Hybrid energy storage optimal configuration method based on multi-time scale regulation and control requirements

Abstract

The invention discloses a hybrid energy storage optimization configuration method based on a multi-time scale regulation and control demand, and the method comprises the steps: collecting the operation data of a power system, and obtaining the power fluctuation and the frequency deviation of the power system under a short time scale through a preset high-proportion renewable energy power system model; on the basis of power fluctuation and frequency deviation, the regulation and control requirements of the power system are divided according to different time scales, different energy storage modules are correspondingly configured, and a multi-level collaborative energy storage regulation and control structure matched with the time scales is formed; based on the power system model and the mathematical model of each energy storage module, constructing a hierarchical nested optimization model oriented to multiple time scales; and selecting a typical scene from the operation data by adopting a time sequence clustering method, and obtaining the optimal capacity and power configuration of each energy storage module based on a hierarchical nested optimization model, and obtaining a cross-time-scale collaborative operation strategy. The global optimal configuration is realized by establishing a multi-time-scale collaborative hierarchical nested optimization architecture.

Description

Technical Field

[0001] This invention relates to the field of power system automation, and in particular to a method for optimizing the configuration of hybrid energy storage based on multi-timescale control requirements. Background Technology

[0002] As the penetration rate of renewable energy sources such as wind power and photovoltaics in the power system continues to increase, the power system is undergoing a transformation towards a high proportion of new energy structures. However, wind and photovoltaic power generation has significant randomness, intermittency, and volatility. The time characteristics of their power output span from rapid frequency fluctuations on the order of seconds and minutes to long-term energy imbalances on the order of days and seasons, constituting complex power fluctuations on multiple time scales, which pose a severe challenge to the real-time power balance and safe and stable operation of the power system. Traditional power systems mainly rely on the regulation capabilities of thermal power units to maintain supply and demand balance. However, thermal power units themselves have inherent characteristics such as slow response speed, limited regulation range, and high start-up and shutdown costs, making it difficult to effectively cope with the rapid, multi-scale dynamic regulation requirements of the system under the condition of high proportion of renewable energy integration.

[0003] To mitigate the uncertainties caused by fluctuations in renewable energy output, various energy storage technologies have been extensively researched and applied. Molten salt thermal energy storage, electrochemical batteries, and flywheel energy storage, among others, offer complementary advantages in their technical characteristics: molten salt thermal energy storage systems feature large capacity and long-term energy storage, suitable for energy transfer and balancing across days and even seasons; battery energy storage systems possess high energy density and fast response speed, capable of handling intraday load regulation and peak shaving; and flywheel energy storage, with its millisecond-level extremely fast response speed and long cycle life, is suitable for frequency regulation and short-term dynamic power support. However, existing research and practice show that although a single type of energy storage technology can improve some of the system's operational performance on a specific time scale, its inherent technical limitations make it difficult to independently meet the comprehensive control needs of the power system at multiple levels and different time scales, including energy balance, intraday peak shaving, and second-level frequency regulation.

[0004] Currently, most existing research focuses on the configuration optimization of a single time scale or a single energy storage type. For example, some literature uses battery energy storage systems to achieve intraday peak shaving and valley filling, but fails to fully consider the long-term energy imbalance caused by seasonal variations in the system; other studies use flywheels or supercapacitors as the main frequency regulators, but their limited energy storage capacity cannot cope with energy deviations over long time scales. This optimization method, limited to a single time scale or a single energy storage type, cannot achieve coordinated optimization of the power system at different time levels, and is prone to problems such as unreasonable allocation of energy storage resources, high overall investment costs, or insufficient system frequency regulation capabilities.

[0005] Furthermore, existing energy storage configuration methods still suffer from significant limitations in mathematical model construction due to simplification. Most methods fail to adequately consider the dynamic coupling relationships between different types of energy storage systems and lack an effective mechanism for connecting long-term scheduling with short-term operation. Since there is an inherent energy constraint between long-term scheduling and short-term operation, independent optimization at each time scale will result in inconsistent optimization results over time, making it difficult to achieve optimal overall system performance. Summary of the Invention

[0006] The purpose of this invention is to provide a hybrid energy storage optimization configuration method based on multi-timescale regulation requirements. By establishing a multi-timescale collaborative hierarchical nested optimization architecture, the global optimal configuration of the hybrid energy storage system is achieved, thereby significantly improving the renewable energy absorption capacity and reducing the overall operating cost while ensuring system frequency stability.

[0007] To address the aforementioned technical problems, a first aspect of this invention provides a method for optimizing the configuration of hybrid energy storage based on multi-timescale regulation requirements, comprising the following steps: Collect power system operation data, and based on the operation data and a preset high-proportion renewable energy power system model, obtain power fluctuations and frequency deviations faced by the power system in a short time scale; Based on the power fluctuations and frequency deviations, the control requirements of the power system are divided according to different time scales, and different energy storage modules are configured accordingly to form a multi-level collaborative energy storage control structure that matches the time scale. Based on the power system model and the mathematical models of each energy storage module, a hierarchical nested optimization model oriented towards multiple time scales is constructed. A time series clustering method is used to select typical scenarios from the operational data, and the optimal capacity and power configuration of each energy storage module, as well as the collaborative operation strategy across time scales, are obtained based on the hierarchical nested optimization model. The power fluctuations include power fluctuations at different time scales caused by renewable energy and load fluctuations, and the mathematical models of each energy storage module include mathematical models characterizing the power, capacity and energy state of each energy storage module.

[0008] Furthermore, the division of power system regulation needs according to different time scales based on the power fluctuations and frequency deviations includes: The power fluctuation spectrum contains fluctuation components with different time constants. The control requirements corresponding to the fluctuation components at the second to minute level are divided into short time scales, the control requirements corresponding to the fluctuation components at the hour to day level are divided into medium time scales, and the control requirements corresponding to the fluctuation components at the multi-day level or above are divided into long time scales.

[0009] Furthermore, the acquisition of the fluctuation components with different time constants contained in the spectrum of the power fluctuation includes: Wavelet packet decomposition was performed on the time series of the power fluctuations to obtain multiple different frequency band sub-sequences; Calculate the energy contribution rate of each frequency band subsequence in the total power fluctuation sequence; Based on the preset energy contribution rate threshold, the characteristic frequency bands that affect the system power balance are selected; The characteristic frequency band is mapped to fluctuation component categories at the second to minute, hour to day, and multiple days or more, according to the time period corresponding to its center frequency.

[0010] Furthermore, the different energy storage modules include molten salt thermal storage modules, battery energy storage modules, and flywheel energy storage modules; the mathematical model of the molten salt thermal storage module is constructed through the following steps: A dual-tank structure model including a high-temperature storage tank and a low-temperature storage tank is established. The high-temperature storage tank is defined to store high-temperature molten salt after conversion by an electric heater, and the low-temperature storage tank stores low-temperature molten salt after energy release. An energy dynamic equation is constructed for the molten salt thermal storage module. The energy dynamic equation describes the energy transfer process between the high-temperature storage tank and the low-temperature storage tank, and takes into account the thermal inertia characteristics of the molten salt circulation system. The energy conversion efficiency parameter of the charging and discharging process is introduced into the energy dynamic equation to characterize the charging efficiency of electrical energy to thermal energy and the discharging efficiency of thermal energy to electrical energy, respectively. The self-discharge rate parameter is introduced into the energy dynamic equation to characterize the heat loss of the molten salt thermal storage system during static storage. A first power constraint model for the molten salt thermal storage module is established, and the constraints on its maximum charging power and maximum discharging power are defined. A first capacity constraint model is established for the molten salt thermal storage module, defining the minimum and maximum thermal storage capacity limits for its high-temperature and low-temperature storage tanks; By integrating the energy dynamic equation, the first power constraint model, and the first capacity constraint model, a mathematical model of the molten salt thermal storage module is obtained.

[0011] Furthermore, the mathematical model of the battery energy storage module is constructed through the following steps. A state of charge (SOC) change model for the battery energy storage module is established. The SOC change model calculates the SOC value at the next moment using the current charging and discharging power and the corresponding charging and discharging efficiency. Define the safe operating range of the state of charge of the battery energy storage module, set the upper and lower thresholds of its state of charge, and automatically limit its charging and discharging behavior when the state of charge reaches the corresponding threshold. A second power constraint model for the battery energy storage module is constructed, stipulating that its maximum charging power and maximum discharging power shall not exceed its rated power value; A second capacity constraint model for the battery energy storage module is established to ensure that its actual stored energy is always within its rated capacity range. Based on the historical operating data of the battery energy storage module, the charge and discharge depth of each charge and discharge cycle is calculated. According to the relationship between the charge and discharge depth and the preset empirical decay index, the equivalent cycle life loss caused by the charge and discharge cycle is calculated. The total equivalent cycle life loss of the battery energy storage module in the optimization period is accumulated and the total equivalent cycle life loss is used as a key parameter for evaluating its operating status. By integrating the state of charge change model, the second power constraint model, the second capacity constraint model, and the equivalent cycle life loss assessment process based on charge-discharge depth, a mathematical model of the battery energy storage module is obtained.

[0012] Furthermore, the equivalent cycle life loss assessment process based on charge / discharge depth also includes: Real-time monitoring of internal temperature changes in the battery energy storage module during charging and discharging; An accelerated life decay model considering both charge / discharge depth and operating temperature is established. This accelerated life decay model characterizes the accelerated decay effect of high-temperature operation on battery cycle life under the same charge / discharge depth. When calculating the equivalent cycle life loss caused by a single charge-discharge cycle, a life decay acceleration factor based on the current operating temperature is introduced simultaneously to obtain the total equivalent cycle life loss combined with the temperature effect.

[0013] Furthermore, the mathematical model of the flywheel energy storage module is constructed through the following steps. A model is established to show the relationship between the rotor kinetic energy and the stored energy of the flywheel energy storage module. The model calculates the instantaneous stored energy through the quantitative relationship between the rotor moment of inertia and the square of the real-time rotational speed. Construct a third power constraint model for the flywheel energy storage module, define its maximum charging power and maximum discharging power limits, and ensure that its power change rate does not exceed the maximum ramp rate allowed by the device; Establish the safe operating range of the flywheel energy storage module's rotational speed, and set an upper and lower threshold for its rotor speed. The lower threshold is used to ensure that the rotor maintains the necessary kinetic energy reserve. An energy state calculation model for the flywheel energy storage module is established. The energy state calculation model determines the ratio of its current available energy to its maximum stored energy by the ratio of real-time rotational speed to rated rotational speed. A droop control response model for the flywheel energy storage module is constructed, a linear adjustment relationship between the system frequency deviation and its output power is established, and the adjustable range of its droop coefficient is defined. A speed feedback correction mechanism is introduced into the droop control response model. When the real-time speed approaches the lower limit threshold, the output power is automatically reduced to ensure that the rotor speed is always maintained within the safe operating range. By integrating the rotor kinetic energy and stored energy relationship model, the third power constraint model, the speed safe operating range constraint, the energy state calculation model, and the droop control response model, a mathematical model of the flywheel energy storage module is obtained.

