization control model is constructed by embedding a trained cost-considered renewable energy power prediction model H into formulas (1) and (2) of an initial power system optimization dispatch model to obtain a data-driven feedback optimization control model.

[0089] z(f)=min x,y [c T x+d T y]

[0090] stAx+By≤L

[0091] Cy=D

[0092] Ey≤F

[0093] Gy≤H * f

[0094] Hx≤1

[0095] Among them, z(f) is the objective function value, f is the eigenvalue, c T is the integer variable coefficient matrix, x is the integer decision variable, d Tis the coefficient matrix of continuous variables, y is a continuous decision variable, A and B are the coefficient matrices of variables x and y in the inequality constraints containing integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of variable y in the equality and inequality constraints of continuous variables, D and F are constant matrices, G is the coefficient matrix of the linear combination of y subject to the upper bound of the linear transformation of the external parameter f through the coefficient matrix H, H * is the prediction vector considering the cost.

[0096] In step S105, the target distributed resources are optimally allocated using a data-driven feedback optimization control model.

[0097] According to the data-driven feedback optimization control method considering distributed resource access proposed in the embodiment of the present invention, the original feedback-free control architecture is optimized based on the data-driven feedback optimization control architecture. By effectively utilizing the existing feature data and training the renewable energy prediction model considering the cost, and finally embedding it into the original model to realize the feedback optimization control model, the feedback optimization control architecture can improve the prediction accuracy of renewable energy, thereby improving the economy and reliability of the entire system.

[0098] Next, a data-driven feedback optimization control device considering distributed resource access according to an embodiment of the present invention will be described with reference to the accompanying drawings.

[0099] Figure 3 It is a block diagram of a data-driven feedback optimization control device considering distributed resource access according to an embodiment of the present invention.

[0100] like Figure 3 As shown, the data-driven feedback optimization control device 30 considering distributed resource access includes: a construction module 301, a selection module 302, a training module 303, an embedding module 304 and an allocation module 305.

[0101] Among them, the construction module 301 is used to establish an initial power system optimization scheduling model for the target distributed resources. The selection module 302 is used to obtain the historical power data of the target distributed resources, and select feature types and representative historical scenarios from the historical power data. The training module 303 is used to construct an empirical risk minimization problem model considering cost, and use feature types and representative historical scenarios to train the empirical risk minimization problem model considering cost to obtain a cost-oriented renewable energy prediction model. The embedding module 304 is used to embed the cost-oriented renewable energy prediction model into the initial power system optimization scheduling model to obtain a data-driven feedback optimization control model. The allocation module 305 is used to optimally allocate the target distributed resources using the data-driven feedback optimization control model.

[0102] In some embodiments, the initial power system optimization scheduling model includes an objective function and constraints, wherein the objective function includes minimizing the start-up and shutdown costs and minimizing the operating costs of the target distributed resources, and the constraints include the flow constraints, safe and stable operation constraints, distributed power supply constraints, and energy storage constraints of the target distributed resources.

[0103] In some embodiments, the selection module 302 includes:

[0104] Obtain historical power data of target distributed resources to construct a linear multivariable function prediction model based on feature vectors of the historical power data;

[0105] Predicting the future power data of the target distributed resources based on the linear multivariable function prediction model and calculating the average standardized regression coefficient;

[0106] The importance of each feature vector in the linear multivariate function prediction model is quantified by the average standardized regression coefficient to select multiple feature types;

[0107] The actual power data of the target distributed resources is obtained, and the difference between the actual power data and the future power data is compared using the Wasserstein distance to identify multiple representative historical scenarios with median Wasserstein distance values.

[0108] In some embodiments, the empirical risk minimization problem model considering cost is:

[0109]

[0110] in, is the SPO loss, and is the predicted power and actual power under a given scenario s, and They are and The optimal target value under .

