Data-driven massive resource regulation and control constraint dimension reduction method and device
By constructing high-dimensional and low-dimensional control constraint models, the problem of massive resource control constraint reduction in virtual power plants is solved, more efficient control optimization is achieved, and the universality and adaptability of control optimization is improved.
Patent Information
- Application Number
- CN202411271855.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2044-09-11
AI Technical Summary
In the prior art, the dimensionality reduction method of massive resource regulation constraints is insufficient in general and adaptability, resulting in high computational complexity and low efficiency in regulation and optimization of virtual power plants.
By building a high-dimensional control constraint model, a regulation simulation data set is generated based on historical operation data of distributed resources, a low-dimensional control constraint model is built, and it is embedded in the upper-level control optimization model to perform control constraints, reducing the computing burden and solution time.
The low-dimensional regulation and constraint model is realized to effectively approximate massive resources, reduce the computational complexity and solution time, improve the versatility and adaptability of regulation and optimization, and enhance the flexibility of virtual power plants.
Smart Images

Figure CN119089174A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of virtual power plant control, and in particular to a data-driven massive resource control constraint dimensionality reduction method and device. Background Art
[0002] Power systems around the world are gradually reducing their reliance on fossil fuels and increasing the installed capacity of wind and photovoltaic power generation resources. In this context, the supply and demand balance of power systems faces unprecedented challenges, because the output of wind and photovoltaic power generation is affected by changeable weather and cannot be flexibly adjusted like traditional thermal power. To meet these challenges, an economical and efficient way is to utilize the flexibility of massive resources on the demand side. However, since the capacity of individual resources is usually small, they usually need to be aggregated through virtual power plants to reach the threshold of participating in the market. The upper power grid directly optimizes and regulates the feasible domain and cost characteristics of the virtual power plant, thereby achieving deep interaction. In order to reduce the computational complexity of the upper-level regulation optimization, the virtual power plant needs to submit its overall regulation constraints to the grid operator instead of directly submitting the original regulation constraints of all resources. This process is the dimensionality reduction of the regulation constraints of massive resources.
[0003] In the related technology, regarding the dimensionality reduction of resource control constraints, the process of obtaining the overall control constraints (i.e., aggregation or dimensionality reduction control constraints) based on the original control constraints of massive resources is essentially not a simple algebraic summation of individual model parameters, but the addition of multiple sets, that is, the Minkowski sum, which currently has no effective general solution method. Therefore, a common method is to use methods such as internal approximation or external approximation to obtain computationally feasible results. For example, scholars at Harvard University in the United States proposed using ellipses to approximate the control constraints of massive resources to speed up subsequent aggregation. However, most methods for dimensionality reduction of control constraints are based on analytical methods, which have some limitations in terms of universality and adaptability. Summary of the invention
[0004] In order to overcome the problems existing in the related art, the present disclosure provides a data-driven massive resource control constraint dimensionality reduction method and device.
[0005] According to the first aspect of the embodiments of the present disclosure, a data-driven method for dimensionality reduction of massive resource control constraints is provided, including: constructing a high-dimensional control constraint model based on the operating parameters of distributed resources; embedding the high-dimensional control constraint model into an upper-level control optimization model, and obtaining a control simulation data set based on the historical operating data of distributed resources participating in the upper-level control; constructing a low-dimensional control constraint model based on the control simulation data set; embedding the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints.
[0006] In some embodiments of the present disclosure, a high-dimensional control constraint model is embedded in an upper-level control optimization model, and a control simulation data set is obtained based on the historical operation data of distributed resources participating in upper-level control, including: generating multiple typical control scenario parameters based on the historical operation data of distributed resources participating in upper-level control; embedding the high-dimensional control constraint model into the upper-level control optimization model, generating optimal control results under different typical control scenarios according to the multiple typical control scenario parameters, and obtaining a control simulation data set.
[0007] In some embodiments of the present disclosure, a low-dimensional control constraint model is constructed based on a control simulation data set, including: taking the optimal control result as the operating result of the low-dimensional control constraint model; fitting the parameters of the low-dimensional control constraint model based on the prediction results obtained by inputting multiple typical control scenario parameters into the low-dimensional control constraint model and the optimal control result as the operating result, and determining the low-dimensional control constraint model.
[0008] In some embodiments of the present disclosure, the high-dimensional control constraint model includes the original control feasible domain P of distributed resources; wherein, Here, P consists of a set of linear constraints on (p, r), the variable with subscript [T] is a T-dimensional vector, and p [T] =(p1,p2,...,p T ), r [T] =(r1,r2,...,r T ), p T The basic energy contribution of the distributed resource in the Tth hour, r T is the frequency modulation capability of the distributed resource in the Tth hour, y [I][T] is the operating variable of the Ith distributed resource, h(·) includes the power and energy constraints of the distributed resource, θ I is the parameter of the Ith distributed resource, I and T are both positive integers; the high-dimensional control constraint model is embedded in the upper-level control optimization model, and the control optimization is expressed as:
[0009]
[0010] st(p [T] ,r [T] )∈P;
[0011] in, is the energy price of distributed resources, is the equivalent frequency regulation capacity price of distributed resources, s perf For performance score, is the frequency regulation capacity price, is the FM mileage price, is the expected frequency modulation mileage at time t, is the cost function of distributed resources;
[0012] For different energy frequency regulation prices By calling the commercial solver to solve the above regulation optimization, the optimal regulation result (p [T] ,r [T] ), and obtain the control simulation data set Where k is the total number of indexes.
