Construction method and device of energy storage cluster regulation and control model, equipment and medium

Abstract

The embodiment of the invention provides a construction method and device of an energy storage cluster regulation and control model, equipment and a medium, and relates to the technical field of energy management. The method comprises the following steps: acquiring a plurality of source load powers and a plurality of energy storage equipment data; performing set construction according to the plurality of source load powers to obtain a confidence set; clustering is carried out according to the multiple pieces of energy storage equipment data to obtain multiple energy storage clusters and an adjustable load capacity curve of each energy storage cluster; and performing model construction according to the confidence set, the adjustable load curve and a plurality of preset power grid parameters to obtain an energy storage cluster regulation and control model. A probability distribution deviation boundary of power fluctuation is quantified by constructing a confidence set, so that a regulation and control plan has robustness for coping with source-load random fluctuation; a plurality of energy storage clusters and adjustable load capacity curves thereof are formed through clustering analysis, so that the energy storage cluster regulation and control model realizes full-level refined regulation and control from clusters to equipment, and the problem of low regulation and control efficiency of the energy storage cluster regulation and control model in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of energy management technology, and in particular to a method, device, equipment, and medium for constructing an energy storage cluster control model. Background Art

[0002] An energy storage cluster is a collaborative operation group formed by aggregating multiple distributed energy storage devices, enabling unified scheduling of power and energy. Energy storage cluster control coordinates the charging and discharging behavior of each energy storage device within the cluster to improve overall operational efficiency. With the distributed transformation of power systems, traditional single-source energy storage control models are unable to meet the needs of collaborative management of large-scale, distributed energy storage resources. Advances in information technology provide technical support for the intelligent control of energy storage clusters, making it a key means of enhancing grid flexibility and stability.

[0003] In existing technologies, a centralized architecture is usually used to build an energy storage cluster control model. A central controller collects operating parameters such as the power and capacity of each energy storage device in real time. A global optimization model is established based on the grid scheduling requirements and equipment constraints. Deterministic optimization algorithms such as mixed integer linear programming or model predictive control are used to calculate the cluster's control plan.

[0004] However, existing technologies suffer from low control efficiency in energy storage cluster control models. This problem stems from the complexity of the energy storage system's operating environment and the inherent characteristics of the control architecture. The random fluctuations in source and load power interact with the computational resource limitations of the centralized control architecture, making it difficult for fixed-parameter control models to adapt to the dynamic changes in actual operating conditions. Furthermore, large-scale distributed energy storage scenarios also face implementation bottlenecks in refined control. These factors collectively lead to reduced control efficiency in energy storage cluster control models. Summary of the Invention

[0005] The embodiments of the present application provide a method, apparatus, device, and medium for constructing an energy storage cluster control model, so as to solve the problem of low control efficiency of energy storage cluster control models in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for constructing an energy storage cluster control model, comprising:

[0007] Acquire multiple source load powers and multiple energy storage device data; wherein the multiple source load powers refer to the power of multiple source load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of the multiple energy storage devices;

[0008] A confidence set is constructed based on the multiple source-load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between a reference probability distribution and a true probability distribution, the reference probability distribution refers to the fluctuation pattern of the multiple source-load powers, and the true probability distribution refers to the fluctuation pattern of the powers of the multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time;

[0009] Clustering the data of the multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each of the energy storage clusters; wherein each of the energy storage clusters includes multiple energy storage devices, and the adjustable load curve is used to represent the load change of each of the energy storage clusters over time;

[0010] A model is constructed based on the confidence set, the adjustable load curve and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control the multiple energy storage clusters and the multiple energy storage devices included in each of the energy storage clusters.

[0011] In a possible implementation, the step of constructing a set based on the multiple source-load powers to obtain a confidence set includes:

[0012] Performing probability distribution fitting according to the multiple source-load powers to obtain the reference probability distribution;

[0013] A set is constructed based on the reference probability distribution and a preset confidence radius to obtain the confidence set; wherein the confidence radius is used to represent the maximum degree of deviation between the reference probability distribution and the true probability distribution.

[0014] In a possible implementation, clustering the data of the plurality of energy storage devices to obtain a plurality of energy storage clusters and an adjustable load curve for each of the energy storage clusters includes:

[0015] Constructing a model based on the plurality of energy storage device data to obtain a plurality of energy storage device models; wherein each of the energy storage device models is used to represent the operating characteristics of each of the energy storage devices;

[0016] Solving each of the energy storage device models to obtain the charge and discharge power of each of the energy storage devices;

[0017] Aggregating the charge and discharge powers of the multiple energy storage devices to obtain the multiple energy storage clusters and the charge and discharge powers of the multiple energy storage clusters;

[0018] The charge and discharge powers of the multiple energy storage clusters are clustered to obtain an adjustable load curve for each of the energy storage clusters.

[0019] In one possible implementation, the energy storage cluster control model includes an upper-level planning model and a lower-level operation model. The model is constructed based on the confidence set, the adjustable load curve, and a plurality of preset grid parameters to obtain the energy storage cluster control model, including:

[0020] A model is constructed based on the confidence set, the adjustable load curve, and a plurality of preset grid parameters to obtain the upper-level planning model; wherein the upper-level planning model is used to represent the interaction process between the plurality of energy storage clusters and the external grid, and the impact of the interaction process on the cost of the external grid;

[0021] Solving the upper-level planning model to obtain a plurality of first plans; wherein the plurality of first plans refer to control plans for the plurality of energy storage clusters;

[0022] A model is constructed based on the multiple first plans, the multiple energy storage device data and the multiple power grid parameters to obtain the lower-level operation model; wherein the lower-level operation model is used to optimize the response strategy and cluster allocation of the multiple energy storage devices under each of the energy storage clusters to minimize the response cost and voltage fluctuation of the multiple energy storage devices under each of the energy storage clusters.

[0023] In a possible implementation, after constructing the model according to the multiple first plans, the multiple energy storage device data, and the multiple grid parameters to obtain the lower-layer operation model, the method further includes:

[0024] Solving the lower-layer operation model to obtain a plurality of second plans; wherein the plurality of second plans refer to control plans for a plurality of energy storage devices under each of the energy storage clusters;

[0025] Obtaining execution results and operation results; wherein the execution results refer to operation data after the multiple energy storage devices under each of the energy storage clusters respectively execute the multiple second plans, and the operation results refer to actual operation data of the multiple energy storage devices under each of the energy storage clusters;

[0026] The deviation between the execution result and the operation result is calculated, and the upper-level planning model and the lower-level operation model are iteratively optimized according to the deviation until the deviation is less than a preset threshold.

[0027] In a possible implementation, the iteratively optimizing the upper-layer planning model and the lower-layer operation model according to the deviation until the deviation is less than a preset threshold includes:

[0028] Inputting the deviation into a preset deviation source determination model to obtain the deviation source;

[0029] Matching the deviation source with a plurality of preset deviation adjustment strategies to obtain a target deviation adjustment strategy; wherein the target deviation adjustment strategy refers to any one of the plurality of deviation adjustment strategies;

[0030] Based on the target deviation adjustment strategy, the upper-level planning model and the lower-level operation model are iteratively optimized until the deviation is less than the threshold.

[0031] In a second aspect, an embodiment of the present application provides a system for constructing an energy storage cluster control model, comprising: a cloud, an edge server, multiple source-load devices, and multiple energy storage devices;

[0032] The cloud is used to implement the method for constructing the energy storage cluster control model according to any one of the first aspects;

[0033] The edge server is configured to collect data on multiple source-load powers and multiple energy storage devices, and send the data on the multiple source-load powers and the multiple energy storage devices to the cloud, as well as send multiple first plans to multiple energy storage clusters, and send multiple second plans to multiple energy storage devices under each of the energy storage clusters; wherein the multiple energy storage clusters, the multiple first plans, and the multiple second plans are calculated according to the method for constructing an energy storage cluster control model according to any one of the first aspects;

[0034] The plurality of source load devices are used to send the plurality of source load powers to the edge server;

[0035] The multiple energy storage devices are used to send the multiple energy storage device data to the edge server, receive the multiple second plans sent by the edge server, and execute the multiple second plans.

[0036] In a third aspect, an embodiment of the present application provides a device for constructing an energy storage cluster control model, including:

[0037] A first acquisition module is configured to acquire multiple source-load powers and multiple energy storage device data; wherein the multiple source-load powers refer to the powers of multiple source-load devices within a preset first time period, where the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to current operating data of multiple energy storage devices;

[0038] a set construction module, configured to construct a set based on the multiple source-load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between a reference probability distribution and a true probability distribution, the reference probability distribution refers to the fluctuation pattern of the multiple source-load powers, and the true probability distribution refers to the fluctuation pattern of the powers of the multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time;

[0039] A clustering module, configured to cluster the plurality of energy storage device data to obtain a plurality of energy storage clusters and an adjustable load curve for each of the energy storage clusters; wherein each of the energy storage clusters includes a plurality of energy storage devices, and the adjustable load curve is configured to represent how the load of each energy storage cluster changes over time;

[0040] A model construction module is used to construct a model based on the confidence set, the adjustable load curve and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control the multiple energy storage clusters and the multiple energy storage devices included in each of the energy storage clusters.

