A method and system for constructing a virtual energy storage model for large power grids

By constructing a virtual energy storage model for large power grids and utilizing power boundaries, power balance, and response rate constraints, the problem of inaccurate regulation capacity of existing virtual energy storage models in large power grids is solved. This achieves a precise mathematical representation of aggregated flexible resources and improves the accuracy and efficiency of power grid dispatch.

CN123116569APending Publication Date: 2026-11-17NORTHWEST BRANCH OF STATE GRID POWER GRID CO +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611243620.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-17
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing virtual energy storage models cannot accurately characterize the strong temporal coupling characteristics and global energy conservation characteristics of aggregated flexible resources in large power grids, leading to an overestimation or underestimation of regulation capabilities, which affects the accuracy and cost of power grid dispatch.

Method used

A virtual energy storage model for large power grids is constructed. By setting power boundary constraints, power balance constraints, and response rate constraints, a complete constraint system is formed. Virtual energy storage parameters are estimated using large power grid operation data, and an inverse reasoning framework of "data-characteristics-parameters" is established to achieve accurate mathematical representation of aggregated flexible resources.

Benefits of technology

It achieves a precise mathematical characterization of the aggregated flexibility resources of large power grids, avoids overestimation or underestimation of regulation capacity, provides a more accurate and reliable flexibility boundary, and provides more efficient utilization of flexibility resources for large power grid dispatch.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123116569A_ABST
    Figure CN123116569A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for constructing a virtual energy storage model for large power grids, relating to the field of power system automation technology. The method includes the following steps: In each time period, the range of decision variables is set to not exceed a set power range allowed for that time period, thus obtaining power boundary constraints; the decision variables for each time period within the target period are summed to obtain the cumulative deviation power, and the cumulative deviation power is set to not exceed a preset tolerance threshold, thus obtaining power balance constraints; the difference between the decision variables of two time periods separated by different time spans is calculated to obtain multiple rates of change, and each rate of change is set to not exceed a corresponding preset rate of change limit, thus obtaining response rate constraints; a virtual energy storage model for the target node is constructed based on the power boundary constraints, power balance constraints, and response rate constraints. This invention fully characterizes the strong coupling characteristics of aggregated flexible resources in the time dimension and the global energy conservation requirements.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power system automation technology, and in particular to a method and system for constructing a virtual energy storage model for large power grids. Background Technology

[0002] As the penetration rate of renewable energy sources such as wind and solar power in the power system continues to increase, their intermittency, volatility, and uncertainty pose severe challenges to the real-time balance and stable operation of the power grid. According to data released by the National Energy Administration, by the end of 2025, my country's cumulative installed capacity of wind and solar power exceeded 1.8 billion kilowatts for the first time, reaching 1.84 billion kilowatts, accounting for 47.3% of the total installed capacity, historically surpassing thermal power capacity by approximately 300 million kilowatts. However, the rapid expansion of installed capacity has also brought increasingly prominent pressure on power consumption. In the first half of 2025, the national overall curtailment rate of solar power rose to 6.6%, and the curtailment rate of wind power reached 5.7%, nearly doubling compared to the same period in 2024. At the same time, the traditional power system's ability to regulate through controllable thermal and hydropower is becoming insufficient. Against this backdrop, energy storage, with its rapid response, bidirectional regulation, and energy time-shifting capabilities, has become a key technological support for smoothing power fluctuations, improving grid flexibility, and reducing wind and solar curtailment rates.

[0003] Current research focuses on two main areas: modeling and optimizing the operation of physical energy storage, and virtual energy storage. Modeling physical energy storage directly utilizes physical parameters such as charging and discharging power and state of charge to construct the feasible region. However, these physical parameters require measurement and cannot adapt to the unmeasurability of parameters after large-scale power grid resource aggregation. Existing virtual energy storage is generally aggregated virtual energy storage, primarily utilizing the direct aggregation characteristics of physical energy storage, especially in electric vehicles and thermal energy storage. Mathematically, it only considers the energy recursion relationship between adjacent time periods and is mostly limited to local scenarios such as microgrids and integrated energy systems.