[0013] In some embodiments of the present disclosure, the low-dimensional control constraint model includes upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints and initial energy, which can be expressed as:
[0014]
[0015]
[0016] Among them, p t,i is the output in time interval t, r t,i is the frequency modulation capacity of time interval t, e t,i is the energy at the end of t, the underlined parameter indicates the lower limit of the corresponding variable, the overlined parameter indicates the upper limit of the corresponding variable, the superscript dis indicates discharging to the grid, the superscript ch indicates charging from the grid, and e 0,i is the initial energy, η is the charge / discharge efficiency, μ and λ are the Lagrange multipliers of the corresponding constraints;
[0017] Based on the low-dimensional regulation constraint model, the constraints for frequency regulation provided by distributed resource i include non-negative capacity constraints, power capacity constraints and maintenance time requirements, which can be expressed as:
[0018]
[0019] Among them, Δt req The time required to maintain maximum frequency modulation output for distributed resources;
[0020] Among them, the sum of the bids for distributed resources is expressed as:
[0021] The optimal control result in the control simulation data set D is taken as the optimal solution of the low-dimensional control constraint model. The control simulation data set D is used to fit the parameters of the low-dimensional control constraint model. The optimization goal is to minimize the distance to the optimal control result:
[0022]
[0023] Here, the superscript a represents the parameters of the low-dimensional control constraint model.
[0024] In some embodiments of the present disclosure, the above method further includes: taking the optimal control result in the control simulation data set D as the optimal solution of the low-dimensional control constraint model, solving it by an iterative algorithm, randomly introducing only the control simulation data of a typical control scenario each time, designing a learning rate update parameter, until the parameters of the low-dimensional control constraint model converge, and obtaining the low-dimensional control constraint model;
[0025] The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints. The scale of resource control constraints after dimensionality reduction is much smaller than the original scale, that is, I a ≤I; the feasible domain of the control constraint is:
[0026]
[0027] According to the second aspect of the embodiment of the present disclosure, a data-driven massive resource control constraint dimensionality reduction device is provided, including: a first model determination unit, used to construct a high-dimensional control constraint model based on the operating parameters of distributed resources; a data set acquisition unit, used to embed the high-dimensional control constraint model into an upper-level control optimization model, and acquire a control simulation data set based on the historical operating data of the distributed resources participating in the upper-level control; a second model determination unit, used to construct a low-dimensional control constraint model according to the control simulation data set; a processing unit, used to embed the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints.
[0028] In some embodiments of the present disclosure, the data set acquisition unit is specifically used to: generate multiple typical control scenario parameters based on the historical operation data of distributed resources participating in upper-level control; embed the high-dimensional control constraint model into the upper-level control optimization model, and generate the optimal control results under different typical control scenarios according to the multiple typical control scenario parameters to obtain a control simulation data set.
[0029] In some embodiments of the present disclosure, the second model determination unit is specifically used to: use the optimal control result as the operating result of the low-dimensional control constraint model; fit the parameters of the low-dimensional control constraint model based on the prediction results obtained by inputting multiple typical control scenario parameters into the low-dimensional control constraint model and the optimal control result as the operating result, and determine the low-dimensional control constraint model.
[0030] In some embodiments of the present disclosure, the high-dimensional control constraint model includes the original control feasible domain P of distributed resources; wherein,
[0031] Here, P consists of a set of linear constraints on (p, r), the variable with subscript [T] is a T-dimensional vector, and p [T] =(p1,p2,...,p T ), r [T] =(r1,r2,...,r T), p T The basic energy contribution of the distributed resource in the Tth hour, r T is the frequency modulation capability of the distributed resource in the Tth hour, y [I][T] is the operating variable of the Ith distributed resource, h(·) includes the power and energy constraints of the distributed resource, θ I is the parameter of the Ith distributed resource, I and T are both positive integers;
[0032] Among them, the high-dimensional control constraint model is embedded in the upper-level control optimization model, and the control optimization is expressed as:
[0033]
[0034] st(p [T] ,r [T] )∈P;
[0035] in, is the energy price of distributed resources, is the equivalent frequency regulation capacity price of distributed resources, s perf For performance score, is the frequency regulation capacity price, is the FM mileage price, is the expected frequency modulation mileage at time t, is the cost function of distributed resources;
[0036] For different energy frequency regulation prices By calling the commercial solver to solve the above regulation optimization, the optimal regulation result (p [T] ,r [T] ), and obtain the control simulation data set Where k is the total number of indexes.
[0037] In some embodiments of the present disclosure, the low-dimensional control constraint model includes upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints and initial energy, which can be expressed as:
[0038]
[0039] Among them, p t,i is the output in time interval t, r t,i is the frequency modulation capacity of time interval t, e t,i is the energy at the end of t, the underlined parameter indicates the lower limit of the corresponding variable, the overlined parameter indicates the upper limit of the corresponding variable, the superscript dis indicates discharging to the grid, the superscript ch indicates charging from the grid, and e 0,i is the initial energy, η is the charge / discharge efficiency, μ and λ are the Lagrange multipliers of the corresponding constraints;
[0040] Based on the low-dimensional regulation constraint model, the constraints for frequency regulation provided by distributed resource i include non-negative capacity constraints, power capacity constraints and maintenance time requirements, which can be expressed as:
[0041]
[0042] Among them, Δt req The time required to maintain maximum frequency modulation output for distributed resources;
[0043] Among them, the sum of the bids for distributed resources is expressed as:
[0044] The optimal control result in the control simulation data set D is taken as the optimal solution of the low-dimensional control constraint model. The control simulation data set D is used to fit the parameters of the low-dimensional control constraint model. The optimization goal is to minimize the distance to the optimal control result:
[0045]
[0046] Here, the superscript a represents the parameters of the low-dimensional control constraint model.
