Energy storage power generation system modeling and target construction method based on digital twinning

Abstract

The embodiment of the application relates to a kind of based on digital twinning energy storage power generation system modeling and target construction method, comprising the following steps: collecting the multidimensional data of energy storage power generation system, obtains energy storage power generation system multidimensional data set;Based on the energy storage power generation system multidimensional data set, construct digital twin model;Based on the operation result of the digital twin model, generate and output target construction and strategy optimization result set.The technical scheme provided in the embodiment of the application, by establishing digital twin model, a virtual system capable of simulating energy storage, photovoltaic, and grid three-party dynamic game and cooperation is constructed, the fusion of multi-source data and multi-level model is realized, so that the system can be based on the various performance and risk targets dynamically constructed, independently execute game rules, and then more accurate data support is provided for the safe, economic, self-optimizing intelligent operation of energy storage power generation system.

Description

Technical Field

[0001] This invention relates to the field of power system energy storage power generation safety and efficiency optimization technology, specifically to a method for modeling and target construction of energy storage power generation systems based on digital twins. Background Technology

[0002] As a critical infrastructure for cybersecurity research, testing, and drills, cyber ranges offer the core value of providing a highly realistic, measurable, and low-risk verification environment for offensive and defensive strategies. For complex energy physics systems such as energy storage and power generation systems, the need for similar modeling and target construction is increasingly prominent.

[0003] Current target range construction primarily relies on a static model combining mechanistic simulation software with empirical threshold management. The model parameters used in modeling are fixed and difficult to correct online. Furthermore, the lack of deep coupling between models based on real-time data and a unified spatiotemporal framework leads to a gradual disconnect between the digital image and the rapidly evolving physical entities. This results in insufficient reliability of simulation results in guiding actual operational decisions, ultimately leading to static or semi-static fixed targets that cannot dynamically respond to microscopic changes in the system's internal state or rapid adjustments to the external environment. Summary of the Invention

[0004] Based on the above-mentioned situation of the prior art, the purpose of this embodiment of the invention is to provide a method for modeling and target construction of energy storage power generation system based on digital twin, so as to solve the above-mentioned technical problems existing in the prior art.

[0005] To achieve the above objectives, according to one aspect of the present invention, a method for modeling and target construction of an energy storage power generation system based on digital twins is provided, comprising the following steps: Collect multi-dimensional data from the energy storage power generation system to obtain a multidimensional dataset of the energy storage power generation system; Based on the multidimensional dataset of the energy storage power generation system, a digital twin model is constructed, which includes a device-level first sub-model, a device-level second sub-model, and a system-level sub-model. Based on the results of the operation of the digital twin model, a target construction and strategy optimization result set is generated and output; The multi-dimensional data includes sensing data of energy storage batteries, power generation data of new energy equipment, and environmental linkage data; the first device-level sub-model is the physical sub-model of energy storage batteries, the second device-level sub-model is the coupling sub-model of new energy power generation, and the system-level sub-model is the multi-agent game interaction sub-model.

[0006] Furthermore, the physical sub-model of the energy storage battery is constructed through the following steps: A physical coupling equation for the internal resistance of an energy storage battery is established, which characterizes the nonlinear fitting relationship between the internal resistance of the energy storage battery and temperature. By adding an internal resistance coefficient to the physical coupling equation, a modified equation for the physical coupling equation is obtained. The coefficients of the modified equation are fitted and corrected using the multi-dimensional data. Substituting the corrected coefficients into the corrected equation of the physical coupling equation yields the physical sub-model of the energy storage battery.

[0007] Furthermore, the coefficients of the modified equation are fitted and corrected using the multi-dimensional data, including the following steps: The internal resistance coefficient is obtained by fitting effective data with temperature values ​​in the low-temperature range from the sensing data. Using a Bayesian optimization iterative method, with the goal of minimizing the error between the internal resistance value predicted by the modified equation and the measured internal resistance value, the nonlinear fitting coefficients and internal resistance coefficients in the modified equation are iteratively corrected to obtain the corrected coefficients.

[0008] Furthermore, the new energy power generation coupling sub-model is constructed through the following steps: The basic architecture for constructing the new energy power generation coupling sub-model is a gradient boosting tree. The new energy power generation coupling sub-model is obtained by iteratively training the power generation data and environmental linkage data.