[0041] In a possible implementation, the set construction module includes:

[0042] a fitting unit, configured to perform probability distribution fitting according to the multiple source load powers to obtain the reference probability distribution;

[0043] A set construction unit is used to construct a set according to the reference probability distribution and a preset confidence radius to obtain the confidence set; wherein the confidence radius is used to represent the maximum deviation between the reference probability distribution and the true probability distribution.

[0044] In a possible implementation, the clustering module includes:

[0045] A first construction unit is configured to construct a model based on the plurality of energy storage device data to obtain a plurality of energy storage device models; wherein each of the energy storage device models is configured to represent an operating characteristic of each of the energy storage devices;

[0046] A first solving unit, configured to solve the model of each energy storage device to obtain the charge and discharge power of each energy storage device;

[0047] an aggregation unit, configured to aggregate the charge and discharge powers of the plurality of energy storage devices to obtain the plurality of energy storage clusters and the charge and discharge powers of the plurality of energy storage clusters;

[0048] The clustering unit is configured to cluster the charge and discharge powers of the plurality of energy storage clusters to obtain an adjustable load curve for each of the energy storage clusters.

[0049] In one possible implementation, the energy storage cluster control model includes an upper-level planning model and a lower-level operation model, and the model construction module includes:

[0050] a second construction unit, configured to construct a model based on the confidence set, the adjustable load curve, and a plurality of preset grid parameters to obtain the upper-level planning model; wherein the upper-level planning model is configured to represent the interaction process between the plurality of energy storage clusters and an external grid, and the impact of the interaction process on the cost of the external grid;

[0051] a second solving unit, configured to solve the upper-level planning model to obtain a plurality of first plans; wherein the plurality of first plans refer to control plans for the plurality of energy storage clusters;

[0052] A third construction unit is configured to construct a model based on the multiple first plans, the multiple energy storage device data, and the multiple grid parameters to obtain the lower-level operation model; wherein the lower-level operation model is configured to optimize the response strategy and cluster allocation of the multiple energy storage devices under each of the energy storage clusters to minimize the response cost and voltage fluctuation of the multiple energy storage devices under each of the energy storage clusters.

[0053] In one possible implementation, the device for constructing the energy storage cluster control model further includes:

[0054] A solution module, configured to solve the lower-layer operation model to obtain a plurality of second plans; wherein the plurality of second plans refer to control plans for the plurality of energy storage devices under each of the energy storage clusters;

[0055] A second acquisition module is configured to acquire execution results and operation results; wherein the execution results refer to operation data after the multiple energy storage devices under each of the energy storage clusters respectively execute the multiple second plans, and the operation results refer to actual operation data of the multiple energy storage devices under each of the energy storage clusters;

[0056] The optimization module is used to calculate the deviation between the execution result and the operation result, and iteratively optimize the upper-level planning model and the lower-level operation model according to the deviation until the deviation is less than a preset threshold.

[0057] In a possible implementation, the optimization module includes:

[0058] a source determination unit, configured to input the deviation into a preset deviation source determination model to obtain a source of the deviation;

[0059] a strategy determination unit, configured to match the deviation source with a plurality of preset deviation adjustment strategies to obtain a target deviation adjustment strategy; wherein the target deviation adjustment strategy refers to any one of the plurality of deviation adjustment strategies;

[0060] An optimization unit is used to iteratively optimize the upper-level planning model and the lower-level operation model based on the target deviation adjustment strategy until the deviation is less than the threshold.

[0061] In a fourth aspect, the present application provides an electronic device, comprising: a processor, and a memory communicatively connected to the processor;

[0062] The memory stores computer-executable instructions;

[0063] When the processor executes the computer-executable instructions stored in the memory, it is used to implement the method for constructing an energy storage cluster control model as described in any one of the first aspects.

[0064] In a fifth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When the computer-executable instructions are executed by a processor, they are used to implement the method for constructing an energy storage cluster control model as described in any one of the first aspects.

[0065] In a sixth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, is used to implement the method for constructing an energy storage cluster control model as described in any one of the first aspects.

[0066] The present application provides a method, device, equipment and medium for constructing an energy storage cluster control model, the method comprising: obtaining multiple source load powers and multiple energy storage device data; wherein the multiple source load powers refer to the power of multiple source load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of multiple energy storage devices; a set is constructed based on the multiple source load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between a reference probability distribution and a true probability distribution, the reference probability distribution refers to the fluctuation law of the multiple source load powers, and the true probability distribution refers to the multiple source load devices The power fluctuation pattern within a preset second time period is determined, and the start time of the second time period is later than the current time; clustering is performed based on the data of the multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each of the energy storage clusters; wherein each of the energy storage clusters includes multiple energy storage devices, and the adjustable load curve is used to represent the load change of each of the energy storage clusters over time; a model is constructed based on the confidence set, the adjustable load curve and multiple preset power grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control the multiple energy storage clusters and the multiple energy storage devices included in each of the energy storage clusters. The method for constructing an energy storage cluster control model of the present application constructs a confidence set based on historical source-load power data to quantify the probability distribution deviation boundary of power fluctuations, replacing the traditional deterministic model's reliance on precise probability distribution, so that the control plan has the robustness to cope with random fluctuations in source and load, and reduces frequent instruction adjustments caused by operating condition deviations; through cluster analysis, multiple energy storage clusters and their adjustable load curves are formed, and the confidence set is combined to quantify the source-load uncertainty, so that the energy storage cluster control model can not only perform overall optimization control at the cluster level, but also perform precise charge and discharge control and response timing management for each energy storage device in each energy storage cluster based on the adjustable load curve, thereby realizing full-level refined control from cluster to device, thereby solving the problem of low control efficiency of energy storage cluster control models in the existing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0068] Figure 1 A schematic diagram of an application scenario of the method for constructing an energy storage cluster control model provided in an embodiment of the present application;

[0069] Figure 2 Schematic diagram of the process of constructing the energy storage cluster control model provided in the embodiment of the present application Figure 1 ;

[0070] Figure 3 Schematic diagram of the process of constructing the energy storage cluster control model provided in the embodiment of the present application Figure 2 ;

[0071] Figure 4 A schematic diagram of the structure of a device for constructing an energy storage cluster control model provided in an embodiment of the present application;

[0072] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0073] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0074] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0075] In the embodiments of the present application, words such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. Those skilled in the art will understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit differences. It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described in this application as "exemplary" or "for example" should not be interpreted as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way. In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more.

[0076] It should be noted that the phrase "at the time of..." in the embodiments of this application can refer to the instantaneous occurrence of a certain situation or a period of time after the occurrence of a certain situation, and this is not specifically limited in this embodiment of this application. Furthermore, the methods, devices, equipment, and media for constructing an energy storage cluster control model provided in the embodiments of this application are merely examples, and the methods, devices, equipment, and media for constructing an energy storage cluster control model may also include more or less content.

[0077] To facilitate a clear description of the technical solutions of the embodiments of the present application, some of the terms and technologies involved in the embodiments of the present application are briefly introduced below:

[0078] Source-load power: This refers to the combined output power of the power source and the power consumed by the load in a power system. "Source" represents the available output power of the power supply unit, while "load" represents the power demand of the power consumption unit. Together, these two parameters constitute the core parameter for power system power balance. In energy storage cluster control, real-time data on source-load power reflects the supply and demand status of the power system.

[0079] A confidence set is a mathematical tool used to describe the permissible deviation between a reference probability distribution and the actual probability distribution. By setting a boundary value, it keeps the difference between the reference distribution and the actual distribution within a reasonable range, allowing the optimization model based on this set to adapt to the uncertain fluctuations in actual operation. In energy storage control, confidence sets eliminate the need to accurately predict future power fluctuations. Simply ensuring that actual fluctuations do not exceed a preset boundary allows for the generation of robust control strategies, effectively reducing the need for frequent adjustments due to inaccurate predictions.

[0080] A reference probability distribution is a probabilistic model derived from analyzing historical source-load power data. It describes the fluctuation patterns of power output and load demand over specific historical periods. This distribution establishes a correspondence between power values ​​and their frequency of occurrence by fitting historical operating data. This provides a baseline for constructing confidence sets, measuring the degree to which future actual operating conditions deviate from historical patterns. In energy storage regulation, the reference probability distribution serves as the foundation for uncertainty quantification, enabling the system to distinguish between historical normal fluctuations and abnormal fluctuations.

[0081] The true probability distribution refers to the actual fluctuations in source and load power in the power system during future operating periods, reflecting the uncertainty of power output and load demand under real-world operating conditions. Unlike reference probability distributions based on historical data, the true probability distribution is an unknown objective entity whose specific form depends on dynamic factors such as actual weather conditions and electricity consumption behavior. It represents the true uncertainty of system operation and is the target that energy storage control strategies ultimately need to adapt to.