[0014] Furthermore, the introduction of a speed feedback correction mechanism into the droop control response model includes: Establish a dynamic mapping relationship between the speed status of the flywheel energy storage module and the droop coefficient. When the real-time speed is close to the rated speed, the standard droop coefficient is adopted, and when the real-time speed is close to the lower limit threshold, the droop coefficient is automatically reduced. In the initial stage of frequency disturbance, the flywheel energy storage module is controlled to respond with the maximum available droop factor, while the state of charge of the battery energy storage module is monitored. When the speed of the flywheel energy storage module drops to the first preset threshold, the frequency regulation function of the battery energy storage module is activated, and its droop coefficient increases linearly from zero. When the speed of the flywheel energy storage module drops to the second preset threshold, it completely exits the frequency regulation response, and the battery energy storage module independently undertakes the subsequent frequency regulation task. By coordinating the determination of the rotational speed and state of charge, the automatic role switching and seamless power transfer between the flywheel energy storage module and the battery energy storage module are achieved during the frequency regulation process.

[0015] Furthermore, the construction of a hierarchical nested optimization model oriented towards multiple time scales, based on the power system model and the mathematical model of each energy storage module, includes: An outer-layer optimization model is established with the goal of minimizing system net load fluctuation and optimizing energy storage configuration capacity to determine the capacity and power configuration parameters of the molten salt thermal storage module and the battery energy storage module. An inner-layer optimization model is established to optimize the frequency regulation power allocation strategy between the flywheel energy storage module and the battery energy storage module with the goal of minimizing the system frequency deviation. A parameter transfer interface is established between the outer optimization model and the inner optimization model to transfer the energy storage configuration parameters determined by the outer optimization to the inner optimization model as the operating boundary. An operation status feedback interface is established in the inner optimization model to feed back the energy state and frequency regulation performance index of the energy storage module obtained from the inner optimization simulation to the outer optimization model. Through the parameter transmission interface and the running status feedback interface, a bidirectional data interaction path is constructed between the outer optimization layer and the inner optimization layer, forming the basic architecture of the hierarchical nested optimization model.

[0016] Furthermore, the step of selecting typical scenarios from the operational data using time-series clustering, obtaining the optimal capacity and power configuration of each energy storage module based on the hierarchical nested optimization model, and the cross-timescale collaborative operation strategy includes: Normalization preprocessing was performed on wind power output data, photovoltaic power output data and load demand data to obtain several time series. Soft dynamic time warping distance was used as a similarity criterion to perform cluster analysis on the time series. Based on the silhouette coefficient, the clustering quality under different numbers of clusters is evaluated, and the optimal number of clusters that can best represent the annual operating characteristics is determined. Clustering is completed based on the optimal number of clusters. The daily running curve of the center of each cluster is selected as a typical scenario, and its annual frequency of occurrence is counted to calculate the corresponding scenario weight. Integrate all the typical scenarios and their scenario weights to construct a multi-scenario set for input to the hierarchical nested optimization model. Input the multi-scenario set into the hierarchical nested optimization model and solve it together using a mixed integer linear programming algorithm and a hierarchical iterative algorithm to simultaneously optimize the energy storage configuration and operation strategy under all typical scenarios. A bidirectional iterative convergence criterion for outer-layer optimization and inner-layer optimization is established. The outer-layer optimization model and the inner-layer optimization model are solved alternately, and the optimization result is judged to meet the preset accuracy requirements based on the convergence criterion. When the convergence criterion is not met, the constraints of the outer optimization model are updated based on the running status data of the inner optimization feedback, and the outer optimization calculation is performed again. When the convergence criterion is met, the solution results are weighted and integrated according to the scenario weights to output the unified optimal capacity configuration, power configuration and collaborative operation strategy of all energy storage modules across the entire time scale.

[0017] Accordingly, a second aspect of the present invention provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the above-described hybrid energy storage optimization configuration method based on multi-timescale regulation requirements.

[0018] Accordingly, a third aspect of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described hybrid energy storage optimization configuration method based on multi-timescale control requirements.

[0019] The above-described technical solutions of the embodiments of the present invention have the following beneficial technical effects: 1. By scientifically dividing the system regulation demand into three time scales—long, medium, and short—and configuring molten salt thermal storage, battery energy storage, and flywheel energy storage modules respectively, a multi-level collaborative energy storage regulation structure precisely matched with the time scale is constructed. Based on the functional allocation mechanism of physical characteristics, the energy storage resources with different characteristics can leverage their strengths and avoid their weaknesses, thus significantly improving the overall flexibility and all-time-domain coordinated control capability of the high-proportion renewable energy power system from the system architecture level, effectively coping with complex power fluctuations from the second level to the seasonal level. 2. The constructed hierarchical nested optimization model for multiple time scales, through the outer layer optimization focusing on minimizing system net load fluctuations and optimizing energy storage capacity, and the inner layer optimization dedicated to minimizing system frequency deviation, successfully solves the problems of target conflict and information disconnect between long-term planning and short-term operation by combining a two-way iterative mechanism. This enables the energy storage system's capacity configuration scheme and real-time operation strategy to be optimized in a coordinated manner, maximizing the investment and operational benefits of energy storage assets while ensuring the stability of the system's dynamic frequency. 3. The energy storage mathematical model deeply integrates refined operational constraints and lifespan degradation models, especially the charge / discharge depth-temperature dual-factor lifespan degradation model for battery energy storage and the electromechanical coupling safety constraint model for flywheel energy storage. This ensures that the optimized configuration results not only meet the system's power and energy requirements but also accurately reflect the performance changes and lifespan losses of energy storage devices during actual operation. Consequently, at the technical solution level, this directly improves the economic efficiency and engineering feasibility of the configuration results, effectively extends the service life of key equipment, and reduces the total lifespan cost. Attached Figure Description

[0020] Figure 1 This is a flowchart of a hybrid energy storage optimization configuration method based on multi-timescale regulation requirements provided in an embodiment of the present invention; Figure 2 A typical daily normalized curve of wind power, photovoltaic power, electricity, heating and cooling loads is provided for an embodiment of the present invention; Figure 3 The frequency fluctuation comparison curves of power systems with and without energy storage participating in frequency regulation are provided for embodiments of the present invention. Detailed Implementation

[0021] 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.

[0022] Please refer to Figure 1 The first aspect of this invention provides a method for optimizing the configuration of hybrid energy storage based on multi-timescale regulation requirements, comprising the following steps: Step S100: Collect power system operation data. Based on the operation data and a pre-set high-proportion renewable energy power system model, obtain power fluctuations and frequency deviations faced by the power system on short-term time scales. Power fluctuations include power fluctuations at different time scales caused by renewable energy and load fluctuations. First, it is necessary to systematically collect multi-source operational data of the power system in the target area, including wind farm output curves, photovoltaic power plant generation capacity, regional load demand data, and operating parameters of thermal power units. The time resolution of this data is typically set at 15-minute intervals, and the time span should cover at least one full year to reflect seasonal characteristics. Based on this operational data and a pre-set high-proportion renewable energy power system model, the power fluctuation characteristics of the system at different time scales are obtained by calculating the time-series changes in net load power (the difference between load demand and renewable energy generation capacity). Simultaneously, by analyzing the statistical characteristics of the system frequency, the frequency deviation risks faced by the power system at short time scales are identified. These power fluctuation and frequency deviation data provide important input for subsequent energy storage configuration.

[0023] Step S200: Based on power fluctuations and frequency deviations, the control requirements of the power system are divided according to different time scales, and different energy storage modules are configured accordingly to form a multi-level collaborative energy storage control structure that matches the time scale.

[0024] Based on the power fluctuation and frequency deviation analysis results obtained in step S100, the power system's regulation needs are scientifically divided according to their time-scale characteristics. Long-timescale regulation needs primarily target power fluctuations with periods exceeding several days, addressing seasonal imbalances in renewable energy output. Medium-timescale regulation needs mainly address hourly power fluctuations within a day, handling peak-valley load differences and intraday fluctuations in renewable energy. Short-timescale regulation needs focus on rapid power fluctuations ranging from seconds to minutes, ensuring system frequency stability. For regulation needs at different timescales, corresponding energy storage modules are configured: large-capacity molten salt thermal energy storage systems for long-timescale needs, energy-type battery energy storage systems for medium-timescale needs, and power-type flywheel energy storage systems for short-timescale needs, forming a multi-level collaborative energy storage regulation structure precisely matched to the time-scale characteristics.

[0025] Step S300: Based on the power system model and the mathematical models of each energy storage module, construct a hierarchical nested optimization model oriented towards multiple time scales. The mathematical model of each energy storage module includes mathematical models characterizing the power, capacity, and energy state of each energy storage module.

[0026] Based on the power system model and the mathematical models of each energy storage module, a hierarchical nested optimization model oriented towards multiple time scales is constructed. The power system model includes power balance equations, operating constraints, and economic objective functions; the mathematical models of each energy storage module characterize their power characteristics, capacity limitations, and energy state changes in detail. The hierarchical nested optimization model adopts a two-layer structure design: the outer layer focuses on long-term economic objectives, determining the capacity configuration scheme of the energy storage system; the inner layer focuses on short-term technical performance, optimizing the operating strategy of the energy storage system. A two-way data interaction mechanism is established between the two optimization models to achieve coordinated optimization through parameter transfer and state feedback, ensuring effective connection between long-term planning and short-term operation.

[0027] Step S400: Select typical scenarios from the operational data using time series clustering method, and obtain the optimal capacity and power configuration of each energy storage module, as well as the collaborative operation strategy across time scales, based on the hierarchical nested optimization model.

[0028] A time-series clustering method was used to select representative typical scenarios from the annual operational data. First, the wind power output, photovoltaic power generation, and load demand data were preprocessed by normalization. Then, cluster analysis was performed based on morphological similarity, and the central daily curve of each cluster was selected as a typical scenario, with its weight calculated. These typical scenarios were input into a hierarchical nested optimization model, and a suitable optimization algorithm was used to solve the problem, obtaining the optimal capacity and power configuration parameters for each energy storage module, including the thermal storage capacity and power level of the molten salt thermal storage system, the rated capacity and power characteristics of the battery energy storage system, and the kinetic energy capacity and power capability of the flywheel energy storage system. Simultaneously, a cross-timescale collaborative operation strategy was generated, clarifying the functional positioning and coordination mechanism of each energy storage system at different time scales.

[0029] The above process, through the establishment of a methodology of multi-timescale partitioning and hierarchical nested optimization, achieves global collaborative configuration and operational optimization of hybrid energy storage systems. This method constructs a collaborative control architecture adaptable to fluctuation characteristics at different time scales by decomposing system regulation requirements according to time characteristics and matching them with corresponding energy storage technologies. A bidirectional iterative mechanism of outer and inner layer optimization solves the coordination problem between long-term planning and short-term operation. Typical scenario extraction and multi-scenario optimization ensure the adaptability of the configuration scheme to year-round operating conditions. Implementation results show that this technical solution can effectively improve the power system's adaptability to renewable energy fluctuations, enhance system operational stability and power supply reliability, and provide a scientific and effective energy storage planning and operation solution for high-proportion renewable energy power systems.

