Energy storage scheduling control method and energy storage system

Abstract

The energy storage scheduling control method and the energy storage system provided by the embodiments of the present application are as follows: a cloud intelligent agent constructs a total revenue target function based on cloud state information and cloud action information, and determines cloud action information when the total revenue target function reaches a maximum value as target cloud action information; after a station intelligent agent obtains the target cloud action information, the station intelligent agent constructs a response target function based on the target cloud action information, station state information and station action information, and determines station action information when the response target function reaches a maximum value as target station action information; and an energy storage unit executes corresponding electric energy scheduling operations based on the target station action information. The embodiments of the present application make the cloud intelligent agent formulate a global electric energy scheduling plan with the goal of maximizing total revenue, and the station intelligent agent optimizes the execution plan by combining the cloud plan and the actual station state, so that the energy storage unit maximizes the revenue and the electric energy scheduling adapts to real-time scheduling.

Description

Technical Field

[0001] This application relates to the field of energy storage, specifically to an energy storage dispatch and control method and an energy storage system. Background Technology

[0002] With the development of new energy technologies, distributed energy storage, as a core resource for peak shaving and valley filling and frequency regulation support of the power grid, is increasingly demonstrating its value in aggregated dispatch. The existing dispersed layout of energy storage units leads to significant challenges in coordinated dispatch. Traditional dispatch methods struggle to simultaneously maximize the multi-objective benefits of the energy market and ancillary services market, and are ill-equipped to effectively address risks arising from market uncertainties such as electricity price fluctuations. Furthermore, the lack of seamless integration between global optimization plans and real-time station responses fails to meet the power grid's dual requirements for dispatch accuracy and response speed. Summary of the Invention

[0003] In view of this, this application provides an energy storage dispatch control method and an energy storage system to help solve the problem of low accuracy in power dispatch of energy storage systems in the prior art.

[0004] In a first aspect, embodiments of this application provide an energy storage system, including:

[0005] A cloud-based intelligent agent is used to construct a total revenue objective function based on cloud-based state information and cloud-based action information, and to determine the cloud-based action information when the total revenue objective function takes the maximum value as the target cloud-based action information. The target cloud-based action information is used to instruct the cloud-based power dispatch plan of the energy storage unit.

[0006] The station-side intelligent agent is communicatively connected to the cloud-side intelligent agent. It is used to acquire the target cloud-side action information, construct a response objective function based on the target cloud-side action information, station-side status information, and station-side action information, and determine the station-side action information when the response objective function is maximized as the target station-side action information. The target station-side action information is used to indicate the station-side power dispatch plan of the energy storage unit.

[0007] The energy storage unit is communicatively connected to the station-end intelligent agent and is used to perform corresponding power dispatching operations based on the target station-end action information.

[0008] In one optional embodiment, the cloud status information includes: market price information and cloud adjustment coefficient;

[0009] The cloud-based intelligent agent is specifically used to construct the total revenue objective function based on energy market revenue, ancillary service market revenue, energy storage equipment degradation cost, and market risk cost.

[0010] The energy market revenue and ancillary service market revenue are determined by the market price information and the cloud action information; the energy storage equipment degradation cost is determined by the cloud adjustment coefficient and the cloud action information; and the market risk cost is determined by the market price information, the cloud adjustment coefficient, and the cloud action information.

[0011] In one optional embodiment, the market price information includes: historical market price and real-time market price, and the cloud action information includes: the first charge / discharge power won at the historical market price and the second charge / discharge power won at the real-time market price;

[0012] The cloud-based intelligent agent is also used to determine historical market revenue based on the historical market price and the first charging / discharging power, to determine real-time market revenue based on the real-time market price and the second charging / discharging power, and to determine the energy market revenue based on the historical market revenue and the real-time market revenue.

[0013] In one optional embodiment, the market price information includes: frequency modulation capacity price and frequency modulation mileage price, and the cloud action information includes: the won frequency modulation capacity and the actual frequency modulation mileage;

[0014] The cloud-based intelligent agent is also used to determine frequency modulation capacity revenue based on the frequency modulation capacity price and the winning bid frequency modulation capacity, to determine frequency modulation mileage revenue based on the frequency modulation mileage price and the actual frequency modulation mileage, and to determine the ancillary service market revenue based on the frequency modulation capacity revenue and the frequency modulation mileage revenue.

[0015] In one optional embodiment, the market price information further includes: inertial response capacity price and inertial response rate price, and the cloud action information includes: the winning bid inertial response capacity and the actual inertial response rate;

[0016] The cloud-based intelligent agent is also used to determine the inertial response capacity revenue based on the inertial response capacity price and the winning bid inertial response capacity, to determine the inertial response rate revenue based on the inertial response speed price and the actual inertial response speed, and to determine the ancillary service market revenue based on the inertial response capacity revenue and the inertial response rate revenue.

[0017] In one optional embodiment, the cloud adjustment coefficient includes an aging cost coefficient and a depth penalty coefficient, and the cloud action information includes the number of charge / discharge cycles and the charge / discharge depth.

[0018] The cloud-based intelligent agent is also used to determine the cycle aging cost based on the aging cost coefficient and the number of charge-discharge cycles, to determine the charge-discharge depth cost based on the depth penalty coefficient and the charge-discharge depth, and to determine the degradation cost of the energy storage device based on the cycle aging cost and the charge-discharge depth cost.

[0019] In one optional embodiment, the cloud adjustment coefficient further includes: a risk aversion coefficient;

[0020] The cloud-based intelligent agent is also used to determine market risk information based on the cloud-based action information and the market price information, and to determine the market risk cost based on the market risk information and the risk aversion coefficient.

[0021] In one optional embodiment, the station status information includes: station adjustment coefficient;

[0022] The station-side intelligent agent is specifically used to construct the response objective function based on scheduling tracking error penalty, operation cost, local equipment degradation cost and auxiliary service market reward.

[0023] The scheduling tracking error penalty is determined by the degree of deviation between the station-side power scheduling plan and the cloud-based power scheduling plan; the operation cost is determined by the station-side adjustment coefficient and the station-side action information; the local equipment degradation cost is determined by the station-side adjustment coefficient and the station-side action information; and the ancillary service market reward is determined by the station-side price and the station-side action information.

