Energy storage control method of energy storage system based on cloud-side cooperation and energy storage system

Abstract

The invention discloses an energy storage control method of an energy storage system based on cloud edge collaboration and the energy storage system. The method comprises the steps that first energy storage control information of energy storage equipment is generated through a cloud server, wherein the energy storage control information is used for indicating and controlling execution of corresponding power charging and discharging operation on the energy storage equipment, controlling of electric energy ownership transfer on the energy storage equipment and controlling of electric energy scheduling on the energy storage equipment; in response to interruption of network connection between the cloud server and the edge end controller, generating second energy storage control information of the energy storage equipment through the edge end controller; and responding to network connection interruption recovery between the cloud server and the edge end controller, and recovering to use the cloud server to generate new first energy storage control information. According to the scheme, the energy storage system can have the self-operation capacity when communication is interrupted, and system energy efficiency optimization and system robustness are achieved.

Description

Technical Field

[0001] This invention relates to the field of new energy technology, and in particular to an energy storage control method and energy storage system based on cloud-edge collaboration. Background Technology

[0002] With the development of new energy technologies, the penetration rate of household distributed photovoltaic and energy storage devices has significantly increased. In current application scenarios, the energy storage system's regulation relies on cloud computing to generate control commands. That is, after analyzing and calculating relevant data through a cloud platform, specific control strategies are then sent to the energy storage system. However, the energy storage system regulation scheme is highly dependent on the network. Once a network outage occurs, the entire energy storage regulation system will fail, exhibiting poor robustness. The centralized data processing method in the cloud is susceptible to sudden events such as network congestion and server failures, which not only leads to insufficient reliability of control information execution but also directly reduces the stability of the energy storage system's operation, making it difficult to meet the continuous energy supply needs of household scenarios. Summary of the Invention

[0003] This invention provides an energy storage control method and energy storage system based on cloud-edge collaboration, enabling the energy storage system to operate autonomously during communication interruptions, thereby achieving system energy efficiency optimization and system robustness.

[0004] According to one aspect of the present invention, an energy storage control method for a cloud-edge collaborative energy storage system is provided, the energy storage system comprising an energy storage device, a cloud server, and an edge controller, the method comprising:

[0005] The cloud server generates first energy storage control information for the energy storage device. The energy storage control information is used to instruct the control to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller via data, and the edge controller is connected to the cloud server via communication.

[0006] In response to a network connection interruption between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing resources that the cloud server can utilize when generating the energy storage control information are greater than those that the edge controller can utilize when generating the energy storage control information. Furthermore, the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller.

[0007] In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server is used again to generate new first energy storage control information.

[0008] According to another aspect of the present invention, an energy storage system is provided, the energy storage system including an energy storage device, a cloud server, and an edge controller, the energy storage system being configured to perform the following operations:

[0009] The cloud server generates first energy storage control information for the energy storage device. The energy storage control information is used to instruct the control to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller via data, and the edge controller is connected to the cloud server via communication.

[0010] In response to a network connection interruption between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing resources that the cloud server can utilize when generating the energy storage control information are greater than those that the edge controller can utilize when generating the energy storage control information. Furthermore, the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller.

[0011] In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server is used again to generate new first energy storage control information.

[0012] The technical solution of this invention, in the process of energy storage control for energy storage systems, allows the cloud server to generate the first energy storage control information for the energy storage devices by leveraging its higher computing power. This effectively avoids control deviations caused by limited computing power at the edge, thereby accurately realizing energy storage control for charging and discharging operations, power ownership transfer, and peak shaving scheduling of the energy storage devices. Although the computing power and control accuracy at the edge are relatively low, when the network connection between the cloud server and the edge controller is interrupted, the edge controller can quickly generate the second energy storage control information locally. This solves the problem of paralysis due to network outages caused by dependence on the cloud server, avoiding power consumption risks or equipment damage caused by uncontrolled energy storage during network outages, and enabling the energy storage system to operate independently during communication interruptions. Moreover, when the network connection between the cloud server and the edge controller is restored, the system can switch back from the edge control mode to the cloud control mode in a timely manner. This avoids insufficient energy storage control accuracy caused by prolonged operation at the edge and ensures that energy storage scheduling quickly returns to a precise and optimized state, guaranteeing the stability and efficiency of the energy storage system during long-term operation.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0015] Figure 1 This is a flowchart illustrating an energy storage control method for a cloud-edge collaborative energy storage system according to an embodiment of the present invention.

[0016] Figure 2 This is a flowchart illustrating another energy storage control method for a cloud-edge collaborative energy storage system provided by an embodiment of the present invention.

[0017] Figure 3 This is a schematic diagram of a cloud-edge collaborative energy storage system provided according to an embodiment of the present invention. Detailed Implementation

[0018] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0019] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0020] Figure 1 This invention provides a flowchart of an energy storage control method for a cloud-edge collaborative energy storage system. This embodiment is applicable to situations where a cloud server and an edge controller in an energy storage system control the energy storage device. This energy storage control method for a cloud-edge collaborative energy storage system can be applied to an energy storage system, which can be implemented in hardware and / or software. The energy storage system can be configured in any device with network communication capabilities.

[0021] like Figure 1 As shown, the energy storage control method for a cloud-edge collaborative energy storage system in this embodiment includes the following processes:

[0022] S110. Generate the first energy storage control information of the energy storage device through the cloud server. The energy storage control information is used to instruct the controller to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller for data connection, and the edge controller is connected to the cloud server for communication connection.

[0023] An energy storage system can include energy storage devices, cloud servers, and edge controllers. For example, an energy storage system might include a 6kW photovoltaic array, a 10kWh residential energy storage device, and a smart load management system. The cloud server can be deployed as a collection of high-performance computing nodes in a remote data center, capable of handling massive data storage and complex model calculations through large-scale clustered hardware. The cloud server supports parallel multi-task processing and can rapidly increase computing power and storage through cluster expansion, with no upper limit on hardware resources. The cloud server is not directly physically connected to the energy storage devices but can interact with the edge controller via the internet (e.g., 4G / 5G, fiber optic).