[0047] In some embodiments of the present disclosure, the optimal control result in the control simulation data set D is used as the optimal solution of the low-dimensional control constraint model, and an iterative algorithm is used to solve it. Each time, only one control simulation data of a typical control scenario is randomly introduced, and the learning rate update parameters are designed until the parameters of the low-dimensional control constraint model converge, thereby obtaining a low-dimensional control constraint model.
[0048] The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints. The scale of resource control constraints after dimensionality reduction is much smaller than the original scale, that is, I a ≤I; the feasible domain of the control constraint is:
[0049]
[0050] According to a third aspect of an embodiment of the present disclosure, an electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described in any one of the first aspects is implemented.
[0051] According to a fourth aspect of an embodiment of the present disclosure, there is provided a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method according to any one of the first aspects is implemented.
[0052] According to a fifth aspect of an embodiment of the present disclosure, a computer program product is provided, comprising a computer program, wherein the computer program implements the method as described in any one of the first aspects when executed by a processor.
[0053] The technical solution provided by the embodiments of the present disclosure may include the following beneficial effects: constructing a high-dimensional control constraint model based on the operating parameters of distributed resources; embedding the high-dimensional control constraint model into the upper-level control optimization model, obtaining a control simulation data set based on the historical operating data of the distributed resources participating in the upper-level control; constructing a low-dimensional control constraint model according to the control simulation data set; embedding the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints. Thus, the high-dimensional constraints of the original massive distributed resources can be approximated by the low-dimensional control constraint model, greatly reducing the computational burden and solution time of the coordinated control of massive resources, and adopting a data-driven technical route, which has better versatility and adaptability than the traditional constraint dimensionality reduction method, and helps to better develop the flexibility of massive resources.
[0054] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.
[0056] Figure 1 It is a flow chart of a data-driven massive resource control constraint dimensionality reduction method according to an exemplary embodiment;
[0057] Figure 2 is a flow chart of another data-driven massive resource control constraint dimensionality reduction method according to an exemplary embodiment;
[0058] Figure 3 is a structural diagram of a data-driven massive resource control constraint dimensionality reduction device according to an exemplary embodiment;
[0059] Figure 4 It is a structural diagram of a device for executing a data-driven massive resource regulation and constraint dimensionality reduction method according to an exemplary embodiment. DETAILED DESCRIPTION
[0060] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.
[0061] The terms used in the disclosed embodiments are only for the purpose of describing specific embodiments, and are not intended to limit the disclosed embodiments. The singular forms "a", "an" and "the" used in the disclosed embodiments and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings.
[0062] It should be understood that although the terms first, second, third, etc. may be used to describe various information in the disclosed embodiments, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the disclosed embodiments, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at" or "when" or "in response to determination".
[0063] In addition, various forms of processes shown in the embodiments of the present disclosure may be used to reorder, add or delete steps. For example, the steps described in the present application may be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present disclosure can be achieved, and this document does not limit this.
[0064] In the related technologies, most of the methods for regulating constrained dimensionality reduction are based on analytical methods, which have some limitations in universality and adaptability.
[0065] In some embodiments, the control constraint aggregation method usually relies on certain assumptions about the operation of each distributed resource. For example, when scholars from Tsinghua University approximated the control constraints of heterogeneous resources, they introduced the assumption that the control constraints of one resource are independent of other resources. However, in the case of virtual power plants aggregating different types of resources participating in the energy and regulation service market, the energy-regulation control constraints of these resources may be coupled, so most control constraint aggregation methods cannot be directly applied. Secondly, the typical aggregated control constraint dimensionality reduction method is associated with the form of each control constraint. For example, scholars from Harvard University in the United States use box constraints to approximate the original control constraints, but these control constraint forms may be inconsistent with the control constraint forms specified by the upper power grid. Therefore, the adaptability of the parsed aggregated control constraint dimensionality reduction method appears to be insufficient.
[0066] Among them, some studies have attempted to use data-driven methods to achieve dimensionality reduction of regulatory constraints. Scholars from Virginia Tech in the United States used the original resource model to label randomly generated operation data as feasible or infeasible, and then used this data to train a convex quadratic classifier to approximate the aggregated regulatory constraints. However, considering the temporal coupling characteristics of resource operation, the amount of training data required will increase exponentially with the increase in time periods, and the trained classifier needs to be converted into the required bidding model form, which introduces additional errors. Therefore, the existing data-driven methods are far from being practical.
[0067] In summary, it is necessary to use data-driven methods to solve the universality and adaptability problems of the analytical control constraint aggregation method, and at the same time, it is necessary to select appropriate approximate models and training methods to implement them. The disclosed embodiment proposes a massive resource energy-frequency aggregation control constraint dimensionality reduction method based on data-driven inverse optimization, using the original massive resource operation model to generate optimal scheduling results under different scenarios, and these scheduling results are regarded as the optimal solutions of the low-dimensional control constraint model after dimensionality reduction, and then used as the input of the inverse optimization problem to fit the parameters of the approximate low-dimensional control constraint model.