[0111] In some embodiments, the cost-oriented renewable energy forecasting model is:

[0112]

[0113]

[0114] Among them, H is a set of decision variables, x1, ....x |H| , y1, ..., y |H| is the decision variable in each scenario, S is the set of all scenarios s, is the SPO loss, and is the predicted power and actual power under a given scenario s, λ is a hyperparameter, ‖H‖1 is the 1-norm of matrix H, c T is the integer variable cost coefficient matrix, x s is an integer variable, d T is the continuous variable cost coefficient matrix, y s is a continuous variable, for The optimal objective value under , A and B are the coefficient matrices of the variables in the inequality constraints containing integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of the variables in the equality constraints and inequality constraints of continuous variables, D and F are constant matrices, and G is the coefficient matrix considering random variables.

[0115] In some embodiments, the data-driven feedback optimization control model is:

[0116] z(f)=min x,y [c T x+d T y]

[0117] stAx+By≤L

[0118] Cy=D

[0119] Ey≤F

[0120] Gy≤H * f

[0121] Hx≤1

[0122] Among them, z(f) is the objective function value, f is the eigenvalue, c T is the integer variable cost coefficient matrix, x is the integer decision variable, d T is the coefficient matrix of continuous variables, y is a continuous decision variable, A and B are the coefficient matrices of variables in inequality constraints involving integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of variables in equality constraints and inequality constraints of continuous variables, D and F are constant matrices, G is the coefficient matrix of the linear combination of y subject to the upper bounds of the external parameters f after linear transformation through the coefficient matrix H, H * is the prediction vector considering the cost.

[0123] It should be noted that the aforementioned explanation of the embodiment of the data-driven feedback optimization control method considering distributed resource access is also applicable to the data-driven feedback optimization control device considering distributed resource access of this embodiment, and will not be repeated here.

[0124] According to the data-driven feedback optimization control device considering the access of distributed resources proposed in the embodiment of the present invention, the original feedback-free control architecture is optimized based on the data-driven feedback optimization control architecture. By effectively utilizing the existing feature data and training the renewable energy prediction model considering the cost, and finally embedding it into the original model to realize the feedback optimization control model, the feedback optimization control architecture can improve the prediction accuracy of renewable energy, thereby improving the economy and reliability of the entire system.

[0125] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0126] Memory 401 , processor 402 , and a computer program stored in the memory 401 and executable on the processor 402 .

[0127] When the processor 402 executes the program, the data-driven feedback optimization control method considering distributed resource access provided in the above embodiment is implemented.

[0128] Furthermore, the electronic device further comprises:

[0129] The communication interface 403 is used for communication between the memory 401 and the processor 402 .

[0130] The memory 401 is used to store computer programs that can be executed on the processor 402 .

[0131] The memory 401 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0132] If the memory 401, the processor 402 and the communication interface 403 are implemented independently, the communication interface 403, the memory 401 and the processor 402 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0133] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can communicate with each other through an internal interface.

[0134] The processor 402 may be a central processing unit (CPU), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0135] An embodiment of the present invention further provides a computer program product, which, when executed by a processor, implements the above-mentioned data-driven feedback optimization control method considering distributed resource access.

[0136] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned data-driven feedback optimization control method considering distributed resource access.

[0137] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0138] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0139] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0140] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0141] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0142] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0143] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0144] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A data-driven feedback optimization control method considering distributed resource access, characterized in that: The following steps are involved: Establish an initial power system optimal dispatch model for target distributed resources; Acquire historical power data of the target distributed resource, and select multiple feature types and multiple representative historical scenarios from the historical power data; Constructing a cost-considered empirical risk minimization problem model, and training the cost-considered empirical risk minimization problem model using the multiple feature types and the multiple representative historical scenarios to obtain a cost-oriented renewable energy prediction model; Embedding the cost-oriented renewable energy prediction model into the initial power system optimization dispatch model to obtain a data-driven feedback optimization control model; The target distributed resources are optimally allocated using the data-driven feedback optimization control model.

2. The data-driven feedback optimization control method considering distributed resource access according to claim 1 is characterized in that: The initial power system optimization scheduling model includes an objective function and constraints, wherein the objective function includes minimizing the start-up and shutdown costs and minimizing the operating costs of the target distributed resources, and the constraints include the flow constraints, safe and stable operation constraints, distributed power supply constraints and energy storage constraints of the target distributed resources.