[0024] The first energy storage control information can be a set of control commands generated by a cloud server for the entire lifecycle of energy storage devices, globally optimized for performing corresponding power charging and discharging operations, energy ownership transfer, and energy dispatch related to energy storage scheduling. Charging and discharging operations can be operations that allow bidirectional flow of electrical energy between energy storage devices and external energy systems (such as the power grid, photovoltaic systems, and loads) through energy conversion and transmission, thereby realizing the storage and release of electrical energy in the energy storage devices. Energy ownership transfer refers to the scheduling operation where the ownership of the electrical energy stored in the energy storage device changes or the usage rights are adjusted between different entities. Energy dispatch can refer to controlling the charging and discharging timing and power of energy storage devices to smooth out fluctuations in the power load of the power grid or local energy systems, responding to the demand of the power grid, controlling the discharge of energy storage devices to replenish energy during peak power consumption to reduce the pressure on the power grid; charging energy storage when using photovoltaic power generation or when electricity prices are low to minimize grid costs and maximize photovoltaic utilization; and controlling the charging of energy storage devices during off-peak power consumption to avoid wasting power on the power grid.

[0025] The energy storage device connects to an edge controller, allowing the edge controller to collect real-time operational data (such as the device's State of Charge (SoC)). This data connection can be a local wired connection or a short-range wireless connection (e.g., RS485). The edge controller also communicates with a cloud server, transmitting the collected operational data. This communication can be a wide area network (WAN) connection (e.g., 4G / 5G, Ethernet), depending on network stability.

[0026] Using the above method, in the process of energy storage control for energy storage systems, the cloud server can generate the first energy storage control information of the energy storage device with higher computing power resources, thereby effectively avoiding control deviations caused by limited computing power at the edge, and thus accurately realizing energy storage control of charging and discharging operations, power ownership transfer and peak shaving scheduling of energy storage devices.

[0027] S120. In response to the interruption of the network connection between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing power resources that the cloud server can call when generating energy storage control information are greater than those that the edge controller can call when generating energy storage control information. Furthermore, the control accuracy of the energy storage device by the energy storage control information generated by the cloud server is greater than the control accuracy of the energy storage device by the energy storage control information generated by the edge controller.

[0028] The second energy storage control information can be a set of control commands generated at the edge for energy storage devices in network outage scenarios, used to perform corresponding power charging and discharging operations, energy ownership transfer, and energy storage scheduling related to power dispatch. The second energy storage control information has a lower priority than the first energy storage control information. The second energy storage control information ensures the basic operation of the energy storage device rather than global optimization throughout its entire lifecycle. The cloud server can call upon large-scale clustered hardware such as multi-core CPUs and GPU clusters that support complex model calculations, while the edge controller only relies on embedded chips that support lightweight computing. Therefore, the computing resources that the cloud server can call upon when generating energy storage control information are greater than those that the edge controller can call upon when generating energy storage control information. The edge controller can be an intelligent energy gateway or energy storage controller deployed on the terminal device side of the energy storage system, with a built-in industrial-grade embedded processor running lightweight algorithm logic.

[0029] The difference in computing power between cloud servers and edge controllers allows cloud servers to handle larger datasets and more complex algorithms, while edge controllers can only process real-time local data and simple rules. Cloud servers can integrate long-term data and optimize energy storage control information from a global perspective, making them suitable for long-term planning and global optimization tasks, and enabling precise control using high computing power. Edge controllers, limited by computing power and data range, can only generate energy storage control information based on short-term data, and their decisions focus more on immediacy and security. They are suitable for real-time response and backup tasks in case of network outages, ensuring the basic operation of the energy storage system with limited computing power.

[0030] Although the computing power and control precision at the edge are relatively low when the network connection between the cloud server and the edge controller is interrupted, the edge controller can quickly generate secondary energy storage control information locally. This solves the problem of paralysis due to network outage caused by dependence on the cloud server, avoids the risk of power consumption or equipment damage caused by energy storage failure during network outage, and enables the energy storage system to operate on its own when communication is interrupted.

[0031] S130, In response to the restoration of the network connection between the cloud server and the edge controller after the interruption, the cloud server is used to generate new first energy storage control information again.

[0032] Once the network connection between the cloud server and the edge controller is restored from the interrupted state, it will automatically switch back to the cloud-led control mode. This means the cloud server will re-generate new primary energy storage control information, replacing the secondary energy storage control information generated by the edge controller. This optional solution ensures that the energy storage system can quickly return to a globally optimized state after network recovery. It avoids the risk of excessive energy storage control deviations due to insufficient accuracy in the edge controller's long-term operation. By switching, the cloud server, with its high computing power and full data, can generate better charging and discharging control information, ensuring the long-term efficient operation of the energy storage system and the continuity of energy storage scheduling.

[0033] As an optional but not limited implementation, the energy storage control method of the cloud-edge collaborative energy storage system in this embodiment may further include the following processes:

[0034] The first energy storage control information of the energy storage device is sent from the cloud server to the edge controller, so that the edge controller can control the energy storage device according to the first energy storage control information; or, the edge controller can control the energy storage device according to the second energy storage control information.

[0035] The initial energy storage control information generated by the cloud server uses the cloud server as the control hub, leveraging its global optimization capabilities to achieve precise scheduling. After the cloud server generates the initial energy storage control information, it sends the information to the edge controller through a communication connection. At this point, the edge controller acts as an intermediary for receiving and executing instructions, without participating in strategy formulation. It is responsible for converting the initial energy storage control information from the cloud server into control commands that the energy storage devices can recognize, driving the energy storage devices to operate according to the energy storage scheduling strategy corresponding to the initial energy storage control information from the cloud server.

[0036] The second energy storage control information generated by the edge controller, with the edge controller as the control core, ensures uninterrupted energy storage scheduling of energy storage devices in the energy storage system even in scenarios with network outages or poor network connectivity. When the network connection between the cloud server and the edge controller is interrupted (e.g., due to network failure or signal loss), the edge controller automatically triggers local control logic, directly generating second energy storage control information based on the operational data related to the energy storage devices obtained by the edge controller to schedule the energy storage devices. While the accuracy of the second energy storage control information generated by the edge controller is lower than the energy storage control commands corresponding to the first energy storage control information generated by the cloud server, it can quickly determine the energy storage scheduling strategy for the energy storage devices through lightweight algorithms. This avoids the loss of control over the energy storage devices due to the absence of cloud control, thus ensuring both the accuracy and economy of the energy storage system's normal operation and mitigating operational risks caused by network dependence.