[0024] In one optional embodiment, the cloud-based power dispatch plan includes: planned charging and discharging power; the station-side power dispatch plan includes: actual charging and discharging power; and the station-side adjustment coefficient includes: tracking error penalty coefficient.

[0025] The station-side intelligent agent is specifically used to determine the difference between the planned charging / discharging power and the actual charging / discharging power, and to determine the scheduling tracking error penalty based on the difference and the tracking error penalty coefficient.

[0026] In one optional embodiment, the station-side adjustment coefficient further includes: an efficiency loss cost coefficient and a charging / discharging efficiency;

[0027] The station-side intelligent agent is also used to determine the operating cost based on the efficiency loss cost coefficient, the charging and discharging efficiency, and the actual charging and discharging power.

[0028] In one optional embodiment, the station-end adjustment coefficient further includes: a local degradation cost coefficient, and the station-end action information includes: changes in health status;

[0029] The station-side intelligent agent is also used to determine the local equipment degradation cost based on the local degradation cost coefficient and the change in health status.

[0030] In one optional embodiment, the station-end adjustment coefficient further includes: a frequency modulation reward coefficient and an inertial response reward coefficient, and the station-end action information includes: station-end frequency modulation mileage or station-end inertial response rate.

[0031] The station-side intelligent agent is further configured to determine the frequency modulation reward based on the frequency modulation reward coefficient and the station-side frequency modulation mileage, determine the inertial response reward based on the inertial response reward coefficient and the station-side inertial response rate, and determine the ancillary service market reward based on the frequency modulation reward and the inertial response reward.

[0032] In one optional embodiment, the energy storage system further includes:

[0033] The market interaction module allows users to obtain the device parameters of the energy storage unit. The device parameters include: rated charge and discharge power threshold, first charge and discharge power threshold, and second charge and discharge power threshold. The first charge and discharge power is determined by the energy capacity of the energy storage unit, and the second charge and discharge power threshold is determined by the health status of the energy storage unit.

[0034] The market interaction module is also used to determine the cloud power dispatch threshold based on the rated charge / discharge power threshold, the first charge / discharge power threshold, and the second charge / discharge power threshold, wherein the cloud power dispatch plan does not exceed the range limited by the cloud power dispatch threshold.

[0035] Secondly, embodiments of this application provide an energy storage scheduling and control method, including:

[0036] The cloud-based intelligent agent constructs a total revenue objective function based on cloud status information and cloud action information, and determines the cloud action information when the total revenue objective function takes the maximum value as the target cloud action information. The target cloud action information is used to instruct the cloud power dispatch plan of the energy storage unit.

[0037] After the station-side intelligent agent obtains the target cloud-side action information, it constructs a response objective function based on the target cloud-side action information, station-side status information, and station-side action information, and determines the station-side action information when the response objective function is maximized as the target station-side action information. The target station-side action information is used to indicate the station-side power dispatch plan of the energy storage unit.

[0038] The energy storage unit performs corresponding power dispatching operations based on the target station's action information.

[0039] Thirdly, embodiments of this application provide an electronic device, including a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the electronic device is triggered to execute the method described in any of the first aspects above.

[0040] Fourthly, embodiments of this application provide a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the method described in any of the first aspects.

[0041] Fifthly, embodiments of this application provide a computer program product comprising executable instructions that, when executed on a computer, cause the computer to perform the method described in any of the first aspects.

[0042] The solution provided in this application involves a cloud-based intelligent agent constructing a total revenue objective function based on cloud state information and cloud action information. The cloud action information that maximizes the total revenue objective function is determined as the target cloud action information, which is used to instruct the energy storage unit's cloud-based power dispatch plan. After acquiring the target cloud action information, the station-based intelligent agent constructs a response objective function based on the target cloud action information, station-based state information, and station-based action information. The station-based action information that maximizes the response objective function is determined as the target station-based action information, which is used to instruct the energy storage unit's station-based power dispatch plan. The energy storage unit executes corresponding power dispatch operations based on the target station-based action information. This application embodiment achieves a global power dispatch plan by the cloud-based intelligent agent with the goal of maximizing total revenue. The station-based intelligent agent optimizes the execution plan by combining the cloud plan with the actual station-based state, ultimately realizing power dispatch that maximizes energy storage unit revenue and adapts to real-time scheduling. Attached Figure Description

[0043] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a schematic diagram illustrating an example of an energy storage scheduling and control method provided in an embodiment of this application.

[0045] Figure 2 A flowchart illustrating an energy storage scheduling and control method provided in an embodiment of this application;

[0046] Figure 3A schematic diagram illustrating another energy storage scheduling and control method provided in this application embodiment;

[0047] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0048] To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0049] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0050] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0051] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0052] Figure 1 This is a schematic diagram of an energy storage system provided in an embodiment of this application. Figure 1 As shown, the energy storage system in this embodiment is an innovative two-layer optimization decision-making architecture, including a cloud-based scheduling optimization layer and a station-based control optimization layer. The cloud-based scheduling planning layer contains a cloud-based intelligent agent, which is the "brain" of the energy storage system and is responsible for macro-level power dispatch planning. Its decision-making timescale focuses on the medium to long term, mainly corresponding to the day-ahead and intraday electricity markets, with a decision cycle typically ranging from 15 minutes to 1 hour. The core objective is to maximize the overall economic benefits of aggregators in the electricity market while effectively managing market risks and ensuring the long-term health of energy storage assets.

[0053] Specifically, the cloud-based intelligent agent uses probabilistic forecasts of electricity prices, renewable energy output, and load to formulate joint bidding strategies for the energy market and ancillary service markets such as frequency regulation and reserve, and optimizes the allocation of the aggregated total energy storage capacity across different markets and time periods. Simultaneously, the cloud-based intelligent agent employs advanced risk measurement methods such as Conditional Value at Risk (CVaR) to control potential economic losses, ultimately generating charging and discharging plan curves and scheduling instructions for energy storage units to be transmitted down to the next layer, providing macro-level guidance for the subsequent execution of the energy storage system.