[0004] Large power grids aggregate diverse, widely distributed, and complexly controlled flexible resources. Their regulation behaviors are interdependent across different time periods and cannot be arbitrarily separated, thus exhibiting strong coupling characteristics in the time dimension. Simultaneously, virtual energy storage needs to satisfy global energy conservation in charging and discharging power, thus possessing power balance characteristics. Existing aggregated virtual energy storage models cannot characterize these two characteristics, potentially overestimating or underestimating the actual regulation capacity of nodes. Overestimation leads to dispatch commands exceeding the true flexibility boundaries of nodes, creating gaps in power balance, frequency regulation, or renewable energy integration; while underestimation results in insufficient utilization of flexibility regulation capacity, increasing the ancillary service regulation costs of the power grid. Summary of the Invention

[0005] Based on the shortcomings of the existing technology, the present invention provides a method and system for constructing a virtual energy storage model for large power grids, which solves the existing problems.

[0006] The present invention adopts the following technical solution: In a first aspect, the present invention provides a method for constructing a virtual energy storage model for large power grids, comprising the following steps: The power deviation of the target node in the large power grid during each period within the target cycle is used as the decision variable. The power deviation is the difference between the actual output value of the node and the planned output value of the day. In each time period, the range of decision variables is set to not exceed the set power range allowed for that time period to obtain power boundary constraints; the decision variables of each time period within the target period are summed to obtain the cumulative deviation power, and the cumulative deviation power is set to not exceed the preset tolerance threshold to obtain power balance constraints; the difference between the decision variables of two time periods separated by different time spans is calculated to obtain multiple rates of change, and each rate of change is set to not exceed the corresponding preset rate of change limit to obtain response rate constraints. A virtual energy storage model for the target node is constructed based on power boundary constraints, energy balance constraints, and response rate constraints.

[0007] Preferably, the set power range includes the maximum deviation and the minimum reverse deviation for the time period.

[0008] Preferably, the target period is one day.

[0009] Preferably, the target node is any power grid node in a large power grid that has aggregated flexible resources.

[0010] Preferably, the virtual energy storage model is as follows: ; ; ; ; … ; In the formula, for t Power deviation over time period and for t The maximum and minimum reverse power deviation during the time period. The cumulative deviation of electricity during the target period. T The number of time periods, The tolerance threshold, , ,…, This represents the limit value for the rate of change over a single time period, two time periods, ..., time periods.

[0011] Preferably, within multiple historical periods, the daily planned output value, actual output value, and corresponding electricity value of the target node for each time period are collected; based on the daily planned output value and actual output value for each time period, multiple power deviations and multiple rates of change are obtained; the maximum power deviation and the minimum reverse power deviation for each time period in multiple historical periods are used as the set power range, and the maximum rate of change is used as the preset rate of change limit value.

[0012] Preferably, the optimization objective is to maximize the difference between the electricity value gain and the cumulative deviation electricity in the historical cycle, and the parameters are estimated and optimized by setting a power range and a preset rate of change limit. The optimization model is then solved to obtain the optimal solution. Based on the optimal solution, a preset tolerance threshold is obtained.

[0013] Secondly, the present invention provides a system for constructing a virtual energy storage model for large power grids, comprising: The acquisition module is used to take the power deviation of the target node in the target period within the target period as the decision variable, wherein the power deviation is the difference between the actual output value of the node and the planned output value of the day-ahead. The setting module is used to set the range of decision variables in each time period so that it does not exceed the set power range allowed in that time period, thereby obtaining power boundary constraints; sum the decision variables of each time period within the target period to obtain the cumulative deviation power, and set the cumulative deviation power not to exceed the preset tolerance threshold, thereby obtaining power balance constraints; and subtract the decision variables of two time periods separated by different time spans to obtain multiple rates of change, and set each rate of change not to exceed the corresponding preset rate of change limit value, thereby obtaining response rate constraints. The module constructs a virtual energy storage model for the target node based on power boundary constraints, energy balance constraints, and response rate constraints.