[0082] Source-load uncertainty refers to the uncertainties on both the generation and load sides of a power system. Source-side uncertainty primarily involves the volatility and randomness of renewable energy, which leads to uncertainty in its output power. Load-side uncertainty primarily manifests itself in the uncertainty of demand-side response, as load demand can fluctuate due to various factors, such as weather, economic activity, and user behavior.

[0083] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent like or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention, as detailed in the appended claims.

[0084] The technical solution of the present invention is described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present invention will be described below with reference to the accompanying drawings.

[0085] In order to clearly understand the technical solution of the present application, the solution of the prior art is first introduced in detail. An energy storage cluster refers to a collaborative operation group formed by aggregating multiple dispersed energy storage devices, which can achieve unified scheduling of power and energy. Energy storage cluster regulation is to coordinate the charging and discharging behavior of each energy storage device in the cluster to improve the overall operating efficiency. With the distributed transformation of the power system, the traditional single energy storage regulation mode is difficult to meet the needs of collaborative management of large-scale and decentralized energy storage resources. The advancement of information technology provides technical support for the intelligent regulation of energy storage clusters, making energy storage cluster regulation an important means to improve the flexibility and stability of the power grid.

[0086] In the existing technology, a centralized architecture is usually used to build an energy storage cluster control model. The power and capacity and other operating parameters of each energy storage device are collected in real time through a central controller. A global optimization model is established based on the grid scheduling requirements and equipment constraints. Deterministic optimization algorithms such as mixed integer linear programming or model predictive control are used to calculate the cluster control plan. However, in the existing technology, the random volatility of source and load power interacts with the computing resource limitations of the centralized control architecture, resulting in the difficulty of the fixed parameter control model to adapt to the dynamic changes in actual operating conditions. At the same time, in large-scale distributed energy storage scenarios, it also faces the implementation bottleneck of refined control. These factors together lead to a reduction in the control efficiency of the energy storage cluster control model. Therefore, the existing technology has the problem of low control efficiency of the energy storage cluster control model.

[0087] To address the low efficiency of existing energy storage cluster control models, research has found that adaptive, hierarchical, and coordinated control of energy storage resources can be implemented while quantifying source-load uncertainty, thereby improving control efficiency. First, by analyzing the distribution characteristics of historical source-load fluctuation data, a time-varying probabilistic boundary model can be constructed to replace fixed thresholds. This approach utilizes a sliding time window to continuously update boundary parameters, enabling the control model to automatically adapt to the dynamic characteristics of wind and solar power output and load changes. This avoids overly conservative scheduling strategies and reduces the frequent instruction revisions caused by forecast bias. Second, a dual clustering criterion can be employed: physical clusters are first divided by electrical distance, and then virtual subclusters are subdivided based on response speed and capacity characteristics. A differentiated control strategy is designed for each subcluster. The upper layer optimizes global power balance, while the lower layer controls device charge and discharge timing, achieving an organic combination of coarse and fine tuning, thus overcoming the computational bottleneck of centralized architectures. Third, an online model calibration system based on actual operating data can be established, dynamically coupling an offline-trained baseline model with real-time operating data. An adaptive weight adjustment algorithm enables the model to rapidly respond to sudden fluctuations while maintaining the optimization framework, improving control accuracy in complex scenarios.

[0088] Specifically, we can first build a source-load dynamic feature extraction model based on a deep time series network, capture the spatiotemporal correlation characteristics of wind and solar output and load fluctuations in real time through a sliding time window, and generate a flexible control boundary with probability guarantee; secondly, use a multi-scale clustering algorithm to divide energy storage resources into virtual clusters with similar response characteristics, and formulate long-term capacity planning, medium-term power allocation and short-term instruction correction strategies at each level; finally, design a data-model hybrid-driven online optimization mechanism, and adaptively adjust the regulation weight and response priority of each cluster through the dynamic game between real-time operating data and preset control strategies, forming an intelligent control system that takes into account economy, reliability and equipment life.

[0089] The embodiments of the present application provide a method, apparatus, device, and medium for constructing an energy storage cluster control model. By constructing a confidence set based on historical source-load power data, the probability distribution deviation boundary of power fluctuations is quantified, replacing the traditional deterministic model's reliance on precise probability distributions. This enables the control plan to be robust against random source-load fluctuations and reduces frequent instruction adjustments due to operating condition deviations. Multiple energy storage clusters and their adjustable load curves are formed through cluster analysis, and the confidence set is combined to quantify source-load uncertainty. This enables the energy storage cluster control model to perform overall optimization control at the cluster level and to perform precise charge and discharge control and response timing management for each energy storage device in each energy storage cluster based on the adjustable load curve, thereby achieving full-level refined control from cluster to device, thereby solving the problem of low control efficiency of energy storage cluster control models in the prior art.

[0090] Based on the above creative findings, the technical solution of the present application is proposed.

[0091] The following describes application scenarios of the method for constructing an energy storage cluster control model provided by an embodiment of the present invention. Figure 1 Schematic diagram of an application scenario of the method for constructing an energy storage cluster control model provided in an embodiment of the present application. Figure 1 As shown, the application scenario includes a cloud 101, an edge server 102, multiple energy storage devices 103, and multiple source-load devices 104. The edge server 102 obtains multiple source-load powers and multiple energy storage device data, and sends the multiple source-load powers and multiple energy storage device data to the cloud 101. The cloud 101 constructs a set based on the multiple source-load powers to obtain a confidence set, clusters the multiple energy storage device data, obtains multiple energy storage clusters and an adjustable load curve for each energy storage cluster, and constructs a model based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain an energy storage cluster control model. The cloud 101 solves the energy storage cluster control model to obtain multiple first plans and multiple second plans. The cloud 101 sends the multiple first plans and multiple second plans to the edge server 102. The edge server 102 sends the multiple first plans and multiple second plans to the multiple energy storage clusters and the multiple energy storage devices 103, respectively. The multiple energy storage clusters and the multiple energy storage devices 103 are controlled according to the multiple first plans and multiple second plans, respectively.

[0092] The embodiments of the present invention are described below with reference to the accompanying drawings.

[0093] Figure 2 Schematic diagram of the process of constructing the energy storage cluster control model provided in the embodiment of the present application Figure 1 .like Figure 2 As shown, in this embodiment, the execution subject of the embodiment of the present invention is the cloud. Then the method for constructing the energy storage cluster control model provided by this embodiment includes the following steps:

[0094] S201. Acquire multiple source-load powers and multiple energy storage device data; wherein, the multiple source-load powers refer to the powers of multiple source-load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of multiple energy storage devices.

[0095] Specifically, sensors and data acquisition systems deployed on source-load devices and energy storage devices can be used to acquire data on multiple source-load powers and multiple energy storage devices. These data are used to monitor the operating status and power output of the equipment in real time, thereby providing basic data support for subsequent confidence set construction and energy storage cluster control models.

[0096] S202. Construct a set based on multiple source-load powers to obtain a confidence set; wherein the confidence set is used to indicate the degree of deviation between the reference probability distribution and the true probability distribution, the reference probability distribution refers to the fluctuation pattern of the power of multiple source-load devices, and the true probability distribution refers to the fluctuation pattern of the power of multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time.

[0097] Specifically, historical source-load power data can be processed through machine learning technology to construct a confidence set. This confidence set is used to quantify the degree of deviation between the reference probability distribution and the true probability distribution, thereby providing the energy storage cluster control model with an uncertainty boundary estimate of future power fluctuations and enhancing the model's robustness and adaptability.

[0098] For example, a kernel density estimation method can be used to fit a reference probability distribution curve based on historical source and load power data. The maximum possible deviation between this distribution and the future actual power fluctuation distribution is then calculated, and the confidence radius is determined. Finally, a confidence set is constructed, with the reference distribution as the center and the confidence radius as the boundary, to encompass a set of probability distributions encompassing all possible fluctuation scenarios. This confidence set characterizes the range of deviation between the future actual power fluctuation distribution and the historical reference distribution at a preset confidence level, providing uncertainty bounds for subsequent robust optimization.

[0099] S203. Clustering is performed based on the data of the multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each energy storage cluster; wherein each energy storage cluster includes multiple energy storage devices, and the adjustable load curve is used to represent the load variation of each energy storage cluster over time.

[0100] Specifically, by applying clustering algorithms to analyze the data of multiple energy storage devices, devices with similar characteristics can be aggregated into multiple energy storage clusters. Based on the data of these clusters, adjustable load curves for each cluster can be generated. These curves are used to describe and predict the load change trends of the cluster, thereby supporting the overall optimization and regulation of the energy storage cluster and the refined management of the device level.