[0054] The station-side control optimization layer comprises station-side intelligent agents and energy storage units, with the station-side intelligent agents communicating with both the cloud-based intelligent agents and the energy storage units. The station-side intelligent agents act as the "cerebellum and nerve endings" of the energy storage system, responsible for micro-level real-time control. Their decision-making timescale focuses on real-time, with cycles ranging from seconds to minutes. Their core objective is to accurately execute the power dispatch plans issued by the cloud-based intelligent agents, while simultaneously responding rapidly to real-time grid fluctuations and prediction errors to ensure the safe and stable operation of the energy storage system.

[0055] Specifically, the station-side intelligent agent will track the charge / discharge plan curves issued by the cloud-based intelligent agent to minimize tracking errors and ensure the implementation of the macro-level plan. Simultaneously, the station-side intelligent agent can rapidly adjust power based on real-time grid frequency, voltage, and other signals to provide ancillary services to the grid, and monitor the SOC, SOH, and temperature of each energy storage unit in real time. If risks such as overcharging, over-discharging, or over-temperature occur, protection strategies will be immediately implemented. Furthermore, while meeting the cloud-based power dispatch plan, the station-side intelligent agent will also perform local optimization of the operation of individual energy storage units (such as minimizing charge / discharge losses), balancing execution accuracy and local operating efficiency. The energy storage units can receive power dispatch plans sent by the station-side intelligent agent in real time and execute corresponding power dispatch operations.

[0056] The execution process of the above energy storage system can be referred to Figure 2 The flowchart shown mainly includes:

[0057] Step 201: The cloud-based intelligent agent constructs a total revenue objective function based on cloud state information and cloud action information, and determines the cloud action information that maximizes the total revenue objective function as the target cloud action information.

[0058] Cloud-based state information primarily refers to objectively existing information in the market or on equipment, such as electricity price information, the state of charge (SOC) value of energy storage units, and the state of health (SOH) value. Cloud-based action information primarily refers to potential power dispatch plans, such as the charging and discharging power, frequency regulation capacity, and inertial response capacity in market bidding. With total market revenue as the objective, different cloud-based action information will typically yield different total revenues under the constraints of cloud-based state information. After constructing a total revenue objective function based on cloud-based state and action information, the cloud-based agent can substitute different cloud-based action information into this objective function and compare the results. When the total revenue objective function reaches its maximum value, the corresponding cloud-based action information represents the most suitable power dispatch plan at present.

[0059] Optionally, the objective function for total revenue can be constructed using the following formula:

[0060]

[0061] Total revenue objective function Defined as motion information in the cloud Under what conditions, the expected reward for all time steps t? The sum, It is a discount factor. It is the reward function at time t, where E[] is the expected value and max is the maximum value.

[0062] In one optional embodiment, the cloud-based state information includes: market price information and cloud-based adjustment coefficients. The cloud-based intelligent agent can construct a total revenue objective function based on energy market revenue, ancillary service market revenue, energy storage device degradation costs, and market risk costs. Energy market revenue mainly refers to the revenue obtained by the energy storage system from charging and discharging during different electricity price periods; the core logic is "buy electricity (charging) when electricity prices are low, sell electricity (discharging) when electricity prices are high." Ancillary service market revenue mainly refers to the additional revenue obtained by the energy storage system through participating in ancillary services such as frequency regulation and inertial response. Energy market revenue and ancillary service market revenue can be determined by market price information and cloud-based action information. Energy storage device degradation costs mainly refer to the aging and loss of energy storage units caused by cyclic charging and discharging, determined by cloud-based adjustment coefficients and cloud-based action information. Market risk costs mainly refer to the economic losses caused by electricity price fluctuations, determined by market price information, cloud-based adjustment coefficients, and cloud-based action information.

[0063] Optional, expected reward The construction can be referenced by the following formula:

[0064]

[0065] For energy market revenue, To support service market revenue, For the degradation cost of energy storage equipment, For market risk costs, , Take positive values. and Taking negative values, in order to maximize the total revenue objective function, the cloud-based intelligent agent should try to maximize revenue while minimizing equipment wear and tear and market risk.

[0066] In this embodiment of the application, by integrating the arbitrage income from energy market charging and discharging, the additional income from the ancillary services market, and deducting the degradation cost of energy storage equipment and the market risk cost caused by electricity price fluctuations, a total income objective function is constructed. This allows for the precise quantification of the income from aggregated dispatching of energy storage, while also taking into account profit targets, lifespan protection of energy storage equipment, and market risk management.

[0067] Specifically, market price information may include historical market prices and real-time market prices, while cloud-based action information may include the first charge / discharge power won at the historical market price and the second charge / discharge power won at the real-time market price. The cloud-based intelligent agent can determine historical market revenue based on historical market prices and the first charge / discharge power, determine real-time market revenue based on real-time market prices and the second charge / discharge power, and determine energy market revenue based on historical market revenue and real-time market revenue.

[0068] Alternatively, the energy market revenue can be constructed using the following formula:

[0069]

[0070] and These are historical market prices and real-time market prices. and These are the discharge and charging powers that were won in bids at historical market prices. and The discharge and charging power are the bids won under real-time market prices. The energy market revenue at time t is calculated using a specific formula by combining historical and real-time market prices with the corresponding winning bids for charging and discharging power. This allows for precise quantification of the arbitrage profits from "buying low and selling high," providing accurate revenue calculation basis for cloud-based intelligent agents to formulate energy market bidding strategies.

[0071] In one optional embodiment, the ancillary service market revenue may involve multiple service items, and its value can be obtained by summing the revenues of the multiple service items. The ancillary service market revenue can be constructed with reference to the following formula:

[0072]

[0073] Where AS is a collection of service items. It is the benefit of participating in the j-th ancillary service at time t.

[0074] The following explanation uses Frequency Response (FR) and Inertial Response (IR) services as examples:

[0075] In frequency modulation (FM) services, market price information may include FM capacity price and FM mileage price, while cloud-based action information may include the awarded FM capacity and actual FM mileage. The cloud-based intelligent agent can determine FM capacity revenue based on the FM capacity price and the awarded FM capacity, determine FM mileage revenue based on the FM mileage price and the actual FM mileage, and determine ancillary service market revenue based on FM capacity revenue and FM mileage revenue.