3. The data-driven feedback optimization control method considering distributed resource access according to claim 1, characterized in that: The acquiring of historical power data of the target distributed resource and selecting a plurality of feature types and a plurality of representative historical scenarios from the historical power data include: Acquire historical power data of the target distributed resource to construct a linear multivariable function prediction model based on feature vectors of the historical power data; Predicting future power data of the target distributed resource based on the linear multivariable function prediction model, and calculating an average standard regression coefficient; quantifying the importance of each feature vector in the linear multivariable function prediction model by the average standard regression coefficient to select the multiple feature types; Actual power data of the target distributed resource is obtained, and a difference between the actual power data and the future power data is compared using a Wasserstein distance to identify the plurality of representative historical scenarios having a median Wasserstein distance value.

4. The data-driven feedback optimization control method considering distributed resource access according to claim 1, characterized in that: The empirical risk minimization problem model considering cost is: in, is the SPO loss, and is the predicted power and actual power under a given scenario s, and They are and The optimal target value under .

5. The data-driven feedback optimization control method considering distributed resource access according to claim 1, characterized in that: The cost-oriented renewable energy prediction model is: Among them, H is a set of decision variables, x1…x |H| , y1…y |H| is the decision variable in each scenario, S is the set of all scenarios s, is the SPO loss, and is the predicted power and actual power under a given scenario s, λ is a hyperparameter, ‖H‖1 is the 1-norm of matrix H, c T is the integer variable cost coefficient matrix, x s is an integer variable, d T is the continuous variable cost coefficient matrix, y s is a continuous variable, for The optimal objective value under , A and B are the coefficient matrices of the variables in the inequality constraints containing integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of the variables in the equality constraints and inequality constraints of continuous variables, D and F are constant matrices, and G is the coefficient matrix considering random variables.

6. The data-driven feedback optimization control method considering distributed resource access according to claim 1, characterized in that: The data-driven feedback optimization control model is: z(f)=min x,y [c T x+d T y] stAx+By≤L Cy=D Ey≤F Gy≤H * f Hx≤1 Among them, z(f) is the objective function value, f is the eigenvalue, c T is the integer variable cost coefficient matrix, x is the integer decision variable, d T is the coefficient matrix of continuous variables, y is a continuous decision variable, A and B are the coefficient matrices of variables in inequality constraints involving integers and continuous variables, L is a constant matrix, C and E are the coefficient matrices of variables in equality constraints and inequality constraints of continuous variables, D and F are constant matrices, G is the coefficient matrix of the linear combination of y subject to the upper bounds of the external parameters f after linear transformation through the coefficient matrix H, H * is the prediction vector considering the cost.

7. A data-driven feedback optimization control device considering distributed resource access, characterized in that: include: A construction module for establishing an initial power system optimal dispatch model for target distributed resources; A selection module, used to obtain historical power data of the target distributed resource, and select feature types and representative historical scenarios from the historical power data; A training module, used for constructing an empirical risk minimization problem model considering cost, and training the empirical risk minimization problem model considering cost by using the feature type and the representative historical scenario to obtain a cost-oriented renewable energy prediction model; An embedding module, used for embedding the cost-oriented renewable energy prediction model into the initial power system optimization dispatch model to obtain a data-driven feedback optimization control model; An allocation module is used to optimally allocate the target distributed resources using the data-driven feedback optimization control model.

8. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the data-driven feedback optimization control method considering distributed resource access as described in any one of claims 1 to 6.

9. A computer program product, characterized in that When the computer program / instructions are executed by a processor, the data-driven feedback optimization control method considering distributed resource access as described in any one of claims 1-6 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the data-driven feedback optimization control method considering distributed resource access as described in any one of claims 1-6.

Citation Information

Patent Citations

  • Power system closed-loop prediction-decision scheduling method based on data driving

    CN116050653A

  • Feature-driven renewable energy closed-loop prediction optimization method

    CN117951473A

  • Robust scheduling method for environmental economic distribution considering relevance of renewable energy sources

    CN118381117A

  • Data-driven three-stage scheduling method for electricity, heat and gas networks based on wind electricity indeterminacy

    US20200313433A1