[0101] For example, the real-time operating parameters of each energy storage device can be extracted, including characteristic quantities such as remaining capacity, charge and discharge power limits, and response rate. A fuzzy clustering algorithm based on electrical distance and response characteristics can then be used to dynamically group devices with similar regulation capabilities into the same cluster. For each cluster, the charge and discharge characteristics and capacity constraints of its member devices are combined to calculate the overall adjustable power range of the cluster at different time periods. Finally, a spline interpolation method is used to generate a smooth adjustable load curve. This curve reflects the cluster's maximum regulation capacity at each moment in the dispatch cycle, while retaining the necessary regulation margin to cope with real-time fluctuations. Throughout this process, the clustering threshold and curve generation parameters can be dynamically adjusted according to the actual needs of the power grid.

[0102] S204. Construct a model based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control multiple energy storage clusters and multiple energy storage devices included in each energy storage cluster.

[0103] Specifically, by integrating confidence sets, adjustable load curves, and preset grid parameters, and utilizing optimization algorithms such as linear programming and nonlinear programming, a storage cluster control model can be constructed. This model is used to optimize the scheduling strategy of the energy storage cluster while taking into account the uncertainty of source-load fluctuations and grid operation constraints, thereby achieving efficient control of energy storage equipment and enhancing grid stability.

[0104] For example, we can first establish a robust optimization objective function based on the confidence set that takes into account the uncertainty of source and load fluctuations, and combine the node voltage constraints and line capacity limitations in the grid parameters to form an upper-level planning model framework. Then, based on the adjustable load curve of each energy storage cluster, we establish the associated constraints between cluster power distribution and equipment response timing, and construct the lower-level operation model structure. Using a two-layer iterative optimization algorithm, the upper-level model outputs the overall control instructions for each cluster, and the lower-level model decomposes the instructions into specific equipment action strategies. The confidence set boundaries and load curve parameters are corrected through a real-time feedback mechanism. The final control model has dynamic response capabilities and can achieve collaborative optimization control from cluster to equipment while meeting the grid safety constraints.

[0105] This embodiment provides a method for constructing an energy storage cluster control model, the method comprising: obtaining multiple source load powers and multiple energy storage device data; wherein the multiple source load powers refer to the powers of multiple source load devices in a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of multiple energy storage devices; constructing a set based on the multiple source load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between the reference probability distribution and the true probability distribution, the reference probability distribution refers to the fluctuation law of the multiple source load powers, and the true probability distribution refers to the power of the multiple source load devices The fluctuation pattern of the energy storage rate within a preset second time period, where the start time of the second time period is later than the current time; clustering is performed based on the data of multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each energy storage cluster; wherein each energy storage cluster includes multiple energy storage devices, and the adjustable load curve is used to represent the load variation of each energy storage cluster over time; a model is constructed based on the confidence set, the adjustable load curve and multiple preset power grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control multiple energy storage clusters and multiple energy storage devices included in each energy storage cluster. A method for constructing an energy storage cluster control model achieves the following technical effects: by constructing a confidence set based on historical source-load power data, the probability distribution deviation boundary of power fluctuations is quantified, replacing the traditional deterministic model's reliance on precise probability distribution, making the control plan robust to random source-load fluctuations and reducing frequent instruction adjustments due to operating condition deviations; by forming multiple energy storage clusters and their adjustable load curves through cluster analysis, and combining the confidence set to quantify source-load uncertainty, the energy storage cluster control model can not only perform overall optimization control at the cluster level, but also perform precise charge and discharge control and response timing management for each energy storage device in each energy storage cluster based on the adjustable load curve, thereby achieving full-level refined control from cluster to device, thereby solving the problem of low control efficiency of energy storage cluster control models in the existing technology.

[0106] In one possible design, S202 constructs a set based on multiple source and load powers to obtain a confidence set, including:

[0107] S2021. Perform probability distribution fitting based on multiple source load powers to obtain a reference probability distribution.

[0108] Specifically, by using machine learning techniques such as maximum likelihood estimation and kernel density estimation to analyze and fit multiple source load power data, we can obtain their reference probability distribution. This reference probability distribution is used to describe the historical fluctuation characteristics of source load power, providing a basis for the subsequent construction of confidence sets, thereby helping to quantify the uncertainty of future power fluctuations.

[0109] For example, to construct a reference probability distribution Assuming that the system operates on a daily basis, cluster i records the power of each source and load device in a specific period of time in the past Q days. Suppose that in time period t, the power of the kth source and load device in day Q of cluster i is By discretizing the historical data, the frequency graph of the source-load device can be constructed and the reference probability distribution can be obtained by fitting. The sample space is divided into R levels, corresponding to the R values ​​that the random variable may take. Finally, the reference probability distribution is obtained. The probability of the rth level in As shown in formula (1):

[0110]

[0111] Among them, 1 {·} is an indicator function, which takes the value 1 when the {·} condition is met, otherwise it takes the value 0. r ,s r+1 ) represents the interval range of the rth discrete level, This is the frequency level at which the sample falls.

[0112] S2022. Construct a set based on the reference probability distribution and a preset confidence radius to obtain a confidence set; wherein the confidence radius is used to represent the maximum degree of deviation between the reference probability distribution and the true probability distribution.

[0113] Specifically, by applying machine learning methods based on the reference probability distribution and combining them with a preset confidence radius, we can construct a set containing all possible true probability distributions, namely the confidence set. This confidence set is used to quantify and limit the maximum deviation between the reference probability distribution and the true probability distribution, thereby providing more robust decision support in the energy storage cluster control model and resisting the uncertainty of source and load power fluctuations.

[0114] For example, define a set of random variables ξ={ξ i,t,k}(i∈I,t∈T,k∈K), where ξ i,t,k represents the power of the kth type of source and load equipment in cluster i during time period t. i,t,k As a set of random variables with the same sample space Ω, each of which obeys the unknown true probability distribution P i,t,k For each cluster’s source-load power distribution in each time period, a confidence set is constructed to ensure that the true distribution is included in the set with a high degree of confidence. The corresponding confidence set of the k-th source-load device in cluster i in time period t is The definition is as shown in formula (2):

[0115]

[0116] in, represents the reference probability distribution constructed by multiple source load powers, is the selected distribution distance metric function, reflecting and P i,t,k The parameter θ is the confidence set radius, which limits the maximum deviation between the reference distribution and the true distribution.

[0117] The technical effect of this solution in this embodiment is: by fitting the reference probability distribution based on the historical source-load power data, and constructing a confidence set in combination with the preset confidence radius, the maximum allowable deviation range between the reference probability distribution and the future true probability distribution is accurately quantified, so that the control model can actively adapt to the random fluctuation characteristics of the source-load power, effectively improve the robustness and adaptability of the control strategy, and reduce frequent scheduling adjustments caused by source-load uncertainty.

[0118] In one possible design, S203 , clustering is performed based on the data of multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each energy storage cluster, including:

[0119] S2031. Construct a model based on the data of multiple energy storage devices to obtain multiple energy storage device models; wherein each energy storage device model is used to represent the operating characteristics of each energy storage device.

[0120] Specifically, by analyzing the data of multiple energy storage devices, physical modeling or data-driven modeling methods, such as equivalent circuit models and machine learning models, can be used to construct operating characteristic models for each energy storage device. These models are used to accurately describe and predict the charging and discharging behavior of energy storage devices under different conditions, providing a basis for subsequent power solutions and cluster optimization.

[0121] S2032. Solve the model of each energy storage device to obtain the charge and discharge power of each energy storage device.

[0122] Specifically, by inputting each energy storage device model into an optimization algorithm or numerical solver, and combining it with current operating conditions and constraints, such as grid demand and device limitations, the optimal charging and discharging power of each energy storage device can be solved. These powers are used to guide the actual operation of the device, thereby achieving effective management of the energy storage device and stable support for the power grid.

[0123] For example, we can first establish a refined mathematical model that takes into account battery characteristics, operating constraints, and equipment status. This model includes key parameters such as state of charge limits, maximum charge and discharge rates, and cycle life loss. Then, we can use a constrained nonlinear optimization algorithm with grid dispatch requirements as the objective function. While meeting the boundary conditions for safe equipment operation, we can solve the charge and discharge power curve. During the solution process, we can monitor operating indicators such as equipment temperature and aging degree in real time, and dynamically adjust the power limit. The final output charge and discharge power plan not only meets the requirements of the upper-level cluster control, but also ensures that individual energy storage devices operate within the appropriate efficiency range.

[0124] S2033: Aggregate the charge and discharge powers of the multiple energy storage devices to obtain multiple energy storage clusters and the charge and discharge powers of the multiple energy storage clusters.

[0125] Specifically, the charging and discharging power data of multiple energy storage devices can be aggregated and integrated, and these devices can be divided into different energy storage clusters based on criteria such as geographical location, device type or operating characteristics. The total charging and discharging power of each cluster can be calculated. These clusters and their power data can be used to optimize cluster-level energy scheduling and management, improving the efficiency and responsiveness of the overall system.