[0076] Specifically, due to fluctuations in user electricity consumption, the power grid may require energy storage systems to stabilize power supply through frequency regulation services. Energy storage systems need to reserve frequency regulation capacity to cope with potential future power fluctuations in the grid; the more reserved frequency regulation capacity, the higher the revenue. The awarded frequency regulation capacity is the "maximum frequency regulation power that can be mobilized at any time" promised by the energy storage system, and the frequency regulation capacity price is the price per unit of frequency regulation capacity given by the grid. The cloud-based intelligent agent multiplies the frequency regulation capacity price and the awarded frequency regulation capacity to obtain the frequency regulation capacity revenue. In actual frequency regulation, the power of the energy storage system participating in frequency regulation services may not reach the reserved upper limit. The actual frequency regulation mileage is the sum of the actual power changes of the energy storage system participating in frequency regulation, and the frequency regulation mileage price is the price per unit of frequency regulation mileage. The cloud-based intelligent agent multiplies the frequency regulation mileage price and the actual frequency regulation mileage to obtain the frequency regulation mileage revenue. Frequency regulation capacity revenue can be considered as basic revenue, and frequency regulation mileage revenue can be considered as additional revenue. The cloud-based intelligent agent adds the two together to obtain the ancillary service market revenue.

[0077] Optionally, the revenue structure for the ancillary services market can be constructed using the following formula:

[0078]

[0079] and These are the FM capacity price and the FM mileage price. M(t) represents the frequency regulation capacity won in the bid, and M(t) represents the actual frequency regulation mileage. Revenue from ancillary services market corresponding to frequency modulation services.

[0080] In inertial response services, market price information also includes: inertial response capacity price and inertial response rate price. Cloud action information includes: awarded inertial response capacity and actual inertial response rate. Cloud-based intelligent agents can determine inertial response capacity revenue based on the inertial response capacity price and awarded inertial response capacity, determine inertial response rate revenue based on the inertial response rate price and actual inertial response rate, and determine ancillary service market revenue based on inertial response capacity revenue and inertial response rate revenue.

[0081] Specifically, when grid power fluctuates drastically, energy storage systems can quickly provide power buffering through inertial response, avoiding control disruptions. The awarded inertial response capacity represents the energy storage system's promised "instantaneous maximum power adjustment capability," and the inertial response capacity price is the unit price. The cloud-based intelligent agent multiplies the awarded inertial response capacity by the inertial response capacity price to obtain the inertial response capacity revenue. The actual inertial response rate represents the actual power adjustment rate, and the inertial response rate price is the unit price. The cloud-based intelligent agent multiplies the actual inertial response rate by the inertial response rate price to obtain the inertial response rate revenue. Finally, the cloud-based intelligent agent adds the inertial response capacity revenue and the inertial response rate revenue to obtain the ancillary service market revenue corresponding to the inertial response service.

[0082] Alternatively, the revenue from the ancillary services market can be constructed using the following formula:

[0083]

[0084] and These are the prices for inertia response capacity and inertia response rate, respectively. It is the inertia response capacity of the winning bid. It is the actual inertial response speed. This refers to the market revenue from ancillary services corresponding to inertia response services.

[0085] In this embodiment, after the energy storage system participates in ancillary services, it can not only obtain capacity and performance benefits through frequency regulation and inertial response, thereby increasing its own economic resources, but also provide the power grid with support such as frequency stability and power buffering, helping the power grid to cope with fluctuations and ensure safe operation, thus improving its own resource utilization efficiency and value.

[0086] In one optional embodiment, the aforementioned cloud-based adjustment coefficient may include an aging cost coefficient and a depth penalty coefficient, and the cloud-based operating information includes the number of charge-discharge cycles and the depth of charge-discharge. During the cyclic charging and discharging process of the energy storage unit, the cloud-based intelligent agent may determine the cyclic aging cost based on the aging cost coefficient and the number of charge-discharge cycles, determine the depth of charge-discharge cost based on the depth penalty coefficient and the depth of charge-discharge, and determine the degradation cost of the energy storage device based on the cyclic aging cost and the depth of charge-discharge cost.

[0087] The degradation cost of energy storage equipment can be constructed using the following formula:

[0088]

[0089] Where α is the aging cost coefficient, and β is the depth penalty coefficient. The sum of the number of charge and discharge cycles. It is a non-linear penalty function that applies when SOC(t) is below a safety threshold. When this happens, the function value will increase dramatically. For example:

[0090]

[0091] In this embodiment of the application, by converting the physical loss of the battery into a calculable economic cost, the cycle aging caused by frequent charging and discharging is avoided, and the irreversible damage to the battery caused by deep discharge is strictly prevented. This allows the cloud-based intelligent agent to make decisions that take into account both short-term gains and long-term lifespan, and ultimately achieve a balance between "maximizing gains" and "extending battery lifespan".

[0092] In one optional embodiment, the cloud adjustment coefficient may include a risk aversion coefficient. The cloud-based intelligent agent can determine market risk information based on cloud action information and market price information, and determine market risk costs based on market risk information and the risk aversion coefficient.

[0093] The market risk cost can be constructed using the following formula:

[0094]

[0095] in, It is the risk aversion coefficient. Loss(t) is the economic loss that may occur at time t, mainly referring to the difference between expected and actual returns caused by market uncertainties (such as electricity prices, load, and policy fluctuations), which needs to be obtained through multi-scenario simulation and evaluation. It is the conditional value of risk at a confidence level α, which is the average of all losses worse than the "worst-case threshold" at a certain confidence level α. Adjustments can alter the bidding strategies of cloud-based intelligent agents, for example... When smaller, They will also become smaller, and cloud-based intelligent agents can adopt aggressive strategies, willing to take risks for high returns. When it is large, This will also increase, and cloud-based intelligent agents can adopt conservative strategies to avoid losses.

[0096] In this embodiment of the application, by quantifying risk costs, the decision-making of cloud-based intelligent agents can be made more scientific, and by adjusting... This allows cloud-based intelligent agents to adapt their decisions to different risk preferences, improving the flexibility of energy storage systems.