[0037] The technical solution of this invention, in the process of energy storage control for energy storage systems, allows the cloud server to generate the first energy storage control information for the energy storage devices by leveraging its higher computing power. This effectively avoids control deviations caused by limited computing power at the edge, thereby accurately realizing energy storage control for charging and discharging operations, power ownership transfer, and peak shaving scheduling of the energy storage devices. Although the computing power and control accuracy at the edge are relatively low, when the network connection between the cloud server and the edge controller is interrupted, the edge controller can quickly generate the second energy storage control information locally. This solves the problem of paralysis due to network outages caused by dependence on the cloud server, avoiding power consumption risks or equipment damage caused by uncontrolled energy storage during network outages, and enabling the energy storage system to operate independently during communication interruptions. Moreover, when the network connection between the cloud server and the edge controller is restored, the system can switch back from the edge control mode to the cloud control mode in a timely manner. This avoids insufficient energy storage control accuracy caused by prolonged operation at the edge and ensures that energy storage scheduling quickly returns to a precise and optimized state, guaranteeing the stability and efficiency of the energy storage system during long-term operation.

[0038] Figure 2This is a flowchart illustrating another energy storage control method for a cloud-edge collaborative energy storage system provided by an embodiment of the present invention. The technical solution of this embodiment further optimizes the process of generating the first energy storage control information of the energy storage device through the cloud server and the second energy storage control information of the energy storage device through the edge controller in the above embodiments based on the technical solution of the above embodiments. This embodiment can be combined with various optional solutions in one or more of the above embodiments.

[0039] like Figure 2 As shown, the energy storage control method for a cloud-edge collaborative energy storage system in this embodiment may include the following processes:

[0040] S210. Call the first processing model configured on the cloud server to predict and generate first reference type information related to the energy storage device, and generate first energy storage control information of the energy storage device based on the first reference type information. The first processing model is a complex model adapted to the computing power of the cloud server to predict and generate reference type information. The first reference type information is used to indicate the reference type information generated by the first processing model within the first time period after the current moment. The reference type information includes the electricity load trend, photovoltaic power generation trend and electricity price change trend related to the energy storage device.

[0041] Among them, energy storage control information is used to instruct and control the corresponding power charging and discharging operations of energy storage devices, control the transfer of power ownership of energy storage devices, and control the power scheduling of energy storage devices. The energy storage devices are connected to the edge controller for data exchange, and the edge controller is connected to the cloud server for communication.

[0042] The first processing model is a prediction model deployed on a cloud server. It leverages the high computing power of the cloud server to predict and generate reference-type information. Utilizing the ample computing resources of the cloud, it supports deep computation and complex logical reasoning of multi-dimensional data. Unlike the lightweight computation of edge controllers, the first processing model can handle larger datasets and more refined prediction tasks. The first processing model can be a complex model trained on long-term data to predict trends in electricity load, photovoltaic power generation, and electricity price changes. For example, it can be built using LSTM or Transformer frameworks.

[0043] The first reference type information is used to indicate the trends in electricity load, photovoltaic power generation, and electricity price changes related to energy storage devices within a first time period starting from the current moment. This first reference type information can be data output by the first processing model to guide energy storage scheduling, representing a prediction of the three major trends in electricity load, photovoltaic power generation, and electricity price within a specific future time period. Specifically, the electricity load trend indicates the direction and magnitude of changes in power consumption of energy storage devices over time; the photovoltaic power generation trend indicates the direction and magnitude of changes in the output power of the photovoltaic system related to the energy storage devices over time; and the electricity price change trend is a comprehensive description of the direction, magnitude, and pattern of dynamic fluctuations in electricity prices over a specific time period. For example, the photovoltaic power generation trend uses a photovoltaic power generation curve. This indicates that the electricity load trend is represented by the household load curve. The trend of electricity price changes is represented by real-time electricity prices. express.

[0044] As an optional but not limited implementation, the first processing model configured on the cloud server is invoked to predict and generate first reference type information related to the energy storage device, including but not limited to the following steps:

[0045] The system acquires first operational data related to the energy storage system through a cloud server. This first operational data is collected by an edge controller and third-party devices and uploaded to the cloud server. The first operational data includes dynamic electricity price data of the power grid, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, health status data of the energy storage device, and meteorological data of the area where the energy storage device is located. Based on the first operational data, the system calls a first processing model configured on the cloud server to predict and generate first reference type information related to the energy storage device.

[0046] The first operational data can include a multi-dimensional set of operational data covering the entire energy storage system scenario when supporting cloud server energy storage scheduling. This first operational data includes data directly related to the energy storage devices, such as household electricity load data and the health status data of the energy storage devices; photovoltaic power generation data related to the energy storage devices; and external environmental and market data (such as dynamic grid electricity price data and meteorological data of the area where the energy storage devices are located), ensuring that the data fully reflects the internal and external conditions of the energy storage system's operation. For example, the first operational data can include data collected by the cloud server over a historical period of 30 days, such as electricity load, photovoltaic power generation, grid electricity price curves, and weather conditions of the energy storage devices. An LSTM network is then used to train a first processing model, which outputs the electricity load trend, photovoltaic power generation trend, and electricity price change trend related to the energy storage devices for the next 24 hours, serving as the first reference type of information.

[0047] The edge controller can be responsible for directly collecting the operation data related to the energy storage system. The third-party device is a device that does not directly participate in the energy storage control but only provides specific external data, and it is a supplementary data collection source for the first operation data. For example, the power market data terminal that collects the dynamic electricity price data of the power grid, the meteorological sensor that collects the regional meteorological data, etc. The role of the third-party device is to fill the external data that cannot be covered by the edge controller and ensure the integrity of the first operation data.

[0048] The dynamic electricity price data of the power grid can reflect the real-time data of the power grid electricity price changing over time, which is the key basis for measuring the economy of energy storage scheduling and contains information such as electricity price values at different time periods and electricity price adjustment nodes. The household electricity load data can reflect the real-time data of the household electricity demand associated with the energy storage device and contains information such as electricity consumption power and electricity consumption time distribution, etc., which is used to judge the scale of local electricity demand that the energy storage device needs to meet. The photovoltaic power generation data reflects the real-time data of the power generation capacity of the photovoltaic system supporting the energy storage device and contains information such as photovoltaic output power and output time characteristics, etc., which is used to predict the amount of renewable energy supply that the energy storage system can obtain. The health status data of the energy storage device can reflect the real-time data of the performance and life status of the energy storage device and contains information such as health status, internal resistance, consistency of cell temperature and voltage, etc., which is used to constrain the model prediction and scheduling strategy to avoid equipment damage caused by scheduling and extend the service life of the equipment. The meteorological data of the location area where the energy storage device is located can reflect the data of the surrounding environmental conditions of the energy storage device and contains information such as light intensity, temperature, wind speed, etc., which is used to assist in predicting the change trend of the photovoltaic power generation data.