[0068] 1) Distributed resource control constraint modeling technology: The output of distributed photovoltaic and wind power is limited by the maximum available power generation. Due to thermal inertia, the temperature of the temperature-controlled load can be analogized to the charging state in energy storage, so as to establish a battery model for the temperature-controlled load. The modeling of the industrial production process is based on the linearized state task network model, including the power limit, material buffer limit, initial value and production target of the industrial production process.
[0069] 2) Constrained dimensionality reduction technique: Constrained dimensionality reduction technique is a method to reduce computational complexity when processing high-dimensional data. Traditional dimensionality reduction techniques such as principal component analysis and linear discriminant analysis focus on the global structure of the data but may ignore the local characteristics of the data. Recent studies such as local linear embedding and isometric mapping better capture the intrinsic geometric structure of the data by preserving the local structure of the data.
[0070] 3) Virtual power plant control optimization technology: Control optimization technology focuses on how to effectively dispatch distributed energy resources to meet the needs of the power grid. Traditional optimization methods include mixed integer linear programming and mixed integer nonlinear programming, which may encounter low computational efficiency when dealing with large-scale problems. With the development of artificial intelligence and machine learning technologies, some new optimization methods, such as reinforcement learning and deep learning, have begun to be applied to virtual power plant control optimization.
[0071] The purpose of the disclosed embodiment is to solve the problem of reducing the dimension of high-dimensional control constraints of massive distributed resources when the virtual power plant needs to submit its overall control constraints to the upper-level control agency. The original high-dimensional control constraints of massive distributed resources are embedded in the upper-level control optimization model problem to generate the optimal control results under different control scenarios to form a control simulation data set; the optimal control result in the control simulation data set is regarded as the running result of the approximate low-dimensional control constraint model after dimensionality reduction, so as to fit the parameters of the approximate low-dimensional control constraint model; the low-dimensional control constraint model obtained by fitting based on the simulation data is embedded in the upper-level control optimization model to achieve control constraint dimensionality reduction. The high-dimensional constraints of the original massive distributed resources can be approximated by the low-dimensional control constraint model, which greatly reduces the computational burden and solution time of the coordinated control of massive distributed resources, and adopts a data-driven technical route, which has better versatility and adaptability than the traditional constraint dimensionality reduction method, and helps to better develop the flexibility of massive resources.
[0072] Figure 1 is a flow chart of a data-driven massive resource control constraint dimensionality reduction method according to an exemplary embodiment. Figure 1 As shown, it should be noted that the data-driven massive resource control constraint dimensionality reduction method of the embodiment of the present disclosure is applied to the data-driven massive resource control constraint dimensionality reduction device. Figure 1 As shown, the method may include the following steps:
[0073] S1: Construct a high-dimensional control constraint model based on the operating parameters of distributed resources.
[0074] In the disclosed embodiment, original operating parameters of each distributed resource are obtained, including the output upper limit of distributed power generation resources, the power and energy capacity of energy storage resources, the adjustment range and energy demand of adjustable loads, etc.
[0075] Among them, based on the operating parameters of each distributed resource, a high-dimensional control constraint model (or refined control constraint model) of massive distributed resources is established, including power constraints, energy constraints, energy demand constraints, capacity constraints involved in regulation, etc.
[0076] S2: Embed the high-dimensional control constraint model into the upper-level control optimization model, and obtain the control simulation data set based on the historical operation data of distributed resources participating in the upper-level control.
[0077] In some embodiments, S2: embedding the high-dimensional control constraint model into the upper-level control optimization model, and obtaining a control simulation data set based on the historical operation data of distributed resources participating in the upper-level control, including: generating multiple typical control scenario parameters based on the historical operation data of distributed resources participating in the upper-level control; embedding the high-dimensional control constraint model into the upper-level control optimization model, generating the optimal control results under different typical control scenarios according to the multiple typical control scenario parameters, and obtaining a control simulation data set.
[0078] In the disclosed embodiment, multiple typical control scenario parameters are generated based on the historical operation data of massive distributed resources participating in upper-level control. The original high-dimensional control constraints of massive resources are embedded into the upper-level control optimization model problem to generate optimal control results under different typical control scenarios, including optimal energy consumption arrangement and frequency regulation capacity, to form a control simulation data set.
[0079] Among them, based on the historical operation data of massive resources participating in upper-level regulation, the parameters of multiple typical regulation scenarios are generated; in theory, to obtain the dispatch results after submitting the original feasible domain, it is necessary to model all units and loads within the scope of the grid operator. However, in practice, the capacity of the resource cluster is very small, so it can be modeled as a price taker. Therefore, under the assumption of perfect competition, the dispatch scenario can be represented by the energy frequency regulation price scenario.
[0080] In the disclosed embodiment, the original high-dimensional control constraints of massive resources are embedded into the upper-level control optimization model problem to generate the optimal control results under typical control scenarios, including the optimal energy consumption arrangement and frequency regulation capacity, forming a control simulation data set.
[0081] Formally, the original control feasible domain of massive resources, that is, the decision space where the control variables are constrained, in some embodiments of the present disclosure, the high-dimensional control constraint model includes the original control feasible domain P of distributed resources.
[0082] in,
[0083] Here, P consists of a set of linear constraints on (p, r), the variable with subscript [T] is a T-dimensional vector, and p [T] =(p1,p2,...,p T ), r [T] =(r1,r2,...,r T ), p T The basic energy contribution of the distributed resource in the Tth hour, r T is the frequency modulation capability of the distributed resource in the Tth hour, y [I][T] is the operating variable of the Ith distributed resource, h(·) includes the power and energy constraints of the distributed resource, θ Iis the parameter of the Ith distributed resource, and both I and T are positive integers.
