An energy storage optimal scheduling uncertainty set construction method fusing estimation and optimization

Abstract

The application discloses a kind of energy storage optimal scheduling uncertainty set construction method of fusion estimation and optimization, comprising the following steps: constructing risk-averse energy storage optimization scheduling model, based on the error vector of price uncertainty Robust Optimization Model is constructed, and uncertainty set needs to cover error distribution under preset probability threshold;Construct the double-layer optimization model with decision as center, optimize related parameters and variables jointly, form nested structure;Execute the size calibration method of statistical feasibility, determine the minimum feasible set radius based on the mahalanobis distance ordering of historical sample, ensure to meet the coverage probability constraint;Calculate the gradient of objective function, apply implicit function theorem and KKT condition to solve gradient, after total gradient is synthesized, gradient descent algorithm is used to iteratively optimize parameters, and the optimal uncertainty set is obtained in combination with early stop mechanism.The method of the application solves the problem of model and strategy disconnection in traditional "first estimation, then optimization" two-stage method, and can effectively improve the robustness and economy of energy storage scheduling.

Description

TECHNICAL FIELD

[0001] The present application relates to renewable energy scheduling technology, in particular to an energy storage optimal scheduling uncertainty set construction method combining estimation and optimization. BACKGROUND

[0002] Due to the volatility and uncertainty of power generation of intermittent renewable energy such as wind energy and solar energy, large-scale access of such energy to the power system brings unprecedented challenges to the stability, safety and scheduling flexibility of the power system. In order to maintain the real-time power balance of the power system, flexible and efficient regulation resources are urgently needed, among which the energy storage system plays an increasingly important role in the modern power system due to its fast response speed, high regulation accuracy and flexible layout. A core scheduling strategy to promote the deployment of the energy storage system is to participate in the power market and cooperate with the power grid for optimal scheduling. The effective implementation of this strategy largely depends on the accurate prediction of future electricity prices. However, with the increase of renewable energy output uncertainty, the intensification of load fluctuations and the increasing complexity of the electricity market mechanism, the electricity price time series presents stronger randomness and non-stationarity. This makes it difficult for traditional point prediction-based energy storage operation strategies to adapt to the increasingly complex actual operating environment. In order to ensure that the energy storage system can efficiently and reliably participate in the optimal scheduling of the power grid, it is necessary to fully consider the influence of the uncertainty of electricity prices in the scheduling decision, so as to improve the robustness and economic efficiency of the scheduling strategy.

[0003] The existing energy storage optimal scheduling model considering the uncertainty of electricity prices mainly adopts the methods of stochastic optimization, robust optimization and distributed robust optimization. Among them, stochastic optimization seeks the maximization of the expected value of the scheduling target, which belongs to the risk-neutral decision method. The risk-neutral decision method may lead to suboptimal or even incorrect operation decisions of the energy storage system, and in some extreme cases, continuous prediction errors may also cause continuous energy storage misoperation, resulting in that the energy storage system does not achieve the ideal scheduling effect in multiple scheduling periods. Therefore, as a risk-averse decision method, robust optimization and distributed robust optimization are widely used. Robust optimization seeks feasible or optimal solutions in the worst case by constructing an uncertainty set, which has strong conservatism and safety. Distributed robust optimization balances between computational complexity and risk controllability by considering a possible probability distribution set (i.e. fuzzy set). Although the above methods have achieved wide application in the theoretical level, the two-stage paradigm of "estimation first and optimization later" commonly used by them has inherent limitations. Specifically, the goal of the uncertainty modeling stage is usually to make the model as close as possible to the historical data characteristics, such as minimizing the distribution estimation error or constructing a set covering the real uncertainty. However, this two-stage approach does not consider the actual needs of the subsequent optimization stage for the structure of the uncertainty, resulting in that the constructed model is difficult to truly serve the optimal scheduling decision.

[0004] The existing energy storage system generally relies on accurate prediction of future electricity prices when participating in the optimal scheduling of the power system. However, with the increasing penetration of renewable energy and the increasing complexity of the market, the electricity price has high uncertainty. Although the methods such as stochastic optimization, robust optimization and distribution robust optimization used in the prior art can handle the electricity price uncertainty to some extent, the "two-stage" modeling process commonly used in the prior art, that is, first constructing an uncertainty model based on historical data and then making an optimization decision, does not fully consider the actual needs of the final decision performance for uncertainty modeling, resulting in a disconnection between the constructed model and the optimization strategy, thereby affecting the economy and robustness of the energy storage system.