[0126] For example, based on the electrical connection relationship and geographical distribution between devices, a clustering algorithm that considers network topology can be used to divide devices with close physical distances and similar electrical characteristics into initial clusters. Then, based on the real-time charging and discharging power capabilities of each device, power superposition calculations can be performed within the cluster to obtain the overall adjustable power range of the cluster. A dynamic consistency verification algorithm can then be used to verify the coordination of the power responses of each device in the cluster, and devices that do not meet the synchronization requirements can be reorganized into clusters. Ultimately, multiple energy storage clusters with stable regulation capabilities are formed, and a total charging and discharging power curve for each cluster during the scheduling cycle is generated. This curve not only retains the overall regulation capability of the cluster, but also ensures that each internal device operates within a safe operating range.

[0127] S2034. Cluster the charge and discharge powers of multiple energy storage clusters to obtain an adjustable load curve for each energy storage cluster.

[0128] Specifically, clustering algorithms can be applied to analyze the charging and discharging power data of multiple energy storage clusters, classify clusters with similar charging and discharging characteristics, and generate adjustable load curves for each energy storage cluster based on these classification results. These curves are used to describe the load regulation capability and response characteristics of the cluster, thereby supporting accurate energy scheduling and optimization decisions.

[0129] For example, we first construct a single energy storage device model based on the characteristics of the energy storage device layer. The energy storage device interacts with the power grid through charging and discharging, and its key calculation parameter is the state of charge. ESDefined as charging positive and discharging negative, assuming the charging efficiency of the energy storage device is n c and discharge efficiency n d It remains unchanged during operation, and the definition formula (3) is:

[0130]

[0131] Where: is the state of charge of the energy storage device at time t, is the rated capacity of the i-th energy storage device, and E i,t are the charging and discharging power and battery capacity of the i-th energy storage device at time t, respectively, k i,t is the operating state of the i-th energy storage device at time t (the charging state is 1, the discharging state is -1, and the non-charging and non-discharging state is 0), is the actual interaction power between the i-th energy storage device and the grid; E i,T and are the actual and expected battery capacity after regulation of the i-th energy storage device, and is the maximum charge / discharge power, maximum battery capacity and minimum battery capacity of the i-th energy storage device, η c and η d are the charging efficiency and discharging efficiency of the energy storage device respectively, Δt is the time step, t∈T, T is the control time window.

[0132] Considering the poor control flexibility and small control potential of distributed energy storage units, distributed energy storage with similar control potential and operational space are aggregated into a whole, and the edge intelligent terminal acts as an intermediary to participate in the dispatch of the distribution network through cloud-edge collaboration. For energy storage devices with similar dispatchable potential, by introducing k i,t Formula (3) is applicable to any control time window, and the requirement that multiple variables belong to the same domain is met. The Minkowski sum is used to aggregate them to obtain the power domain and electricity domain boundaries of the distributed energy storage cluster, as shown in Formula (4):

[0133]

[0134] Where: is the charging and discharging power of the energy storage cluster ν at time t, and the charging and discharging state k of each energy storage device i,t and charge and discharge rated power Directly related, and is the upper and lower power limit of the energy storage cluster ν at time t, I ν is the total amount of energy storage in the distributed energy storage cluster ν.

[0135] Finally, the power regulation data of the distributed energy storage cluster is divided into multiple clusters by K-Means (K-means clustering algorithm). Let the cluster to which the distributed energy storage regulation day belongs be C m , the number of adjustable load curves of the cluster is l, and the cluster center is Z m , t=[t start ,t end ]The load curve during the regulation period is x t , define the adjustable load curve ΔP of the i-th energy storage cluster in period t i (t) can be expressed as shown in formula (5):

[0136]

[0137] Where: x t is the load baseline for the adjustment period, x ψ is the load curve with the longest Euclidean distance within the same cluster as the load baseline, x q is the load curve in the same cluster as the load baseline, ΔP i (t) min is the minimum adjustable load curve, which is expressed as the difference between the baseline load during the adjustment period and the lowest load in the cluster to which it belongs, ΔP i (t) max is the maximum adjustable load curve, expressed as the difference between the baseline load during the adjustment period and the cluster center of the cluster with the minimum load value. Equation (5) represents the energy storage clustering form, which reduces the variable dimension during collaborative optimization. The greater the number of distributed energy storage in the cluster, the larger its dispatchable operating domain and the higher its flexibility.

[0138] The technical effect of this solution in this embodiment is: by constructing an energy storage device model and solving the charging and discharging power, and then clustering and aggregating based on the power characteristics, an energy storage cluster with a clear and adjustable load curve is formed. This not only retains the operating characteristics of a single energy storage device, but also reduces the complexity of the optimization problem through clustering processing, so that the control model can take into account both computing efficiency and control accuracy, and realize multi-level coordinated control of energy storage devices from single units to clusters.

[0139] In one possible design, the energy storage cluster control model includes an upper-level planning model and a lower-level operation model. In step S204, a model is constructed based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain the energy storage cluster control model, including:

[0140] S2041. Construct a model based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain an upper-level planning model; wherein the upper-level planning model is used to represent the interaction process between multiple energy storage clusters and the external grid, and the impact of the interaction process on the cost of the external grid.

[0141] Specifically, by integrating confidence sets, adjustable load curves, and preset grid parameters, an upper-level planning model can be constructed using optimization modeling techniques such as linear programming and mixed integer programming. This model is used to simulate and analyze the interaction between multiple energy storage clusters and the external power grid, and evaluate the impact of different control strategies on grid operating costs, thereby providing a basis for formulating efficient energy storage control plans.

[0142] For example, the upper-level planning model is mainly used to determine the output power of the distributed energy storage cluster and the amount of electricity purchased from the upper level to minimize the total operating cost of the distribution network. The output power of the distributed energy is defined by the optimization variable. Therefore, the objective function and multiple constraints of the upper-level planning model can be defined as shown in Equation (6):

[0143]

[0144] sT ele,t +P sto,o,t +P DG,g,t =P cut,u (t)+P trf,u (t)+P sel,t

[0145]

[0146] P cut,u (t) = α cut,u (t)μ cut,u (t)P pri,cut,u (t)

[0147]

[0148] Where: C ele,t and P ele,t are the electricity price and power purchased by the distribution system from the upper power grid, C sto,o,t and P sto,o,t are the unit response cost and response power of the distributed energy storage cluster, C DG,g,t and P DG,g,t are the unit power generation cost and power generation of distributed power generation, C u,t is the user electricity price, P cut,u (t) and P trf,u (t) are the load power that users can reduce and transfer, P sel,t is the electricity sold on the user side, P pri,cut,u (t) and P cut,u (t) are the power before and after load reduction, α cut,u (t) is the 1 / 0 flag parameter for whether to cut, μ cut,u (t) is the reduction coefficient, τ, and They are the maximum total load reduction time, reduction start time, and the maximum and minimum time limits for a single load reduction. Represents the probability distribution P of the source-load scenario i,t,k Under expectations, Refers to the user's maximum response load capacity.

[0149] S2042. Solve the upper-level planning model to obtain multiple first plans; wherein the multiple first plans refer to control plans for multiple energy storage clusters.

[0150] Specifically, the upper-level planning model can be solved by applying optimization algorithms, such as linear programming solvers and genetic algorithms, to determine the optimal control strategy under given constraints, thereby generating multiple first plans. These plans are used to guide the control activities of each energy storage cluster to achieve interaction with the power grid, reduce operating costs and improve the overall efficiency of the system.

[0151] For example, in the second plan generation stage, we can first construct a robust optimization problem based on the confidence set that considers the worst fluctuation scenario, set the minimization of the grid operation cost as the objective function, and incorporate safety conditions such as node voltage constraints and line transmission capacity. Then, we use the decomposition and coordination algorithm to decompose the large-scale optimization problem into multiple sub-problems, and solve the power allocation plan of each energy storage cluster through alternating iterations. In the solution process, slack variables are introduced to deal with uncertainty constraints, and convergence criteria are set to ensure the feasibility of the solution. Finally, the optimal power adjustment plan for each cluster within the scheduling period is output as the first plan.

[0152] During the second plan execution phase, each energy storage cluster controller, based on the refined control instructions it receives, first diagnoses and evaluates the real-time operating status of each energy storage device within the cluster. It then dynamically allocates specific output tasks to each device, combining parameters such as the device's remaining capacity and charge / discharge efficiency. A hierarchical control strategy is employed during execution, with the cluster's master controller providing overall coordination. Each device precisely executes power adjustment instructions, and its response is continuously monitored via a real-time data acquisition network. When a deviation between the actual output and the planned value is detected, an adaptive adjustment mechanism is immediately activated, intelligently optimizing the power distribution ratio between devices while ensuring the overall cluster control objectives. Safety protection logic is also established to anticipate and intervene in potential abnormal situations such as device overloads and capacity limits, ensuring a safe and reliable control process. The system can also establish a feedback channel to transmit execution results back to the upper-level system in real time, providing a basis for subsequent optimization of the control plan.