[0049] By adopting the above method, due to the limitations of the deployment scenario and hardware resources, the edge controller cannot independently obtain external data such as the dynamic electricity price of the power grid and regional meteorology. However, with the participation of the third-party device and the aggregation function of the cloud, through the collaborative collection of the edge controller and the third-party device, multi-dimensional data such as the power grid, household, photovoltaic, device, and meteorology are integrated into the first operation data, breaking through the limitations of a single data source. The comprehensive data input avoids prediction deviations caused by data loss.

[0050] S220. In response to the network connection interruption between the cloud server and the edge controller, call the second processing model configured on the edge controller to predict and generate the second reference type information related to the energy storage device, and generate the second energy storage control information of the energy storage device according to the second reference type information. The second processing model is a lightweight model adapted to the computing power of the edge controller to predict and generate the reference type information. The second reference type information is used to indicate the reference type information generated by the second processing model within the second time period starting from the current moment. The value of the first time period is greater than the value of the second time period.

[0051] Among them, the computing power resources that the cloud server can call when generating energy storage control information are greater than those that the edge controller can call when generating energy storage control information, and the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller.

[0052] The second processing model is a prediction model deployed on the edge controller. It is adapted to the low computing power and lightweight design of the edge. The second processing model aims to consume less computing resources and have a fast operation speed. By simplifying the algorithm architecture and compressing the parameter scale, it can achieve fast prediction on the limited hardware resources of the edge controller (such as embedded chips and small-capacity storage). Unlike the complex models of cloud servers, it prioritizes response speed rather than extreme accuracy.

[0053] The second processing model is designed for the low computing power characteristics of the edge. By simplifying algorithms (such as using lightweight models like ARIMA and MiniGRU), the computational complexity is reduced, resulting in a response speed far faster than the minutes-level response speed of complex models in the cloud. This resource adaptability ensures that the edge controller can quickly output effective control commands under limited hardware resources, without causing itself to lag or other functions to fail due to excessive resource consumption by model computation, thus balancing response speed and resource consumption.

[0054] The second reference type information can be a predictive dataset for short-cycle scheduling output by the second processing model. The second duration defines the prediction time span (e.g., 1-6 hours), adapting to the real-time response requirements of the edge controller. The second reference type information also includes electricity load trends, photovoltaic power generation trends, and electricity price change trends, but its data dimensions and prediction granularity are simpler than the first reference type information on the cloud server, balancing prediction speed and basic decision-making needs. The first duration value is greater than the second duration value, ensuring that the prediction time span indicated in the first reference type information on the cloud server (e.g., 24-72 hours) is longer than the prediction time span indicated in the second reference type information on the edge controller (e.g., 1-6 hours). The cloud server supports long-cycle global planning, while the edge controller only needs to meet short-cycle real-time fallback requirements.

[0055] The second reference type of information uses the second duration as the prediction time span, focusing on key trend predictions in the short term. This avoids the deviations caused by data uncertainty in long-term predictions and better meets the real-time backup needs of the edge. It only needs to predict the electricity consumption, photovoltaic, and electricity price trends in the next few hours to formulate effective short-term control strategies without having to consider complex long-term variables. This reduces the difficulty and error of prediction, improves the reliability of control commands during grid outages, and reduces equipment damage or power imbalance caused by prediction deviations.

[0056] By using a mechanism that triggers a local model during network interruption, the dependence of energy storage systems on cloud servers is resolved. When the network is interrupted, the edge controller does not need to wait for the cloud to recover. It can immediately generate predictive information and control commands through the second processing model to drive the energy storage equipment to maintain basic operation, avoiding the risk of equipment operation due to loss of cloud control, and giving the energy storage system the ability to operate autonomously without being paralyzed when the network is interrupted.

[0057] As an optional but not limited implementation, the second processing model configured on the edge controller is invoked to predict and generate second reference type information related to the energy storage device, including but not limited to the following steps:

[0058] The second operating data related to the energy storage system is obtained through the edge controller. The second operating data is collected by the edge controller and uploaded to the cloud server. The second operating data includes grid dynamic electricity price data, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, and the state of charge of the energy storage device. Based on the second operating data, the second processing model configured on the edge controller is called to predict and generate second reference type information related to the energy storage device.

[0059] The second type of operational data can be operational data related to the energy storage device that the edge controller can directly collect from the energy storage device and adapt to local lightweight processing. For example, the second type of operational data may include the photovoltaic power of the photovoltaic device related to the energy storage device. Power consumption of energy storage equipment The second operational data includes the state-of-charge (SoC) of energy storage devices and dynamic electricity prices from the grid. Unlike multi-source data integrated in the cloud, the second operational data focuses on data that can be acquired locally by the edge controller without relying on third-party devices, ensuring rapid local collection and use. The second operational data only includes the core dimensions that can be collected locally at the edge, eliminating complex external data required by the cloud (such as regional meteorological data and in-depth parameters of device health status), resulting in smaller data volume and lower processing difficulty. At the same time, the second processing model is lightweight and can complete prediction calculations based on the second operational data without consuming too much computing power and storage resources at the edge. This ensures that the edge can efficiently complete data processing and prediction with limited hardware resources, without affecting other functions due to resource overload.

[0060] As an optional but not limited implementation, the first energy storage control information of the energy storage device is generated based on the first reference type information, including but not limited to the following steps:

[0061] The first reference type information is input into the third processing model configured on the cloud server. The third processing model generates the first energy storage control information of the energy storage device. The third processing model is a reinforcement learning model that uses the reference type information as the objective function to achieve the target energy storage scheduling requirements as the optimization direction, and outputs energy storage control information that adapts to the operation requirements of the energy storage device through reinforcement learning strategy learning and iterative optimization.

[0062] Accordingly, second energy storage control information for the energy storage device is generated based on the second reference type information, including but not limited to the following steps:

[0063] The system detects whether different reference type information in the second reference type information meets the preset energy storage control conditions associated with each reference type information. Each preset energy storage control condition associated with each reference type information is associated with energy storage control information adapted to each reference type information. Based on whether each reference type information meets the preset energy storage control conditions associated with each reference type information, the system generates the second energy storage control information for the energy storage device. The computational complexity of generating the first energy storage control information based on the first reference type information is greater than the computational complexity of generating the second energy storage control information based on the first reference type information, and the response speed of generating the first energy storage control information based on the first reference type information is greater than the response speed of generating the second energy storage control information based on the first reference type information.