[0097] In one optional embodiment, the decision-making of the cloud-based intelligent agent also needs to meet preset constraints, including:

[0098] (1) The SOC of the energy storage unit must not exceed the threshold: ;

[0099] (2) Charge and discharge power constraints (charge and discharge states are mutually exclusive):

[0100] , and It can be either 0 or 1, and the values ​​are different, to satisfy the mutual exclusion of charging and discharging.

[0101] (3) SOC dynamic equation: ,in, and These are charging efficiency and discharging efficiency, respectively. It represents the total energy storage capacity. This equation is used to show that the SOC at the next moment is determined by the current state and the charging and discharging behavior.

[0102] (4) Bidding constraints in the ancillary services market: For auxiliary service power, This is the maximum power for auxiliary services.

[0103] Step 202: After the station-side intelligent agent obtains the target cloud-side action information, it constructs a response objective function based on the target cloud-side action information, station-side status information, and station-side action information, and determines the station-side action information when the response objective function is maximized as the target station-side action information.

[0104] The target cloud-based action information can be viewed as a cloud-based power dispatch plan output by the cloud-based intelligent agent. Station-side status information may include: station-side adjustment coefficients, constraints, and equipment status. Station-side action information is used to instruct the station-side power dispatch plan, mainly including: actual power dispatch or service auxiliary operations. The station-side intelligent agent needs to track the cloud-based power dispatch plan as closely as possible to ensure that the actual power dispatch is consistent with it; on the other hand, it can make adaptive adjustments based on real-time changes at the station.

[0105] Based on the above requirements, the station-side intelligent agent can construct a response objective function based on scheduling tracking error penalty, operating costs, local equipment degradation costs, and ancillary service market rewards. The scheduling tracking error penalty ensures that the power executed by the station is as close as possible to the planned power issued by the cloud; its value is determined by the degree of deviation between the station-side power dispatch plan and the cloud power dispatch plan. Operating costs mainly include charge / discharge efficiency losses and standby losses; their values ​​are determined by the station-side regulation coefficient and station-side action information. Local equipment degradation costs focus on the micro-degradation of individual energy storage units during real-time operation, related to charge / discharge cycles and the rate of change of state of health (SOH); their values ​​are determined by the station-side regulation coefficient and station-side action information. The ancillary service market rewards reflect the rewards for the energy storage system participating in ancillary services; their values ​​are determined by the station-side regulation coefficient and station-side action information.

[0106] Optionally, the objective function for response can be constructed using the following formula:

[0107]

[0108] in, It is a discount factor. It is the reward function at time t. This refers to the station-side action information. The station-side agent can determine the station-side action information that maximizes the expected value of the reward function as the target station-side action information.

[0109] Optional, The construction can be referenced by the following formula:

[0110]

[0111] in, Penalty for scheduling tracking errors, For operating costs, For the cost of partial equipment degradation, Rewards for supporting service markets.

[0112] In this embodiment of the application, by maximizing the accumulated expected reward within the second-level real-time control cycle at the station, the precise execution of the cloud-based power dispatch plan, the minimization of local operating costs and equipment losses, and the maximization of contributions to grid ancillary services are achieved, thus achieving a multi-objective optimization balance.

[0113] In one optional embodiment, the cloud-based power dispatch plan may include: planned charging and discharging power; the station-side power dispatch plan may include: actual charging and discharging power; and the station-side adjustment coefficient may include: a tracking error penalty coefficient. The station-side intelligent agent may first determine the difference between the planned charging and discharging power and the actual charging and discharging power, and then determine the dispatch tracking error penalty based on the difference and the tracking error penalty coefficient.

[0114] Optionally, the scheduling tracking error penalty can be constructed using the following formula:

[0115]

[0116] in, To track the error penalty coefficient, It is the actual charging and discharging power of the energy storage unit at time t. It is the planned charging and discharging power issued by the cloud. Penalty for scheduling tracking errors.

[0117] In the embodiments of this application, The specific values ​​can influence the scheduling tendency of cloud-based intelligent agents, when When the deviation between the actual charging / discharging power and the planned charging / discharging power is large, it will result in a significant penalty, and the station-side intelligent system will tend to stably track the cloud-based power dispatch plan. When the scale is smaller, the station-side intelligent system has more autonomy, and the station-side benefits are usually higher, but it will affect the overall power dispatch planning. The above formula can balance the station-side's tracking stability of the cloud-based power dispatch plan with its own operational autonomy, and can adapt to the power dispatch needs in different scenarios.

[0118] In one optional embodiment, the station-side adjustment coefficient further includes: an efficiency loss cost coefficient and a charge / discharge efficiency. The station-side intelligent agent can determine the operating cost based on the efficiency loss cost coefficient, the charge / discharge efficiency, and the actual charge / discharge power. Its construction can refer to the following formula:

[0119]

[0120] in, It is the efficiency loss cost coefficient. and These are the actual charging power and the actual discharging power. and These are charging efficiency and discharging efficiency, respectively. The operating cost is used to quantify the energy loss cost during the real-time operation of an energy storage system. The core is to convert the energy waste during the charging and discharging process into a calculable economic cost.

[0121] In this embodiment, the operational cost is incorporated into the response objective function, enabling the station-side intelligent agent to prioritize newer batteries with higher charging and discharging efficiency and reduce ineffective charging and discharging actions. By transforming the invisible heat generation and internal resistance losses in the energy storage unit into explicit economic costs, the station-side can meet scheduling requirements while proactively selecting low-loss operating modes during real-time control, ultimately achieving a balance between precise execution and cost savings.

[0122] In one optional embodiment, the station-side adjustment coefficient includes a local degradation cost coefficient, and the station-side action information includes changes in health status. The station-side intelligent agent can determine the local equipment degradation cost based on the local degradation cost coefficient and the changes in health status. Its construction can refer to the following formula:

[0123]

[0124] in, It is the local degradation cost coefficient. It represents the change in health status at time t, which can be calculated through real-time monitoring and the SOH prediction model. To reduce the cost of localized equipment degradation, this approach focuses on the individual losses of each energy storage unit, accurately identifying which battery ages faster. This allows for targeted optimization of control strategies, preventing excessive losses in a single energy storage unit from dragging down the overall lifespan.