[0014] Compared with the prior art, the above-mentioned at least one technical solution adopted by the present invention can achieve the following beneficial effects: This invention forms a complete constraint system by organically combining power deviation at each time period as the decision variable with power boundary constraints, power balance constraints, and response rate constraints. Power boundary constraints limit the physical limits of the decision variables at each time period; power balance constraints ensure that the cumulative deviation power over the entire target period meets energy conservation requirements; and response rate constraints simultaneously restrict the rate of change at different time spans. This explicitly mathematically fixes the interdependence of aggregated flexibility resources across different time periods. For the first time, this invention achieves a precise mathematical representation of the strong temporal coupling characteristics and global energy conservation characteristics of aggregated flexibility resources at the large power grid level. This effectively avoids the overestimation or underestimation of regulation capacity caused by traditional models that only consider the recursive relationship between adjacent time periods, providing a more accurate and reliable flexibility boundary for large power grid dispatch. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a method for constructing a virtual energy storage model for large power grids according to the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Existing research on virtual energy storage is largely limited to localized scenarios such as microgrids and integrated energy systems, and relies on the direct aggregation characteristics of physical energy storage such as electric vehicles and thermal energy storage. No virtual energy storage model applicable to the large-scale power grid has yet been established. The flexibility of large-scale power grid aggregates (such as numerous distributed resources, flexible loads, and renewable energy clusters) exhibits significant strong time-coupled characteristics (flexibility is interdependent across different time periods and cannot be arbitrarily separated) and power balance constraints (the charging and discharging power of virtual energy storage must satisfy global energy conservation). Therefore, it is urgent to break through the equivalent modeling theory of "flexibility → virtual energy storage" at the large-scale power grid, reveal the mapping relationship between the flexibility of the aggregate and the dynamic operating boundary of virtual energy storage, and construct a mathematical representation system for virtual energy storage that includes time coupling constraints and power balance constraints. This will lay the model foundation for utilizing the flexibility of the large power grid itself to replace physical energy storage.

[0019] Aggregated virtual energy storage is a smart energy system that aggregates one or more resources, such as adjustable loads, energy storage, and power sources, from different locations to achieve autonomous, coordinated, and optimized control, and participate in power system operation and electricity market transactions. It is a cross-space, wide-area integrator of power generation, grid, load, and storage. The aggregation targets of aggregated virtual energy storage include adjustable loads, distributed power sources, and energy storage resources. Adjustable load resources should be power users with adjustable capabilities and connected to the aggregation operation platform. Aggregated virtual energy storage should have the ability to aggregate and manage adjustable resources and participate in grid interaction services and electricity market transactions, such as auxiliary peak shaving, auxiliary frequency regulation, and demand response.

[0020] The modeling of traditional aggregated virtual energy storage specifically includes the following steps: Variable definition: for t Discharge power during the period for t Charging power during the period for t State of charge during a period of time P max This represents the maximum charging and discharging power. SOC min and SOC max These represent the minimum and maximum values ​​of the state of charge.

[0021] Mathematical modeling: (1a); (2a); (3a); (4a); Aggregated virtual energy storage possesses the charging and discharging characteristics of physical energy storage, resembling a large energy storage pool. Equations (1a) and (2a) state that the charging and discharging power cannot exceed the limit, equation (3a) represents the capacity value of the energy storage pool after charging and discharging, and equation (4a) represents the capacity limit of the energy storage pool. Only time is considered in the time scale. t and t +1 The coupling relationship between the two time periods.

[0022] Reference Figure 1 Construct a flexible virtual energy storage model: definition for t Power deviation over time period The cumulative deviation of electricity throughout the day. and for tThe maximum and minimum reverse power during the time period. , ,…, The rate of change of the cumulative deviation is limited to the rate of change over one time period, two time periods, ..., time periods. To limit the deviation in power consumption throughout the day, T This represents the number of items in the entire time period.

[0023] Establish a runtime characteristic model: (1b); (2b); (3b); (4b); … (5b); Flexible virtual energy storage utilizes the absorption or release of a certain power deviation to ensure that the net regulated electricity consumption for the entire day is zero, i.e., energy balance. Formula (1b) indicates that the power deviation value of absorption and release should be within a given range, similar to the traditional (1a) and (2a); (2b) indicates that the deviation electricity consumption for the day should be 0, but absolute 0 does not exist, so it is limited to a very small number. ε (3b)-(5b) represent the response characteristics of flexible virtual energy storage, which is completely different from traditional aggregated virtual energy storage. This indicates that flexible virtual energy storage is not actual physical energy storage and needs to meet a certain response speed when responding, with different speed limits at different time scales. It can be seen that flexible virtual energy storage is coupled at various time scales (i.e., from one adjacent time period to the next adjacent time period), and has strong time coupling characteristics compared to traditional aggregated virtual energy storage.

[0024] This invention proposes a novel concept of "nodal virtual energy storage": a virtual element is introduced at each grid node. This element, based on its planned output, possesses a certain deviation adjustment capability, and its cumulative deviation adjustment amount is zero over a daily timescale. The operating constraints of this virtual element are mathematically identical to those of physical energy storage elements, thus exhibiting energy storage-like characteristics. However, it is not an actual physical energy storage device, but rather an abstract model implemented through a control strategy; therefore, it is defined as virtual energy storage.