[0064] The third processing model is a cloud-based reinforcement learning model. It optimizes the objective function used to achieve the target energy storage scheduling requirements, constructing a decision-making loop of state, action, and reward through reinforcement learning algorithms. It transforms the first reference type information into the system state, uses energy storage control operations as output actions, and uses the achievement of the objective function as the reward signal. Through continuous learning and optimization strategies, it ensures that the output control information achieves the preset energy storage scheduling requirements. The predicted first reference type information is input into the axis of the third processing model (such as the reinforcement learning model built by DDPG) to optimize the objective function used to achieve the target energy storage scheduling requirements.

[0065] ;

[0066] in, This represents the electricity price at time t; This represents the power sold to the grid at time t; This represents the power purchased from the grid at time t; , This represents the weighting coefficient, which is set according to user preferences. ; Estimated carbon credit consumption; , representing the energy storage state decision at each moment (such as charging and discharging power).

[0067] The objective function for the target energy storage dispatch demand can be a mathematical function that quantifies the core requirements of energy storage dispatch. It serves as the optimization criterion for the third-processing model and typically includes multi-dimensional objectives such as economic efficiency (e.g., maximizing revenue, minimizing costs), safety (e.g., minimizing battery wear), and system synergy (e.g., maximizing grid peak shaving contribution). These objectives are weighted to form a comprehensive optimization index, guiding the third-processing model to generate the optimal control strategy. For example, the objective function may include, but is not limited to, the following: a function to minimize electricity purchase costs; a function to maximize electricity sales revenue; and a user comfort function. , , Penalty coefficient; Carbon emission reduction estimation function: , where ϵ(t) is the carbon emission factor per unit of electricity purchased from the grid.

[0068] The first reference type information generated by the first processing model is used as input and passed to the third processing model pre-configured on the cloud server. Guided by the preset target energy storage scheduling requirements, the third processing model uses a reinforcement learning strategy learning mechanism and multiple rounds of iterative optimization to finally output the first energy storage control information that adapts to the full life cycle operation requirements of the energy storage device. Adapting to the operation requirements of the energy storage device means that the first energy storage control information must conform to the hardware constraints of the energy storage device, avoid generating control commands that exceed the device's capabilities, and also take into account the long-term lifespan of the device.

[0069] The preset energy storage control conditions can be rule thresholds pre-configured by the edge controller and corresponding to each reference type information in the second reference type information. These thresholds serve as the basis for determining whether to trigger energy storage control commands. Different reference type information is associated with different energy storage control conditions. For each reference type information, when a reference type information meets its corresponding preset energy storage control condition, the corresponding energy storage control information is directly output (e.g., if the electricity price exceeds a threshold, discharge is triggered; if the load exceeds the power limit, self-consumption discharge is prioritized), without complex calculations. The edge controller checks each dimension of the second reference type information one by one, determining whether each reference type information meets its corresponding preset energy storage control condition. Based on the detection results of each dimension of the reference type information, it matches the corresponding preset energy storage control information, ultimately integrating them to form the second energy storage control information.

[0070] The computational complexity of generating the first energy storage control information based on the first reference type information is higher because it requires complex operations such as policy iteration and objective function optimization of the reinforcement learning model. The computational complexity of generating the second energy storage control information based on the second reference type information is lower. It only requires simple logical judgment of data detection and rule matching, and heuristic optimization (such as dynamic threshold control) can be used to quickly generate the second energy storage control information.

[0071] S230, in response to the restoration of the network connection between the cloud server and the edge controller after the interruption, resume the generation of new first energy storage control information using the cloud server.

[0072] As an optional but not limited implementation, the energy storage control method of the cloud-edge collaborative energy storage system in this embodiment may include the following process:

[0073] In response to the restoration of network connection between the cloud server and the edge controller, the operating parameters of the second processing model configured on the edge controller during the network connection interruption are sent to the cloud server for archiving via the edge controller; based on the operating parameters of the second processing model during the network connection interruption, the first processing model configured on the cloud server is corrected via the cloud server to update the first processing model.

[0074] The operating parameters of the second processing model refer to the key data generated during the operation of the second processing model when the network connection between the cloud server and the edge controller is interrupted. This includes the second operating data input to the second processing model, the second reference type information output, the executed control strategy records, and the model's own computational parameters. This data is the core data reflecting the actual operating status of the edge controller and the decision-making logic of the second processing model during the network connection interruption. The cloud server can perform structured storage and classification management of the received operating parameters from the edge controller during the network outage, forming a traceable historical database. This facilitates subsequent troubleshooting and operational review, and also provides data support for the calibration of the first processing model. The calibration of the first processing model refers to the cloud server analyzing the deviation between the previous predictions and actual operation of the first processing model based on the operating parameters of the edge controller during the network connection interruption, and correcting the deviation by adjusting model parameters and optimizing algorithm logic.

[0075] Uploading and archiving operational parameters during network outages at the edge fills the gaps in operational data during network interruptions, ensuring complete and traceable operational data throughout the energy storage system's entire lifecycle. In subsequent maintenance, if scheduling anomalies occur, the archived data can be used to trace model decisions and equipment status during the outage period, quickly pinpointing the root cause of the problem. Correcting the first-processing model based on real operational parameters from the edge effectively corrects prediction biases caused by a lack of actual outage data, making the model more closely aligned with complex and ever-changing real-world operating scenarios. The corrected first-processing model generates more accurate first-reference type information.

[0076] As an optional but not limited implementation, the energy storage control method of the cloud-edge collaborative energy storage system in this embodiment may further include the following processes:

[0077] Based on the correction results of the first processing model configured on the cloud server, an update strategy and model parameters for the second processing model are generated and sent to the edge controller to update the second processing model configured on the edge controller; and, in response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the third operating data related to the energy storage system is synchronized to the edge controller through the cloud server. The third operating data includes grid dynamic electricity price data, photovoltaic power generation data related to the energy storage device, and meteorological data of the area where the energy storage device is located.

[0078] As an optional but not limited implementation, the energy storage control method of the cloud-edge collaborative energy storage system in this embodiment may further include the following processes:

[0079] In response to the triggered operation of the target energy storage scheduling demand, the objective function used to achieve the target energy storage scheduling demand is determined from multiple objective functions by the cloud server, and the objective function adopted by the third processing model is determined.