[0125] In one optional embodiment, the station-side adjustment coefficient further includes: a frequency modulation reward coefficient and an inertial response reward coefficient, and the station-side action information includes: station-side frequency modulation mileage or station-side inertial response rate. The station-side agent can determine the frequency modulation reward based on the frequency modulation reward coefficient and station-side frequency modulation mileage, determine the inertial response reward based on the inertial response reward coefficient and station-side inertial response rate, and determine the ancillary service market reward based on the frequency modulation reward and inertial response reward. The ancillary service market reward can be constructed using the following formula:

[0126]

[0127] The supplementary service market rewards are obtained by summing the rewards from each service item. As a frequency tuning reward, As a reward for inertial response, For station-side frequency modulation mileage, The station-end inertial response rate, This is the frequency modulation reward coefficient. Inertia response reward coefficient.

[0128] Ancillary service market rewards reflect the additional local rewards earned by energy storage units when providing ancillary services in real-time response to grid demand. Frequency regulation reward coefficients and inertial response reward coefficients are adjustable variables used to determine the balance between "local service" and "global tracking" for the station-side agent. When the reward coefficients are large, the station-side agent will increase the proportion of ancillary services to obtain more revenue; however, excessive pursuit of revenue may deviate from the cloud-based power dispatch plan and lead to increased equipment losses. When the reward coefficients are small, the rewards obtained by the station-side agent from ancillary services are insufficient to offset other negative benefits, and the overall response objective function may become smaller; therefore, the station-side agent will tend to reduce the proportion of ancillary services.

[0129] In this embodiment, the relationship between the station's participation in auxiliary services and the tracking of the global scheduling plan and equipment wear is balanced by adjusting the reward coefficient, thereby achieving accurate quantification and multi-objective optimization of local service contribution rewards.

[0130] The station-side intelligent agent integrates the above multiple dimensions to obtain the maximum value of the response objective function, and determines the station-side action information at this time as the target station-side action information, i.e., the station-side power dispatch plan.

[0131] In one optional embodiment, the decision-making of the station-side intelligent agent also needs to meet preset constraints, including:

[0132] (1) Energy storage unit SOC constraint: (where i represents the i-th energy storage unit)

[0133] (2) Charging and discharging power limits: , and Let represent the minimum and maximum charging power of the i-th energy storage unit, respectively. and Let represent the minimum and maximum discharge power of the i-th energy storage unit, respectively.

[0134] (3) Power change rate limit (gradient): The power of the energy storage unit cannot "rise and fall abruptly." The power change between two adjacent real-time control cycles (e.g., with a 1-second interval) must be less than its maximum ramp rate. , This represents the power difference between adjacent cycles.

[0135] (4) Power deviation constraint: ,in, This is the maximum allowable power deviation. Station-side control must, to a certain extent, follow the plan issued by the cloud to ensure the achievement of overall optimization goals. This belongs to the station-side power dispatch plan. It belongs to the cloud-based power dispatch plan.

[0136] Step 203: The energy storage unit performs corresponding power dispatching operations based on the target station's action information.

[0137] The energy storage unit responds to the station-side power dispatch plan output by the station-side intelligent agent and performs corresponding power dispatch operations, ultimately achieving maximum energy storage unit revenue and power dispatch that is adapted to real-time dispatch.

[0138] In an optional embodiment, the energy storage system may further include a market interaction module. (See also...) Figure 3The market interaction module belongs to the cloud-based scheduling and planning layer and is responsible for deep and intelligent interaction with the electricity market, including but not limited to the formulation of market bidding strategies, risk management (especially deep discharge penalties), and flexible participation in the ancillary service market.

[0139] Optionally, the market interaction module can obtain the device parameters of the energy storage unit, including: a rated charge / discharge power threshold, a first charge / discharge power threshold, and a second charge / discharge power threshold. The first charge / discharge power is determined by the energy storage unit's capacity, and the second charge / discharge power threshold is determined by the energy storage unit's health status. The market interaction module can determine a cloud-based power dispatch threshold based on the rated charge / discharge power threshold, the first charge / discharge power threshold, and the second charge / discharge power threshold, wherein the cloud-based power dispatch plan does not exceed the range defined by the cloud-based power dispatch threshold.

[0140] Specifically, the market interaction module can assist the cloud-based intelligent agent in limiting the bidding range of available capacity at time t+1, i.e., the upper and lower limits of charging and discharging power. Available capacity It is a dynamic variable that is affected by a variety of factors, including: (1) Energy storage unit SOC and SOH; (2) Temperature and power limitations: physical limitations such as battery operating temperature and maximum charge and discharge power; (3) Historical operating data: charge and discharge characteristics and capacity decay patterns learned through historical data; (4) Uncertainty prediction: prediction of uncertainties such as future load and photovoltaic output, and dynamic adjustment of the confidence interval of available capacity.

[0141] Upper and lower limits of available capacity and Modeling can be done in the following ways:

[0142]

[0143] in, and These are the rated maximum discharge power and rated maximum charging power of the energy storage unit. It is the current average SOC of the energy storage unit. and These are the minimum and maximum permissible SOC of the energy storage unit. It is the total capacity of the energy storage units. It is the time step of the planning cycle. and This is the maximum charge / discharge power after taking into account the influence of SOH.

[0144] In the above formulas, both the discharge power and the charging power have three limiting values, including:

[0145] Equipment physical limitations: and ;

[0146] SOC security limits: ;

[0147] Health status limits: .

[0148] The discharge power is positive, and the charging power is negative. To meet their respective three limits, the discharge power should be the minimum of the three limits, and the charging power should be the maximum of the three limits. By implementing these limits, penalties for exceeding bidding capacity can be avoided, while maximizing the use of available capacity to gain market benefits.

[0149] Optionally, the market interaction module can also be used to assist the cloud-based intelligent agent in optimizing the total revenue objective function, as shown in the following formula:

[0150]

[0151] This is a deep discharge penalty, used to impose an additional penalty when the state of charge (SOC) of the energy storage unit falls below a safe threshold, thereby preventing over-discharge and extending battery life. Its construction formula is as follows:

[0152] penalty function It has non-linear characteristics; when the SOC approaches or falls below the safety threshold... At that time, the punishment increased rapidly. It is the deep discharge penalty coefficient. This is the discharge threshold.