[0009] Furthermore, the new energy power generation coupling sub-model is a photovoltaic power generation coupling sub-model; the power generation data includes the irradiance of the photovoltaic module, the temperature of the photovoltaic module, and the real-time output value of the photovoltaic module; the environmental linkage data includes the ambient temperature; the new energy power generation coupling sub-model is iteratively trained using the power generation data and the environmental linkage data, including: The irradiance of the photovoltaic module, the temperature of the photovoltaic module, and the ambient temperature are used as input data for the photovoltaic power generation coupling sub-model, and the real-time output value of the photovoltaic module is used as output data for the photovoltaic power generation coupling sub-model. The root mean square error is used as the evaluation index to iteratively train the photovoltaic power generation coupling sub-model.

[0010] Furthermore, the method also includes: Constraints are set to constrain the predicted output value of the photovoltaic power generation coupling sub-model.

[0011] Furthermore, the method also includes: The probability of power curtailment is calculated based on the real-time power output prediction of the photovoltaic modules obtained from the photovoltaic power generation coupling sub-model.

[0012] Furthermore, the multi-agent game interaction sub-model is constructed through the following steps: Construct an infrastructure based on a multi-entity rule base, where the multi-entities include energy storage batteries, new energy equipment, and power grids; Construct the initial interaction rules for the multi-agent rule base and the initial thresholds during the interaction process; Based on historical data over a predetermined period, the initial interaction rules and initial thresholds are revised and iterated. Based on the revised and iterated interaction rules and thresholds, the multi-agent game interaction sub-model is obtained.

[0013] Furthermore, the method also includes: Input the output data of the current energy storage battery physical sub-model and the new energy power generation coupling sub-model, as well as the environmental linkage data, into the multi-agent game interaction sub-model to obtain the charging and discharging commands of the energy storage battery.

[0014] Furthermore, the target construction and strategy optimization result set includes an executable rule base and a set of dynamic threshold targets bound to the executable rule base.

[0015] In summary, this invention provides a method for modeling and target construction of an energy storage power generation system based on digital twins, including the following steps: collecting multi-dimensional data of the energy storage power generation system to obtain a multi-dimensional dataset of the energy storage power generation system; constructing a digital twin model based on the multi-dimensional dataset of the energy storage power generation system; and generating and outputting a target construction and strategy optimization result set based on the operating results of the digital twin model. The technical solution provided by this invention, by establishing a digital twin model, constructs a virtual system capable of simulating the dynamic game and collaboration among energy storage, photovoltaics, and the power grid, realizing the fusion of multi-source data and multi-level models. This enables the system to autonomously execute game rules based on dynamically constructed performance and risk targets, thereby providing more accurate data support for the safe, economical, and self-optimizing intelligent operation of the energy storage power generation system. Attached Figure Description

[0016] Figure 1 This is a flowchart of the modeling and target construction method for energy storage power generation system based on digital twin provided in the embodiments of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0018] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the element or object listed following the word and its equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0019] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings. An embodiment of the present invention provides a method for modeling and target construction of an energy storage power generation system based on digital twins. Figure 1 The flowchart illustrates the modeling and target construction method for energy storage power generation systems based on digital twins provided in this embodiment of the invention. Figure 1 As shown, the method includes the following steps: S202. Collect multi-dimensional data from the energy storage power generation system to obtain a multi-dimensional dataset of the energy storage power generation system. The multi-dimensional data includes sensing data from the energy storage battery, power generation data from the new energy equipment, and environmental linkage data. The multi-dimensional dataset of the energy storage power generation system includes sensing data sequences, power generation data sequences, and spatiotemporal tag data. Sensing data from the energy storage battery can be acquired through sensors deployed inside the energy storage battery, including battery cell temperature, charge / discharge rate, and battery internal resistance. The collected data can be formatted as a sensing data sequence, for example, "collection time - battery cluster number - battery cell number - temperature value - charge / discharge rate value - battery internal resistance value". In this embodiment of the invention, the new energy equipment is a photovoltaic power generation device. Power generation data includes the irradiance of the photovoltaic module, the temperature of the photovoltaic module, and the real-time output value of the corresponding photovoltaic module. The collected data can be spatiotemporally correlated with geographical coordinates and collection time to form a power generation data sequence. Environmental linkage data includes the ambient temperature of a predetermined area and the peak-valley electricity price signal of the power grid. The predetermined area can be marked with a regional tag, which can be divided into tags according to the geographical partitions of the energy storage battery and the new energy equipment, corresponding to the battery cluster number and the geographical coordinates of the photovoltaic power generation equipment. The collected data can be used to create spatiotemporal labeled data, which provides environmental condition labels and electricity price time period labels for the aforementioned sensing data and power generation data. The sampling frequency of the above multi-dimensional data can be set according to actual needs. Generally, the sampling frequency of sensing data is the highest, followed by power generation data. Environmental linkage data, due to its relatively slow changes, can use a lower sampling frequency, such as minutes or hours as the sampling period unit.