[0030] Specifically, step S100 involves collecting power system operation data. Based on this data and a pre-set high-proportion renewable energy power system model, power fluctuations and frequency deviations faced by the power system in a short timescale are obtained, including: Step S110: Collect power output data of wind farms, power generation data of photovoltaic power plants, load demand data of regional power grids, and operating parameters of thermal power units, including at least minimum technical output, ramp rate, and start-stop characteristics.

[0031] In the specific implementation process, it is necessary to systematically collect multi-source operational data of the power system. This includes output data obtained from wind farm monitoring systems, power generation data obtained from photovoltaic power plant energy management systems, regional load demand data obtained from the power grid dispatch center, and unit operating parameters collected from the distributed control systems of thermal power plants. Among these, the operating parameters of thermal power units need to be recorded in detail, including their minimum stable technical output, maximum output capacity, upward and downward ramp rate limits, cold and hot start-up time characteristics, and corresponding start-up and shutdown cost curves.

[0032] Step S120: Normalize the collected wind power output data, photovoltaic power generation data and load demand data to eliminate the influence of dimensions and unify the time resolution to the preset scheduling cycle.

[0033] Standardizing the collected raw data is a crucial step in ensuring model accuracy. First, wind power output, photovoltaic power generation, and load demand data are normalized to a unified baseline to eliminate dimensional differences. Then, the temporal resolution of all data is adjusted to a preset scheduling cycle, typically set at 15-minute intervals, to meet the actual needs of power system dispatching. This process requires considering timestamp alignment issues from different data sources. Missing data is appropriately filled using interpolation between adjacent time periods or substitution with similar days to ensure a complete and consistent time-series dataset for subsequent modeling.

[0034] Step S130: Based on the preprocessed data, establish a power system power balance equation that includes wind power, photovoltaic power, load and thermal power units. The power balance equation represents the supply and demand relationship between renewable energy power generation and load demand at any given time.

[0035] Based on the preprocessed data, an accurate power system power balance equation is established. This equation must satisfy a dynamic balance between generation and load at any given time, specifically expressed as the sum of wind power output, photovoltaic power output, thermal power unit output, and energy storage charging / discharging power equaling load demand and network losses. To accurately describe scenarios with a high proportion of renewable energy integration, the equation must specifically consider the stochastic characteristics and spatiotemporal correlation of wind and solar power output, as well as the uncertainty of load forecasting errors. Establishing this core balance relationship provides a mathematical model foundation for subsequent analysis of system power deficits and surpluses, and for assessing energy storage configuration needs.

[0036] Step S140: Introduce operating constraints characterizing the system's regulation capability into the power balance equation, including upper and lower limits of thermal power unit output, ramp rate constraints, and system spinning reserve constraints.

[0037] Physical constraints characterizing the actual operating characteristics of the system are introduced based on the power balance equation. Constraints for thermal power units include their upper and lower limits of technical output, which reflect the stable operating range of the generator unit; ramp rate constraints limit the unit's output variation limit per unit time, directly affecting the system's regulation flexibility; and system spinning reserve constraints require maintaining sufficient online reserve capacity to cope with sudden power shortages.

[0038] Step S150: Construct the optimization objective function of the power system model. The optimization objective function includes a wind and solar curtailment penalty term that characterizes the level of renewable energy consumption, and a power balance adjustment term that characterizes the system regulation demand.

[0039] The objective function for constructing the power system model comprises two key components. The wind and solar curtailment penalty term sets a unit curtailment penalty cost, prompting the system to maximize the absorption of renewable energy generation and reduce clean energy waste during the optimization process. The power balance regulation term quantifies the regulation cost incurred by the system to maintain real-time power balance, guiding the optimization process to seek the most economically optimal operating mode.

[0040] Step S160: The power balance equation, operating constraints and optimization objective function are integrated to obtain a high-proportion renewable energy power system model for energy storage configuration analysis.

[0041] The aforementioned power balance equations, system operation constraints, and optimization objective functions are systematically integrated to form a complete high-proportion renewable energy power system model. This model can simulate the dynamic behavior of the system under different operating scenarios, quantitatively assess the impact of wind and solar power output fluctuations on the system power balance, and accurately calculate the regulation resources required to maintain the safe and stable operation of the system. This integrated model serves as the analytical basis for subsequent hybrid energy storage optimization configuration, providing a reliable simulation platform for scientifically determining the capacity and power configuration of various types of energy storage.

[0042] Specifically, step S200, based on the power fluctuations and frequency deviations, divides the power system's regulation needs according to different time scales, including: Step S210: Obtain the fluctuation components with different time constants contained in the spectrum of power fluctuations, divide the control requirements corresponding to the fluctuation components at the second to minute level into short time scales, divide the control requirements corresponding to the fluctuation components at the hour to day level into medium time scales, and divide the control requirements corresponding to the fluctuation components at the multi-day level or above into long time scales.

[0043] In the specific implementation process, signal processing technology was used to analyze the spectral characteristics of the power fluctuation time series. By performing wavelet packet decomposition on the system's net load power change curve, the original power signal was decomposed into multiple subsequences of different frequency bands, each subsequence corresponding to a fluctuation component with a specific time constant. Based on the time period corresponding to the center frequency of each frequency band in the decomposition results, the high-frequency fluctuation components at the second to minute level were classified as short-term time-scale control needs, which are mainly caused by random load changes and instantaneous fluctuations in wind and solar power output; the medium-frequency fluctuation components at the hour to day level were classified as medium-term time-scale control needs, which mainly originate from intraday load periodic changes and the diurnal periodic characteristics of wind and solar power output; and the low-frequency fluctuation components at the multi-day level or above were classified as long-term time-scale control needs, which are mainly affected by long-term factors such as weather system transitions and seasonal changes. This classification method based on physical characteristics provides a scientific basis for the subsequent targeted allocation of energy storage resources.

[0044] Furthermore, after dividing the power system regulation requirements according to different time scales based on the power fluctuations and frequency deviations in step S200, it also includes: Step S220: Determine the power support demand for short-term time-scale regulation based on the statistical characteristics of frequency deviation, determine the power regulation demand for medium-term time-scale regulation based on the characteristics of daily net load curve, and determine the capacity support demand for long-term time-scale regulation based on the annual renewable energy output characteristics.

[0045] Based on the division results of step S210, the specific regulation requirements for each time scale are further quantified. For short time scales, by statistically analyzing the standard deviation and maximum deviation of historical frequency deviation data, the instantaneous power support capacity required to maintain frequency stability is determined. This requirement mainly addresses the initial recovery phase after the system experiences disturbances. For medium time scales, by analyzing the peak-to-valley difference rate and volatility of typical daily net load curves, the energy regulation capacity required to smooth intraday power fluctuations is determined. This requirement mainly addresses the supply-demand imbalance problem at different times of the day. For long time scales, by analyzing the seasonal variation of the annual renewable energy output curve and the distribution of periods of continuous power surplus or shortage, the large-capacity energy storage support required to achieve cross-day or cross-seasonal energy transfer is determined. This requirement mainly addresses the long-term energy management challenges caused by the uneven seasonal distribution of wind and solar resources.

[0046] Step S230 couples the power support requirements, power regulation requirements, and capacity support requirements in the time dimension to form a complete regulation and control requirement system covering the second level to the seasonal level.

[0047] By systematically coupling the regulation needs across three time scales along the time dimension, a complete regulation demand system is constructed. Short-term power support needs provide the system with rapid dynamic response capabilities, ensuring frequency stability; medium-term power regulation needs provide intraday energy dispatch capabilities, optimizing operational economy; and long-term capacity support needs provide long-term energy balance guarantees, improving the absorption of renewable energy. These three levels of regulation needs are not independent but interconnected through time boundary conditions. That is, the initial state of optimization at a shorter time scale is determined by the optimization results at the previous time scale, forming a full-time-scale coordination mechanism from the second level to the season level. This coupling relationship ensures the consistency of regulation strategies across time scales, laying a theoretical foundation for constructing a collaborative and efficient hybrid energy storage system.

[0048] Furthermore, the power fluctuation spectrum obtained in step S210 includes fluctuation components with different time constants, including: Step S211: Perform wavelet packet decomposition on the time series of power fluctuations to obtain multiple different frequency band subsequences.

[0049] When performing wavelet packet decomposition on the power fluctuation time series, the Daubechies wavelet basis function, suitable for non-stationary signal analysis, is first selected. Based on the characteristics of the power fluctuation signal, the decomposition level is determined to be 8 levels, decomposing the original power fluctuation sequence into 256 subsequences of different frequency bands. During the decomposition process, the Mallat fast algorithm is used to achieve step-by-step decomposition of the signal. Each level of decomposition separates the signal into a low-frequency approximation part and a high-frequency detail part. Through this multi-resolution analysis method, the localized characteristics of power fluctuations in both the time and frequency domains can be obtained simultaneously. The depth of wavelet packet decomposition needs to be adjusted according to the main periodic components of power fluctuations in the actual system to ensure that the decomposed subsequences can fully represent various fluctuation characteristics from the second level to the seasonal level, providing complete frequency band information for subsequent analysis.

[0050] Step S212: Calculate the energy contribution rate of each frequency band subsequence in the total power fluctuation sequence.

[0051] After wavelet packet decomposition, the energy contribution rate of each frequency band subsequence needs to be calculated to assess its importance. The energy contribution rate is calculated using Parseval's theorem. First, the sum of squares of each frequency band subsequence is calculated as the energy of that frequency band. Then, the ratio of the energy of each frequency band to the total energy of the total power fluctuation sequence is calculated. In actual calculations, the weight differences in the impact of different frequency band energies on the system also need to be considered. For specific frequency bands that may cause system stability problems, their weight coefficients can be appropriately increased.

[0052] Step S213: Based on the preset energy contribution rate threshold, select the characteristic frequency bands that affect the system power balance.

[0053] Based on the energy contribution rate calculation results, a preset energy contribution rate threshold is used for characteristic frequency band screening. The threshold setting needs to consider the actual operating requirements and safety standards of the system. Typically, frequency bands with an energy contribution rate exceeding 1% are defined as characteristic frequency bands; for particularly important systems, even stricter threshold standards may be used. During the screening process, not only the energy contribution rate of a single frequency band needs to be considered, but also the energy distribution characteristics between adjacent frequency bands need to be analyzed to avoid energy dispersion caused by frequency band division.

[0054] Step S214: Map the characteristic frequency band to the fluctuation component categories of seconds to minutes, hours to days, and multiple days or more, according to the time period corresponding to its center frequency.