[0005] In addition, with the increasing volatility of electricity market prices, the energy storage system faces a series of problems such as incorrect decision-making of energy storage operation caused by electricity price prediction error, excessive conservatism or insufficient risk control of existing uncertainty modeling methods, disconnection between uncertainty set construction and actual decision-making needs, lack of customized modeling framework for energy storage operation characteristics, high computational complexity affecting actual application efficiency, and poor market fluctuation adaptability when participating in the optimal scheduling of the power grid, which seriously affects the economy and operation robustness of the energy storage system participating in the optimal scheduling of the power grid.

[0006] Therefore, it is urgent to propose an integrated modeling method that closely combines uncertainty modeling and optimization decision-making and is oriented towards decision performance, which can ensure the effective operation of the energy storage system with controllable risk in the optimal scheduling of the power grid even if the electricity price prediction error is inevitable.

[0007] It should be noted that the information disclosed in the above background section is only for understanding the background of the present application, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0008] The main purpose of the present application is to overcome the defects in the above background art, and to provide a fusion estimation and optimization energy storage optimal scheduling uncertainty set construction method.

[0009] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0010] A fusion estimation and optimization energy storage optimal scheduling uncertainty set construction method, comprising the following steps:

[0011] S1, constructing a risk-averse energy storage optimal scheduling model: based on the electricity price uncertainty error vector, a robust optimization model is constructed to minimize the worst-case task loss, wherein the uncertainty set needs to cover the error distribution under a preset probability threshold;

[0012] S2, constructing a decision-centered double-layer optimization model: jointly optimizing the uncertainty set parameters and the energy storage decision variables to form a nested decision structure to minimize the scheduling objective function;

[0013] S3, performing a statistically feasible size calibration method: dynamically determining the minimum feasible set radius based on the Mahalanobis distance ranking of historical samples to ensure that the uncertainty set meets the preset coverage probability constraint;

[0014] S4, calculating the gradient of the objective function: applying the implicit function theorem and KKT condition to analytically solve the gradient of the decision variable with respect to the set parameter, combining the radius gradient to synthesize the total gradient of the objective function, and using the gradient descent algorithm to iteratively optimize the set parameter, combined with the early stopping mechanism to obtain the optimal uncertainty set.

[0015] Further, step S1 specifically includes:

[0016] The energy storage scheduling model is converted into a robust optimization form, and an ellipsoidal uncertainty set is defined to represent the statistical characteristics and correlation of the price error;

[0017] The robust optimization model is converted into an equivalent convex optimization form that can be analytically solved by taking the dual transformation of the inner maximization problem;

[0018] The uncertainty set needs to meet the preset confidence requirement of covering the real distribution of the error vector.

[0019] Further, step S2 specifically includes:

[0020] An explicit association between the objective function and the uncertainty set parameter is established, and the optimal set parameter is determined by minimizing the objective value of the scheduling decision;

[0021] The double-layer model includes a nested structure of outer set parameter optimization and inner energy storage decision optimization, and is limited by the coverage probability constraint.

[0022] Further, step S3 specifically includes:

[0023] The Mahalanobis distance of the historical sample relative to the set center is calculated and arranged in ascending order to generate a scalar sequence;

[0024] Based on the statistical characteristics of the binomial distribution and the preset confidence level, the minimum radius threshold that meets the coverage probability is selected from the scalar sequence;

[0025] The radius parameter is represented as a function of the set center and the shape parameter.

[0026] Further, the Mahalanobis distance calculation is specifically:

[0027] The covariance matrix inverse matrix of the ellipsoidal set measures the degree of deviation of the sample, and its scalar conversion preserves the correlation characteristics between error vectors.

[0028] Further, step S4 specifically comprises:

[0029] The inner-layer optimization problem is represented as a differentiable system satisfying KKT conditions, and the gradient of the decision variable with respect to the set parameter is analyzed by implicit function theorem;

[0030] The gradient of the joint objective function with respect to the decision variable, the gradient of the radius parameter, and the gradient of the radius with respect to the set parameter are combined to form the total gradient;

[0031] The gradient calculation includes a segmented derivative processing of the Mahalanobis distance sorting result.

