Wind storage combined frequency modulation method and device based on wind power participation coefficient and energy storage charging and discharging balance

Abstract

The invention discloses a wind storage combined frequency modulation method and device based on a wind power participation coefficient and energy storage charge-discharge balance. The method comprises the following steps: firstly, calculating a kinetic energy margin and a power margin of a wind turbine generator in real time, generating a wind power participation coefficient by adopting LOGISTIC mapping, calculating a normalized SOC of an energy storage unit according to a grouping mode to obtain a balance degree, dynamically adjusting an energy storage droop coefficient based on the balance degree, and triggering charge and discharge role interchange when a single SOC approaches an upper limit and a lower limit. And then a combined state space model including wind power, energy storage, thermal power and controllable load is further constructed, model prediction control MPC or quadratic programming is adopted in a prediction time domain, constraint optimization distribution is performed on active power output of various resources, and finally a control instruction is issued in a rolling time domain mode to realize real-time closed-loop adjustment. According to the method, the fan operation safety and the energy storage service life are both considered, the maximum frequency deviation of the system is remarkably reduced by coordinating dynamic participation of wind power and energy storage, frequency oscillation is restrained, energy storage deep circulation is reduced, and good engineering implementation and remarkable economic benefits are achieved.

Description

Technical Field

[0001] This invention relates to power system frequency control and new energy grid-connected dispatch technology, specifically to a wind-storage joint frequency regulation method and device based on wind power participation coefficient and energy storage charge-discharge balance. Background Technology

[0002] With the large-scale integration of renewable energy sources, wind power is playing an increasingly important role in the power system. Wind power has advantages such as being clean and abundant in resources, but its grid connection and dispatch, as well as system frequency stability, face the following typical problems:

[0003] 1. Weakened physical coupling between wind power and grid frequency. Modern wind turbines mostly employ variable-speed generators and power electronic converters, achieving electrical decoupling between rotor speed and grid frequency. This reduces the natural inertia contribution of wind power, making the system more vulnerable to frequency response disturbances. As wind power penetration increases, the overall system inertia decreases, and the frequency drop rate accelerates, making traditional frequency regulation systems, primarily based on thermal power, unable to independently cope with large disturbances.

[0004] 2. Dual constraints of wind turbine kinetic energy and converter capacity. Wind turbines typically rely on rotor kinetic energy release or virtual inertia and droop control for frequency support. However, rotor kinetic energy reserves are limited, and frequent or excessive release can lead to increased mechanical stress, deviation of the rotational speed from the safe range, and even shutdown or equipment damage. At the same time, the upper limit of the converter's active power also limits the wind turbine's additional frequency regulation capability at high wind speeds.

[0005] 3. Lifespan and Charge / Discharge Balance Issues of Energy Storage Systems. Battery-based energy storage offers rapid response and can compensate for the limitations of wind power in frequency regulation. However, the charge-discharge cycles of energy storage lead to aging, and over-discharging or over-charging of individual cells shortens lifespan and increases maintenance costs. Furthermore, without a reasonable grouping and balancing mechanism, the cumulative charge-discharge imbalance after frequency events can cause long-term performance degradation. Existing methods often treat energy storage as "infinite" or simply constrain it from the perspective of SOC (State of Charge), lacking engineering strategies that consider both inter-group balance and SOH (State of Health) adaptation.

[0006] 4. Limitations of Existing Joint Frequency Regulation Strategies. Research on wind-storage joint frequency regulation mainly focuses on two directions: utilizing the virtual inertia / droop of wind turbines in conjunction with energy storage to undertake primary frequency regulation; and employing complex model predictive control (MPC) or hierarchical optimization for joint scheduling of wind power and energy storage. The first type of method is effective in reducing frequency peaks but is prone to excessive consumption of wind turbine power; the second type of method, although having good optimization performance, is usually computationally complex, highly dependent on models and predictions, and often requires solving real-time and equipment interoperability issues when implemented in engineering. In addition, many studies have neglected the coordination relationship between wind power participation and energy storage charge-discharge balance, resulting in uneven allocation of wind and storage roles, accelerated energy storage degradation, or decreased frequency regulation effectiveness.

[0007] Based on the above technical background, the industry urgently needs a frequency modulation scheme that can meet the following requirements:

[0008] (i) It can quantify in real time the safe and available frequency regulation capability of each wind turbine at the current operating point;

[0009] (ii) It can perform charge and discharge scheduling of energy storage groups in a balanced drive manner to avoid excessive cycling of individual cells and take into account both SOC and SOH;

[0010] (iii) It can incorporate the real-time capabilities of wind power and energy storage into a unified state-space model for engineering real-time optimization.

[0011] (iv) The method should include engineering protection measures such as dead zone, reversal and priority rules to improve practicality and reliability. Summary of the Invention

[0012] Based on the above problems, this invention proposes a wind-storage joint frequency regulation method and device based on the wind power participation coefficient and the energy storage charging and discharging balance. This method can quantify the wind turbine frequency regulation capability in real time, maintain the energy storage balance, and adapt to engineering optimization solutions.