[0025] Furthermore, the concept of virtual energy storage can be used to describe the flexibility exhibited by system nodes in response to dispatch commands: it can both absorb or release a certain power deviation (similar to the charging and discharging behavior of energy storage) and ensure that the net regulation power is zero throughout the day (avoiding energy accumulation or loss). This characteristic enables it to participate in ancillary services such as power balancing, frequency regulation, or renewable energy consumption of the power grid without introducing actual energy storage devices, thus providing a low-cost and easily implemented flexible resource representation method for system operation.

[0026] Therefore, any combination of components with bidirectional grid regulation capabilities, and which meet certain operational constraints, can be considered as a "virtual energy storage".

[0027] Flexible virtual energy storage is not a physical entity; its "virtual" nature means that relevant parameters (such as virtual charging and discharging power, virtual power balance, and virtual rate of change) cannot be directly obtained through traditional physical measurement methods. Existing research completely ignores the modeling difficulties caused by this parameter unmeasurability. How to utilize the aggregated flexibility information contained in large power grid operation data to indirectly estimate the equivalent parameters of virtual energy storage through data-driven methods is a key technical challenge that has not yet been addressed. It is necessary to establish a "data-characteristic-parameter" inverse reasoning framework to solve the problem of identifying the dynamic characteristics of virtual energy storage under the condition of lacking physical measurements, and to achieve real-time quantification and operable characterization of the flexibility of large power grid aggregates.

[0028] Parameter estimation for flexible virtual energy storage: Assuming virtual energy storage is established at a certain node of the power grid, then the parameters... , , , ,…, , The estimation method is as follows: (1) For the actual power grid at this node N The results of the celestial movement are used for... N The sampling process involves collecting the planned output value, actual output value, and actual electricity value at that node. The planned output value refers to the planned curve generated by the power market operator one day prior to the operating day through market clearing, used to guide the generation, charging, discharging, or consumption regulation of generating units and new operating entities during each period of the operating day. The electricity value is the value of the electricity price, which refers to the marginal price of electricity in market clearing, reflecting the impact of increasing or decreasing one unit of electricity at a specific node or time period on system operating costs and constraint feasibility.

[0029] (2) No. i The sampled values ​​range from 0 to... T The planned output value for the day is , … .

[0030] (3) No. i The sampled values ​​range from 0 to... T The actual output value is , … Each sampling period is one day, and the number of data points sampled each time is the number of time periods included in a day. T It can be 24 (hours) or 96 (every 15 minutes).

[0031] (4) Construct the first i The deviation value of the second sampling .

[0032] (5) .

[0033] (6) .

[0034] (7) No. i The sampled values ​​range from 0 to... T The value of electricity is , , ..., Based on the parameters obtained from (5)-(6), an optimization model is established: ; ; ; ; … ; in, For this node at the th i Heavenly t The power deviation over a given time period is the decision variable. The deviation penalty is given for a given power dispatch. The above optimization model is a very simple linear programming model, and the optimal solution of this optimization model can be obtained using classical algorithms. .

[0035] (7) Parameters The estimated value can be characterized as .

[0036] This invention proposes constructing virtual energy storage by leveraging the operational flexibility of large power grid nodes. This eliminates the need for physical energy storage entities, introducing a virtual element at each grid node. This element, based on planned output, possesses a certain degree of deviation adjustment capability, and its cumulative deviation adjustment is zero over a daily timescale. The operational constraints of this virtual element are mathematically identical to those of physical energy storage elements, thus exhibiting energy storage-like characteristics. However, it is not an actual physical energy storage device, but rather an abstract model implemented through control strategies; hence, it is defined as "virtual energy storage." This concept of flexible virtual energy storage can be used to describe the flexibility exhibited by system nodes in responding to dispatch commands: it can absorb or release certain power deviations (similar to the charging and discharging behavior of energy storage), while ensuring that the net adjustment amount is zero throughout the day (avoiding energy accumulation or loss). This characteristic allows it to participate in ancillary services such as power balancing, frequency regulation, or renewable energy consumption in the power grid without introducing actual energy storage devices, thereby providing a low-cost and easily implemented method for representing flexible resources in system operation.