[0032] Further, step S4 specifically comprises:

[0033] An adaptive gradient descent algorithm is used to update the set center parameter;

[0034] The size calibration calculation feasible radius is performed based on the current parameter in each iteration;

[0035] When the target function value does not improve for a preset number of times, the iteration is terminated and the optimal parameter is output.

[0036] Further, the gradient descent algorithm specifically is:

[0037] An adaptive matrix estimation algorithm is used to adjust the parameter update step, and a greedy early stopping mechanism is used to avoid overfitting.

[0038] A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the fusion estimation and optimization of the uncertainty set construction method of optimal energy storage scheduling.

[0039] A computer program product includes a computer program, and the computer program is executed by a processor to implement the fusion estimation and optimization of the uncertainty set construction method of optimal energy storage scheduling.

[0040] The present application has the following beneficial effects:

[0041] The application proposes a storage optimal scheduling uncertainty set construction method fusing estimation and optimization, which is a risk-avoiding decision-centered uncertainty set construction method. Through the decision gradient descent method designed based on the implicit function theorem and combined with the statistically feasible size calibration method, the uncertainty estimation and optimization decision are organically fused, thereby improving the operation robustness and economy of the storage system in the electricity price uncertain environment. The method constructs an integrated modeling framework that can closely combine uncertainty modeling and optimization decision and is oriented to the final decision performance. By deeply fusing the electricity price uncertainty modeling and optimization decision process, the problem of disconnection between model and strategy in the traditional two-stage method of 'estimation first and optimization later' is solved. This decision performance-oriented integrated modeling method can still guarantee the effective operation of the storage system in the optimal dispatch of the power grid with controllable risk under the condition that the electricity price prediction error is inevitable, thereby effectively dealing with the operation risk caused by the electricity price prediction error and improving the operation robustness and grid economy of the storage system optimal dispatch in the complex market environment. Meanwhile, the method is based on the statistically effective size calibration technology and combined with the differentiable optimization framework and gradient analysis method to dynamically adjust the uncertainty set parameters, so as to directly serve the improvement of the operation economy and robustness of the storage system. In actual application, the method can effectively reduce the risk of wrong operation caused by prediction error under the premise of ensuring the preset safety probability, enhance the adaptability of the system to market fluctuations, and has good calculation efficiency and online deployment potential. The application not only helps to improve the effective utilization of storage assets in the power grid system, but also provides key technical support for promoting the safe and stable operation of high-proportion renewable energy systems, and provides a more intelligent and stable technical path for the storage participating in the optimal dispatch of the power grid under the background of high-proportion renewable energy.

[0042] Other beneficial effects in the embodiments of the application will be further described below. BRIEF DESCRIPTION OF DRAWINGS

[0043] Figure 1 The flowchart of the storage optimal scheduling uncertainty set construction method fusing estimation and optimization. DETAILED DESCRIPTION

[0044] The embodiments of the application are described in detail below. It should be emphasized that the following description is only exemplary and is not intended to limit the scope of the application and its applications.

[0045] In addition, the terms "first", "second", etc. are used only for descriptive purposes and are not to be construed as indicating or implying relative importance or an indicated number of technical features. Thus, features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of embodiments of the present application, the meaning of "a plurality" is two or more, unless otherwise explicitly and specifically limited.

[0046] Referring to Figure 1 Embodiments of the present application provide a storage optimal scheduling uncertainty set construction method combining estimation and optimization, comprising the following steps:

[0047] Step S1, constructing a risk-averse storage optimal scheduling model: based on the price uncertainty error vector, a robust optimization model is constructed to minimize the worst-case task loss, wherein the uncertainty set needs to cover the error distribution under the preset probability threshold.

[0048] In some embodiments, step S1 specifically includes: converting the storage scheduling model into a robust optimization form, defining an ellipsoid type uncertainty set to represent the statistical characteristics and correlation of the price error; by taking the dual transformation of the inner maximization problem, the robust optimization model is converted into a convex optimization form that can be analytically solved; the uncertainty set needs to meet the preset confidence requirement of covering the real distribution of the error vector.

[0049] Step S2, constructing a decision-centered double-layer optimization model: jointly optimizing the uncertainty set parameters and the storage decision variables to form a nested decision structure, guided by minimizing the scheduling objective function.

[0050] In some embodiments, step S2 specifically includes: establishing an explicit association between the objective function and the uncertainty set parameters, and determining the optimal set parameters by minimizing the target value of the scheduling decision; the double-layer model includes a nested structure of outer set parameter optimization and inner storage decision optimization, and is limited by the coverage probability constraint.