[0080] The triggering operation for target energy storage dispatch demand can refer to the behavior of initiating dispatch demand, including proactive operations (such as selecting a mode via an app) and automatic triggering (such as grid peak-valley switching or sudden changes in photovoltaic output). These triggers the selection of the objective function, ensuring that the dispatch strategy can be dynamically adjusted according to demand. Multiple objective functions correspond to different energy storage dispatch demand modes pre-configured on the cloud server. Each objective function corresponds to a different energy storage dispatch demand mode for the energy storage device, namely, energy saving priority mode, environmental protection priority mode, reserve power mode, and profit priority mode, as shown in Table 1 below.

[0081] Table 1

[0082] model Weighting coefficient settings Energy conservation priority α=0,β=0 Environmental protection first β = high value Backup power priority Strengthen SoC minimum value constraints or penalties Profitability First Increase the weighting of electricity sales revenue γ(t)

[0083] In this embodiment of the energy storage system, the cloud server is used to collect and store multi-dimensional operational data from the home energy storage system, including dynamic grid electricity price data, household electricity load data, photovoltaic power generation data, weather, and battery energy storage device health status data. Based on long-term data training, a prediction model (such as LSTM or Transformer) is used to predict household electricity load trends and photovoltaic power generation trends. Multi-objective optimization algorithms (such as reinforcement learning) are executed to generate high-precision energy storage device control strategies. Updated strategies and model parameters are periodically distributed to edge devices. The edge controller is deployed on the user-end smart energy gateway or energy storage controller; it collects local operational data in real time (photovoltaic power, load, SoC, electricity price, etc.); it performs lightweight prediction and rapid optimization locally to generate current scheduling instructions; it can operate independently to ensure basic energy scheduling functions when the network is disconnected, reducing reliance on cloud responses; and it supports user preference settings, such as switching between modes like "cost priority," "low-carbon priority," and "comfort priority."

[0084] The technical solution of this invention, in the process of energy storage control for energy storage systems, allows the cloud server to generate the first energy storage control information for the energy storage devices by leveraging its higher computing power. This effectively avoids control deviations caused by limited computing power at the edge, thereby accurately realizing energy storage control for charging and discharging operations, power ownership transfer, and peak shaving scheduling of the energy storage devices. Although the computing power and control accuracy at the edge are relatively low, when the network connection between the cloud server and the edge controller is interrupted, the edge controller can quickly generate the second energy storage control information locally. This solves the problem of paralysis due to network outages caused by dependence on the cloud server, avoiding power consumption risks or equipment damage caused by uncontrolled energy storage during network outages, and enabling the energy storage system to operate independently during communication interruptions. Moreover, when the network connection between the cloud server and the edge controller is restored, the system can switch back from the edge control mode to the cloud control mode in a timely manner. This avoids insufficient energy storage control accuracy caused by prolonged operation at the edge and ensures that energy storage scheduling quickly returns to a precise and optimized state, guaranteeing the stability and efficiency of the energy storage system during long-term operation.

[0085] Figure 3 This invention provides a schematic diagram of a cloud-edge collaborative energy storage system. This embodiment is applicable to situations where a cloud server and an edge controller control the energy storage device in a cloud-edge collaborative energy storage system. The energy storage system can be implemented in hardware and / or software and can be configured in any device with network communication capabilities.

[0086] like Figure 3As shown, the cloud-edge collaborative energy storage system of this embodiment includes an energy storage device 310, a cloud server 330, and an edge controller 320. The energy storage system is configured to perform the following operations:

[0087] The cloud server generates first energy storage control information for the energy storage device. The energy storage control information is used to instruct the control to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller via data, and the edge controller is connected to the cloud server via communication.

[0088] In response to a network connection interruption between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing resources that the cloud server can utilize when generating the energy storage control information are greater than those that the edge controller can utilize when generating the energy storage control information. Furthermore, the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller.

[0089] In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server is used again to generate new first energy storage control information.

[0090] Based on the above embodiments, optionally, the first energy storage control information of the energy storage device is generated through the cloud server, including:

[0091] The system calls a first processing model configured on the cloud server to predict and generate first reference type information related to the energy storage device. Based on the first reference type information, it generates first energy storage control information for the energy storage device. The first processing model is a complex model adapted to the computing power of the cloud server to predict and generate reference type information. The first reference type information is used to indicate the reference type information generated by the first processing model within a first time period after the current moment. The reference type information includes the electricity load trend, photovoltaic power generation trend, and electricity price change trend related to the energy storage device.

[0092] The edge controller generates second energy storage control information for the energy storage device, including:

[0093] The second processing model configured on the edge controller is invoked to predict and generate second reference type information related to the energy storage device. Based on the second reference type information, second energy storage control information of the energy storage device is generated. The second processing model is a lightweight model adapted to the computing power of the edge controller to predict and generate reference type information. The second reference type information is used to indicate the reference type information generated by the second processing model within a second duration starting from the current moment. The value of the first duration is greater than the value of the second duration.

[0094] Based on the above embodiments, optionally, a first processing model configured on the cloud server is invoked to predict and generate first reference type information related to the energy storage device, including:

[0095] The first operational data related to the energy storage system is obtained through the cloud server. The first operational data is collected by the edge controller and third-party devices and uploaded to the cloud server. The first operational data includes dynamic electricity price data of the power grid, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, health status data of the energy storage device, and meteorological data of the area where the energy storage device is located.

[0096] Based on the first operating data, the first processing model configured on the cloud server is invoked to predict and generate the first reference type information related to the energy storage device.

[0097] Based on the above embodiments, optionally, a second processing model configured on the edge controller is invoked to predict and generate second reference type information related to the energy storage device, including:

[0098] The edge controller acquires second operating data related to the energy storage system. The second operating data is collected by the edge controller and uploaded to the cloud server. The second operating data includes grid dynamic electricity price data, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, and the state of charge of the energy storage device.

[0099] Based on the second operating data, the second processing model configured on the edge controller is invoked to predict and generate the second reference type information related to the energy storage device.

[0100] Based on the above embodiments, optionally, the generation of first energy storage control information for the energy storage device according to the first reference type information includes:

[0101] The first reference type information is input into the third processing model configured on the cloud server, and the first energy storage control information of the energy storage device is generated through the third processing model. The third processing model is a reinforcement learning model that uses the reference type information as the objective function to achieve the target energy storage scheduling requirements as the optimization direction, and outputs energy storage control information that adapts to the operation requirements of the energy storage device through reinforcement learning strategy learning and iterative optimization.

