Virtual power plant-oriented energy storage power station combined regulation and control method and system

Abstract

According to the virtual power plant-oriented energy storage power station combined regulation and control method and system, historical data of an energy storage power station, a photovoltaic power station and a wind power station are collected, a data set is constructed, and an energy prediction model is established; predicting energy prediction data in a future time period by using the trained energy prediction model; obtaining real-time price and supply and demand data of the electricity market; performing time period division on the data through an extreme random tree model ERT, and distributing real-time electricity price interval labels; combining the data subjected to time period division and label distribution into a scheduling data set; according to the scheduling data set, defining an optimization target, and constructing a virtual power plant multi-energy collaborative scheduling model; energy scheduling constraint conditions are constructed, a scheduling scheme is obtained, control instructions are sent to an energy storage power station, a photovoltaic power station and a wind power station, and real-time control over all energy main bodies is achieved. The power market fluctuation can be responded in real time, the multi-energy collaborative scheduling of the virtual power plant is optimized, the energy configuration efficiency is improved, and the operation cost is reduced.

Description

Technical Field

[0001] This invention belongs to the field of power system operation and maintenance management technology, and particularly relates to a joint control method and system for energy storage power stations for virtual power plants. Background Technology

[0002] With the adjustment of the energy structure, the proportion of distributed energy sources such as solar and wind power generation in the power system is constantly increasing. As an effective means of integrating distributed energy sources, virtual power plants aggregate multiple distributed energy sources through information technology and intelligent control technology, achieving unified scheduling and management, and participating in electricity market transactions.

[0003] In existing technologies, for example, patent application CN119130071B designs a strategy determination module. This module acquires multi-dimensional optimization objectives and determines corresponding predictive factors for each objective. The strategy determination unit analyzes the attributes of each optimization objective to provide multi-dimensional power demand forecasts for energy storage power plants. Another example is patent application CN118300143A, which decouples the centralized energy storage peak-shaving model into a load-side energy storage virtual power plant power optimal gamete problem and a distribution network master station peak-shaving scheduling sub-problem. These sub-problems are optimized using the alternating direction multiplier method to finally obtain the peak-shaving response power of the centralized energy storage power plant and the load-side energy storage virtual power plant. This scheme achieves coordinated scheduling of multiple types of energy storage by establishing a two-layer scheduling framework. Yet another example is patent application CN119518891A, which provides a coordinated optimization scheduling method for energy storage power plants, applicable to virtual power plants. This method involves the virtual power plant reporting the operating parameters of the energy storage power plant and receiving scheduling instructions. The system determines the maximum charging and discharging power based on these parameters and uses this information for optimized scheduling of the energy storage power station. The scheduling instructions include the charging or discharging power of the virtual power plant in the next scheduling period, thereby achieving coordinated scheduling of the energy storage power station.

[0004] However, existing virtual power plant technologies mainly rely on traditional dispatch optimization methods, typically using only static data prediction and a single optimization objective for energy dispatch. While these methods can provide basic dispatch functions, they have significant shortcomings in handling real-time changes in the electricity market, the synergistic effects between different energy resources, and electricity price fluctuations. Summary of the Invention

[0005] Current technologies largely rely on historical data to predict future electricity demand and supply, but they are slow to respond to fluctuations in real-time electricity prices and rapid changes in the supply and demand relationship in the electricity market, making it impossible to adjust dispatch strategies in a timely manner, resulting in low efficiency in the allocation of energy resources. To address the shortcomings of existing technologies, this invention provides a method and system for the joint control of energy storage power stations for virtual power plants.

[0006] The present invention adopts the following technical solution.

[0007] This invention proposes a joint control method for energy storage power stations oriented towards virtual power plants, including: S1: Collect historical data from energy storage power stations, photovoltaic power stations, and wind power stations; preprocess the historical data; divide the processed historical data and construct a dataset; S2, Based on big data analysis methods, establish an energy forecasting model; train the energy forecasting model using a dataset; use the trained energy forecasting model to predict energy forecast data for future periods; S3: Obtain real-time price and supply-demand data of the electricity market; Divide the energy forecast data and the real-time price and supply-demand data of the electricity market into time periods using the Extreme Random Tree (ERT) model, and assign real-time electricity price range labels to the data in each time period; Combine the energy forecast data divided into time periods and labeled with labels and the real-time price and supply-demand data of the electricity market into a scheduling dataset; S4. Based on the scheduling dataset, considering the costs of electricity trading and resource scheduling, define the optimization objective and construct a virtual power plant multi-energy collaborative scheduling model; construct energy scheduling constraints for the virtual power plant multi-energy collaborative scheduling model; and obtain the scheduling scheme based on the constraints. S5, according to the scheduling plan, sends control commands to energy storage power stations, photovoltaic power stations and wind power stations to achieve real-time control of each energy source.