[0084] Among them, the high-dimensional control constraint model is embedded in the upper-level control optimization model, and the control optimization is expressed as:
[0085]
[0086] st(p [T] ,r [T] )∈P;
[0087] in, is the energy price of distributed resources, is the equivalent frequency regulation capacity price of distributed resources, s perf For performance score, is the frequency modulation capacity price, is the FM mileage price, is the expected frequency modulation mileage at time t, is the cost function of distributed resources;
[0088] For different energy frequency regulation prices By calling the commercial solver to solve the above regulation optimization, the optimal regulation result (p [T] ,r [T] ), and obtain the control simulation data set Where k is the total number of indexes.
[0089] S3: Construct a low-dimensional regulatory constraint model based on the regulatory simulation dataset.
[0090] In some embodiments, S3: constructing a low-dimensional control constraint model based on a control simulation data set, including: taking the optimal control result as the operating result of the low-dimensional control constraint model; fitting the parameters of the low-dimensional control constraint model based on the prediction results obtained by inputting parameters of multiple typical control scenarios into the low-dimensional control constraint model and the optimal control result as the operating result, and determining the low-dimensional control constraint model.
[0091] In the disclosed embodiment, a low-dimensional control constraint model is designed to construct an inverse optimization problem of inverting model parameters based on the control results; as a default approximate low-dimensional control constraint model, a low-order virtual energy storage model can be used to approximate the original control constraints, including upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints, etc.
[0092] Among them, a low-dimensional control constraint model is designed as the default approximate control constraint model. A low-order virtual energy storage model can be used to approximate the original control constraints, including the upper and lower limits of energy storage charging and discharging power, the upper and lower limits of energy storage energy, energy storage energy conversion constraints, initial energy, etc.
[0093] Based on the virtual energy storage model, the constraints for frequency regulation provided by resource i include non-negative capacity, power capacity limit and maintenance time requirement, where Δt req Indicates the time required for the resource to maintain its maximum frequency output, for example, 15 minutes.
[0094] S4: Embed the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints.
[0095] In some embodiments of the present disclosure, the low-dimensional control constraint model includes upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints and initial energy, which can be expressed as:
[0096]
[0097]
[0098] Among them, p t,i is the output in time interval t, r t,i is the frequency modulation capacity of time interval t, e t,i is the energy at the end of t, the underlined parameter indicates the lower limit of the corresponding variable, the overlined parameter indicates the upper limit of the corresponding variable, the superscript dis indicates discharging to the grid, the superscript ch indicates charging from the grid, and e 0,i is the initial energy, η is the charge / discharge efficiency, μ and λ are the Lagrange multipliers of the corresponding constraints;
[0099] Based on the low-dimensional regulation constraint model, the constraints for frequency regulation provided by distributed resource i include non-negative capacity constraints, power capacity constraints and maintenance time requirements, which can be expressed as:
[0100]
[0101] Among them, Δt req The time required to maintain maximum frequency modulation output for distributed resources;
[0102] Among them, the sum of the bids for distributed resources is expressed as:
[0103] The optimal control result in the control simulation data set D is taken as the optimal solution of the low-dimensional control constraint model. The control simulation data set D is used to fit the parameters of the low-dimensional control constraint model. The optimization goal is to minimize the distance to the optimal control result:
[0104]
[0105] Here, the superscript a represents the parameters of the low-dimensional control constraint model.
[0106] Among them, an inverse optimization problem is constructed to infer the parameters of the low-dimensional control constraint model based on the control results; the idea is to regard the scheduling results generated in the data set D as the optimal solution of the optimization problem formulated by the approximate model (adding some noise), and fit the problem parameters by using D. This can be expressed as a data-driven inverse optimization problem, with constraints including strong dual conditions, feasibility of the original problem, feasibility of the dual problem, etc., and the optimization goal is to minimize the distance to the optimal control result.
[0107] The optimal control result in the control simulation data set is regarded as the operating result of the approximate low-dimensional control constraint model after dimensionality reduction (also called the approximate control constraint model), so as to fit the parameters of the approximate low-dimensional control constraint model; the low-dimensional control constraint model fitted based on the simulation data is embedded in the upper-level control optimization model to achieve dimensionality reduction of the control constraints.
[0108] In the disclosed embodiment, the optimal control result in the control simulation data set is regarded as the operating result of the approximate low-dimensional control constraint model after dimensionality reduction, so as to fit the approximate low-dimensional control constraint model parameters; wherein, the optimization problem is large in scale and there are nonlinear terms in the strong duality conditions, which are difficult to solve. Therefore, an iterative algorithm is adopted to solve it, and only the control simulation data of a typical scene is randomly introduced each time, and the learning rate is designed to update the parameters until the model parameters converge. The obtained low-dimensional control constraint model can be used to approximate the original high-dimensional control constraint model.