[0102] The second energy storage control information of the energy storage device is generated based on the second reference type information, including:

[0103] The system detects whether different reference type information in the second reference type information meets the preset energy storage control conditions associated with each reference type information. The preset energy storage control conditions associated with each reference type information are associated with energy storage control information adapted to each reference type information.

[0104] Based on whether each reference type of information meets the preset energy storage control conditions associated with each reference type of information, the second energy storage control information of the energy storage device is generated;

[0105] Specifically, the computational complexity of generating the first energy storage control information based on the first reference type information is greater than the computational complexity of generating the second energy storage control information based on the first reference type information, and the response speed of generating the first energy storage control information based on the first reference type information is greater than the response speed of generating the second energy storage control information based on the first reference type information.

[0106] Optionally, based on the above embodiments, the energy storage system of this embodiment further includes:

[0107] In response to the restoration of the network connection between the cloud server and the edge controller, the operating parameters of the second processing model configured on the edge controller during the network connection interruption are sent to the cloud server for archiving via the edge controller.

[0108] Based on the operating parameters of the second processing model during the network connection interruption, the first processing model configured on the cloud server is corrected by the cloud server to update the first processing model.

[0109] Optionally, based on the above embodiments, the energy storage system of this embodiment further includes:

[0110] Based on the correction results of the first processing model configured on the cloud server, an update strategy and model parameters for the second processing model are generated and sent to the edge controller to update the second processing model configured on the edge controller; and, in response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the third operating data related to the energy storage system is synchronized to the edge controller through the cloud server. The third operating data includes grid dynamic electricity price data, photovoltaic power generation data related to the energy storage device, and meteorological data of the area where the energy storage device is located.

[0111] Optionally, based on the above embodiments, the energy storage system of this embodiment further includes:

[0112] In response to the triggering operation of the target energy storage scheduling requirement, the objective function used by the third processing model is determined from multiple objective functions by the cloud server to achieve the target energy storage scheduling requirement.

[0113] Optionally, based on the above embodiments, the method further includes:

[0114] The cloud server sends the first energy storage control information of the energy storage device to the edge controller, so that the edge controller can perform energy storage control on the energy storage device according to the first energy storage control information; or...

[0115] The edge controller performs energy storage control on the energy storage device according to the second energy storage control information.

[0116] The technical solution of this invention, in the process of energy storage control for energy storage systems, allows the cloud server to generate the first energy storage control information for the energy storage devices by leveraging its higher computing power. This effectively avoids control deviations caused by limited computing power at the edge, thereby accurately realizing energy storage control for charging and discharging operations, power ownership transfer, and peak shaving scheduling of the energy storage devices. Although the computing power and control accuracy at the edge are relatively low, when the network connection between the cloud server and the edge controller is interrupted, the edge controller can quickly generate the second energy storage control information locally. This solves the problem of paralysis due to network outages caused by dependence on the cloud server, avoiding power consumption risks or equipment damage caused by uncontrolled energy storage during network outages, and enabling the energy storage system to operate independently during communication interruptions. Moreover, when the network connection between the cloud server and the edge controller is restored, the system can switch back from the edge control mode to the cloud control mode in a timely manner. This avoids insufficient energy storage control accuracy caused by prolonged operation at the edge and ensures that energy storage scheduling quickly returns to a precise and optimized state, guaranteeing the stability and efficiency of the energy storage system during long-term operation.

[0117] The cloud-edge collaborative energy storage system provided in this embodiment of the invention can execute the energy storage control method of the cloud-edge collaborative energy storage system provided in any of the above embodiments of the invention. It has the corresponding functions and beneficial effects of executing the energy storage control method of the cloud-edge collaborative energy storage system. For details, please refer to the relevant operations of the energy storage control method of the cloud-edge collaborative energy storage system in the foregoing embodiments.

[0118] It is worth noting that the various units and modules included in the above-mentioned device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of the present invention.

[0119] According to embodiments of the present invention, the process described in the above-described flowchart can be applied to energy storage systems. For example, an embodiment of the present invention includes an energy storage system comprising: at least one battery module; at least one battery management system, wherein the at least one battery management system corresponds one-to-one with the at least one battery module, and the battery management system is electrically connected to the corresponding battery module. The energy storage system includes an energy storage device, a cloud server, and an edge controller. The energy storage system is used to execute the program code of the energy storage control method of the cloud-edge collaborative energy storage system shown in the flowchart.

[0120] In such embodiments, the energy storage system provided in this application is applicable to various application scenarios, such as grid-connected power generation and energy storage, off-grid photovoltaic and energy storage (for powering electrical equipment in homes, RVs, and yachts), wind power storage and energy generation, and electric equipment. The specific application scenario can be determined according to the actual application scenario, and no limitations are imposed here. The following will use the off-grid photovoltaic and energy storage field as an example for explanation. Other application scenarios are basically similar and will not be described in detail.

[0121] In off-grid photovoltaic (PV) and energy storage (ESS) applications, a complete PV-ESS system includes at least a photovoltaic (PV) power generation system, a power conversion system, an energy storage system, and a power consumption system. The PV power generation system consists of several solar panels connected in series and parallel to convert solar energy into electrical energy. The power conversion system injects the electrical energy generated by the PV power generation system into the energy storage system for storage. The power consumption system then adapts the stored electrical energy to the power required by the electrical equipment. The aforementioned power conversion system can typically be implemented using a DC / DC converter with MPPT (Multi-Level Photovoltaic) functionality, while the power consumption system can typically be implemented using a DC / DC converter or a DC / AC converter. This section focuses on the energy storage system. An energy storage system is usually composed of multiple battery packs connected together. Connecting battery packs in series increases the output voltage of the battery pack, while connecting them in parallel achieves a larger battery capacity. Therefore, to obtain an energy storage system with a target voltage level and capacity, users will connect multiple battery packs in series and parallel to obtain a high-voltage, high-capacity energy storage system for both energy storage and power supply. In addition, for convenient centralized control, each battery pack in an energy storage system is typically equipped with a communication interface, allowing the battery packs to communicate with each other via methods such as CAN bus, RS485 bus, and RS232 bus. Currently, with the further development of the new energy industry and the expansion of application scenarios, higher requirements are being placed on various performance aspects of energy storage systems, including system stability and consistency, waterproofing, dustproofing, and safety levels.