[0051] Step S3, performing a statistically feasible size calibration method: based on the Mahalanobis distance sorting of historical samples, dynamically determining the minimum feasible set radius to ensure that the uncertainty set meets the preset coverage probability constraint.

[0052] In some embodiments, step S3 specifically includes: calculating the Mahalanobis distance of the historical sample relative to the set center and generating a scalar sequence in ascending order; based on the binomial distribution statistical characteristics and the preset confidence level, the minimum radius threshold that meets the coverage probability is selected from the scalar sequence; the radius parameter is represented as a function of the set center and the shape parameter.

[0053] In some embodiments, the Mahalanobis distance calculation is specifically based on the covariance matrix inverse matrix of the ellipsoid set to measure the deviation degree of the sample, and the scalar conversion preserves the correlation characteristics between the error vectors.

[0054] Step S4, calculating the gradient of the objective function: applying the implicit function theorem and KKT condition to analytically solve the gradient of the decision variable with respect to the set parameter, combining the radius gradient to synthesize the total gradient of the objective function, and using the gradient descent algorithm to iteratively optimize the set parameter, and combining the early stopping mechanism to obtain the optimal uncertainty set.

[0055] In some embodiments, step S4 specifically includes: representing the inner optimization problem as a differentiable system satisfying the KKT condition, and analytically solving the gradient of the decision variable with respect to the set parameter by the implicit function theorem; combining the gradient of the objective function with respect to the decision variable, the gradient of the radius parameter, and the gradient of the radius with respect to the set parameter to synthesize the total gradient; the gradient calculation includes piecewise derivative processing of the Mahalanobis distance sorting result.

[0056] In some embodiments, step S4 specifically further includes: updating the set center parameter using an adaptive gradient descent algorithm; performing size calibration to calculate the feasible radius based on the current parameter in each iteration; and terminating the iteration and outputting the optimal parameter when the objective function value does not improve for a preset number of times.

[0057] In further embodiments, the gradient descent algorithm specifically includes: adjusting the parameter update step size using an adaptive matrix estimation algorithm, and avoiding overfitting through a greedy early stopping mechanism.

[0058] The main technical advantage of the present application is that by deeply integrating the electricity price uncertainty modeling and optimization decision process, an innovative decision-centered uncertainty set construction method is proposed, which effectively overcomes the model and strategy disconnection problem caused by the traditional "estimation first and optimization second" two-stage paradigm; based on the statistical feasible size calibration technology, the minimum set radius is dynamically determined to ensure the error coverage ability under the preset safety probability, and at the same time, combined with the differentiable optimization framework and implicit function gradient analysis means, the adaptive adjustment of the uncertainty set parameter is realized, which significantly improves the dispatching economy and operation robustness of the energy storage system in the highly random environment of electricity price; while reducing the risk of misoperation caused by prediction error and enhancing the market fluctuation adaptation ability, the method also has the advantages of high computational efficiency and online deployment potential, providing key technical support for the intelligent dispatching of energy storage assets in high-proportion renewable energy power grids.

[0059] The following further describes specific embodiments of the present application and examples of algorithm implementation.

[0060] A method for constructing an uncertainty set for optimal energy storage scheduling that integrates estimation and optimization is proposed. Aiming at risk aversion, it implements a decision-centered uncertainty set construction. It uses a decision-oriented gradient descent method based on implicit function theorem and combines it with a statistically feasible sizing method to organically integrate uncertainty estimation and optimization decision-making, thereby improving the operational robustness and economic efficiency of energy storage systems participating in grid scheduling under uncertain electricity prices. Figure 1 , the method specifically comprises the following steps:

[0061] 1) Establishing a risk-averse optimal scheduling model for energy storage

[0062] This method focuses on the operation mechanism of energy storage systems participating in day-ahead grid dispatch. Energy storage owners, as price takers, participate in grid dispatch by only declaring power consumption without specifying a transaction price. t is the index of the market clearing period, T Represents the total number of time periods within the optimization cycle. The optimal scheduling problem of energy storage can be modeled as follows:

[0063]

[0064] Where, represents the market electricity price; Represents the interaction power between energy storage and the grid; represents the charging power of the energy storage, Indicates the discharge power of energy storage; Indicates the charging rated power of the energy storage, Indicates the discharge rated power of the energy storage; Indicates the state of charge of the energy storage; represents the charging efficiency of energy storage, Indicates the discharge efficiency of energy storage; and Indicates the upper and lower limits of energy storage.