[0020] The collected data can undergo preprocessing, including data cleaning of the sensing data, calibration of the power generation data, and construction of spatiotemporal correlation indexes for various data types. Data cleaning and calibration can employ conventional data processing methods. The spatiotemporal correlation index can use the collection time as the primary key, associating the battery cluster number of the sensing data, the geographical coordinates of the power generation data, and the regional label of the environmental linkage data to generate a unified data ID, thereby accurately matching the above multi-dimensional data. These data collection steps, by acquiring multi-dimensional real-time data from energy storage batteries, photovoltaic power generation equipment, and the system environment, and constructing spatiotemporal correlation indexes, form a complete digital mirror of the physical system, providing a precise and synchronized data foundation for constructing a virtual mapping.

[0021] S204. Based on the aforementioned multidimensional dataset of the energy storage power generation system, a digital twin model is constructed. In this embodiment of the invention, the digital twin model includes a device-level first sub-model, a device-level second sub-model, and a system-level sub-model. The device-level first sub-model is a physical sub-model of the energy storage battery, the device-level second sub-model is a new energy power generation coupling sub-model, and the system-level sub-model is a multi-agent game interaction sub-model.

[0022] The physical sub-model of the energy storage battery can be constructed using the following steps: S20411. Establish a physical coupling equation for the internal resistance of an energy storage battery. This physical coupling equation characterizes the nonlinear fitting relationship between the internal resistance of a single energy storage battery cell and its temperature. The basic equation of the physical coupling equation can be expressed as: in, This represents the internal resistance of a single energy storage battery cell at a temperature of T. The battery internal resistance at standard room temperature can be obtained from the manufacturer's datasheet. , The internal resistance temperature coefficient, also known as the nonlinear fitting coefficient, represents the temperature range within a typical temperature range. The default value is: , .

[0023] In this embodiment of the invention, for low-temperature operating conditions (0 The following (potential lithium deposition problem in energy storage batteries) can be addressed by adding an internal resistance coefficient to the above basic equations, resulting in a modified equation for the physical coupling equations: in, This represents the internal resistance coefficient. It can be represented as: in, , This represents the coefficients to be fitted.

[0024] S20412. The coefficients of the above-mentioned correction equation are fitted and corrected using the collected multi-dimensional data. Temperature values ​​within the low-temperature range (e.g., -20°C) are selected from the above-mentioned sensing data. ~0 The effective data is used as fitting data. The temperature and internal resistance values ​​in the fitting data are substituted into the corrected equation of the physical coupling equation, and the result is obtained by fitting using, for example, the least squares method. The initial value.

[0025] According to certain optional embodiments, a Bayesian optimization iterative method can also be used, with the optimization objective being to minimize the error between the predicted and measured internal resistance values, and to adjust the coefficients in the correction equation. , and Iterative correction is performed. During the iterative correction process, environmental temperature data corresponding to the fitted data in the environmental linkage data is used. Based on this environmental temperature data, the fitted data is stratified, and differentiating weights are assigned to the fitted data at different environmental temperatures. Then, an objective function is established for iterative correction to obtain the corrected coefficients. For example, based on environmental temperature, three levels are defined, with the first level having an environmental temperature of [-20...]. -15 The second layer ambient temperature is [-15]. -5 The ambient temperature of the third layer is [-5]. ,0 The weighting coefficients for the fitted data corresponding to the first, second, and third layer ambient temperatures are set as follows: , and The weighting coefficient can be set to This is to ensure that data from extreme low-temperature operating conditions dominate the optimization of coefficients. The objective function can be expressed as: in, This represents the predicted internal resistance value corresponding to the i-th layer. This represents the actual internal resistance value corresponding to the i-th layer. The optimization objective is to minimize the value of the objective function.

[0026] S20413. Substitute the corrected coefficients into the corrected equation of the physical coupling equation to obtain the physical sub-model of the energy storage battery. Based on the physical sub-model, input the currently collected sensing data and environmental linkage data of the energy storage battery to obtain the predicted internal resistance value. Based on the predicted internal resistance value, the lithium deposition risk level and available capacity of the energy storage battery can also be obtained. According to the predicted internal resistance value and the initial internal resistance value, the internal resistance change rate can be obtained, and the lithium deposition risk level can be set in combination with the ambient temperature, as shown in Table 1.