[0055] The selected characteristic frequency bands are mapped and classified according to the time period corresponding to their center frequency. The period corresponding to the center frequency is calculated using a frequency-period conversion formula. Characteristic frequency bands with periods ranging from 1 second to 15 minutes are mapped to second- to minute-level fluctuation components, which mainly affect the frequency stability and transient response characteristics of the system. Characteristic frequency bands with periods ranging from 15 minutes to 24 hours are mapped to hour- to day-level fluctuation components, which mainly affect the intraday scheduling and energy balance of the system. Characteristic frequency bands with periods exceeding 24 hours are mapped to fluctuation components at multiple days or more, which mainly affect the long-term operation mode and energy storage configuration requirements of the system.

[0056] Specifically, in step S300, the different energy storage modules include molten salt thermal storage modules, battery energy storage modules, and flywheel energy storage modules.

[0057] Accordingly, the mathematical model of the molten salt thermal storage module is constructed through the following steps: Step S311: Establish a dual-tank structure model including a high-temperature tank and a low-temperature tank, defining that the high-temperature tank stores high-temperature molten salt after conversion by an electric heater, and the low-temperature tank stores low-temperature molten salt after energy release.

[0058] When constructing the mathematical model of the molten salt thermal energy storage module, the first step is to establish a precise description of its physical structure. The molten salt thermal energy storage system employs a dual-tank structure, comprising a high-temperature tank and a low-temperature tank. The high-temperature tank stores the high-temperature molten salt obtained by converting electrical energy through an electric heater, with the temperature typically maintained at around 565 degrees Celsius. The low-temperature tank stores the molten salt whose temperature has decreased to approximately 290 degrees Celsius after energy release through a heat exchanger. The two tanks are connected by a molten salt pump and piping system, forming a complete molten salt circulation loop. During the modeling process, it is necessary to define in detail the geometric dimensions of the two tanks, the characteristics of the insulation layer, and the physicochemical properties of the molten salt. These parameters directly affect the system's thermal storage capacity and heat loss characteristics.

[0059] Step S312: Construct the energy dynamic equation of the molten salt thermal storage module. The energy dynamic equation describes the energy transfer process between the high-temperature storage tank and the low-temperature storage tank, and considers the thermal inertia characteristics of the molten salt circulation system.

[0060] Based on a dual-tank structural model, mathematical equations describing the dynamic energy changes of a molten salt thermal storage system are constructed. These equations need to accurately characterize the energy transfer process between the high-temperature and low-temperature tanks, including energy transport due to molten salt flow, heat conduction losses from the tank walls, and heat radiation losses to the environment. When establishing the equations, the thermal inertia of the molten salt circulation system must be considered, as this is due to the transmission delay of molten salt flowing in the pipes and the dynamic response characteristics caused by temperature stratification within the tanks. Thermal inertia causes a certain lag in the actual power response of the system relative to control commands. This lag effect needs to be represented in the mathematical model through first-order inertial elements or transmission delay elements to ensure that the model accurately reflects the dynamic behavior of the actual system.

[0061] Step S313: Introduce energy conversion efficiency parameters for the charging and discharging process into the energy dynamic equation to characterize the charging efficiency from electrical energy to thermal energy and the discharging efficiency from thermal energy to electrical energy, respectively.

[0062] Introducing energy conversion efficiency parameters for the charging and discharging processes into the energy dynamics equations is crucial for ensuring model accuracy. Charging efficiency parameters characterize the conversion efficiency from electrical energy to thermal energy, primarily including the electrothermal conversion efficiency of the electric heater and the energy losses of related auxiliary equipment (such as molten salt pumps), typically ranging from 95% to 98%. Discharging efficiency parameters characterize the conversion efficiency from thermal energy to electrical energy, primarily including the heat transfer efficiency of the heat exchanger, the thermal-to-work conversion efficiency of the steam turbine, and the electromechanical conversion efficiency of the generator, with the overall efficiency typically ranging from 45% to 50%.

[0063] Step S314: Introduce the self-discharge rate parameter into the energy dynamic equation to characterize the heat loss of the molten salt thermal storage system during static storage.

[0064] Molten salt thermal storage systems experience heat loss through various pathways during static storage, which needs to be characterized in mathematical models using the self-discharge rate parameter. Self-discharge primarily includes heat conduction loss from the tank walls, heat dissipation loss due to incomplete insulation, and energy consumption by the molten salt pumps to maintain the system's minimum cycle time. The magnitude of the self-discharge rate is closely related to factors such as the temperature difference between the inside and outside of the tank, the performance of the insulation layer, and environmental conditions, and is typically quantified as the hourly temperature drop or the percentage of energy loss per hour.

[0065] Step S315: Establish the first power constraint model of the molten salt thermal storage module and define the constraints on its maximum charging power and maximum discharging power.

[0066] Establishing a power constraint model for molten salt thermal energy storage modules requires starting from the actual operational limitations of the system. The maximum charging power is limited by the rated power of the electric heater, the grid connection capacity, and the maximum allowable heating rate of the molten salt system; the maximum discharging power is constrained by the heat transfer capacity of the heat exchanger, the maximum steam inlet of the steam turbine, and the rated output of the generator. When determining the power constraints, it is also necessary to consider the minimum time interval between the system's transition from charging to discharging states, as well as the maximum rate of power change. These constraints collectively define the power regulation range and dynamic response capability of the molten salt thermal energy storage system.

[0067] Step S316: Establish the first capacity constraint model of the molten salt thermal storage module and define the minimum and maximum thermal storage capacity limits of its high-temperature and low-temperature storage tanks.

[0068] The establishment of a capacity constraint model requires comprehensive consideration of the physical volume limitations, minimum safe liquid level requirements, and operating temperature ranges of both high-temperature and low-temperature storage tanks. The maximum heat storage capacity of the high-temperature tank is determined by its tank volume and maximum allowable temperature, while the minimum heat storage capacity must ensure that the system can maintain normal circulation even at the lowest operating liquid level. The constraints for the low-temperature tank are similar, but the freezing point limitation of the molten salt must be considered. Furthermore, the capacity constraints of the two tanks are coupled, meaning that the reduction in molten salt in the high-temperature tank must equal the increase in molten salt in the low-temperature tank; this mass balance relationship needs to be reflected in the capacity constraints.

[0069] Step S317: Integrate the energy dynamic equation, the first power constraint model, and the first capacity constraint model to obtain the mathematical model of the molten salt thermal storage module.

[0070] The energy dynamics equations, power constraint models, and capacity constraint models are systematically integrated to form a complete mathematical model of the molten salt thermal energy storage module. During integration, it is necessary to ensure parameter consistency and unit uniformity among the various sub-models, while also addressing model switching issues at different time scales. For long-term planning problems, simplified models can be used to improve computational efficiency; for short-term operational optimization, detailed models are required to ensure accuracy. The integrated mathematical model should accurately reflect the static characteristics and dynamic response of the molten salt thermal energy storage system, including key characteristics such as energy storage capacity, power limitations, efficiency characteristics, and operational constraints, providing a reliable model foundation for subsequent system-level optimization.

[0071] Accordingly, the mathematical model of the battery energy storage module is constructed through the following steps: Step S321: Establish a state of charge (SOC) change model for the battery energy storage module. The SOC change model calculates the SOC value for the next moment using the current charging / discharging power and the corresponding charging / discharging efficiency.

[0072] The state of charge (SOC) model for battery energy storage modules is based on the principle of energy conservation. It calculates the SOC value at the next moment by using the current charging / discharging power and corresponding charging / discharging efficiency. In the modeling process, charging efficiency considers the overall efficiency from energy absorption from the grid to battery storage, including converter losses and internal battery heat loss; discharging efficiency covers the energy loss from battery energy release to grid connection. The SOC calculation needs to consider the battery's rated capacity, initial SOC, and coulombic and energy efficiencies during charging and discharging. For lithium-ion batteries, the impact of charge / discharge rate on efficiency also needs to be considered, establishing a nonlinear relationship model between efficiency and charging / discharging power. This model accurately reflects the energy state change trajectory of the battery under different operating conditions, providing key state parameters for system optimization and scheduling.

[0073] Step S322: Define the safe operating range of the state of charge of the battery energy storage module, set the upper and lower thresholds of its state of charge, and automatically limit its charging and discharging behavior when the state of charge reaches the corresponding threshold.

[0074] Defining the safe operating range of the state of charge (SOC) of a battery energy storage module is crucial for ensuring the safe and stable operation of the system. The upper threshold is typically set between 90% and 95% to avoid accelerated battery aging and safety risks caused by overcharging; the lower threshold is generally set between 10% and 20% to prevent irreversible damage caused by over-discharge. When setting the thresholds, the characteristics of the battery chemistry must be considered. For lithium iron phosphate batteries, the operating range can be appropriately relaxed, while ternary lithium batteries require stricter control. When the SOC approaches the boundary threshold, the system smoothly transitions to a boundary protection state through a power limiting algorithm, avoiding sudden power drops caused by hard truncation. Furthermore, the setting of the operating range must balance system regulation requirements with battery life, maximizing battery lifespan while ensuring sufficient regulation margin.

[0075] Step S323: Construct a second power constraint model for the battery energy storage module, stipulating that its maximum charging power and maximum discharging power must not exceed its rated power value.

[0076] The power constraint model for battery energy storage modules needs to comprehensively consider the limitations of inverter capacity, battery rate characteristics, and thermal management system. Maximum charging power is limited by the battery's maximum acceptable charging current, which dynamically adjusts with changes in battery temperature, state of charge, and state of health. Maximum discharging power is constrained by battery internal resistance, heat dissipation capacity, and grid connection conditions. When establishing the power constraint model, it is necessary to distinguish between continuous power and short-term power constraints. Continuous power is determined based on thermal equilibrium conditions, while short-term power considers the battery's transient response capability. The power constraint model also needs to include constraints on the rate of power change to avoid adverse effects of sudden power fluctuations on battery life and grid stability, ensuring a smooth power transition.

[0077] Step S324: Establish a second capacity constraint model for the battery energy storage module to ensure that its actual stored energy is always within its rated capacity range.

[0078] The capacity constraint model ensures that the actual stored energy of the battery energy storage system remains within a safe operating range. This model is based on the battery's rated capacity and considers the impact of factors such as temperature and aging on the actual usable capacity. During modeling, it is necessary to distinguish between the theoretical capacity and the actual usable capacity, as the actual usable capacity gradually decreases with increasing cycle count. The capacity constraint also includes energy retention constraints, ensuring that the system's energy state changes at any given time conform to physical laws; that is, the current energy state equals the previous energy state plus the charging energy minus the discharging energy minus self-discharge losses. Furthermore, the impact of different charge / discharge rates on the effective capacity needs to be considered, establishing a coupling relationship model between capacity and power.