[0122] In an optional but not limited implementation, this application also provides a photovoltaic power generation system, which includes: a photovoltaic power generation module, a controller, and an energy storage system according to any one of the above embodiments; the photovoltaic power generation module is used to convert solar energy into direct current (DC) power; and the controller is used to store the DC power in the energy storage system.

[0123] It should be understood that the various processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein. The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An energy storage control method for a cloud-edge collaborative energy storage system, characterized in that, The energy storage system includes energy storage devices, a cloud server, and an edge controller; the method includes: The cloud server generates first energy storage control information for the energy storage device. The energy storage control information is used to instruct the control to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller via data, and the edge controller is connected to the cloud server via communication. In response to a network connection interruption between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing resources that the cloud server can utilize when generating the energy storage control information are greater than those that the edge controller can utilize when generating the energy storage control information. Furthermore, the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller. In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server is used again to generate new first energy storage control information.

2. The method according to claim 1, characterized in that, The cloud server generates the first energy storage control information for the energy storage device, including: The system calls a first processing model configured on the cloud server to predict and generate first reference type information related to the energy storage device. Based on the first reference type information, it generates first energy storage control information for the energy storage device. The first processing model is a complex model adapted to the computing power of the cloud server to predict and generate reference type information. The first reference type information is used to indicate the reference type information generated by the first processing model within a first time period after the current moment. The reference type information includes the electricity load trend, photovoltaic power generation trend, and electricity price change trend related to the energy storage device. The edge controller generates second energy storage control information for the energy storage device, including: The second processing model configured on the edge controller is invoked to predict and generate second reference type information related to the energy storage device. Based on the second reference type information, second energy storage control information of the energy storage device is generated. The second processing model is a lightweight model adapted to the computing power of the edge controller to predict and generate reference type information. The second reference type information is used to indicate the reference type information generated by the second processing model within a second duration starting from the current moment. The value of the first duration is greater than the value of the second duration.

3. The method according to claim 2, characterized in that, The first processing model configured on the cloud server is invoked to predict and generate first reference type information related to the energy storage device, including: The first operational data related to the energy storage system is obtained through the cloud server. The first operational data is collected by the edge controller and third-party devices and uploaded to the cloud server. The first operational data includes dynamic electricity price data of the power grid, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, health status data of the energy storage device, and meteorological data of the area where the energy storage device is located. Based on the first operating data, the first processing model configured on the cloud server is invoked to predict and generate the first reference type information related to the energy storage device.

4. The method according to claim 2, characterized in that, The second processing model configured on the edge controller is invoked to predict and generate second reference type information related to the energy storage device, including: The edge controller acquires second operating data related to the energy storage system. The second operating data is collected by the edge controller and uploaded to the cloud server. The second operating data includes grid dynamic electricity price data, household electricity load data related to the energy storage device, photovoltaic power generation data related to the energy storage device, and the state of charge of the energy storage device. Based on the second operating data, the second processing model configured on the edge controller is invoked to predict and generate the second reference type information related to the energy storage device.

5. The method according to claim 2, characterized in that, Generate first energy storage control information for the energy storage device based on the first reference type information, including: The first reference type information is input into the third processing model configured on the cloud server, and the first energy storage control information of the energy storage device is generated through the third processing model. The third processing model is a reinforcement learning model that uses the reference type information as the objective function to achieve the target energy storage scheduling requirements as the optimization direction, and outputs energy storage control information that adapts to the operation requirements of the energy storage device through reinforcement learning strategy learning and iterative optimization. The second energy storage control information of the energy storage device is generated based on the second reference type information, including: The system detects whether different reference type information in the second reference type information meets the preset energy storage control conditions associated with each reference type information. The preset energy storage control conditions associated with each reference type information are associated with energy storage control information adapted to each reference type information. Based on whether each reference type of information meets the preset energy storage control conditions associated with each reference type of information, the second energy storage control information of the energy storage device is generated; Specifically, the computational complexity of generating the first energy storage control information based on the first reference type information is greater than the computational complexity of generating the second energy storage control information based on the first reference type information, and the response speed of generating the first energy storage control information based on the first reference type information is greater than the response speed of generating the second energy storage control information based on the first reference type information.

6. The method according to claim 2, characterized in that, The method further includes: In response to the restoration of the network connection between the cloud server and the edge controller, the operating parameters of the second processing model configured on the edge controller during the network connection interruption are sent to the cloud server for archiving via the edge controller. Based on the operating parameters of the second processing model during the network connection interruption, the first processing model configured on the cloud server is corrected by the cloud server to update the first processing model.

7. The method according to claim 6, characterized in that, The method further includes: Based on the correction results of the first processing model configured on the cloud server, an update strategy and model parameters for the second processing model are generated and sent to the edge controller to update the second processing model configured on the edge controller. In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server synchronizes third operating data related to the energy storage system to the edge controller. The third operating data includes dynamic electricity price data of the power grid, photovoltaic power generation data related to the energy storage device, and meteorological data of the area where the energy storage device is located.

8. The method according to claim 5, characterized in that, The method further includes: In response to the triggering operation of the target energy storage scheduling requirement, the objective function used by the third processing model is determined from multiple objective functions by the cloud server to achieve the target energy storage scheduling requirement.

9. The method according to claim 1, characterized in that, The method further includes: The cloud server sends the first energy storage control information of the energy storage device to the edge controller, so that the edge controller can perform energy storage control on the energy storage device according to the first energy storage control information; or... The edge controller performs energy storage control on the energy storage device according to the second energy storage control information.

10. A cloud-edge collaborative energy storage system, characterized in that, The energy storage system includes energy storage devices, a cloud server, and an edge controller. The energy storage system is configured to perform the following operations: The cloud server generates first energy storage control information for the energy storage device. The energy storage control information is used to instruct the control to perform corresponding power charging and discharging operations on the energy storage device, control the transfer of power ownership of the energy storage device, and control the power scheduling of the energy storage device. The energy storage device is connected to the edge controller via data, and the edge controller is connected to the cloud server via communication. In response to a network connection interruption between the cloud server and the edge controller, the edge controller generates second energy storage control information for the energy storage device. The computing resources that the cloud server can utilize when generating the energy storage control information are greater than those that the edge controller can utilize when generating the energy storage control information. Furthermore, the control accuracy of the energy storage device generated by the cloud server is greater than that of the energy storage device generated by the edge controller. In response to the restoration of the network connection between the cloud server and the edge controller after an interruption, the cloud server is used again to generate new first energy storage control information.

Images