[0109] In some embodiments of the present disclosure, the above method further includes: taking the optimal control result in the control simulation data set D as the optimal solution of the low-dimensional control constraint model, solving it by an iterative algorithm, randomly introducing only the control simulation data of a typical control scenario each time, designing a learning rate update parameter, until the parameters of the low-dimensional control constraint model converge, and obtaining the low-dimensional control constraint model;
[0110] The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints. The scale of resource control constraints after dimensionality reduction is much smaller than the original scale, that is, I a ≤I; the feasible domain of the control constraint is:
[0111]
[0112] For example, Figure 2As shown, in the embodiment of the present disclosure, a data-driven method for dimensionality reduction of massive resource control constraints is provided. First, the original operating parameters of each distributed resource are obtained, and a refined control constraint model is established (see the high-dimensional control constraint model in the above method). Then, based on the high-dimensional control constraints, control simulation results in different typical scenarios are generated. Then, a low-dimensional control constraint model is designed (see the low-dimensional control constraint model in the above method), and an inverse optimization problem of deducing parameters is constructed. The low-dimensional control constraint model parameters are fitted based on the control simulation results. Finally, the obtained low-dimensional control constraints are embedded in the upper-level control optimization model.
[0113] The disclosed embodiment establishes a data-driven massive resource control constraint dimensionality reduction method based on the distributed resource control constraint modeling and massive constraint dimensionality reduction technology in the virtual power plant control technology. Compared with the methods in the related technology, this method uses a data-driven method to solve the universality and adaptability problems of the analytical control constraint aggregation method, and at the same time needs to select a suitable approximate model and training method to implement it. By applying the disclosed embodiment, the high-dimensional constraints of the original massive distributed resources can be approximated with a low-dimensional control constraint model, thereby greatly reducing the computational burden and solution time of the optimization and control of massive distributed resources, and thus has important practical significance and good application prospects.
[0114] The disclosed embodiment constructs a high-dimensional control constraint model based on the operating parameters of distributed resources; embeds the high-dimensional control constraint model into the upper-level control optimization model, obtains a control simulation data set based on the historical operating data of the distributed resources participating in the upper-level control; constructs a low-dimensional control constraint model based on the control simulation data set; embeds the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints. Thus, the high-dimensional constraints of the original massive distributed resources can be approximated by the low-dimensional control constraint model, greatly reducing the computational burden and solution time of the coordinated control of massive resources, and adopts a data-driven technical route, which has better versatility and adaptability than the traditional constraint dimensionality reduction method, and helps to better develop the flexibility of massive resources.
[0115] Figure 3 1 is a structural diagram of a data-driven massive resource control constraint dimensionality reduction device 1 according to an exemplary embodiment. Figure 3 The device 1 includes a first model determination unit 11, a data set acquisition unit 12, a second model determination unit 13 and a processing unit 14.
[0116] Among them, the first model determination unit 11 is used to build a high-dimensional control constraint model based on the operating parameters of the distributed resources.
[0117] The data set acquisition unit 12 is used to embed the high-dimensional control constraint model into the upper-level control optimization model, and acquire the control simulation data set based on the historical operation data of the distributed resources participating in the upper-level control.
[0118] The second model determination unit 13 is used to construct a low-dimensional control constraint model according to the control simulation data set.
[0119] The processing unit 14 is used to embed the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints.
[0120] In some embodiments, the data set acquisition unit 12 is specifically used to: generate multiple typical control scenario parameters based on the historical operation data of distributed resources participating in upper-level control; embed the high-dimensional control constraint model into the upper-level control optimization model, and generate the optimal control results under different typical control scenarios according to the multiple typical control scenario parameters to obtain a control simulation data set.
[0121] In some embodiments, the second model determination unit 13 is specifically used to: use the optimal control result as the operating result of the low-dimensional control constraint model; based on the prediction results obtained by inputting multiple typical control scenario parameters into the low-dimensional control constraint model and the optimal control result as the operating result, fit the parameters of the low-dimensional control constraint model to determine the low-dimensional control constraint model.
[0122] In some embodiments of the present disclosure, the high-dimensional control constraint model includes the original control feasible domain P of distributed resources; wherein,
[0123] Here, P consists of a set of linear constraints on (p, r), the variable with subscript [T] is a T-dimensional vector, and p [T] =(p1,p2,...,p T ), r [T] =(r1,r2,...,r T ), p T The basic energy contribution of the distributed resource in the Tth hour, r T is the frequency modulation capability of the distributed resource in the Tth hour, y [I][T] is the operating variable of the Ith distributed resource, h(·) includes the power and energy constraints of the distributed resource, θ I is the parameter of the Ith distributed resource, I and T are both positive integers;
[0124] Among them, the high-dimensional control constraint model is embedded in the upper-level control optimization model, and the control optimization is expressed as:
[0125]
[0126] st(p [T] ,r[T] )∈P;
[0127] in, is the energy price of distributed resources, is the equivalent frequency regulation capacity price of distributed resources, s perf For performance score, is the frequency modulation capacity price, is the FM mileage price, is the expected frequency modulation mileage at time t, is the cost function of distributed resources;
[0128] For different energy frequency regulation prices By calling the commercial solver to solve the above regulation optimization, the optimal regulation result (p [T] ,r [T] ), and obtain the control simulation data set Where k is the total number of indexes.