[0079] Step S325: Based on the historical operating data of the battery energy storage module, the charge and discharge depth of each charge and discharge cycle is counted. According to the relationship between the charge and discharge depth and the preset empirical decay index, the equivalent cycle life loss caused by the charge and discharge cycle is calculated. The total equivalent cycle life loss of the battery energy storage module in the optimization cycle is accumulated and counted. The total equivalent cycle life loss is used as a key parameter for evaluating its operating status.

[0080] The core innovation of battery mathematical models lies in the assessment of equivalent cycle life loss based on charge / discharge depth. This assessment process first extracts the depth characteristics of each charge / discharge cycle from historical operating data, where charge / discharge depth is defined as the magnitude of change in the state of charge during that cycle. Then, based on the empirical decay index obtained from battery aging experiments, a quantitative relationship between charge / discharge depth and single-cycle life loss is established, typically exhibiting an exponential function relationship, meaning that deep charge / discharge significantly accelerates lifespan decay. When calculating the total equivalent cycle life loss, the linear cumulative damage theory is employed, weighting and summing the equivalent losses of each cycle, with weights considering the effects of cycle sequence and recovery. The total equivalent cycle life loss, as a key state parameter, directly reflects the battery's health state and provides an important basis for optimizing operating strategies.

[0081] Step S326: Integrate the state of charge change model, the second power constraint model, the second capacity constraint model, and the equivalent cycle life loss assessment process based on charge and discharge depth to obtain the mathematical model of the battery energy storage module.

[0082] This system integrates the state-of-charge (POC) change model, power constraint model, capacity constraint model, and lifetime degradation assessment process to form a complete mathematical model for battery energy storage modules. During integration, it is necessary to address the coupling relationships between the various sub-models; for example, power constraints dynamically change with accumulated lifetime degradation, and capacity constraints adjust as the health state declines. The integrated model employs a unified parameter system and state variables to ensure internal consistency. This mathematical model can simultaneously describe the static characteristics and dynamic response of the battery, accurately predicting performance and lifetime changes under different operating strategies, providing high-precision battery behavior predictions for system-level optimization.

[0083] Furthermore, the equivalent cycle life loss assessment process based on charge / discharge depth in step S326 also includes: Step S3261: Monitor the internal temperature changes of the battery energy storage module in real time during the charging and discharging process.

[0084] During the operation of a battery energy storage system, a complete temperature monitoring system is required to track temperature changes inside the battery in real time. This monitoring system collects temperature data during the battery charging and discharging process by placing thermocouples or fiber optic temperature sensors at key locations within the battery module, sampling at least once per second. Monitoring points need to cover hot spots in the battery module, cooling system inlets and outlets, and ambient temperature reference points to ensure a comprehensive reflection of the battery's temperature distribution. The collected temperature data is processed and recorded in real time by the battery management system, establishing a complete temperature-over-time curve. This temperature data is not only used for control decisions by the battery thermal management system but also provides crucial input parameters for subsequent lifespan assessments, ensuring that the lifespan assessment model accurately reflects the impact of actual operating conditions on the battery's health status.

[0085] Step S3262: Establish a lifetime degradation acceleration model that considers both charge / discharge depth and operating temperature. The lifetime degradation acceleration model characterizes the accelerated degradation effect of high temperature operation on battery cycle life under the same charge / discharge depth.

[0086] Based on extensive battery aging experimental data, a lifespan degradation acceleration model considering both charge / discharge depth and operating temperature was established. This model describes the effect of temperature on the chemical reaction rate using the Arrhenius equation and characterizes the effect of charge / discharge depth on battery stress levels using the Eyring model, forming a complete two-factor mathematical model for lifespan degradation. Under the same charge / discharge depth conditions, high-temperature operation significantly accelerates the loss of active materials inside the battery, electrolyte decomposition, and the growth of the solid electrolyte interfacial film. These degradation mechanisms collectively lead to accelerated capacity decay and rapid increase in internal resistance. The model parameters were determined through accelerated aging tests conducted at different temperatures (0℃ to 55℃) and different charge / discharge depths (10% to 100%), establishing a quantitative relationship between temperature and lifespan degradation rate, providing a theoretical basis for lifespan prediction under different operating conditions.

[0087] In step S3263, when calculating the equivalent cycle life loss caused by a single charge-discharge cycle, a life decay acceleration factor based on the current operating temperature is introduced simultaneously to obtain the total equivalent cycle life loss combined with the temperature effect.

[0088] When calculating the equivalent cycle life loss caused by a single charge-discharge cycle, a life decay acceleration factor based on the current operating temperature needs to be introduced simultaneously. This acceleration factor is determined comprehensively based on the battery's average temperature, maximum temperature, and duration of high temperature during that cycle, and the specific acceleration coefficient is obtained by querying a preset temperature-life decay relationship curve. During the calculation, a baseline cycle life loss is first calculated based on the depth of charge and discharge. Then, this loss value is multiplied by the temperature acceleration factor to obtain the actual cycle life loss considering the temperature effect. The loss values ​​of all cycles are accumulated over time, considering the mutual influence between different cycles and the battery's self-recovery effect, ultimately yielding a more accurate total equivalent cycle life loss. This total loss metric more accurately reflects the battery's health degradation under current operating conditions.

[0089] Specifically, the mathematical model of the flywheel energy storage module is constructed through the following steps: Step S331: Establish a model for the relationship between the rotor kinetic energy and the stored energy of the flywheel energy storage module. The model calculates the instantaneous stored energy through the quantitative relationship between the rotor's moment of inertia and the square of the real-time rotational speed.

[0090] The model for the relationship between rotor kinetic energy and stored energy in a flywheel energy storage module is based on the principles of rigid body rotational dynamics. This model calculates instantaneous stored energy through the quantitative relationship between the flywheel rotor's moment of inertia and the square of its real-time rotational speed. The moment of inertia depends on the density distribution and geometric design of the flywheel material. In actual modeling, it is necessary to accurately measure or calculate the axial moment of inertia of the flywheel rotor, while also considering the additional inertia of the support system. The calculation of stored energy requires distinguishing between the theoretical maximum stored energy and the actual usable energy. The theoretical value is determined by the safe maximum rotational speed, which is determined by the rotor material's strength limit, while the usable energy in actual operation is limited by the system control strategy and safety margin. Furthermore, the model also needs to consider the variations in bearing losses and wind resistance losses at different rotational speeds, as these losses directly affect the system's net stored energy.

[0091] Step S332: Construct the third power constraint model of the flywheel energy storage module, define its maximum charging power and maximum discharging power limits, and ensure that its power change rate does not exceed the maximum ramp rate allowed by the device.

[0092] The power constraint model for flywheel energy storage modules needs to comprehensively consider the capacity of the power electronic converter, the torque characteristics of the motor, and rotor dynamics limitations. The maximum charging power is limited by the maximum input current of the motor when operating as an electric motor. This limitation is closely related to the speed; in the low-speed range, it is typically limited by the maximum torque, while in the high-speed range, it is limited by the back electromotive force. The maximum discharging power is constrained by the maximum output capacity of the motor when operating as a generator, the performance of the cooling system, and the grid connection conditions. The power change rate constraint mainly considers the rotor's ability to withstand mechanical stress. Excessive power change can lead to accelerated fatigue due to alternating stress on the rotor, and can also cause torsional vibration problems in the transmission system. Power constraint models typically use piecewise linear or nonlinear functions to describe these complex constraints.

[0093] Step S333: Establish the safe operating range of the flywheel energy storage module's rotational speed, and set the upper and lower thresholds of its rotor speed. The lower threshold is used to ensure that the rotor maintains the necessary kinetic energy reserves.

[0094] The setting of the safe operating speed range is based on multiple factors, including the mechanical strength, dynamic stability, and energy efficiency of the flywheel system. The upper threshold is typically set at 80%-90% of the maximum safe speed corresponding to the material's allowable stress, preserving sufficient safety margin to cope with unexpected operating conditions. The lower threshold needs to consider the minimum kinetic energy reserve required for the system to maintain stable operation, generally set at 30%-40% of the rated speed, ensuring a certain power support capability is maintained during frequency regulation. When setting the speed range, the efficiency characteristics of different speed ranges also need to be considered; generally, the mid-to-high speed range has better energy conversion efficiency. The safe speed range is dynamically adjusted according to the aging of the equipment and changes in environmental conditions to ensure safe operation of the system throughout its entire lifespan.

[0095] Step S334: Establish an energy state calculation model for the flywheel energy storage module. The energy state calculation model determines the ratio of its current available energy to its maximum stored energy by the ratio of real-time rotational speed to rated rotational speed.

[0096] The energy state calculation model for the flywheel energy storage module determines the ratio of currently available energy to maximum stored energy by comparing the real-time rotational speed with the rated rotational speed. This model is based on the square relationship between kinetic energy and rotational speed, defining the energy state as the ratio of kinetic energy at the current rotational speed to that at the rated rotational speed. In practical implementation, factors such as the accuracy of rotational speed measurement, the influence of temperature on the material's elastic modulus, and the impact of bearing preload variations on system stiffness need to be considered, as these factors all affect the accurate correspondence between rotational speed and stored energy. The energy state calculation also needs to consider the differences in available energy across different rotational speed ranges. When approaching the lower threshold, the actual available energy will decrease accordingly because the system needs to retain a minimum kinetic energy reserve. This model provides crucial state information for optimal system scheduling.

[0097] Step S335: Construct a droop control response model for the flywheel energy storage module, establish a linear adjustment relationship between the system frequency deviation and its output power, and define the adjustable range of its droop coefficient.

[0098] The droop control response model for flywheel energy storage modules establishes a linear adjustment relationship between system frequency deviation and output power. Its core is defining an appropriate droop coefficient. Setting the droop coefficient requires consideration of the system's inertial time constant, frequency regulation requirements, and coordination with other frequency regulation resources. The adjustable range design allows the flywheel to adapt to different operating modes and system conditions. A larger droop coefficient provides rapid support during rapid frequency changes, while a smaller droop coefficient avoids over-adjustment during minor frequency fluctuations. The droop control model also needs to consider dead-zone settings to avoid unnecessary responses to frequency fluctuations within the normal range. Furthermore, the model needs to include a feedforward control loop to improve the predictive response capability to frequency change trends.

[0099] Step S336: Introduce a speed feedback correction mechanism into the droop control response model. When the real-time speed approaches the lower limit threshold, automatically reduce its output power to ensure that the rotor speed is always maintained within the safe operating range.

[0100] Introducing a speed feedback correction mechanism into the droop control response model is a key measure to ensure the safe operation of the flywheel. When the real-time speed approaches the lower threshold, the system automatically reduces the droop coefficient through the speed feedback loop, thereby limiting the output power and preventing further speed drop that could jeopardize system stability. The correction mechanism employs a gradual adjustment strategy, smoothly reducing power output as the speed enters the warning range. As the speed approaches the critical value, the correction intensity gradually increases until it completely exits the frequency regulation state. This mechanism also needs to consider its adaptability under different operating conditions; in system emergencies, the restrictions can be appropriately relaxed to prioritize grid safety. The coordinated operation of speed feedback correction and droop control ensures that the flywheel maintains a safe operating state while fully utilizing its frequency regulation efficiency.