[0008] Furthermore, in S2, the energy forecast data includes the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations.

[0009] Furthermore, in S3, the energy forecast data is time-aligned with the real-time electricity price and supply-demand data of the electricity market. The aligned data is then used to construct the ERT sample dataset S, which includes N samples and M features. Each sample represents a time node, and each time node corresponds to 8 features, namely the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations in the energy forecast data, as well as the real-time electricity price and supply-demand data at the corresponding time node.

[0010] Furthermore, the target variables for ERT are set as time period divisions and the corresponding electricity price range labels for each time period; Each tree in ERT is partitioned based on features and target variables, and the GINI value is used at each node to select the optimal split. For the samples at each time point The GINI value is calculated as follows: ; in, Indicates the first interval label classification result The frequency of occurrence of each category This represents the total number of interval classification results.

[0011] Furthermore, the samples at each time point The GINI value is used as the splitting threshold to split the dataset S into two subsets. and Next, the GINI values ​​of the two subsets are calculated and then summed with weights to obtain the total GINI value after partitioning. Total GNI value The calculation is as follows: ; in, , Two subsets and The number of samples, where N is the total number of samples in dataset S; Obtain the total GINI value for all time points. Select the smallest total GINI value. The corresponding time node is used as the split point to divide the dataset S into two subsets, and the same operation is recursively performed on each subset until the maximum tree depth is reached. The ERT model is used to obtain energy forecast data, real-time electricity prices and supply and demand data for each time period, and assign an electricity price range label to each time period.

[0012] Furthermore, in S4, the optimization objective is defined as maximizing benefits, as follows: ; in, For the total benefits of the virtual power plant, The total number of time periods. Time period Real-time electricity prices in the electricity market Time period The amount of electricity output, Time period The amount of electricity input, Time period Operating costs Electricity price range label The weighting coefficients are used to adjust the impact of range-based electricity prices. For the profit and loss of electricity trading.

[0013] Furthermore, the weighting coefficient of the electricity price range label Determined in the following ways: ; Among them, High Price refers to more than 1.5 times the historical average electricity price, Mediun Price refers to 0.5 to 1.5 times the historical average electricity price, and Low Price refers to less than 0.5 times the historical average electricity price.

[0014] Furthermore, in S4, energy dispatch constraints include energy balance constraints, energy storage charging and discharging constraints, power generation capacity constraints, and grid constraints. The formula for energy balance constraints is: ; in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station For the charging and discharging power of energy storage power stations, Time period The amount of electricity input, Time period The electrical load; The formula for energy storage charge and discharge constraints is: ; in, Time period The energy storage battery's capacity, and Time periods The minimum and maximum capacity of the energy storage battery.

[0015] Furthermore, the formula for the power generation capacity constraint is: ; in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station Time period The maximum power generation capacity of the photovoltaic power station Time period The maximum power generation capacity of the wind power station; The formula for power grid constraints is: ; in, Time period The amount of electricity input, and Time periods The minimum and maximum power input.

[0016] This invention also proposes a joint control system for energy storage power stations oriented towards virtual power plants, including a data acquisition and transmission module, a data processing and analysis module, a joint scheduling and decision-making module, and an execution control module, characterized in that: The data acquisition and transmission module uses sensor technology to collect data information from energy storage power stations and virtual power plants distributed energy sources and transmits it to the data processing and analysis module. The data processing and analysis module uses big data technology to build energy forecasting models and analyze and process the uploaded information. The joint scheduling and decision-making module analyzes and processes the data information and transmits it to the joint scheduling and decision-making module to run the virtual power plant multi-energy collaborative scheduling model to formulate joint scheduling strategies for energy storage power stations and distributed energy. The execution control module sends the final execution strategy to the execution control module, which then sends the final control commands to the energy storage power station and the new energy power station and monitors the execution process of the commands.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The Extremely Random Tree (ERT) of this invention effectively avoids overfitting in high-dimensional data by randomly selecting features and partitioning nodes, and improves the overall model's stability by integrating the prediction results of multiple decision trees. Regarding time period segmentation, the ERT can accurately identify the relationship between electricity demand patterns and electricity price changes within different time periods. By dividing electricity price ranges, the model can accurately predict electricity market price fluctuations and, when integrated with other energy data, optimize energy supply decisions.