[0027] Table 1 Available capacity can be calculated using the following formula: in, Indicates available capacity. Indicates the rated capacity of the energy storage battery. Indicates the rate of change of internal resistance. This represents the capacity attenuation coefficient, which is set to 0.8 in this application.

[0028] In this embodiment of the invention, the new energy power generation coupling sub-model is a photovoltaic power generation coupling sub-model, which can be constructed using the following steps: S20421. Gradient Boosting Tree (GBDT) is adopted as the basic architecture of the photovoltaic power generation coupling sub-model. The photovoltaic power generation coupling sub-model is iteratively trained using power generation data and environmental linkage data to obtain a trained model. From the aforementioned power generation data, the irradiance, temperature, and real-time output value of the photovoltaic modules are selected. From the aforementioned environmental linkage data, the ambient temperature is selected as the training data for the photovoltaic power generation coupling sub-model. The irradiance, temperature, and ambient temperature of the photovoltaic modules are the input data of the photovoltaic power generation coupling sub-model, and the real-time output value of the photovoltaic modules is the output data. The root mean square error is used as the evaluation index to iteratively train the sub-model to obtain a trained photovoltaic power generation coupling sub-model.

[0029] S20422. Based on the physical laws governing photovoltaic (PV) power generation, constraints are set to constrain the predicted output value of the PV power generation coupling sub-model. The output constraint of the PV module can be expressed as: in, This indicates the upper limit of the output power of the photovoltaic module. Indicates the rated power of the photovoltaic module. Indicates standard irradiance. This represents the irradiance in the power generation data. If the output predicted by the photovoltaic power generation coupled sub-model is greater than this upper limit, the predicted output will be corrected to this upper limit; if the output predicted by the photovoltaic power generation coupled sub-model is less than 0, the predicted output will be corrected to 0.

[0030] S20423. Based on the real-time power output prediction value of the photovoltaic module obtained from the photovoltaic power generation coupling sub-model, the curtailment probability can be calculated. The curtailment probability refers to the probability that, within a predetermined time interval, the actual power output of new energy power generation equipment cannot be fully connected to the grid and is forced to be abandoned due to insufficient grid capacity, limited energy storage regulation, or other reasons. For example, the functional relationship between the curtailment probability and the power output fluctuation value can be obtained by fitting historical data. Based on the real-time power output prediction value of the photovoltaic module, the power output fluctuation value of the photovoltaic module within a predetermined time period (e.g., 15 minutes) can be calculated. : in, This represents the predicted maximum output value within the predetermined time period. This represents the minimum predicted output value within a predetermined time period. This represents the predicted average output value over a predetermined time period. After fitting, the following functional relationship can be obtained: in, Indicates the probability of power curtailment. Indicates the threshold of absorption capacity. This represents the fitting coefficient.

[0031] The multi-agent game interaction sub-model can be constructed using the following steps: S20431, the multi-agent game interaction sub-model is based on a multi-agent rule base architecture, in which the multiple agents include energy storage batteries, new energy equipment and power grid. The decision basis for energy storage batteries is the risk level and available capacity output by the energy storage battery physical sub-model. The decision basis for new energy equipment is the output prediction and curtailment probability output by the new energy power generation coupling sub-model. The decision basis for power grid is the peak-valley electricity price and peak-shaving power demand in the environmental linkage data.

[0032] S20432. Constructing initial interaction rules for a multi-agent rule base and initial thresholds during the interaction process, which may include, for example: When the curtailment probability output by the new energy power generation coupling sub-model is greater than the curtailment probability threshold, the power grid issues a charging command to the energy storage battery. When the risk level output by the physical sub-model of the energy storage battery is ≤ Medium, the energy storage battery starts charging at a rate of 0.5C to prioritize the consumption of electricity from new energy devices; where 1C rate means that the battery can be fully charged or discharged within 1 hour; when the risk level output by the physical sub-model of the energy storage battery is High, the energy storage battery automatically reduces the charge / discharge rate to 0.3C, and the grid simultaneously reduces the demand for new energy consumption to prioritize the safety of the energy storage battery. When the environmental linkage data is during off-peak hours and the available energy storage capacity is less than the energy storage capacity threshold, the energy storage system starts charging; during peak hours, the energy storage system starts discharging.