[0129] In some embodiments of the present disclosure, the low-dimensional control constraint model includes upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints and initial energy, which can be expressed as:
[0130]
[0131] Among them, p t,i is the output in time interval t, r t,i is the frequency modulation capacity of time interval t, e t,i is the energy at the end of t, the underlined parameter indicates the lower limit of the corresponding variable, the overlined parameter indicates the upper limit of the corresponding variable, the superscript dis indicates discharging to the grid, the superscript ch indicates charging from the grid, and e 0,i is the initial energy, η is the charge / discharge efficiency, μ and λ are the Lagrange multipliers of the corresponding constraints;
[0132] Based on the low-dimensional regulation constraint model, the constraints for frequency regulation provided by distributed resource i include non-negative capacity constraints, power capacity constraints and maintenance time requirements, which can be expressed as:
[0133]
[0134] Among them, Δt req The time required to maintain maximum frequency modulation output for distributed resources;
[0135] Among them, the sum of the bids for distributed resources is expressed as:
[0136] The optimal control result in the control simulation data set D is taken as the optimal solution of the low-dimensional control constraint model. The control simulation data set D is used to fit the parameters of the low-dimensional control constraint model. The optimization goal is to minimize the distance to the optimal control result:
[0137]
[0138] Here, the superscript a represents the parameters of the low-dimensional control constraint model.
[0139] In some embodiments of the present disclosure, the optimal control result in the control simulation data set D is used as the optimal solution of the low-dimensional control constraint model, and an iterative algorithm is used to solve it. Each time, only one control simulation data of a typical control scenario is randomly introduced, and the learning rate update parameters are designed until the parameters of the low-dimensional control constraint model converge, thereby obtaining a low-dimensional control constraint model.
[0140] The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints. The scale of resource control constraints after dimensionality reduction is much smaller than the original scale, that is, I a ≤I; the feasible domain of the control constraint is:
[0141]
[0142] By implementing the disclosed embodiment, the data-driven massive resource control constraint dimensionality reduction device includes a first model determination unit, which is used to construct a high-dimensional control constraint model based on the operating parameters of distributed resources; a data set acquisition unit, which is used to embed the high-dimensional control constraint model into the upper-level control optimization model, and acquire the control simulation data set based on the historical operation data of the distributed resources participating in the upper-level control; a second model determination unit, which is used to construct a low-dimensional control constraint model according to the control simulation data set; and a processing unit, which is used to embed the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints. The high-dimensional constraints of the original massive distributed resources can be approximated by the low-dimensional control constraint model, which greatly reduces the computational burden and solution time of the coordinated control of massive resources, and adopts a data-driven technical route, which has better versatility and adaptability than the traditional constraint dimensionality reduction method, and helps to better develop the flexibility of massive resources.
[0143] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0144] Figure 41 is a block diagram of an apparatus 700 for executing a data-driven massive resource control constraint dimensionality reduction method according to an exemplary embodiment. For example, the apparatus 700 may be an electronic device, such as a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.
[0145] Reference Figure 4 , the device 700 may include one or more of the following components: a processing component 702 , a memory 704 , a power component 706 , a multimedia component 708 , an audio component 710 , an input / output (I / O) interface 712 , a sensor component 714 , and a communication component 716 .
[0146] The processing component 702 generally controls the overall operation of the device 700, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 702 may include one or more processors 720 to execute instructions to complete all or part of the steps of the above-mentioned method. In addition, the processing component 702 may include one or more modules to facilitate the interaction between the processing component 702 and other components. For example, the processing component 702 may include a multimedia module to facilitate the interaction between the multimedia component 708 and the processing component 702.
[0147] The memory 704 is configured to store various types of data to support operations on the device 700. Examples of such data include instructions for any application or method operating on the device 700, contact data, phone book data, messages, pictures, videos, etc. The memory 704 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.
[0148] The power component 706 provides power to the various components of the device 700. The power component 706 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 700.
[0149] The multimedia component 708 includes a screen that provides an output interface between the device 700 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 708 includes a front camera and / or a rear camera. When the device 700 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.
[0150] The audio component 710 is configured to output and / or input audio signals. For example, the audio component 710 includes a microphone (MIC), and when the device 700 is in an operating mode, such as a call mode, a recording mode, and a speech recognition mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 704 or sent via the communication component 716. In some embodiments, the audio component 710 also includes a speaker for outputting audio signals.
[0151] I / O interface 712 provides an interface between processing component 702 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.
[0152] The sensor assembly 714 includes one or more sensors for providing various aspects of status assessment for the device 700. For example, the sensor assembly 714 can detect the open / closed state of the device 700, the relative positioning of components, such as the display and keypad of the device 700, and the sensor assembly 714 can also detect the position change of the device 700 or a component of the device 700, the presence or absence of user contact with the device 700, the orientation or acceleration / deceleration of the device 700, and the temperature change of the device 700. The sensor assembly 714 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 714 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 714 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
[0153] The communication component 716 is configured to facilitate wired or wireless communication between the device 700 and other devices. The device 700 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 716 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 716 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
[0154] In an exemplary embodiment, the apparatus 700 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above methods.
[0155] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 704 including instructions, and the instructions can be executed by the processor 720 of the device 700 to perform the above method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.
[0156] In an exemplary embodiment, a computer program product is also provided, comprising a computer program, which implements the above method when executed by the processor 720 of the device 700.
[0157] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The description and examples are to be considered exemplary only, and the true scope and spirit of the present invention are indicated by the following claims.
[0158] It should be understood that the present invention is not limited to the exact construction that has been described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.
Claims
1. A data-driven massive resource control constraint dimensionality reduction method, characterized in that: include: Construct a high-dimensional control constraint model based on the operating parameters of distributed resources; The high-dimensional control constraint model is embedded in the upper-level control optimization model, and a control simulation data set is obtained based on the historical operation data of the distributed resources participating in the upper-level control; Constructing a low-dimensional regulation constraint model according to the regulation simulation data set; The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints.