[0018] 2. The virtual power plant multi-energy collaborative dispatch model of this invention, combined with the application of electricity price range labels, can significantly improve the efficiency and sustainability of energy dispatch. By assigning electricity price range labels to each time period, the model can fully consider the real-time dynamic changes in the electricity market during the dispatch process, thereby optimizing the dispatch strategies for various energy resources.

[0019] 3. The virtual power plant multi-energy collaborative scheduling model of this invention considers multi-dimensional optimization of electricity trading and resource scheduling, which can effectively reduce the operating costs of virtual power plants and enhance market competitiveness. The scheduling strategy combined with interval labeling can ensure the system's economy, stability, and sustainability, and improve the quality of energy services for users. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method for joint control of energy storage power stations for virtual power plants according to the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0022] Example 1 This invention proposes a joint control method for energy storage power stations oriented towards virtual power plants, such as... Figure 1 As shown, it includes: S1: Collect historical data from energy storage power stations, photovoltaic power stations, and wind power stations; preprocess the historical data; divide the processed historical data and construct a dataset.

[0023] Specifically, historical data from the past 90 days was collected from energy sources such as energy storage power stations, photovoltaic power stations, and wind power stations. This data included the charge / discharge status, battery capacity, operating temperature, power generation, and load of energy storage power stations; irradiance, panel temperature, panel tilt and azimuth angles, panel aging degree, power generation, and load of photovoltaic power stations; and wind speed and direction, turbine speed, blade angle, air pressure and temperature, power generation, and load of wind power stations. The historical data underwent preprocessing, including data cleaning and outlier removal. The processed historical data was then used as a dataset, divided into training, validation, and test sets.

[0024] S2, based on big data analysis methods, establish an energy forecasting model; train the energy forecasting model using a dataset; and use the trained energy forecasting model to predict energy forecast data for future periods.

[0025] Specifically, energy forecasting models are established using big data analytics techniques such as time series analysis and machine learning. These models are then trained using datasets. The trained models are used to predict energy forecast data for future periods, including the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations.

[0026] S3: Obtain real-time price and supply-demand data of the electricity market; Divide the energy forecast data and the real-time price and supply-demand data of the electricity market into time periods using the Extremely Random Tree (ERT) model, and assign real-time electricity price range labels to the data in each time period; Combine the energy forecast data divided into time periods and labeled with labels and the real-time price and supply-demand data of the electricity market into a scheduling dataset.

[0027] Specifically, the relationship between energy forecast data and real-time electricity prices and supply and demand data in the electricity market is often complex and non-linear. For example, wind and solar power generation fluctuate greatly and can be affected by weather changes, while electricity market prices are typically strongly influenced by supply and demand. Therefore, it is impossible to accurately model the relationship between these variables using simple linear regression or other traditional linear models. Furthermore, the feature space of energy forecast data and real-time electricity prices and supply and demand data in the electricity market is very high. High-dimensional data leads to a significant increase in the complexity of subsequent models, further resulting in overfitting. Therefore, an extreme random tree model is used for further data processing.

[0028] Specifically, Extremely Randomized Trees (ERT) uses an ensemble of multiple trees to process energy forecasting data and the nonlinear relationships in real-time electricity price and supply and demand data in the electricity market. It learns on its own to select the most appropriate time division point. At the same time, ERT can handle high-dimensional feature data and avoid the problem of overfitting.

[0029] Specifically, energy forecast data is time-aligned with real-time electricity price and supply-demand data from the electricity market. The aligned data is then used to construct an ERT sample dataset S, which includes N samples and M features. Each sample represents a time node, and each time node corresponds to 8 features: the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations in the energy forecast data, as well as the real-time electricity price and supply-demand data for the corresponding time node.

[0030] Energy forecasting data and real-time electricity price and supply-demand data from the electricity market are time-series data. When processing these data, the goal of the ERT model is to find the most suitable time segmentation, ensuring optimal classification of the data within each time period across electricity price ranges. Therefore, the objective variables of ERT are the time period segmentation and the corresponding electricity price range label for each time period.