[0065] The inherent uncertainty in market price forecasts creates economic risks, prompting risk-averse decision makers to seek robust bidding strategies. These forecast uncertainties are represented by the error vector In order to effectively manage the uncertainty in the optimal scheduling, the model defined by (1)-(5) is reformulated as a robust optimization problem. The charging decision goal is to achieve the error vector The true distribution of Minimize the worst-case task loss within the confidence region:

[0066]

[0067] Where, represents the feasible domain of charging decisions defined by constraints (2)-(5),ξ denotes the error vector in the set of uncertainty, and λ is the regularization coefficient, denotes the established set of uncertainty, p denotes the decision variable. The set of uncertainty needs to guarantee the ability to cover the true distribution of the error vector ε denotes the error tolerance.

[0068] In practical scenarios, the exact distribution of the error vector is unknown, and the set of uncertainty is determined based on a set of historical data samples . The proposed method adopts an ellipsoidal set of uncertainty as shown in (7) because it has the dual advantages of computational tractability and the ability to model the correlation between uncertain variables. The proposed method is also applicable to other convex and differentiable sets of uncertainty, such as box-type and polyhedral sets.

[0069]

[0070] where the parameter tuple determines the shape and size of the ellipsoid, where L is a lower triangular matrix, is the mean of the error vector ξ , and r is a scalar.

[0071] By taking the dual of the inner maximization problem and using strong duality, the robust optimization problem (6) can be transformed into an equivalent tractable form:

[0072]

[0073] When the set of uncertainty ideally satisfies the uncertainty coverage condition, i.e. , then the objective function value of model (8) is equal to the following conditional risk probability value:

[0074]

[0075] where denotes the objective function in formula (6).

[0076] 2) Establish a decision-centered set of uncertainty

[0077] ​​The scheduling decision of energy storage is affected by the set of uncertain parameters as shown in equation (8). However, the existing methods adopt a "two-stage" modeling procedure, i.e., constructing an uncertainty model based on historical data first and then making an optimization decision, which does not fully consider the actual needs of uncertainty modeling for the performance of the final decision, leading to a disconnection between the constructed model and the optimization strategy, thus affecting the economy and robustness of the energy storage system. To solve this problem, this paper proposes a decision-centric uncertainty set (DCUS), which combines parameter estimation with bid profitability, enhancing the matching between uncertainty modeling and economic performance.

[0078] The goal of the proposed method is to determine the parameters of the DCUS to minimize the objective function while ensuring the coverage of the uncertainty set for the uncertain variables. The mathematical model is represented as follows:

[0079]

[0080] The double nested structure of the above model and the coverage probability constraint (12) make it difficult to solve. To effectively solve the model, the invention first uses a statistically feasible size calibration method to ensure that the coverage probability constraint (12) is met. According to equation (7), once and L are determined, the parameter r determines the size of the uncertainty set. A larger r value allows the DCUS to cover more sample points, but at the cost of increasing conservatism. Conversely, a smaller r value leads to a less conservative solution, but may not satisfy equation (12). Therefore, the goal of the proposed size calibration method is to find the smallest feasible r value that satisfies equation (12). To achieve this, the Mahalanobis distance MD ( ξ ) is defined as follows:

[0081] .

[0082] Each sample vector is converted to a scalar by the Mahalanobis distance, and arranged in ascending order into a new set . The required parameter r is estimated by the following formula:

[0083]

[0084] In the formula, represents the confidence level; denotes the index corresponding to the optimal value.

[0085] After ensuring the probability constraint (12) by the statistical feasible sizing method described above, the parameters r can be expressed as a function of denoted as . The gradient of can then be used to optimize the objective function in equation (10) to solve the model. The gradient is calculated as follows:

[0086]

[0087] where denotes the objective function in (10); and The gradients denoted by and can be obtained directly by mathematical analysis. The main challenge in calculating the gradients shown in equation (15) is to calculate and

[0088] . The proposed method uses a differentiable optimization layer to calculate , and the gradient calculation is achieved by applying the KKT conditions and the implicit function theorem. For clarity in subsequent derivations, equation (11) is expressed in the following general form:

[0089]

[0090] h ( p ) represents equality constraints, g ( p ) represents inequality constraints.