[0033] The initial value of the curtailment probability threshold can be set to 20%, and the initial value of the energy storage capacity threshold can be set to 80%.

[0034] S20433. Collect historical data within a predetermined period (e.g., the past 3 months), statistically analyze the operational performance under different rule thresholds, and revise and iterate the thresholds in the initial interaction rules mentioned above. For example, statistically analyze the renewable energy consumption rate and energy storage battery safety accident rate corresponding to a 20% curtailment probability threshold. If the average consumption rate is 78% and there are no safety accidents when the curtailment probability threshold is 20%, maintain this probability threshold; if the consumption rate is lower than 75%, lower the probability threshold to increase the priority of consumption. The gradient descent method can be used to adjust the probability threshold. For example, let the consumption rate target be... The actual consumption rate is The adjustment amount for the power curtailment probability threshold can be expressed as: in, This indicates the adjustment amount for the power curtailment probability threshold. This represents the learning rate, and the initial value can be set to 0.1.

[0035] S20434, the multi-agent game interaction sub-model is based on a modified and iterative multi-agent rule base. Running the constructed digital twin model, the energy storage battery physical sub-model and the new energy power generation coupling sub-model obtain output data based on their respective input data. The output data of the current energy storage battery physical sub-model and the new energy power generation coupling sub-model, along with environmental linkage data, are input into the multi-agent game interaction sub-model, which outputs charging and discharging commands for the energy storage battery.

[0036] In the above technical solution of this invention, a digital twin model is constructed, and key quantitative indicators and rule thresholds representing the system state, risk, and economy, serving as the basis for multi-agent interactive decision-making, are built within the digital twin model as targets for various aspects of the system. These targets are then optimized and adjusted using key judgment conditions in the multi-agent rule base of the multi-agent game interaction sub-model. By establishing the multi-agent game interaction sub-model, a virtual system capable of simulating the dynamic game and collaboration among energy storage, photovoltaics, and the power grid is constructed. This achieves the fusion of multi-source data and multi-level models, enabling the system to autonomously execute game rules based on dynamically constructed performance and risk targets (such as available capacity and output prediction), thereby providing more accurate data support for the safe, economical, and self-optimizing intelligent operation of the energy storage power generation system.

[0037] S206. Based on the operational results of the aforementioned digital twin model, generate and output a target construction and strategy optimization result set. Based on the digital twin model constructed and iteratively optimized in the above steps, generate and output a target construction and strategy optimization result set. This result set includes an executable rule base and a set of dynamic threshold targets bound to the executable rule base. This result set can, for example, be used to guide the operation and control of the energy management system (EMS) of an energy storage power generation system.

[0038] In this step, the multi-agent interaction rules and related thresholds, validated and iteratively optimized using historical data, are standardized and encapsulated to generate a structured executable rule base file. This executable rule base specifies the triggering logic for each entity (including energy storage batteries, new energy equipment, and the power grid) within each energy storage and power generation system under different input conditions (e.g., risk level, curtailment probability, electricity price signals). Simultaneously, a dynamic threshold target set bound to the executable rule base is generated. This set includes at least strategy rule targets and safety preset targets. Strategy rule targets include, for example, optimized curtailment probability trigger thresholds and energy storage capacity charging thresholds; safety constraint targets include, for example, internal resistance change rate thresholds corresponding to high, medium, and low risk levels under different ambient temperature ranges. These thresholds can be parameter sets with accompanying usage conditions and confidence intervals, which can be called by the upper-level system or automatically fine-tuned under certain conditions.

[0039] In summary, this invention relates to a method for modeling and target construction of an energy storage power generation system based on digital twins, comprising the steps of: collecting multi-dimensional data of the energy storage power generation system to obtain a multidimensional dataset of the energy storage power generation system; constructing a digital twin model based on the multidimensional dataset of the energy storage power generation system; and generating and outputting a target construction and strategy optimization result set based on the operating results of the digital twin model. The technical solution provided by this invention, by establishing a digital twin model, constructs a virtual system capable of simulating the dynamic game and collaboration among energy storage, photovoltaics, and the power grid, realizing the fusion of multi-source data and multi-level models. This enables the system to autonomously execute game rules based on dynamically constructed performance and risk targets, thereby providing more accurate data support for the safe, economical, and self-optimizing intelligent operation of the energy storage power generation system.