[0101] Step S337: Integrate the rotor kinetic energy and stored energy relationship model, the third power constraint model, the speed safe operating range constraint, the energy state calculation model, and the droop control response model to obtain the mathematical model of the flywheel energy storage module.

[0102] Integrating the various sub-models into a complete mathematical model of the flywheel energy storage module requires addressing the coupling relationships between different models and ensuring parameter consistency. During integration, it's crucial to ensure that the rotor dynamics model, power constraint model, speed safety model, energy state model, and droop control model employ a consistent system of state variables and parameters. The integrated mathematical model should accurately describe the static characteristics and dynamic response of the flywheel energy storage system, including key features such as energy storage capacity, power regulation range, efficiency characteristics, safety constraints, and control system behavior. The model also needs to consider its applicability across different time scales; a quasi-static model can be used for long-term planning problems, while a detailed dynamic model is required for short-term operational optimization and real-time control. Validation of the integrated model requires comparative analysis of actual operating data to ensure that the model's predictive accuracy meets engineering application requirements.

[0103] Furthermore, the introduction of a speed feedback correction mechanism into the droop control response model in step S336 includes: Step S3361: Establish a dynamic mapping relationship between the speed state of the flywheel energy storage module and the droop coefficient. When the real-time speed is close to the rated speed, the standard droop coefficient is used. When the real-time speed is close to the lower limit threshold, the droop coefficient is automatically reduced.

[0104] Establishing a dynamic mapping relationship between the flywheel energy storage module's rotational speed and its droop coefficient is a core control strategy to ensure the safe and stable operation of the system. This mapping relationship is constructed based on piecewise linear or nonlinear functions. When the real-time rotational speed is between 85% and 100% of the rated speed, the system adopts a standard droop coefficient, at which point the flywheel has sufficient kinetic energy reserves and can provide complete frequency regulation capabilities. When the rotational speed drops to between 60% and 85% of the rated speed, the system begins to linearly reduce the droop coefficient, limiting power output to preserve necessary kinetic energy reserves. When the rotational speed further drops to between 40% and 60% of the rated speed, the system adopts a significantly reduced droop coefficient, providing only limited power support. This dynamic mapping relationship is achieved by real-time monitoring of the rotational speed and querying a preset rotational speed-droop coefficient correspondence table, ensuring that the flywheel fully utilizes its frequency regulation function while avoiding the risk of excessively low rotational speed due to over-discharge.

[0105] Step S3362: In the initial stage of frequency disturbance, control the flywheel energy storage module to respond with the maximum available droop coefficient, while monitoring the state of charge of the battery energy storage module.

[0106] In the initial stage of a system frequency disturbance, the control strategy requires the flywheel energy storage module to respond to frequency changes with the maximum available droop factor. This stage typically refers to the 0.5 to 2 seconds following the occurrence of the frequency disturbance. The flywheel, utilizing its rapid response characteristics, immediately provides power support once the frequency exceeds the dead zone. Simultaneously, the system begins monitoring the state of charge (SOC) of the battery energy storage module, assessing its available frequency regulation capacity and response capability. Monitoring parameters include the battery's current SOC, available charge / discharge power, and health indicators. This data is acquired in real-time by the battery management system and transmitted to the central controller, providing a basis for subsequent coordinated control decisions. This rapid response in this stage provides crucial first-time support for system frequency stability.

[0107] Step S3363: When the speed of the flywheel energy storage module drops to the first preset threshold, the frequency regulation function of the battery energy storage module is activated, and its droop coefficient is linearly increased from zero.

[0108] When the flywheel energy storage module's rotational speed drops to the first preset threshold (usually set to 60% of the rated speed), the system activates the battery energy storage module's frequency regulation function. The droop coefficient of the battery energy storage module increases linearly from zero according to a preset slope, which is determined based on the system's frequency recovery requirements and the battery's power regulation capability. During startup, the response delay characteristics of the battery energy storage system need to be considered; a ramp-up startup method is typically used to avoid sudden power changes. Simultaneously, the flywheel energy storage module continues its frequency regulation function, but its droop coefficient further decreases as the rotational speed decreases, and the output power gradually decreases. The power outputs of the two energy storage systems complement each other during this stage, ensuring a smooth transition of the total frequency regulation power.

[0109] In step S3364, when the speed of the flywheel energy storage module drops to the second preset threshold, it completely exits the frequency regulation response, and the battery energy storage module independently undertakes the subsequent frequency regulation task.

[0110] When the flywheel energy storage module's speed drops to the second preset threshold (usually set to 45% of the rated speed), the system controls the flywheel to completely exit frequency regulation response. The exit process employs a smooth transition strategy, gradually reducing power output as the speed approaches the second threshold until it completely exits when the speed reaches the threshold. At this point, the battery energy storage module has established a stable frequency regulation power output, and its droop coefficient has reached the preset maximum value, enabling it to independently handle the system's frequency regulation task. After exiting frequency regulation, the flywheel enters the charging recovery phase, accelerating its recovery to the rated speed using grid power or renewable energy when the system frequency allows, preparing for the next frequency disturbance.

[0111] Step S3365: By coordinating the judgment of rotational speed and state of charge, the automatic role switching and seamless power transfer between the flywheel energy storage module and the battery energy storage module during the frequency regulation process are realized.

[0112] By coordinating the judgment of rotational speed and state of charge (SOC), the automatic role switching and seamless power transfer between the two energy storage modules during frequency regulation are achieved. The coordinating judgment logic is based on multi-objective optimization principles, simultaneously considering the flywheel's rotational speed, the battery's SOC, the system frequency deviation, and the frequency change trend. When the flywheel's rotational speed decreases while the battery's SOC is sufficient, the system automatically triggers the control transfer process; when the battery's SOC is insufficient but the flywheel still has frequency regulation capability, the switching threshold is adjusted appropriately. The power transfer process employs a master-slave control architecture, using a feedforward-feedback composite control algorithm to ensure continuous power output and avoid power gaps or surges caused by control switching.

[0113] By introducing a speed feedback correction mechanism and establishing a coordinated control strategy for the flywheel and battery, optimized coordination and seamless switching of the hybrid energy storage system during frequency regulation are achieved. This mechanism fully leverages the rapid response advantage of flywheel energy storage and the continuous support capability of battery energy storage, effectively extending equipment lifespan while ensuring system frequency stability. The coordinated control strategy ensures optimal power allocation under different operating conditions, improving the overall operating efficiency and reliability of the hybrid energy storage system and providing more flexible and reliable technical support for power system frequency stability.

[0114] Specifically, step S300 involves constructing a hierarchical nested optimization model oriented towards multiple time scales based on the power system model and the mathematical models of each energy storage module, including: Step S341: Establish an outer-layer optimization model, with the goal of minimizing system net load fluctuation and optimizing energy storage configuration capacity, and determine the capacity and power configuration parameters of the molten salt thermal storage module and the battery energy storage module.

[0115] The outer-layer optimization model is established with the core objectives of minimizing system net load fluctuations and optimizing energy storage capacity. Through analysis of the power system's annual operating data, the optimal capacity and power configuration parameters for molten salt thermal storage modules and battery energy storage modules are determined. Minimizing net load fluctuations is achieved by minimizing the variance between load and renewable energy output, which helps the energy storage system effectively mitigate intraday and seasonal power fluctuations. Optimizing energy storage capacity is achieved by balancing energy storage investment costs and operational benefits, ensuring the economic rationality of the configuration scheme. During the optimization process, the outer-layer model considers key parameters such as the thermal storage capacity and charge / discharge power of the molten salt thermal storage system, and the rated capacity and maximum charge / discharge power of the battery energy storage system. Linear programming or mixed-integer linear programming algorithms are used to solve the problem, ensuring that the obtained configuration scheme meets system regulation requirements while possessing good economic indicators.

[0116] Step S342: Establish an inner-layer optimization model and optimize the frequency regulation power allocation strategy between the flywheel energy storage module and the battery energy storage module with the goal of minimizing the system frequency deviation.

[0117] The inner-layer optimization model focuses on minimizing system frequency deviation. Given the energy storage configuration parameters determined by the outer-layer optimization, it optimizes the frequency regulation power allocation strategy between the flywheel energy storage module and the battery energy storage module. Minimizing frequency deviation is achieved by minimizing the sum of squares or the sum of absolute values ​​of the frequency deviations, requiring the energy storage system to respond quickly and accurately to changes in system frequency. Based on a detailed system dynamic model, the inner-layer model considers the rapid response characteristics of flywheel energy storage and the continuous support capability of battery energy storage. An optimization algorithm determines the optimal power allocation ratio between the two energy storage systems under different frequency disturbance scenarios. The optimization process also considers operational constraints of the energy storage, including the safe rotational speed range of the flywheel, the state of charge limit of the battery, and the power change rate constraint, ensuring the engineering feasibility of the frequency regulation strategy.

[0118] Step S343: Establish a parameter transfer interface between the outer optimization model and the inner optimization model to transfer the energy storage configuration parameters determined by the outer optimization to the inner optimization model as the operating boundary.

[0119] A parameter transfer interface is established between the outer and inner optimization models. This interface is responsible for accurately transferring the energy storage configuration parameters determined by the outer optimization to the inner optimization model as operational boundary conditions. The transferred parameters include key configuration parameters such as the rated capacity and maximum charge / discharge power of the molten salt thermal storage system, and the rated capacity and maximum charge / discharge power of the battery energy storage system. The interface design adopts standardized data formats and communication protocols to ensure the accuracy and real-time performance of parameter transfer. After each outer optimization iteration, the parameter transfer interface automatically updates the boundary conditions of the inner optimization model, providing the latest configuration parameter input for the inner optimization. This mechanism ensures effective coordination between long-term configuration planning and short-term operational optimization.

[0120] Step S344: Establish an operation status feedback interface in the inner optimization model to feed back the energy status and frequency regulation performance indicators of the energy storage module obtained from the inner optimization simulation to the outer optimization model.

[0121] An operational status feedback interface is established within the inner-layer optimization model. This interface is responsible for feeding back the energy state and frequency regulation performance indicators of the energy storage modules derived from the inner-layer optimization simulation to the outer-layer optimization model. The feedback data includes key operational indicators such as the rotational speed trajectory of the flywheel energy storage, the state of charge change curve of the battery energy storage, the system frequency response characteristics, and the frequency regulation contribution of each energy storage unit. The feedback interface extracts characteristic parameters of the operational status through data processing algorithms and converts them into identifiable constraints or objective function correction terms for the outer-layer optimization model. This feedback mechanism enables the outer-layer optimization to dynamically adjust the configuration scheme based on the inner-layer operational results, achieving coordinated optimization of configuration and operation.

[0122] Step S345: Through the parameter transmission interface and the running status feedback interface, a bidirectional data interaction path is constructed between the outer optimization layer and the inner optimization layer, forming the basic architecture of the hierarchical nested optimization model.