[0091] The KKT conditions for model (16) can be expressed in the following matrix form:

[0092]

[0093] where denotes the primal and dual variables of the model, p denotes the decision variables, v denotes the dual variables related to the inequality constraints g ( p ), μ denotes the dual variables related to the equality constraints h ( p ); is a diagonal matrix composed of vectors

[0094] According to the implicit function theorem, for a continuously differentiable function The existence of the following derivative relationship:

[0095]

[0096] where, F the Jacobian matrix of

[0097] According to the implicit function theorem shown in equation (18), solving it can get :

[0098]

[0099] After calculating , the next step is to calculate .The expression of is as follows:

[0100]

[0101] The second equation is true because In the set , the formula is unique, so when otherwise it is 0.

[0102] By taking the above steps, the gradient of the objective function (10) can be calculated under the premise of satisfying the constraint (12), and then sampling the algorithm 1 shown in the following table 1 can obtain the uncertainty set parameters of the decision center:

[0103] Table 1

[0104]

[0105] ​In summary, the application proposes a kind of energy storage optimal scheduling uncertainty set construction method of fusion estimation and optimization, and the uncertainty set is constructed for risk-averse decision center, and the uncertainty estimation and optimization decision are organically fused by the decision gradient descent method based on implicit function theorem design and combined with the statistically feasible size calibration method;The modeling method based on decision center solves the problem of model and strategy disconnection in traditional "first estimation, then optimization" two-stage method by deep fusion of price uncertainty modeling and optimization decision process, and based on statistically effective size calibration technology, combined with the micro-optimal framework and gradient analysis means, the uncertainty set parameters are dynamically adjusted, so as to directly serve to improve the operation robustness and economy of energy storage system in the environment of price uncertainty participating in power grid scheduling;In practical application, the method can effectively reduce the risk of wrong operation caused by prediction error under the premise of ensuring the preset safety probability, enhance the adaptability of system to market fluctuation, and has good calculation efficiency and online deployment potential, provide a more intelligent and robust technical path for energy storage participating in power grid scheduling under the background of high proportion of renewable energy.

[0106] The embodiment of the application further provides a storage medium for storing a computer program, which is executed to perform at least the method described above.

[0107] The embodiment of the application further provides a control device, which includes a processor and a storage medium for storing a computer program; wherein the processor is used to execute the computer program to perform at least the method described above.

[0108] The embodiment of the application further provides a processor, which executes a computer program to perform at least the method described above.

[0109] The storage medium can be implemented by any type of nonvolatile storage device, or a combination thereof. The nonvolatile memory can be a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Ferromagnetic Random Access Memory (FRAM), a Flash memory, a magnetic surface storage, an optical disc or a Compact Disc Read-Only Memory (CD-ROM). The magnetic surface storage can be a disk memory or a tape memory. The storage medium described in the embodiments of the present application is intended to include, but is not limited to, these and any other suitable type of memory.

[0110] In several embodiments provided by the present application, it should be understood that the disclosed system and method can be implemented in other manners. The described device embodiments are merely schematic, and the division of the units is merely a logical function division. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling or direct coupling or communication connection between the components can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0111] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e., can be located in one place or distributed on a plurality of network units; some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0112] In addition, each functional unit in the embodiments of the present application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in the form of hardware, or in the form of hardware plus software functional units.

[0113] Those skilled in the art can understand that all or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware, and the foregoing program can be stored in a computer readable storage medium, and the program performs the steps of the above-mentioned method embodiments when executed; and the foregoing storage medium includes a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disc or an optical disc and various storage medium capable of storing program codes.

[0114] Alternatively, the integrated unit of the present application can be stored in a computer readable storage medium if it is realized in the form of a software function module and sold or used as an independent product. Based on such understanding, the technical solutions of the embodiments of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium, includes several instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the methods described in the embodiments of the present application. The foregoing storage medium includes a mobile storage device, a ROM, a RAM, a magnetic disc or an optical disc and various storage medium capable of storing program codes.

[0115] The methods disclosed in the several method embodiments provided by the present application can be combined arbitrarily without conflict to obtain new method embodiments.

[0116] The features disclosed in the several product embodiments provided by the present application can be combined arbitrarily without conflict to obtain new product embodiments.