[0153] S2043. Construct a model based on multiple first plans, multiple energy storage device data, and multiple grid parameters to obtain a lower-level operation model; wherein the lower-level operation model is used to optimize the response strategy and cluster allocation of multiple energy storage devices under each energy storage cluster to minimize the response cost and voltage fluctuation of multiple energy storage devices under each energy storage cluster.

[0154] Specifically, by combining multiple first plans, specific data of energy storage devices, and operating parameters of the power grid, an optimization modeling method can be used to construct a lower-level operating model. This model is used to refine and optimize the device response strategy and resource allocation within each energy storage cluster to minimize the device response cost and voltage fluctuations, thereby ensuring the efficient operation of the energy storage system and the stability of the power grid.

[0155] For example, the upper-level planning model can be used to obtain the target response time, response capacity and other parameters of each energy storage cluster. The decision variables of the lower-level operation model are the energy storage cluster allocation of each energy storage device and its response time period in the energy storage cluster. and Two state variables are used as decision variables, and Both contain the allocation and response start time information of each energy storage device, which are defined as follows:

[0156] 0-1 variable, representing the allocation result and starting response period of energy storage device i. If energy storage device i is allocated to energy storage cluster g and just starts responding in period t, then otherwise

[0157] 0-1 variable, representing the allocation result and current response status of energy storage device i. If energy storage device i is allocated to energy storage cluster g and is in the response state at time t, then otherwise

[0158] where t∈Ω T , i∈Ω I , g∈Ω G Ω T ,Ω I and Ω G They are response time period set, response energy storage device set and energy storage cluster set.

[0159] Node voltage is an important indicator for measuring grid stability. The access of distributed renewable energy will aggravate voltage fluctuations. Therefore, this application uses voltage stability to evaluate the ability of distributed energy storage to smooth voltage fluctuations during operation. At the same time, energy storage equipment actually has a priority in the regulation process. Energy storage equipment with lower average cost and better technical performance should be called first. Therefore, this application sets weight factor parameters to reflect the difference in the probability of calling each energy storage device when it actually responds. In summary, the model aims to minimize the estimated total response cost and voltage stability of each energy storage device in the substation, which can be expressed as shown in formula (7):

[0160]

[0161] Where: c i is the unit capacity response cost of energy storage device i, p i is the response capacity of energy storage device i, ω g is the weight factor of energy storage device i, is the average voltage within the response period, U n,t is the voltage value at the grid node at time t.

[0162] The constraints of the lower-level running model are as follows:

[0163] 1) Energy storage device allocation and call frequency constraints

[0164] Any energy storage device i can be assigned to at most one energy storage cluster and can only respond once during the entire period. These two constraints can be expressed in one expression, as shown in Equation (8):

[0165]

[0166] 2) Association constraints between energy storage device initial response period variables and response state variables

[0167] For any energy storage device i, and There is a constraint relationship, as shown in formula (9):

[0168]

[0169] when When , formula (3) is equivalent to when and When , formula (3) is equivalent to when and When , formula (3) always holds. Combining formulas (2) and (3), we can express and The relationship between them.

[0170] 3) Energy storage device response duration constraints

[0171]

[0172] Where, T i max is the maximum sustainable response time of energy storage device i.

[0173] 5) Energy storage cluster response capacity constraints

[0174] The aggregated devices in the energy storage cluster must meet the overall response capacity requirements. Right now

[0175]

[0176] Where, γ i is the response credibility of energy storage device i.

[0177] 6) Energy storage cluster response rate constraints

[0178] After aggregation, each device in the energy storage cluster must meet its overall response rate and recovery rate requirements, that is,

[0179]

[0180] Where: r i U and r i D are the response rate and recovery rate of energy storage device i, and are the overall response rate and recovery rate requirements of the energy storage cluster g respectively.

[0181] The above constitutes the mathematical model for distributed energy storage cluster control proposed in this application. This model is a standard 0-1 planning model and can be solved using an optimization toolkit.

[0182] The technical effect of this solution in this embodiment is as follows: by constructing a two-layer optimization architecture consisting of an upper-layer planning model and a lower-layer operation model, the upper layer achieves global economic optimization of the interaction between the energy storage cluster and the power grid based on confidence sets and adjustable load curves, while the lower layer performs fine-grained regulation of the response strategy and cluster allocation for each energy storage device, ensuring the overall operational efficiency at the power grid level while achieving precise control at the device level, thereby improving the systematicity and coordination of energy storage cluster regulation in an environment of source and load uncertainty.

[0183] Figure 3 Schematic diagram of the process of constructing the energy storage cluster control model provided in the embodiment of the present application Figure 2 In this embodiment, Figure 2Based on the provided embodiment, the method for constructing the energy storage cluster control model is further explained. The method for constructing the energy storage cluster control model includes:

[0184] S301. Acquire multiple source-load powers and multiple energy storage device data; wherein, the multiple source-load powers refer to the powers of multiple source-load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of multiple energy storage devices.

[0185] S302. Construct a set based on multiple source-load powers to obtain a confidence set; wherein the confidence set is used to indicate the degree of deviation between the reference probability distribution and the true probability distribution, the reference probability distribution refers to the fluctuation pattern of the power of multiple source-load devices, and the true probability distribution refers to the fluctuation pattern of the power of multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time.

[0186] S303. Clustering is performed based on the data of multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each energy storage cluster; wherein each energy storage cluster includes multiple energy storage devices, and the adjustable load curve is used to represent the load variation of each energy storage cluster over time.

[0187] S304. Construct a model based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control multiple energy storage clusters and multiple energy storage devices included in each energy storage cluster.

[0188] S301-S304 are similar to S201-S204 and will not be described in detail in this embodiment.

[0189] S305: Solve the lower-level operation model to obtain multiple second plans; wherein the multiple second plans refer to control plans for multiple energy storage devices under each energy storage cluster.

[0190] Specifically, numerical optimization algorithms, such as dynamic programming and mixed integer linear programming, can be applied to solve the underlying operating model to determine the control strategy for each energy storage device under specific operating conditions, thereby generating multiple second plans. These plans are used to guide the specific operation of the energy storage devices to achieve efficient energy management and response optimization within the cluster, ensuring that the system achieves the expected performance goals in actual operation.

[0191] For example, in the second plan generation stage, the first plan issued by the upper layer can first be decomposed into power regulation tasks for each energy storage cluster. Combined with the real-time status data of the equipment, an optimization model that takes into account response speed, regulation accuracy and loss cost is established. A distributed computing method is used to allow each cluster to independently solve the power allocation plan for each internal device, and coordinate the timing coordination across clusters through a consistency algorithm. A rolling time domain optimization strategy is introduced in the solution process, and the power instructions are dynamically corrected according to device feedback. Finally, a second plan is generated that is accurate to the charging and discharging timing and power intensity of a single device.

[0192] During the second plan execution phase, each cluster can further decompose its assigned power instructions to its respective energy storage devices. A priority scheduling algorithm is used to determine the output curve for each device based on parameters such as the device's current state of charge, health, and operating limits. Local controllers perform refined power control while monitoring device operating status, automatically initiating protective adjustments when out-of-limit issues occur. The system establishes a multi-layered data acquisition network, providing real-time feedback on device-level execution to upper-level models, forming a complete closed-loop control system from the grid to the device.

[0193] S306. Obtain execution results and operation results; wherein, the execution results refer to the operation data after the multiple energy storage devices under each energy storage cluster respectively execute the multiple second plans, and the operation results refer to the actual operation data of the multiple energy storage devices under each energy storage cluster.

[0194] Specifically, through real-time monitoring and data acquisition systems, the operating data of each energy storage device after executing multiple second plans can be collected as execution results, and the status and performance data of the device in the actual operating environment can be recorded as operation results. These results are used to evaluate the execution effect of the plan and the actual response of the system, and provide necessary feedback information for subsequent model calibration and optimization.

[0195] S307: Calculate the deviation between the execution result and the operation result, and iteratively optimize the upper-level planning model and the lower-level operation model according to the deviation until the deviation is less than a preset threshold.

[0196] Specifically, by comparing the execution results with the operating results, the differences in key performance indicators can be calculated, and the deviation information can be used to adjust and correct the parameters and structure of the upper-level planning model and the lower-level operating model. Iterative optimization algorithms such as gradient descent and Bayesian optimization can be used to continuously update the model until the deviation is less than the preset threshold. This process is used to improve the accuracy and robustness of the model and ensure the efficient and stable operation of the energy storage system in a dynamic environment.

[0197] The technical effect of this solution in this embodiment is: by establishing a closed-loop control mechanism of plan execution-data collection-deviation calculation-model optimization, the deviation between the second plan execution result and the actual operation data is compared in real time, and the upper and lower model parameters are dynamically iteratively optimized based on the deviation, so that the energy storage cluster control system has online self-correction capabilities, effectively eliminating the control errors caused by source and load fluctuations and changes in equipment characteristics, and improving the adaptability and robustness of the system in a dynamic operating environment.