[0153] Optional, There are two possible scenarios:

[0154] when At that time, the energy storage unit is relatively safe. (No penalty);

[0155] when At that time, energy storage units are at risk of loss.

[0156] ,

[0157] in, It is an exponent greater than 1, used to control the steepness of the penalty. It is a preset safety SOC lower limit, which is usually higher than the recommended minimum SOC to provide additional protection margin.

[0158] In the embodiments of this application, The key value lies in early warning and preventing deep discharge. This is in contrast to the degradation costs of energy storage devices. different, Judgment value Higher than This is equivalent to adding a buffer zone between the safe zone and the damaged zone. Without it... Even considering The cost of cloud-based intelligent agents may also lead them to disregard the need for short-term gains (such as charging more at a low price and discharging more at a high price) in pursuit of short-term profits. The limitations. When the two are combined, at the cloud level, To ensure the plan is not overly aggressive, a safety margin for the State of Operations (SOC) should be established in advance to reduce subsequent execution risks. At the execution level, This ensures that wear and tear from each charge and discharge cycle is factored into the cost, preventing stations from ignoring the true aging of the battery in order to keep up with the schedule and earn auxiliary service rewards.

[0159] The aforementioned energy storage system achieves collaborative optimization based on a two-layer architecture: a cloud-based scheduling optimization layer and a station-based control optimization layer. The cloud-based intelligent agent combines market information and equipment status to integrate energy and ancillary service market revenues, deducting equipment degradation and market risk costs, and formulates medium- to long-term bidding and scheduling plans. It reserves a SOC safety buffer through deep discharge penalties to mitigate aggressive risks and meet multiple constraints. The station-based intelligent agent responds in real-time at the second / minute level, accurately executing cloud plans while balancing local operating efficiency and single-device protection, meeting the safety constraints of individual energy storage units. A market interaction module further accurately estimates available capacity and adapts to market rules. Ultimately, this achieves a multi-objective balance of maximizing aggregator revenue, extending equipment lifespan, and ensuring stable grid operation, and can be flexibly adapted to different scenarios through parameter adjustments.

[0160] Corresponding to the above embodiments, this application also provides an electronic device, which serves as the control device for the above-mentioned energy storage system to realize the energy storage scheduling and control method. Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device 400 may include a processor 401, a memory 402, and a communication unit 403. These components communicate through one or more buses. Those skilled in the art will understand that the structure of the electronic device shown in the figure does not constitute a limitation on the embodiment of this application. It may be a bus-shaped structure or a star-shaped structure, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0161] The communication unit 403 is used to establish a communication channel, enabling the electronic device to communicate with other devices. It receives user data from other devices or sends user data to other devices.

[0162] The processor 401 serves as the control center of the electronic device, connecting various parts of the device via interfaces and lines. It executes software programs, instructions, and / or modules stored in the memory 402, and calls data stored in the memory to perform various functions and / or process data. The processor may be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 401 may consist only of a central processing unit (CPU). In this embodiment, the CPU may have a single processing core or include multiple processing cores.

[0163] The memory 402 is used to store the execution instructions of the processor 401. The memory 402 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0164] When the execution instructions in memory 402 are executed by processor 401, the electronic device 400 is able to perform some or all of the steps in the above embodiments.

[0165] In a specific implementation, this application also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, it may include some or all of the steps of the various embodiments of the energy storage scheduling and control method provided in this application. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0166] In a specific implementation, this application also provides a computer program product, wherein the computer program product includes executable instructions, which, when executed on a computer, cause the computer to perform some or all of the steps in the various embodiments of the energy storage scheduling and control method provided in this application.

[0167] This application also provides a non-transitory computer-readable storage medium that stores computer instructions, which cause the computer to execute the energy storage scheduling and control method provided in this application.

[0168] The aforementioned non-transitory computer-readable storage medium may be any combination of one or more computer-readable media. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or flash memory, optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in connection with an instruction execution system, apparatus, or device.

[0169] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including—but not limited to—electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of transmitting, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0170] The program code contained on a computer-readable medium may be transmitted using any suitable medium, including—but not limited to—wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0171] Those skilled in the art will clearly understand that the techniques in the embodiments of this application can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of this application, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application or some parts of the embodiments.

[0172] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments and terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

Claims

1. An energy storage system, characterized by, The cloud agent is configured to construct a total revenue target function based on cloud state information and cloud action information, and determine target cloud action information when the total revenue target function reaches a maximum value, the target cloud action information being used to indicate a cloud energy scheduling plan of the energy storage unit. The station agent is configured to, after obtaining the target cloud action information, construct a response target function based on the target cloud action information, station state information and station action information, and determine target station action information when the response target function reaches a maximum value, the target station action information being used to indicate a station energy scheduling plan of the energy storage unit. The energy storage unit is configured to perform a corresponding energy scheduling operation based on the target station action information. The cloud state information includes market price information and a cloud adjustment coefficient. The cloud agent is specifically configured to construct the total revenue target function based on energy market revenue, auxiliary service market revenue, energy storage device degradation cost and market risk cost, wherein the energy market revenue and the auxiliary service market revenue are determined by the market price information and the cloud action information, the energy storage device degradation cost is determined by the cloud adjustment coefficient and the cloud action information, and the market risk cost is determined by the market price information, the cloud adjustment coefficient and the cloud action information. The station state information includes a station adjustment coefficient. The station agent is specifically configured to construct the response target function based on scheduling tracking error penalty, operation cost, local device degradation cost and auxiliary service market reward, wherein the scheduling tracking error penalty is determined by a deviation degree of the station energy scheduling plan relative to the cloud energy scheduling plan, the operation cost is determined by the station adjustment coefficient and the station action information, the local device degradation cost is determined by the station adjustment coefficient and the station action information, and the auxiliary service market reward is determined by the station adjustment coefficient and the station action information. The market price information includes historical market price and real-time market price, and the cloud action information includes a first charge-discharge power bid under the historical market price and a second charge-discharge power bid under the real-time market price.