[0123] By leveraging the collaborative work of parameter transfer and operational status feedback interfaces, a bidirectional data interaction path is constructed between the outer and inner optimization layers. This path forms the basic architecture of a hierarchical nested optimization model, enabling bidirectional interaction between long-term planning and short-term operation. During optimization iteration, the outer model provides configuration parameters to guide the inner model's operation, while the inner model provides feedback on operational status to correct the outer planning, forming a closed-loop optimization mechanism. The bidirectional data interaction employs asynchronous communication, allowing the outer and inner layers to solve in parallel at different time scales, with interface coordination ensuring consistency in the optimization direction. This architecture effectively solves the problem of disconnect between long-term planning and short-term operation in traditional optimization methods, providing a technical foundation for the global optimization of hybrid energy storage systems.

[0124] Specifically, step S400 uses time-series clustering to select typical scenarios from the operational data, obtains the optimal capacity and power configuration of each energy storage module based on a hierarchical nested optimization model, and establishes a cross-timescale collaborative operation strategy, including: Step S410: Normalize and preprocess the wind power output data, photovoltaic power output data, and load demand data to obtain several time series. Use soft dynamic time warping distance as a similarity criterion to perform cluster analysis on the time series.

[0125] In the selection of typical scenarios, the annual wind power output data, photovoltaic power output data, and load demand data are first preprocessed using normalization to transform the original data of different dimensions into a unified benchmark range. The normalization process employs a min-max normalization method, mapping each data sequence to the [0,1] interval to eliminate the impact of magnitude differences on cluster analysis. The preprocessed data forms multiple complete time series, and soft dynamic time warping distance is used as a similarity criterion for cluster analysis. Soft dynamic time warping distance can effectively handle the scaling and deformation of time series on the time axis, overcoming the limitation of traditional Euclidean distance's sensitivity to the phase of time series, and more accurately capturing the morphological similarity of wind and solar load curves. The clustering algorithm uses an improved k-medoids method, which minimizes the intra-cluster distance through iterative optimization, ensuring that the daily operating curves within the same cluster have highly similar fluctuation characteristics.

[0126] Step S420: Evaluate the clustering quality under different numbers of clusters based on the silhouette coefficient, and determine the optimal number of clusters that can best represent the annual operating characteristics.

[0127] The silhouette coefficient is used to evaluate the clustering quality under different numbers of clusters and determine the optimal number of clusters. The silhouette coefficient assesses the compactness and separation of clustering results by calculating the average distance between each sample point and other sample points in the same cluster, as well as the average distance between each sample point and sample points in other clusters. Specifically, for each possible clustering number k value (typically from 2 to 10), the average silhouette coefficient of all sample points is calculated, and the k value that maximizes the average silhouette coefficient is selected as the optimal number of clusters. The evaluation process also needs to consider the physical meaning of the clustering results to ensure that the selected number of clusters can reasonably represent the typical operational characteristics under different seasons and weather types throughout the year, avoiding insufficient representativeness caused by over-subdivision or over-aggregation.

[0128] Step S430: Complete the clustering based on the optimal number of clusters, select the central daily running curve from each cluster as a typical scenario, and calculate the annual frequency of its occurrence to determine the corresponding scenario weight.

[0129] The final clustering is completed based on the determined optimal number of clusters. From each cluster, a center day's running curve is selected as the typical scenario for that cluster. The selection criterion for the center day is the date with the smallest average distance from the running curves of other days in the same cluster, ensuring that the selected typical scenario best represents the overall characteristics of that cluster. For each cluster, the annual frequency of its included dates is statistically analyzed, and the corresponding scenario weight is calculated. The calculation of scenario weights considers the distribution of running days in different seasons, assigning higher weights to clusters containing more running days, ensuring that frequently occurring running scenarios are prioritized during optimization. Finally, a scenario set containing typical scenario curves and their corresponding weights is formed, providing an input basis for multi-scenario optimization.

[0130] To reduce the computational complexity of the year-round operational data, this embodiment employs a time series clustering algorithm based on soft-DTW (Dynamic Time Warping) distance metric to extract several representative typical daily scenarios from the year-round wind power, photovoltaic, and load data. The normalized output curves of renewable energy and load for one typical day are shown below. Figure 2 As shown, the curve illustrates the variation characteristics of wind power, photovoltaic, electricity load, heating load, and cooling load within a typical day, providing a crucial data foundation for the multi-scenario set used as input for the hierarchical nested optimization model.

[0131] Step S440: Integrate all typical scenarios and their scenario weights to construct a multi-scenario set for input to the hierarchical nested optimization model. Input the multi-scenario set into the hierarchical nested optimization model and use a mixed integer linear programming algorithm and a hierarchical iterative algorithm to solve the problem together, and simultaneously optimize the energy storage configuration and operation strategy under all typical scenarios.

[0132] A multi-scenario set is constructed by integrating all typical scenarios and their weights for input to a hierarchical nested optimization model. This set includes typical operating curves under different seasons and weather conditions, with each scenario accompanied by its annual occurrence probability weight. This multi-scenario set is then input into the hierarchical nested optimization model, which employs a combined mixed-integer linear programming algorithm and a hierarchical iterative algorithm for solution. The mixed-integer linear programming algorithm handles discrete decision variables in energy storage configuration (such as equipment deployment and operating status), while the hierarchical iterative algorithm coordinates the solution processes of the outer and inner optimization layers. The optimization process simultaneously considers the system operating requirements under all typical scenarios, ensuring that the resulting energy storage configuration scheme can adapt to various possible operating conditions throughout the year.

[0133] Step S450: Establish a bidirectional iterative convergence criterion for outer-layer optimization and inner-layer optimization. Solve the outer-layer optimization model and the inner-layer optimization model alternately, and determine whether the optimization result meets the preset accuracy requirements based on the convergence criterion.

[0134] A bidirectional iterative convergence criterion is established for outer-layer and inner-layer optimization, achieving dynamic equilibrium between the two optimization layers through alternating solutions. The convergence criterion includes three indicators: the rate of change of the objective function, the degree of constraint violation, and the stability of state variables. Specifically, the annualized total cost change rate of the outer-layer optimization must be less than a preset threshold, the frequency deviation index of the inner-layer optimization must meet stability requirements, and the energy storage operation state must remain stable during continuous iterations. In the alternating solution process, the outer-layer configuration parameters are first fixed to solve the inner-layer operational optimization. Then, the outer-layer configuration optimization is corrected based on the feedback from the inner-layer operational results. This process is repeated until all convergence conditions are met. This process ensures a dynamic equilibrium between long-term configuration planning and short-term operational strategies.

[0135] Step S460: When the convergence criterion is not met, update the constraints of the outer optimization model based on the running status data of the inner optimization feedback, and recalculate the outer optimization.

[0136] When the optimization result does not meet the convergence criterion, the constraints of the outer optimization model are updated based on the operational status data fed back from the inner optimization, and the outer optimization calculation is performed again. The constraints that need to be updated mainly include the operational boundaries of the energy storage device, such as the reasonable range of battery state of charge, the safe range of flywheel speed, and the energy balance limit of molten salt thermal storage. The update process considers the operational bottlenecks found in the inner optimization, and appropriately adjusts the feasible domain range of the configuration parameters to avoid configuration schemes that are theoretically optimal but difficult to operate in practice. During the re-optimization process, the outer model resolves the energy storage configuration parameters under the corrected constraints, and the inner model re-optimizes the operation strategy based on the new configuration parameters, forming a closed-loop optimization cycle.

[0137] Step S470: When the convergence criterion is met, the solution results are weighted and integrated according to the scenario weights to output the unified optimal capacity configuration, power configuration and collaborative operation strategy of all energy storage modules under the whole time scale.

[0138] When the optimization results satisfy all convergence criteria, the solution results are weighted and integrated according to the weights of each typical scenario to output the final optimized configuration scheme. The weighted integration process averages the optimization results for each typical scenario based on its annual occurrence probability, resulting in a unified optimal configuration that comprehensively considers various operating conditions. The output results include the rated capacity and maximum power of the molten salt thermal storage module, the rated capacity and maximum power of the battery energy storage module, the rated capacity and maximum power of the flywheel energy storage module, and the collaborative operation strategies of each energy storage unit at different time scales. The operation strategies clearly define the functional positioning and coordination mechanisms of each energy storage unit in day-ahead planning, intraday scheduling, and real-time control, ensuring the efficient and stable operation of the system across all time scales.

[0139] Simulation results show that the proposed method can effectively improve the utilization rate of new energy power generation, reduce wind and solar curtailment, significantly reduce the annualized operating cost of the system, and improve the frequency stability of the power grid. Verification in a typical regional power grid shows that, compared with the traditional single energy storage optimization configuration method, this invention can increase the utilization rate of new energy by about 8%, reduce the system operating cost by about 10%, and reduce the frequency fluctuation amplitude by more than 40%. The frequency fluctuation comparison curves of the power system with and without energy storage participating in frequency regulation are shown below. Figure 3 As shown, this fully demonstrates the significant effect of the present invention in improving system frequency stability.

[0140] Accordingly, a second aspect of the present invention provides an electronic device, including: at least one processor and a memory connected to the at least one processor. The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, cause the at least one processor to perform the aforementioned hybrid energy storage optimization configuration method based on multi-timescale control requirements.

[0141] Accordingly, a third aspect of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described hybrid energy storage optimization configuration method based on multi-timescale control requirements.

[0142] The embodiments of the present invention aim to protect the above-mentioned hybrid energy storage optimization configuration method based on multi-timescale regulation requirements, which has the following effects: 1. By scientifically dividing the system regulation demand into three time scales—long, medium, and short—and configuring molten salt thermal storage, battery energy storage, and flywheel energy storage modules respectively, a multi-level collaborative energy storage regulation structure precisely matched with the time scale is constructed. Based on the functional allocation mechanism of physical characteristics, the energy storage resources with different characteristics can leverage their strengths and avoid their weaknesses, thus significantly improving the overall flexibility and all-time-domain coordinated control capability of the high-proportion renewable energy power system from the system architecture level, effectively coping with complex power fluctuations from the second level to the seasonal level. 2. The constructed hierarchical nested optimization model for multiple time scales, through the outer layer optimization focusing on minimizing system net load fluctuations and optimizing energy storage capacity, and the inner layer optimization dedicated to minimizing system frequency deviation, successfully solves the problems of target conflict and information disconnect between long-term planning and short-term operation by combining a two-way iterative mechanism. This enables the energy storage system's capacity configuration scheme and real-time operation strategy to be optimized in a coordinated manner, maximizing the investment and operational benefits of energy storage assets while ensuring the stability of the system's dynamic frequency. 3. The energy storage mathematical model deeply integrates refined operational constraints and lifespan degradation models, especially the charge / discharge depth-temperature dual-factor lifespan degradation model for battery energy storage and the electromechanical coupling safety constraint model for flywheel energy storage. This ensures that the optimized configuration results not only meet the system's power and energy requirements but also accurately reflect the performance changes and lifespan losses of energy storage devices during actual operation. Consequently, at the technical solution level, this directly improves the economic efficiency and engineering feasibility of the configuration results, effectively extends the service life of key equipment, and reduces the total lifespan cost.