[0037] Based on the same concept, the present invention also provides a system for constructing a virtual energy storage model for large power grids, including an acquisition module, a setting module and a construction module.

[0038] The acquisition module is used to take the power deviation of the target node in the large power grid at each time period within the target period as the decision variable. The power deviation is the difference between the actual output value of the node and the planned output value of the day.

[0039] The setting module is used to set the range of decision variables in each time period so that it does not exceed the set power range allowed in that time period, thereby obtaining power boundary constraints; sum the decision variables of each time period within the target period to obtain the cumulative deviation power, and set the cumulative deviation power not to exceed the preset tolerance threshold, thereby obtaining power balance constraints; and subtract the decision variables of two time periods separated by different time spans to obtain multiple rates of change, and set each rate of change not to exceed the corresponding preset rate of change limit value, thereby obtaining response rate constraints.

[0040] The building module constructs a virtual energy storage model for the target node based on power boundary constraints, energy balance constraints, and response rate constraints.

[0041] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0042] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for constructing a virtual energy storage model for large power grids, characterized in that, Includes the following steps: The power deviation of the target node in the large power grid during each period within the target cycle is used as the decision variable. The power deviation is the difference between the actual output value of the node and the planned output value of the day. In each time period, the range of decision variables is set to not exceed the set power range allowed for that time period to obtain power boundary constraints; the decision variables of each time period within the target period are summed to obtain the cumulative deviation power, and the cumulative deviation power is set to not exceed the preset tolerance threshold to obtain power balance constraints; the difference between the decision variables of two time periods separated by different time spans is calculated to obtain multiple rates of change, and each rate of change is set to not exceed the corresponding preset rate of change limit to obtain response rate constraints. A virtual energy storage model for the target node is constructed based on power boundary constraints, energy balance constraints, and response rate constraints.

2. The method for constructing a virtual energy storage model for a large power grid as described in claim 1, characterized in that, The set power range includes the maximum deviation and the minimum reverse deviation for each time period.

3. The method for constructing a virtual energy storage model for a large power grid as described in claim 1, characterized in that, The target period is one day.

4. The method for constructing a virtual energy storage model for a large power grid as described in claim 1, characterized in that, The target node is any power grid node in a large power grid that has aggregated flexible resources.

5. The method for constructing a virtual energy storage model for a large power grid as described in claim 1, characterized in that, The virtual energy storage model is shown below: ; ; ; ; … ; In the formula, for t Power deviation over time period and for t The maximum and minimum reverse power deviation during the time period. The cumulative deviation of electricity during the target period. T The number of time periods, The tolerance threshold, , ,…, This represents the limit value for the rate of change over a single time period, two time periods, ..., time periods.

6. The method for constructing a virtual energy storage model for a large power grid as described in claim 1, characterized in that, Within multiple historical periods, the daily planned output value, actual output value, and corresponding electricity value of the target node for each time period are collected. Based on the daily planned output value and actual output value for each time period, multiple power deviations and multiple rates of change are obtained. The maximum power deviation and the minimum reverse power deviation for each time period in multiple historical periods are used as the set power range, and the maximum rate of change is used as the preset rate of change limit value.

7. The method for constructing a virtual energy storage model for a large power grid as described in claim 6, characterized in that, The optimization objective is to maximize the difference between the value of electricity and the cumulative deviation in electricity volume over a historical period. A parameter estimation optimization model is constructed with constraints set by a power range and a preset rate of change limit. The optimization model is then solved to obtain the optimal solution. A preset tolerance threshold is obtained based on the optimal solution.

8. A system for constructing a virtual energy storage model for large power grids, characterized in that, include: The acquisition module is used to take the power deviation of the target node in the target period within the target period as the decision variable, wherein the power deviation is the difference between the actual output value of the node and the planned output value of the day-ahead. The setting module is used to set the range of decision variables in each time period so that it does not exceed the set power range allowed in that time period, thereby obtaining power boundary constraints; sum the decision variables of each time period within the target period to obtain the cumulative deviation power, and set the cumulative deviation power not to exceed the preset tolerance threshold, thereby obtaining power balance constraints; and subtract the decision variables of two time periods separated by different time spans to obtain multiple rates of change, and set each rate of change not to exceed the corresponding preset rate of change limit value, thereby obtaining response rate constraints. The module constructs a virtual energy storage model for the target node based on power boundary constraints, energy balance constraints, and response rate constraints.