2. The energy storage system of claim 1, wherein, The cloud agent is further configured to determine historical market revenue based on the historical market price and the first charge-discharge power, determine real-time market revenue based on the real-time market price and the second charge-discharge power, and determine the energy market revenue based on the historical market revenue and the real-time market revenue. The market price information includes frequency modulation capacity price and frequency modulation mileage price, and the cloud action information includes a bid frequency modulation capacity and an actual frequency modulation mileage.

3. The energy storage system of claim 1, wherein, ​ The cloud agent is further configured to determine a frequency modulation capacity benefit based on the frequency modulation capacity price and the winning frequency modulation capacity, determine a frequency modulation mileage benefit based on the frequency modulation mileage price and the actual frequency modulation mileage, and determine the ancillary service market benefit based on the frequency modulation capacity benefit and the frequency modulation mileage benefit.

4. The energy storage system of claim 1, wherein, The market price information further includes an inertia response capacity price and an inertia response rate price, and the cloud action information includes a winning inertia response capacity and an actual inertia response rate. The cloud agent is further configured to determine an inertia response capacity benefit based on the inertia response capacity price and the winning inertia response capacity, determine an inertia response rate benefit based on the inertia response rate price and the actual inertia response rate, and determine the ancillary service market benefit based on the inertia response capacity benefit and the inertia response rate benefit.

5. The energy storage system of claim 1, wherein, The cloud adjustment coefficient further includes an aging cost coefficient and a depth penalty coefficient, and the cloud action information includes a number of charging and discharging times and a charging and discharging depth. The cloud agent is further configured to determine a cyclic aging cost based on the aging cost coefficient and the number of charging and discharging times, determine a charging and discharging depth cost based on the depth penalty coefficient and the charging and discharging depth, and determine the energy storage device degradation cost based on the cyclic aging cost and the charging and discharging depth cost.

6. The energy storage system of claim 1, wherein, The cloud adjustment coefficient further includes a risk aversion coefficient. The cloud agent is further configured to determine market risk information based on the cloud action information and the market price information, and determine the market risk cost based on the market risk information and the risk aversion coefficient.

7. The energy storage system of claim 1, wherein, The cloud power dispatching plan includes a planned charging and discharging power, the station-side power dispatching plan includes an actual charging and discharging power, and the station-side adjustment coefficient includes a tracking error penalty coefficient. The station-side agent is specifically configured to determine a difference degree of the planned charging and discharging power and the actual charging and discharging power, and determine the dispatching tracking error penalty based on the difference degree and the tracking error penalty coefficient.

8. The energy storage system of claim 1, wherein, The station-side adjustment coefficient further includes an efficiency loss cost coefficient and a charging and discharging efficiency. The station-side agent is further configured to determine the operation cost based on the efficiency loss cost coefficient, the charging and discharging efficiency, and the actual charging and discharging power.

9. The energy storage system of claim 1, wherein, The station-side adjustment coefficient further includes a local degradation cost coefficient, and the station-side action information includes a health condition change amount. The station-side agent is further configured to determine the local device degradation cost based on the local degradation cost coefficient and the health condition change amount.

10. The energy storage system of claim 1, wherein, The station-side adjustment coefficient further includes a frequency modulation reward coefficient and an inertia response reward coefficient, and the station-side action information includes a station-side frequency modulation mileage or a station-side inertia response rate. The station-side agent is further configured to determine a frequency modulation reward based on the frequency modulation reward coefficient and the station-side frequency modulation mileage, determine an inertia response reward based on the inertia response reward coefficient and the station-side inertia response rate, and determine the ancillary service market reward based on the frequency modulation reward and the inertia response reward.

11. The energy storage system of claim 1, wherein, The energy storage system further includes The market interaction module is configured to obtain device parameters of the energy storage unit, the device parameters comprising: a rated charging and discharging power threshold, a first charging and discharging power threshold, and a second charging and discharging power threshold, the first charging and discharging power threshold being determined by the power of the energy storage unit, and the second charging and discharging power threshold being determined by the health condition of the energy storage unit; The market interaction module is further configured to determine a cloud-side power scheduling threshold based on the rated charging and discharging power threshold, the first charging and discharging power threshold, and the second charging and discharging power threshold, wherein the cloud-side power scheduling plan is within a range defined by the cloud-side power scheduling threshold.

12. A method for energy storage dispatching control, characterized in that, The method is applied to the energy storage system according to any one of claims 1 to 11, and the method comprises: The cloud-side agent constructs a total revenue objective function based on cloud-side state information and cloud-side action information, and determines target cloud-side action information when the total revenue objective function reaches a maximum value, the target cloud-side action information being used to indicate a cloud-side power scheduling plan of the energy storage unit; After the station-side agent obtains the target cloud-side action information, the station-side agent constructs a response objective function based on the target cloud-side action information, station-side state information, and station-side action information, and determines target station-side action information when the response objective function reaches a maximum value, the target station-side action information being used to indicate a station-side power scheduling plan of the energy storage unit; The energy storage unit performs a corresponding power scheduling operation based on the target station-side action information; The cloud-side state information comprises: market price information and a cloud-side regulation coefficient; The cloud-side agent constructs the total revenue objective function based on energy market revenue, auxiliary service market revenue, energy storage device degradation cost, and market risk cost, wherein the energy market revenue and the auxiliary service market revenue are determined by the market price information and the cloud-side action information, the energy storage device degradation cost is determined by the cloud-side regulation coefficient and the cloud-side action information, and the market risk cost is determined by the market price information, the cloud-side regulation coefficient, and the cloud-side action information; The station-side state information comprises: a station-side regulation coefficient; The station-side agent constructs the response objective function based on scheduling tracking error penalty, operation cost, local device degradation cost, and auxiliary service market reward, wherein the scheduling tracking error penalty is determined by a deviation degree of the station-side power scheduling plan from the cloud-side power scheduling plan, the operation cost is determined by the station-side regulation coefficient and the station-side action information, the local device degradation cost is determined by the station-side regulation coefficient and the station-side action information, and the auxiliary service market reward is determined by the station-side regulation coefficient and the station-side action information.

Images