[0198] In one possible design, S307, iteratively optimizing the upper-level planning model and the lower-level operation model according to the deviation until the deviation is less than a preset threshold, includes:

[0199] S3071. Input the deviation into a preset deviation source determination model to obtain the deviation source.

[0200] Specifically, a deviation source determination model can be constructed. The model uses a machine learning algorithm to input the calculated deviation into it, analyze the pattern and characteristics of the deviation, and identify possible sources of deviation, such as inaccurate model parameters, changes in the external environment, etc. This process is used to clarify the root cause of the deviation and provide a basis for the subsequent selection of appropriate deviation adjustment strategies, thereby improving the optimization efficiency and effectiveness of the model.

[0201] S3072. Match the deviation source with multiple preset deviation adjustment strategies to obtain a target deviation adjustment strategy; wherein the target deviation adjustment strategy refers to any one of the multiple deviation adjustment strategies.

[0202] Specifically, a strategy matching system can be established to compare and match the identified sources of deviation with multiple preset deviation adjustment strategies, and use methods such as rule engines or decision trees to select the most appropriate adjustment strategy as the target deviation adjustment strategy. This process is used to ensure that the most effective optimization measures are taken for different types of deviations, thereby improving the adjustment efficiency of the model and the overall performance of the system.

[0203] S3073. Based on the target deviation adjustment strategy, iteratively optimize the upper-level planning model and the lower-level operation model until the deviation is less than the threshold.

[0204] Specifically, the upper-level planning model and the lower-level operation model can be gradually corrected and optimized by applying the optimization methods and parameter adjustment steps defined in the target deviation adjustment strategy. Iterative algorithms such as gradient descent and genetic algorithms are used to continuously update the parameters and structure of the model until the calculated deviation is less than the preset threshold. This process is used to improve the accuracy and adaptability of the model, ensuring that the energy storage system can effectively respond to changes and maintain efficient operation in actual operation.

[0205] The technical effect of this solution in this embodiment is to achieve precise and targeted optimization of model parameters through intelligent diagnosis of deviation sources and a strategy matching mechanism. First, the model identifies the main error causes using the deviation sources, and then automatically matches the corresponding adjustment strategy. This makes the model iterative optimization process targeted and efficient, avoiding the computational burden of blindly adjusting global parameters while improving model convergence speed and control accuracy, ultimately achieving rapid deviation convergence and stable system operation.

[0206] An embodiment of the present application also provides a system for constructing an energy storage cluster control model, including: a cloud, an edge server, multiple source-load devices, and multiple energy storage devices.

[0207] Cloud for implementation Figure 2 or Figure 3 A method for constructing an energy storage cluster control model.

[0208] The edge server is used to collect multiple source load power and multiple energy storage device data, and send the multiple source load power and multiple energy storage device data to the cloud, and send multiple first plans to multiple energy storage clusters, and send multiple second plans to multiple energy storage devices under each energy storage cluster; wherein, multiple energy storage clusters, multiple first plans and multiple second plans are based on Figure 2 or Figure 3 It is calculated based on the construction method of the energy storage cluster control model.

[0209] The plurality of source load devices are used to send a plurality of source load powers to the edge server.

[0210] The multiple energy storage devices are used to send multiple energy storage device data to the edge server, receive the multiple second plans sent by the edge server, and execute the multiple second plans.

[0211] Figure 4 This is a schematic diagram of the structure of the device for constructing the energy storage cluster control model provided in the embodiment of the present application. Figure 4 As shown, the construction device of the energy storage cluster control model includes:

[0212] The first acquisition module 401 is used to obtain multiple source-load powers and multiple energy storage device data; wherein, the multiple source-load powers refer to the power of multiple source-load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of multiple energy storage devices.

[0213] The set construction module 402 is used to construct a set based on multiple source load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between the reference probability distribution and the true probability distribution, the reference probability distribution refers to the fluctuation pattern of the power of multiple source load devices, and the true probability distribution refers to the fluctuation pattern of the power of multiple source load devices within a preset second time period, and the start time of the second time period is later than the current time.

[0214] Clustering module 403 is used to cluster the data of multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each energy storage cluster; wherein each energy storage cluster includes multiple energy storage devices, and the adjustable load curve is used to represent the load change of each energy storage cluster over time.

[0215] The model construction module 404 is used to construct a model based on the confidence set, the adjustable load curve and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control multiple energy storage clusters and the multiple energy storage devices included in each energy storage cluster.

[0216] In a possible implementation, the set construction module 402 includes:

[0217] The fitting unit is used to perform probability distribution fitting based on multiple source load powers to obtain a reference probability distribution.

[0218] The set construction unit is used to construct a set according to the reference probability distribution and a preset confidence radius to obtain a confidence set; wherein the confidence radius is used to represent the maximum deviation between the reference probability distribution and the true probability distribution.

[0219] In a possible implementation, the clustering module 403 includes:

[0220] The first construction unit is used to construct a model based on the data of multiple energy storage devices to obtain multiple energy storage device models; wherein each energy storage device model is used to represent the operating characteristics of each energy storage device.

[0221] The first solving unit is used to solve the model of each energy storage device to obtain the charging and discharging power of each energy storage device.

[0222] an aggregation unit, configured to aggregate the charge and discharge power of multiple energy storage devices to obtain multiple energy storage clusters and the charge and discharge power of the multiple energy storage clusters;

[0223] The clustering unit is used to cluster the charging and discharging powers of multiple energy storage clusters to obtain an adjustable load curve for each energy storage cluster.

[0224] In one possible implementation, the energy storage cluster control model includes an upper-level planning model and a lower-level operation model. The model construction module 404 includes:

[0225] The second construction unit is used to construct a model based on the confidence set, the adjustable load curve, and multiple preset grid parameters to obtain an upper-level planning model; wherein the upper-level planning model is used to represent the interaction process between multiple energy storage clusters and the external grid, and the impact of the interaction process on the cost of the external grid.

[0226] The second solving unit is used to solve the upper-level planning model to obtain multiple first plans; wherein the multiple first plans refer to control plans of multiple energy storage clusters.

[0227] The third construction unit is used to construct a model based on the multiple first plans, the multiple energy storage device data and the multiple grid parameters to obtain a lower-level operation model; wherein the lower-level operation model is used to optimize the response strategy and cluster allocation of the multiple energy storage devices under each energy storage cluster to minimize the response cost and voltage fluctuation of the multiple energy storage devices under each energy storage cluster.

[0228] In one possible implementation, the device for constructing an energy storage cluster control model further includes:

[0229] The solving module is used to solve the lower-level operation model to obtain multiple second plans; wherein the multiple second plans refer to the control plans of multiple energy storage devices under each energy storage cluster.

[0230] The second acquisition module is used to obtain execution results and operation results; wherein, the execution results refer to the operation data after the multiple energy storage devices under each energy storage cluster respectively execute multiple second plans, and the operation results refer to the actual operation data of the multiple energy storage devices under each energy storage cluster.

[0231] The optimization module is used to calculate the deviation between the execution result and the operation result, and iteratively optimize the upper-level planning model and the lower-level operation model according to the deviation until the deviation is less than the preset threshold.

[0232] In a possible implementation, the optimization module includes:

[0233] The source determination unit is used to input the deviation into a preset deviation source determination model to obtain the deviation source.

[0234] The strategy determination unit is used to match the deviation source with a plurality of preset deviation adjustment strategies to obtain a target deviation adjustment strategy; wherein the target deviation adjustment strategy refers to any one of the plurality of deviation adjustment strategies.

[0235] The optimization unit is used to iteratively optimize the upper-level planning model and the lower-level operation model based on the target deviation adjustment strategy until the deviation is less than a threshold.

[0236] The energy storage cluster control model construction device provided in this embodiment can be executed Figure 2 and Figure 3 The technical solution of the embodiment of the method for constructing an energy storage cluster control model is shown, and its implementation principle and technical effect are similar to those of the embodiment of the method for constructing an energy storage cluster control model. Figure 2 and Figure 3 The embodiment of the method for constructing an energy storage cluster control model shown is similar and will not be described in detail here.

[0237] Figure 5 Schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. Figure 5 As shown, the electronic device 50 includes: at least one processor 510 and a memory 520. The electronic device also includes a communication component 530. The processor 510, the memory 520 and the communication component 530 are connected via a bus 540.

[0238] During the specific implementation process, at least one processor 510 executes the computer-executable instructions stored in the memory 520, so that the at least one processor 510 is used to implement the method for constructing an energy storage cluster control model of the above embodiment.

[0239] The specific implementation process of the processor 510 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0240] In the above embodiment, it should be understood that the processor 510 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention can be directly implemented by a hardware processor or implemented by a combination of hardware and software modules in the processor.