[0117] The features disclosed in the several method or device embodiments provided by the present application can be combined arbitrarily without conflict to obtain new method embodiments or device embodiments.

[0118] The above is a further detailed description of the present application in combination with specific preferred embodiments, and the specific implementation of the present application cannot be limited to these descriptions. For those skilled in the art of the present application, without departing from the concept of the present application, a number of equivalent substitutions or obvious modifications can be made, and the performance or use is the same, which should be regarded as belonging to the protection scope of the present application.

Claims

1. A method for constructing an uncertainty set for optimal energy storage scheduling that integrates estimation and optimization, characterized by: The following steps are involved: S1. Build a risk-averse energy storage optimization scheduling model: Based on the electricity price uncertainty error vector, a robust optimization model is constructed to minimize the worst-case task loss. The uncertainty set must cover the error distribution under a preset probability threshold. S2. Construct a decision-centered two-level optimization model: Guided by minimizing the scheduling objective function, jointly optimize the uncertainty set parameters and energy storage decision variables to form a nested decision structure; S3. Execute a statistically feasible size calibration method: dynamically determine the minimum feasible set radius based on the Mahalanobis distance sorting of historical samples to ensure that the uncertainty set meets the preset coverage probability constraint; S4. Calculate the objective function gradient: Apply the implicit function theorem and KKT conditions to analytically solve the gradient of the decision variable with respect to the set parameters, combine the radius gradient to synthesize the total gradient of the objective function, and use the gradient descent algorithm to iteratively optimize the set parameters. Combined with the early stopping mechanism, the optimal uncertainty set is obtained.

2. The method according to claim 1, characterized in that Step S1 specifically includes: The energy storage scheduling model is transformed into a robust optimization form, and an ellipsoidal uncertainty set is defined to characterize the statistical characteristics and correlation of electricity price errors. By taking the dual transformation of the inner maximization problem, the robust optimization model is transformed into an equivalent convex optimization form that can be solved analytically. The uncertainty set must meet a preset confidence requirement for covering the true distribution of the error vector.

3. The method according to claim 1 or 2, characterized in that Step S2 specifically includes: Establish an explicit association between the objective function and the uncertainty set parameters, and determine the optimal set parameters by minimizing the target value of the scheduling decision; The two-layer optimization model includes a nested structure of outer layer set parameter optimization and inner layer energy storage decision optimization, and is subject to coverage probability constraints.

4. The method according to any one of claims 1 to 2, characterized in that Step S3 specifically includes: Calculate the Mahalanobis distance of historical samples relative to the center of the set and arrange them in ascending order to generate a scalar sequence; Based on the statistical characteristics of the binomial distribution and the preset confidence level, the minimum radius threshold that satisfies the coverage probability is selected from the scalar sequence; The radius parameter is expressed as a function of the center of the set and the shape parameter.

5. The method according to claim 4, characterized in that The Mahalanobis distance calculation is specifically as follows: The inverse matrix of the covariance matrix based on the ellipsoid set measures the degree of sample deviation, and its scalarization transformation retains the correlation characteristics between error vectors.

6. The method according to any one of claims 1 to 2, characterized in that Step S4 specifically includes: The inner optimization problem is expressed as a differentiable system that satisfies the KKT condition, and the gradient of the decision variable with respect to the set parameter is analyzed by the implicit function theorem. Combine the gradient of the objective function with respect to the decision variable, the gradient with respect to the radius parameter, and the gradient of the radius with respect to the set parameter to synthesize the total gradient; Gradient calculation includes piecewise derivation of the Mahalanobis distance sorting results.

7. The method according to any one of claims 1 to 2, characterized in that Step S4 specifically includes: Adaptive gradient descent algorithm is used to update the set center parameters; Each iteration performs size calibration based on the current parameters to calculate the feasible radius; When the objective function value does not improve for consecutive preset rounds, the iteration is terminated and the optimal parameters are output.

8. The method according to claim 7, characterized in that The gradient descent algorithm is specifically: An adaptive moment estimation algorithm is used to adjust the parameter update step size, and a greedy early stopping mechanism is used to avoid overfitting.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for constructing an uncertainty set of energy storage optimal scheduling by integrating estimation and optimization as described in any one of claims 1 to 8 is implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for constructing an uncertainty set of energy storage optimal scheduling by integrating estimation and optimization as described in any one of claims 1 to 8 is implemented.

Images