[0031] Furthermore, ERT constructs a set of decision trees that are independent of overfitting by randomly selecting features and split thresholds. Each tree is split based on features and the target variable, and the optimal split is selected at each node using the GINI value.

[0032] Specifically, for the samples at each time point The GINI value is calculated as follows: ; in, Indicates the first interval label classification result The frequency of occurrence of each category This represents the total number of interval classification results.

[0033] Samples at each time point The GINI value is used as the splitting threshold to split the dataset S into two subsets. and Next, the GINI values ​​of the two subsets are calculated and then summed with weights to obtain the total GINI value after partitioning. Total GNI value The calculation is as follows: ; in, , Two subsets and The number of samples is N, where N is the total number of samples in dataset S.

[0034] Obtain the total GINI value for all time points. Select the smallest total GINI value. The corresponding time points are used as split points to divide the dataset S into two subsets, and the same operation is recursively performed on each subset until the maximum tree depth is reached. In this embodiment, the maximum tree depth is selected using a rule of thumb or cross-validation. Each tree assigns an electricity price range label to the samples in the subset.

[0035] The ERT model is used to obtain energy forecast data, real-time electricity prices, and supply and demand data for each time period, and a price range label is assigned to each time period. Then, the energy forecast data, real-time electricity prices, supply and demand information, and corresponding price range labels for these time periods are combined into a scheduling dataset.

[0036] Furthermore, during energy dispatch, in order to maximize energy efficiency, balance supply and demand, and ensure the economy, stability and sustainability of the dispatch scheme, a virtual power plant multi-energy collaborative dispatch model is constructed to output the dispatch scheme.

[0037] Combining data from the scheduling dataset, the virtual power plant multi-energy collaborative scheduling model comprehensively considers the costs of electricity trading and resource scheduling, and the optimization objective is defined as maximizing efficiency, as follows: ; in, For the total benefits of the virtual power plant, The total number of time periods. Time period Real-time electricity prices in the electricity market Time period The amount of electricity output, Time period The amount of electricity input, Time period Operating costs Electricity price range label The weighting coefficients are used to adjust the impact of range-based electricity prices. For the profit and loss of electricity trading.

[0038] Electricity price range labels will affect electricity trading decisions in the dispatch model. In this embodiment, when the market electricity price is in the high price range (i.e., In high-price ranges, energy storage power stations will prioritize discharging electricity, while in low-price ranges, electricity will be purchased more from the grid, and energy storage devices will be used more for charging. The weighting coefficient for the electricity price range label... It can be determined in the following ways: ; Among them, High Price refers to more than 1.5 times the historical average electricity price, Mediun Price refers to 0.5 to 1.5 times the historical average electricity price, and Low Price refers to less than 0.5 times the historical average electricity price.

[0039] By precisely dividing time periods and assigning electricity price range labels, the volatility of the electricity market and changes in electricity demand across different time periods can be effectively captured. This method enables the dispatch system to respond in real time to dynamic changes in the electricity market, optimize energy dispatch strategies, and improve the efficiency of energy resource utilization. Furthermore, extreme random trees, by integrating multiple decision trees, improve model stability and prediction accuracy, reduce the risk of overfitting, and ensure the reliability of the dispatch strategy.

[0040] S4. Based on the scheduling dataset, considering the costs of electricity trading and resource scheduling, define the optimization objective and construct a virtual power plant multi-energy collaborative scheduling model; construct energy scheduling constraints for the virtual power plant multi-energy collaborative scheduling model; and obtain the scheduling scheme based on the constraints.

[0041] Specifically, energy dispatch constraints include energy balance constraints, energy storage charging and discharging constraints, power generation capacity constraints, and grid constraints.

[0042] Specifically, energy balance constraints ensure the balance between energy supply and demand in each time period: ; in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station For the charging and discharging power of energy storage power stations, Time period The amount of electricity input, Time period The electrical load.

[0043] Energy storage charge and discharge constraints refer to the fact that the charge and discharge power of the energy storage battery should be within the allowable range, and the state of charge of the energy storage battery should meet the actual requirements: ; in, Time period The energy storage battery's capacity, and Time periods The minimum and maximum capacity of the energy storage battery.

[0044] Power generation capacity constraints refer to the requirement that the power generation capacity of photovoltaic and wind power generation equipment should be within its capacity range. ; in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station Time period The maximum power generation capacity of the photovoltaic power station Time period The maximum power generation capacity of the wind power station.