[0143] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0144] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0145] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0146] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0147] 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 scope of protection of the claims of the present invention.

Claims

1. A method for optimizing the configuration of hybrid energy storage based on multi-timescale regulation needs, characterized in that, Includes the following steps: Collect power system operation data, and based on the operation data and a preset high-proportion renewable energy power system model, obtain power fluctuations and frequency deviations faced by the power system in a short time scale; Based on the power fluctuations and frequency deviations, the control requirements of the power system are divided according to different time scales, and different energy storage modules are configured accordingly to form a multi-level collaborative energy storage control structure that matches the time scale. Based on the power system model and the mathematical models of each energy storage module, a hierarchical nested optimization model oriented towards multiple time scales is constructed. A time series clustering method is used to select typical scenarios from the operational data, and the optimal capacity and power configuration of each energy storage module, as well as the collaborative operation strategy across time scales, are obtained based on the hierarchical nested optimization model. The power fluctuations include power fluctuations at different time scales caused by renewable energy and load fluctuations, and the mathematical models of each energy storage module include mathematical models characterizing the power, capacity and energy state of each energy storage module.

2. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 1, characterized in that, The method of dividing the power system regulation requirements according to different time scales based on the power fluctuations and frequency deviations includes: The power fluctuation spectrum contains fluctuation components with different time constants. The control requirements corresponding to the fluctuation components at the second to minute level are divided into short time scales, the control requirements corresponding to the fluctuation components at the hour to day level are divided into medium time scales, and the control requirements corresponding to the fluctuation components at the multi-day level or above are divided into long time scales.

3. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 2, characterized in that, The fluctuation components with different time constants included in the spectrum of the power fluctuation are as follows: Wavelet packet decomposition was performed on the time series of the power fluctuations to obtain multiple different frequency band sub-sequences; Calculate the energy contribution rate of each frequency band subsequence in the total power fluctuation sequence; Based on the preset energy contribution rate threshold, the characteristic frequency bands that affect the system power balance are selected; The characteristic frequency band is mapped to fluctuation component categories at the second to minute, hour to day, and multiple days or more, according to the time period corresponding to its center frequency.

4. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 3, characterized in that, The different energy storage modules include molten salt thermal energy storage modules, battery energy storage modules, and flywheel energy storage modules; the mathematical model of the molten salt thermal energy storage module is constructed through the following steps: A dual-tank structure model including a high-temperature storage tank and a low-temperature storage tank is established. The high-temperature storage tank is defined to store high-temperature molten salt after conversion by an electric heater, and the low-temperature storage tank stores low-temperature molten salt after energy release. An energy dynamic equation is constructed for the molten salt thermal storage module. The energy dynamic equation describes the energy transfer process between the high-temperature storage tank and the low-temperature storage tank, and takes into account the thermal inertia characteristics of the molten salt circulation system. The energy conversion efficiency parameter of the charging and discharging process is introduced into the energy dynamic equation to characterize the charging efficiency of electrical energy to thermal energy and the discharging efficiency of thermal energy to electrical energy, respectively. The self-discharge rate parameter is introduced into the energy dynamic equation to characterize the heat loss of the molten salt thermal storage system during static storage. A first power constraint model for the molten salt thermal storage module is established, and the constraints on its maximum charging power and maximum discharging power are defined. A first capacity constraint model is established for the molten salt thermal storage module, defining the minimum and maximum thermal storage capacity limits for its high-temperature and low-temperature storage tanks; By integrating the energy dynamic equation, the first power constraint model, and the first capacity constraint model, a mathematical model of the molten salt thermal storage module is obtained.

5. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 4, characterized in that, The mathematical model of the battery energy storage module is constructed through the following steps: A state of charge (SOC) change model for the battery energy storage module is established. The SOC change model calculates the SOC value at the next moment using the current charging and discharging power and the corresponding charging and discharging efficiency. Define the safe operating range of the state of charge of the battery energy storage module, set the upper and lower thresholds of its state of charge, and automatically limit its charging and discharging behavior when the state of charge reaches the corresponding threshold. A second power constraint model for the battery energy storage module is constructed, stipulating that its maximum charging power and maximum discharging power shall not exceed its rated power value; A second capacity constraint model for the battery energy storage module is established to ensure that its actual stored energy is always within its rated capacity range. Based on the historical operating data of the battery energy storage module, the charge and discharge depth of each charge and discharge cycle is calculated. According to the relationship between the charge and discharge depth and the preset empirical decay index, the equivalent cycle life loss caused by the charge and discharge cycle is calculated. The total equivalent cycle life loss of the battery energy storage module in the optimization period is accumulated and the total equivalent cycle life loss is used as a key parameter for evaluating its operating status. By integrating the state of charge change model, the second power constraint model, the second capacity constraint model, and the equivalent cycle life loss assessment process based on charge-discharge depth, a mathematical model of the battery energy storage module is obtained.

6. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 5, characterized in that, The equivalent cycle life loss assessment process based on charge / discharge depth also includes: Real-time monitoring of internal temperature changes in the battery energy storage module during charging and discharging; An accelerated life decay model considering both charge / discharge depth and operating temperature is established. This accelerated life decay model characterizes the accelerated decay effect of high-temperature operation on battery cycle life under the same charge / discharge depth. When calculating the equivalent cycle life loss caused by a single charge-discharge cycle, a life decay acceleration factor based on the current operating temperature is introduced simultaneously to obtain the total equivalent cycle life loss combined with the temperature effect.

7. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 4, characterized in that, The mathematical model of the flywheel energy storage module is constructed through the following steps: A model is established to show the relationship between the rotor kinetic energy and the stored energy of the flywheel energy storage module. The model calculates the instantaneous stored energy through the quantitative relationship between the rotor moment of inertia and the square of the real-time rotational speed. Construct a third power constraint model for the flywheel energy storage module, define its maximum charging power and maximum discharging power limits, and ensure that its power change rate does not exceed the maximum ramp rate allowed by the device; Establish the safe operating range of the flywheel energy storage module's rotational speed, and set an upper and lower threshold for its rotor speed. The lower threshold is used to ensure that the rotor maintains the necessary kinetic energy reserve. An energy state calculation model for the flywheel energy storage module is established. The energy state calculation model determines the ratio of its current available energy to its maximum stored energy by the ratio of real-time rotational speed to rated rotational speed. A droop control response model for the flywheel energy storage module is constructed, a linear adjustment relationship between the system frequency deviation and its output power is established, and the adjustable range of its droop coefficient is defined. A speed feedback correction mechanism is introduced into the droop control response model. When the real-time speed approaches the lower limit threshold, the output power is automatically reduced to ensure that the rotor speed is always maintained within the safe operating range. By integrating the rotor kinetic energy and stored energy relationship model, the third power constraint model, the speed safe operating range constraint, the energy state calculation model, and the droop control response model, a mathematical model of the flywheel energy storage module is obtained.

8. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 7, characterized in that, The introduction of a speed feedback correction mechanism into the droop control response model includes: Establish a dynamic mapping relationship between the speed status of the flywheel energy storage module and the droop coefficient. When the real-time speed is close to the rated speed, the standard droop coefficient is adopted, and when the real-time speed is close to the lower limit threshold, the droop coefficient is automatically reduced. In the initial stage of frequency disturbance, the flywheel energy storage module is controlled to respond with the maximum available droop factor, while the state of charge of the battery energy storage module is monitored. When the speed of the flywheel energy storage module drops to the first preset threshold, the frequency regulation function of the battery energy storage module is activated, and its droop coefficient increases linearly from zero. When the speed of the flywheel energy storage module drops to the second preset threshold, it completely exits the frequency regulation response, and the battery energy storage module independently undertakes the subsequent frequency regulation task. By coordinating the determination of the rotational speed and state of charge, the automatic role switching and seamless power transfer between the flywheel energy storage module and the battery energy storage module are achieved during the frequency regulation process.

9. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to any one of claims 1-8, characterized in that, Based on the power system model and the mathematical models of each energy storage module, a hierarchical nested optimization model oriented towards multiple time scales is constructed, including: An outer-layer optimization model is established with the goal of minimizing system net load fluctuation and optimizing energy storage configuration capacity to determine the capacity and power configuration parameters of the molten salt thermal storage module and the battery energy storage module. An inner-layer optimization model is established to optimize the frequency regulation power allocation strategy between the flywheel energy storage module and the battery energy storage module with the goal of minimizing the system frequency deviation. A parameter transfer interface is established between the outer optimization model and the inner optimization model to transfer the energy storage configuration parameters determined by the outer optimization to the inner optimization model as the operating boundary. An operation status feedback interface is established in the inner optimization model to feed back the energy state and frequency regulation performance index of the energy storage module obtained from the inner optimization simulation to the outer optimization model. Through the parameter transmission interface and the running status feedback interface, a bidirectional data interaction path is constructed between the outer optimization layer and the inner optimization layer, forming the basic architecture of the hierarchical nested optimization model.

10. The method for optimal configuration of hybrid energy storage based on multi-timescale regulation requirements according to claim 9, characterized in that, The method employs time-series clustering to select typical scenarios from the operational data, and obtains the optimal capacity and power configuration for each energy storage module based on the hierarchical nested optimization model, as well as a cross-timescale collaborative operation strategy, including: Normalization preprocessing was performed on wind power output data, photovoltaic power output data and load demand data to obtain several time series. Soft dynamic time warping distance was used as a similarity criterion to perform cluster analysis on the time series. Based on the silhouette coefficient, the clustering quality under different numbers of clusters is evaluated, and the optimal number of clusters that can best represent the annual operating characteristics is determined. Clustering is completed based on the optimal number of clusters. The daily running curve of the center of each cluster is selected as a typical scenario, and its annual frequency of occurrence is counted to calculate the corresponding scenario weight. Integrate all the typical scenarios and their scenario weights to construct a multi-scenario set for input to the hierarchical nested optimization model. Input the multi-scenario set into the hierarchical nested optimization model and solve it together using a mixed integer linear programming algorithm and a hierarchical iterative algorithm to simultaneously optimize the energy storage configuration and operation strategy under all typical scenarios. A bidirectional iterative convergence criterion for outer-layer optimization and inner-layer optimization is established. The outer-layer optimization model and the inner-layer optimization model are solved alternately, and the optimization result is judged to meet the preset accuracy requirements based on the convergence criterion. When the convergence criterion is not met, the constraints of the outer optimization model are updated based on the running status data of the inner optimization feedback, and the outer optimization calculation is performed again. When the convergence criterion is met, the solution results are weighted and integrated according to the scenario weights to output the unified optimal capacity configuration, power configuration and collaborative operation strategy of all energy storage modules across the entire time scale.

Images