[0241] The memory 520 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk storage.

[0242] The bus 540 may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus 540 may be classified as an address bus, a data bus, a control bus, etc. For ease of illustration, the bus 540 in the drawings of this application is not limited to a single bus or a single type of bus.

[0243] The above-mentioned functions implemented by the electronic device and the main control device have introduced the solutions provided by the embodiments of the present invention. It can be understood that in order to implement the above-mentioned functions, the electronic device or the main control device includes hardware structures and / or software modules corresponding to the execution of each function. In combination with the units and algorithm steps of the various examples described in the embodiments disclosed in the embodiments of the present invention, the embodiments of the present invention can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed in the form of hardware or computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the technical solution of the embodiments of the present invention.

[0244] The present application also provides a computer-readable storage medium storing computer-executable instructions. When executed by a processor, the computer-executable instructions implement the method for constructing an energy storage cluster control model described in the aforementioned embodiment. In the specific implementation of the aforementioned method for constructing an energy storage cluster control model, each module may be implemented as a processor.

[0245] The above-mentioned readable storage medium can be implemented by any type of volatile or non-volatile memory 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. The readable storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0246] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in an electronic device or a main control device.

[0247] An embodiment of the present application further provides a computer program product, including a computer program. When the computer program is executed by a processor, it is used to implement a method for constructing an energy storage cluster control model of the above embodiment.

[0248] The computer program is stored in a readable storage medium. At least one processor can read the computer program from the readable storage medium, and at least one processor can execute the computer program to perform the solution provided in any of the above embodiments.

[0249] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0250] So far, the technical solution of the present application has been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the scope of protection of the present application is obviously not limited to these specific embodiments. The above embodiments are only used to illustrate the technical solution of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solution to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for constructing an energy storage cluster control model, characterized in that: include: Acquire multiple source load powers and multiple energy storage device data; wherein the multiple source load powers refer to the power of multiple source load devices within a preset first time period, the end time of the first time period is earlier than the current time, and the multiple energy storage device data refer to the current operating data of the multiple energy storage devices; A confidence set is constructed based on the multiple source-load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between a reference probability distribution and a true probability distribution, the reference probability distribution refers to the fluctuation pattern of the multiple source-load powers, and the true probability distribution refers to the fluctuation pattern of the powers of the multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time; Clustering the data of the multiple energy storage devices to obtain multiple energy storage clusters and an adjustable load curve for each of the energy storage clusters; wherein each of the energy storage clusters includes multiple energy storage devices, and the adjustable load curve is used to represent the load change of each of the energy storage clusters over time; A model is constructed based on the confidence set, the adjustable load curve and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control the multiple energy storage clusters and the multiple energy storage devices included in each of the energy storage clusters.

2. The method for constructing an energy storage cluster control model according to claim 1, characterized in that: The step of constructing a set based on the multiple source-load powers to obtain a confidence set includes: Performing probability distribution fitting according to the multiple source load powers to obtain the reference probability distribution; A set is constructed based on the reference probability distribution and a preset confidence radius to obtain the confidence set; wherein the confidence radius is used to represent the maximum degree of deviation between the reference probability distribution and the true probability distribution.

3. The method for constructing an energy storage cluster control model according to claim 1, characterized in that: The clustering of the plurality of energy storage device data to obtain a plurality of energy storage clusters and an adjustable load curve of each of the energy storage clusters includes: Model building is performed based on the plurality of energy storage device data to obtain a plurality of energy storage device models; wherein each of the energy storage device models is used to represent the operating characteristics of each of the energy storage devices; Solving each of the energy storage device models to obtain the charge and discharge power of each of the energy storage devices; Aggregating the charge and discharge powers of the multiple energy storage devices to obtain the multiple energy storage clusters and the charge and discharge powers of the multiple energy storage clusters; The charge and discharge powers of the multiple energy storage clusters are clustered to obtain an adjustable load curve for each of the energy storage clusters.

4. The method for constructing an energy storage cluster control model according to claim 1, characterized in that: The energy storage cluster control model includes an upper-level planning model and a lower-level operation model. The model is constructed based on the confidence set, the adjustable load curve, and a plurality of preset grid parameters to obtain the energy storage cluster control model, including: A model is constructed based on the confidence set, the adjustable load curve, and a plurality of preset grid parameters to obtain the upper-level planning model; wherein the upper-level planning model is used to represent the interaction process between the plurality of energy storage clusters and the external grid, and the impact of the interaction process on the cost of the external grid; Solving the upper-level planning model to obtain a plurality of first plans; wherein the plurality of first plans refer to control plans for the plurality of energy storage clusters; A model is constructed based on the multiple first plans, the multiple energy storage device data and the multiple power grid parameters to obtain the lower-level operation model; wherein the lower-level operation model is used to optimize the response strategy and cluster allocation of the multiple energy storage devices under each of the energy storage clusters to minimize the response cost and voltage fluctuation of the multiple energy storage devices under each of the energy storage clusters.

5. The method for constructing an energy storage cluster control model according to claim 4, characterized in that: After constructing the model according to the plurality of first plans, the plurality of energy storage device data, and the plurality of grid parameters to obtain the lower-layer operation model, the method further includes: Solving the lower-layer operation model to obtain a plurality of second plans; wherein the plurality of second plans refer to control plans for a plurality of energy storage devices under each of the energy storage clusters; Obtaining execution results and operation results; wherein the execution results refer to operation data after the multiple energy storage devices under each of the energy storage clusters respectively execute the multiple second plans, and the operation results refer to actual operation data of the multiple energy storage devices under each of the energy storage clusters; The deviation between the execution result and the operation result is calculated, and the upper-level planning model and the lower-level operation model are iteratively optimized according to the deviation until the deviation is less than a preset threshold.

6. The method for constructing an energy storage cluster control model according to claim 5, characterized in that: The iteratively optimizing the upper-layer planning model and the lower-layer operation model according to the deviation until the deviation is less than a preset threshold includes: Inputting the deviation into a preset deviation source determination model to obtain the deviation source; Matching the deviation source with a plurality of preset deviation adjustment strategies to obtain a target deviation adjustment strategy; wherein the target deviation adjustment strategy refers to any one of the plurality of deviation adjustment strategies; Based on the target deviation adjustment strategy, the upper-level planning model and the lower-level operation model are iteratively optimized until the deviation is less than the threshold.

7. A system for constructing an energy storage cluster control model, comprising: Cloud, edge servers, multiple source-load devices, and multiple energy storage devices; The cloud is used to implement the method for constructing the energy storage cluster control model according to any one of claims 1 to 6; The edge server is used to collect data on multiple source-load powers and multiple energy storage devices, and send the multiple source-load powers and the multiple energy storage device data to the cloud, as well as send multiple first plans to multiple energy storage clusters, and send multiple second plans to multiple energy storage devices under each of the energy storage clusters; wherein the multiple energy storage clusters, the multiple first plans, and the multiple second plans are calculated according to the method for constructing an energy storage cluster control model according to any one of claims 1 to 6; The plurality of source load devices are used to send the plurality of source load powers to the edge server; The multiple energy storage devices are used to send the multiple energy storage device data to the edge server, receive the multiple second plans sent by the edge server, and execute the multiple second plans.

8. A device for constructing an energy storage cluster control model, characterized in that: include: an acquisition module, configured to acquire data of multiple source-load powers and multiple energy storage devices; wherein the multiple source-load powers refer to the powers of multiple source-load devices within a preset first time period, the end time of the first time period being earlier than the current time, and the multiple energy storage device data refer to current operating data of multiple energy storage devices; a set construction module, configured to construct a set based on the multiple source-load powers to obtain a confidence set; wherein the confidence set is used to represent the degree of deviation between a reference probability distribution and a true probability distribution, the reference probability distribution refers to the fluctuation pattern of the multiple source-load powers, and the true probability distribution refers to the fluctuation pattern of the powers of the multiple source-load devices within a preset second time period, where the start time of the second time period is later than the current time; A clustering module, configured to cluster the plurality of energy storage device data to obtain a plurality of energy storage clusters and an adjustable load curve for each of the energy storage clusters; wherein each of the energy storage clusters includes a plurality of energy storage devices, and the adjustable load curve is configured to represent how the load of each energy storage cluster changes over time; A model construction module is used to construct a model based on the confidence set, the adjustable load curve and multiple preset grid parameters to obtain an energy storage cluster control model; wherein the energy storage cluster control model is used to control the multiple energy storage clusters and the multiple energy storage devices included in each of the energy storage clusters.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; When the processor executes the computer-executable instructions stored in the memory, it is used to implement the method for constructing the energy storage cluster control model according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method for constructing an energy storage cluster control model according to any one of claims 1 to 6.

11. A computer program product, characterized in that It includes a computer program, which, when executed by a processor, is used to implement the method for constructing an energy storage cluster control model according to any one of claims 1 to 6.

Images