[0045] Grid constraints refer to the constraints on the power input and output of the power grid to prevent overload. ; in, Time period The amount of electricity input, and Time periods The minimum and maximum power input.

[0046] The optimal solution is obtained by applying the above constraints, and it serves as the scheduling scheme.

[0047] The virtual power plant multi-energy collaborative dispatch model can comprehensively consider multiple factors such as electricity market price fluctuations, resource supply, and costs to formulate the optimal dispatch scheme, thereby effectively reducing the operating costs of virtual power plants. By optimizing the dispatch strategies of various energy sources, the model not only improves the allocation efficiency of energy resources but also enhances the competitiveness of virtual power plants in the electricity market. Furthermore, the implementation of this model improves the flexibility and sustainability of energy dispatch, ensuring the economic efficiency and stability of the system.

[0048] S5, according to the scheduling plan, sends control commands to energy storage power stations, photovoltaic power stations and wind power stations to achieve real-time control of each energy source.

[0049] Based on the dispatch plan formulated by the joint dispatch decision-making level, executable control commands are sent to energy storage power stations, photovoltaic power stations, wind power stations, etc. These commands are transmitted to the target equipment via a 5G private network or fiber optic communication, enabling real-time control of each energy entity. Simultaneously, the execution process is monitored to ensure the accurate execution of control commands.

[0050] In this embodiment, a virtual power plant comprises a 10MW / 20MWh energy storage station and five 2MW wind turbines. During real-time monitoring, it was discovered that one of the wind turbines experienced a fault, reducing its power output from 2MW to 1MW, while the real-time electricity price increased from 0.4 yuan / kWh to 0.8 yuan / kWh. Therefore, according to the joint dispatch method of this invention, the discharge power of the energy storage station is adjusted from 0 to 2MW to compensate for the decrease in wind turbine power output. Simultaneously, the monitoring frequency of the energy storage station's battery temperature is increased from once every 30 minutes to once every 10 minutes to ensure the safe operation of the energy storage station. New constraints are introduced to limit the power output of the faulty wind turbine to no more than 1.5MW to prevent further equipment damage. Through the adjustment of the above parameters and steps, the virtual power plant can better adapt to changes in operating conditions, improve energy utilization efficiency and economic benefits, and ensure the safe and stable operation of the equipment.

[0051] When there is sufficient sunlight but low electricity load, excess electricity from photovoltaic power plants is stored in energy storage power plants; at night or when sunlight is insufficient, the electricity stored in the energy storage power plants is released to meet electricity demand. At the same time, in conjunction with the power generation of wind power plants, comprehensive scheduling is carried out to achieve a stable energy supply.

[0052] Example 2 This invention also proposes a joint control system for energy storage power stations oriented towards virtual power plants, comprising: The data acquisition and transmission module uses sensor technology to collect data information from energy storage power stations and virtual power plants distributed energy sources and transmits it to the data processing and analysis module. The data processing and analysis module uses big data technology to build energy forecasting models and analyze and process the uploaded information. The joint scheduling and decision-making module analyzes and processes the data information and transmits it to the joint scheduling and decision-making module to run the virtual power plant multi-energy collaborative scheduling model to formulate joint scheduling strategies for energy storage power stations and distributed energy. The execution control module sends the final execution strategy to the execution control module, which then sends the final control commands to the energy storage power station and the new energy power station and monitors the execution process of the commands.

[0053] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A joint control method for energy storage power stations oriented towards virtual power plants, characterized in that, include: S1: Collect historical data from energy storage power stations, photovoltaic power stations, and wind power stations; perform data preprocessing on the historical data; The processed historical data is divided into segments to construct a dataset; S2, Based on big data analysis methods, establish an energy forecasting model; train the energy forecasting model using a dataset; use the trained energy forecasting model to predict energy forecast data for future periods; S3: Obtain real-time price and supply-demand data of the electricity market; Divide the energy forecast data and the real-time price and supply-demand data of the electricity market into time periods using the Extreme Random Tree (ERT) model, and assign real-time electricity price range labels to the data in each time period; Combine the energy forecast data divided into time periods and labeled with labels and the real-time price and supply-demand data of the electricity market into a scheduling dataset; S4. Based on the scheduling dataset, considering the costs of electricity trading and resource scheduling, define the optimization objective and construct a virtual power plant multi-energy collaborative scheduling model; construct energy scheduling constraints for the virtual power plant multi-energy collaborative scheduling model; and obtain the scheduling scheme based on the constraints. S5, according to the scheduling plan, sends control commands to energy storage power stations, photovoltaic power stations and wind power stations to achieve real-time control of each energy source.

2. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 1, characterized in that: In S2, energy forecast data includes the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations.

3. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 1, characterized in that: In S3, energy forecast data is time-aligned with real-time electricity price and supply-demand data from the electricity market. The aligned data is then used to construct the ERT sample dataset S, which includes N samples and M features. Each sample represents a time node, and each time node corresponds to 8 features: the power generation and load of energy storage power stations, photovoltaic power stations, and wind power stations in the energy forecast data, as well as the real-time electricity price and supply-demand data for the corresponding time node.

4. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 3, characterized in that: The target variables for ERT are time period divisions and the corresponding electricity price range labels for each time period. Each tree in ERT is partitioned based on features and target variables, and the GINI value is used at each node to select the optimal split. For the samples at each time point The GINI value is calculated as follows: ； in, Indicates the first interval label classification result The frequency of occurrence of each category This represents the total number of interval classification results.

5. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 4, characterized in that: Samples at each time point The GINI value is used as the splitting threshold to split the dataset S into two subsets. and Next, the GINI values ​​of the two subsets are calculated and then summed with weights to obtain the total GINI value after partitioning. Total GNI value The calculation is as follows: ； in, , Two subsets and The number of samples, where N is the total number of samples in dataset S; Obtain the total GINI value for all time points. Select the smallest total GINI value. The corresponding time node is used as the split point to divide the dataset S into two subsets, and the same operation is recursively performed on each subset until the maximum tree depth is reached. The ERT model is used to obtain energy forecast data, real-time electricity prices and supply and demand data for each time period, and assign an electricity price range label to each time period.

6. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 5, characterized in that: In S4, the optimization objective is defined as maximizing benefits, as follows: ； in, For the total benefits of the virtual power plant, The total number of time periods. Time period Real-time electricity prices in the electricity market Time period The amount of electricity output, Time period The amount of electricity input, Time period Operating costs Electricity price range label The weighting coefficients are used to adjust the impact of range-based electricity prices. For the profit and loss of electricity trading.

7. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 6, characterized in that: Weighting coefficient of electricity price range label Determined in the following ways: ； Among them, High Price refers to more than 1.5 times the historical average electricity price, Mediun Price refers to 0.5 to 1.5 times the historical average electricity price, and Low Price refers to less than 0.5 times the historical average electricity price.

8. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 1, characterized in that: In S4, energy dispatch constraints include energy balance constraints, energy storage charging and discharging constraints, power generation capacity constraints, and grid constraints. The formula for energy balance constraints is: ； in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station For the charging and discharging power of energy storage power stations, Time period The amount of electricity input, Time period The electrical load; The formula for energy storage charge and discharge constraints is: ； in, Time period The energy storage battery's capacity, and Time periods The minimum and maximum capacity of the energy storage battery.

9. The joint control method for energy storage power stations oriented towards virtual power plants according to claim 8, characterized in that: The formula for the power generation capacity constraint is: ； in, Time period The power generation capacity of the photovoltaic power station Time period The power generation capacity of the wind power station Time period The maximum power generation capacity of the photovoltaic power station Time period The maximum power generation capacity of the wind power station; The formula for power grid constraints is: ； in, Time period The amount of electricity input, and Time periods The minimum and maximum power input.

10. A joint control system for energy storage power stations oriented towards virtual power plants, comprising a data acquisition and transmission module, a data processing and analysis module, a joint scheduling and decision-making module, and an execution control module, characterized in that: The data acquisition and transmission module uses sensor technology to collect data information from energy storage power stations and virtual power plants distributed energy sources and transmits it to the data processing and analysis module. The data processing and analysis module uses big data technology to build energy forecasting models and analyze and process the uploaded information. The joint scheduling and decision-making module analyzes and processes the data information and transmits it to the joint scheduling and decision-making module to run the virtual power plant multi-energy collaborative scheduling model to formulate joint scheduling strategies for energy storage power stations and distributed energy. The execution control module sends the final execution strategy to the execution control module, which then sends the final control commands to the energy storage power station and the new energy power station and monitors the execution process of the commands.

Images