2. The method according to claim 1, characterized in that The method of embedding the high-dimensional control constraint model into an upper-level control optimization model and acquiring a control simulation data set based on historical operation data of the distributed resources participating in upper-level control includes: Generate multiple typical control scenario parameters based on the historical operation data of the distributed resources participating in the upper-level control; The high-dimensional control constraint model is embedded into the upper-level control optimization model, and the optimal control results under different typical control scenarios are generated according to the multi-typical control scenario parameters to obtain the control simulation data set.
3. The method according to claim 2, characterized in that The step of constructing a low-dimensional control constraint model according to the control simulation data set includes: Using the optimal control result as the operation result of the low-dimensional control constraint model; Based on the prediction results obtained by inputting the parameters of the multiple typical control scenarios into the low-dimensional control constraint model and the optimal control result as the operation result, the parameters of the low-dimensional control constraint model are fitted to determine the low-dimensional control constraint model.
4. The method according to any one of claims 1 to 3, characterized in that The high-dimensional control constraint model includes the original control feasible domain P of distributed resources; wherein, Here, P consists of a set of linear constraints on (p, r), the variable with subscript [T] is a T-dimensional vector, and p [T] =(p1,p2,...,p T ), r [T] =(r1,r2,...,r T ), p T is the basic energy output of the distributed resource in hour T, r T is the frequency modulation capability of the distributed resource in the Tth hour, y [I][T] is the operating variable of the Ith distributed resource, h(·) includes the power and energy constraints of the distributed resource, θ I is the parameter of the Ith distributed resource, I and T are both positive integers; The high-dimensional control constraint model is embedded in the upper-level control optimization model, and the control optimization is expressed as: s.t.(p [T] ,r [T] )∈P; in, is the energy price of distributed resources, is the equivalent frequency regulation capacity price of distributed resources, s perf For performance score, is the frequency regulation capacity price, is the FM mileage price, is the expected frequency modulation mileage at time t, is the cost function of distributed resources; For different energy frequency regulation prices By calling the commercial solver to solve the above regulation optimization, the optimal regulation result (p [T] ,r [T] ), and obtain the control simulation data set Where k is the total number of indexes.
5. The method according to claim 4, characterized in that The low-dimensional control constraint model includes upper and lower limits of energy storage charging and discharging power, upper and lower limits of energy storage energy, energy storage energy conversion constraints and initial energy, which can be expressed as: Among them, p t,i is the output in time interval t, r t,i is the frequency modulation capacity of time interval t, e t,i is the energy at the end of t, the underlined parameter indicates the lower limit of the corresponding variable, the overlined parameter indicates the upper limit of the corresponding variable, the superscript dis indicates discharging to the grid, the superscript ch indicates charging from the grid, and e 0,i is the initial energy, η is the charge / discharge efficiency, μ and λ are the Lagrange multipliers of the corresponding constraints; Based on the low-dimensional regulation constraint model, the constraints for frequency regulation provided by distributed resource i include non-negative capacity constraints, power capacity constraints and maintenance time requirements, which are expressed as: Among them, Δt req The time required to maintain maximum frequency modulation output for distributed resources; Among them, the sum of the bids for distributed resources is expressed as: The optimal control result in the control simulation data set D is taken as the optimal solution of the low-dimensional control constraint model. The control simulation data set D is used to fit the parameters of the low-dimensional control constraint model. The optimization goal is to minimize the distance to the optimal control result: Here, the superscript a represents the parameters of the low-dimensional control constraint model.
6. The method according to claim 5, characterized in that The method further comprises: The optimal control result in the control simulation data set D is used as the optimal solution of the low-dimensional control constraint model, and an iterative algorithm is used to solve it. Only one control simulation data of a typical control scenario is randomly introduced each time, and the learning rate update parameters are designed until the parameters of the low-dimensional control constraint model converge, thereby obtaining the low-dimensional control constraint model; The low-dimensional control constraint model is embedded in the upper-level control optimization model to perform control constraints. The resource control constraint scale after dimensionality reduction is much smaller than the original scale, i.e., I a ≤I; the feasible domain of the control constraint is:
7. A data-driven massive resource control constraint dimensionality reduction device, characterized in that: include: A first model determination unit, used to construct a high-dimensional control constraint model based on the operating parameters of the distributed resources; A data set acquisition unit, used to embed the high-dimensional control constraint model into an upper-level control optimization model, and acquire a control simulation data set based on the historical operation data of the distributed resources participating in the upper-level control; A second model determination unit, configured to construct a low-dimensional control constraint model according to the control simulation data set; The processing unit is used to embed the low-dimensional control constraint model into the upper-level control optimization model to perform control constraints.
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 when the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.
10. A computer program product, comprising a computer program, characterized in that The computer program implements the method according to any one of claims 1 to 6 when executed by a processor.
Citation Information
Patent Citations
Electricity-gas virtual power plant double-layer optimization scheduling method considering energy storage and demand response
CN115483718A
Distributed resource aggregation regulation and control method, device and equipment for virtual power plant
CN117013597A
Virtual power plant dispatching boundary description method and device based on flexible operation domain
CN117424292A
Virtual power plant real-time peak regulation optimization regulation and control method and system based on average field game
CN117833372A
Virtual power plant optimal regulation and control method and system based on first-order random dominance constraint
CN118353089A
Cited By
Virtual power plant adjustable load aggregation and real-time response regulation and control method and system
CN122246782A
Virtual power plant adjustable load aggregation and real-time response regulation method and system
CN122246782B