[0040] It should be understood that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of this invention, technical features of the above embodiments or different embodiments can also be combined, steps can be implemented in any order, and many other variations exist regarding different aspects of one or more embodiments of the invention as described above; for the sake of brevity, they are not provided in the details. The specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

Claims

1. A method for modeling and target construction of energy storage power generation systems based on digital twins, characterized in that, Including the following steps: Collect multi-dimensional data from the energy storage power generation system to obtain a multidimensional dataset of the energy storage power generation system; Based on the multidimensional dataset of the energy storage power generation system, a digital twin model is constructed, which includes a device-level first sub-model, a device-level second sub-model, and a system-level sub-model. Based on the results of the operation of the digital twin model, a target construction and strategy optimization result set is generated and output; The multi-dimensional data includes sensing data of energy storage batteries, power generation data of new energy equipment, and environmental linkage data; the first device-level sub-model is the physical sub-model of energy storage batteries, the second device-level sub-model is the coupling sub-model of new energy power generation, and the system-level sub-model is the multi-agent game interaction sub-model.

2. The method according to claim 1, characterized in that, The physical sub-model of the energy storage battery is constructed through the following steps: A physical coupling equation for the internal resistance of an energy storage battery is established, which characterizes the nonlinear fitting relationship between the internal resistance of the energy storage battery and temperature. By adding an internal resistance coefficient to the physical coupling equation, a modified equation for the physical coupling equation is obtained. The coefficients of the modified equation are fitted and corrected using the multi-dimensional data. Substituting the corrected coefficients into the corrected equation of the physical coupling equation yields the physical sub-model of the energy storage battery.

3. The method according to claim 2, characterized in that, The process of fitting and correcting the coefficients of the modified equation using the multi-dimensional data includes the following steps: The internal resistance coefficient is obtained by fitting effective data with temperature values ​​in the low-temperature range from the sensing data. Using a Bayesian optimization iterative method, with the goal of minimizing the error between the internal resistance value predicted by the modified equation and the measured internal resistance value, the nonlinear fitting coefficients and internal resistance coefficients in the modified equation are iteratively corrected to obtain the corrected coefficients.

4. The method according to claim 1, characterized in that, The new energy power generation coupling sub-model is constructed through the following steps: The basic architecture for constructing the new energy power generation coupling sub-model is a gradient boosting tree. The new energy power generation coupling sub-model is obtained by iteratively training the power generation data and environmental linkage data.

5. The method according to claim 4, characterized in that, The new energy power generation coupling sub-model is a photovoltaic power generation coupling sub-model; the power generation data includes the irradiance of the photovoltaic module, the temperature of the photovoltaic module, and the real-time output value of the photovoltaic module; the environmental linkage data includes the ambient temperature; the new energy power generation coupling sub-model is iteratively trained using the power generation data and the environmental linkage data, including: The irradiance of the photovoltaic module, the temperature of the photovoltaic module, and the ambient temperature are used as input data for the photovoltaic power generation coupling sub-model, and the real-time output value of the photovoltaic module is used as output data for the photovoltaic power generation coupling sub-model. The root mean square error is used as the evaluation index to iteratively train the photovoltaic power generation coupling sub-model.

6. The method according to claim 5, characterized in that, The method further includes: Constraints are set to constrain the predicted output value of the photovoltaic power generation coupling sub-model.

7. The method according to claim 6, characterized in that, The method further includes: The probability of power curtailment is calculated based on the real-time power output prediction of the photovoltaic module obtained from the photovoltaic power generation coupling sub-model.

8. The method according to claim 1, characterized in that, The multi-agent game interaction sub-model is constructed through the following steps: Construct an infrastructure based on a multi-entity rule base, where the multi-entities include energy storage batteries, new energy equipment, and power grids; Construct the initial interaction rules for the multi-agent rule base and the initial thresholds during the interaction process; Based on historical data over a predetermined period, the initial interaction rules and initial thresholds are revised and iterated. Based on the revised and iterated interaction rules and thresholds, the multi-agent game interaction sub-model is obtained.

9. The method according to claim 8, characterized in that, The method further includes: Input the output data of the current energy storage battery physical sub-model and the new energy power generation coupling sub-model, as well as the environmental linkage data, into the multi-agent game interaction sub-model to obtain the charging and discharging commands of the energy storage battery.

10. The method according to any one of claims 1-9, characterized in that, The target construction and strategy optimization result set includes an executable rule base and a set of dynamic threshold targets bound to the executable rule base.

Images