[0013] The present invention adopts the following technical solution:

[0014] A wind-storage joint frequency regulation method based on wind power participation coefficient and energy storage charge-discharge balance is used to coordinate and control wind turbines, energy storage, and other frequency regulation resources in the power system in real time to maintain system frequency stability. The method includes the following steps:

[0015] Real-time acquisition of operating status data for each wind turbine, including rotor speed, current output power, and converter capacity limits, is performed. Based on this operating status data, the kinetic energy margin and power margin of the wind turbine are calculated. The available frequency regulation capability of the wind turbine is determined based on these kinetic energy margin and power margin, and then mapped to the wind power participation factor using the LOGISTIC function. ;

[0016] The energy storage system is configured as a group containing multiple energy storage units. The state of charge (SOC) of each energy storage unit is collected in real time, and the normalized balance of the energy storage group is calculated. Based on the aforementioned balance Dynamically adjust the droop coefficient of this energy storage group And when the SOC of any energy storage unit is detected to reach the preset upper or lower limit, the strategy of switching the charging and discharging roles of each unit within the energy storage group is triggered.

[0017] Construct a joint state-space model that includes the dynamics of wind power, energy storage, thermal power, and controllable loads, and use the wind power participation coefficient as the basis for this model. and droop coefficient As model input or constraint;

[0018] Based on the aforementioned joint state-space model and the real-time acquired wind power participation coefficient and droop coefficient Within the preset prediction time domain, the optimal control sequence is solved using model predictive control or quadratic programming methods to minimize the comprehensive objective function within the prediction time domain.

[0019] The control command at the first moment of the optimal control sequence obtained by the solution is sent to the corresponding actuators of wind turbines, energy storage, thermal power and controllable loads to implement real-time adjustment, and the feedback status data of each actuator is collected to complete the closed-loop control.

[0020] Furthermore, the kinetic energy margin of the wind turbine unit Defined as:

[0021] ;

[0022] in, This represents the current rotor kinetic energy of the wind turbine. For rotational inertia, The rotor speed, and These are the maximum and minimum available kinetic energies calculated based on the safe upper and lower limits of the fan speed.

[0023] Furthermore, the power margin of the wind turbine unit Defined as:

[0024] ;

[0025] in, This is the current output power. , This represents the upper and lower power limits of the wind turbine under the constraints of the converter and control.

[0026] Furthermore, the wind power participation factor Determined by the following formula:

[0027] ;

[0028] in, , These are real-valued parameters used for modulation sensitivity and threshold.

[0029] Furthermore, the energy storage group balance The calculation is as follows:

[0030] ;

[0031] in For energy storage units The state of charge is used to achieve a normalized measure of balance; for groups of more than two units, a [specific term] can be defined. This is the variance or range index of the normalized SOC for each unit.

[0032] Furthermore, the balance-based Dynamically adjust the droop coefficient of this energy storage group Specifically, this includes: the energy storage droop coefficient. According to balance Dynamic adjustments are made using segmented or continuous mapping relationships; when Less than a certain lower threshold When this occurs, energy storage groups prioritize entering charging mode or increase the charging droop coefficient; when Greater than a certain upper limit threshold When this occurs, the energy storage group prioritizes entering discharge mode or increases the discharge droop coefficient; when When in the middle range, energy storage participates in frequency regulation with a conventional droop coefficient.

[0033] Furthermore, the state variables of the joint state-space model include system frequency, wind turbine output power deviation, energy storage output power deviation, energy storage SOC, thermal power unit output power deviation, and tie-line power deviation+, and the discrete-time state equations are obtained through linearization / discretization:

[0034]

[0035] It is the state variable at time k+1; It is the state variable at time k; It is the input variable at time k; It is the perturbation variable at time k; It is a state variable matrix; It is the input variable matrix; It is the perturbation variable matrix.

[0036] Furthermore, the objective function of the MPC includes a frequency deviation squared term, a tie-line power deviation term, and a control input / change cost term, denoted as:

[0037]

[0038] It is the objective function; The time index variable is N; N is the total duration of interest. yes Frequency deviation at any given moment; yes Line power deviation at any given time; yes Deviation of input variables at any given time; It is the weighting coefficient of the frequency deviation term; It is the weighting coefficient for the line power deviation term; These are the weighting coefficients for the input variable deviation term;

[0039] Furthermore, the optimization process is subject to constraints such as active power upper and lower limits, ramp rate constraints, energy storage SOC constraints, and frequency regulation dead zone constraints. For prediction in the time domain, the range is 10 to 50; These are weighting coefficients, which are used in simulation / engineering calibration to balance frequency performance with control costs.

[0040] A wind-storage joint frequency regulation device based on wind power participation factor and energy storage charge-discharge balance, used to implement the above method, the system comprising:

[0041] The wind power regulation module is used to acquire the operating status data of each wind turbine in real time, including the turbine rotor speed, current output power and converter capacity limit, and calculate the kinetic energy margin and power margin of the wind turbine based on the operating status data; determine the available frequency regulation capability of the wind turbine based on the kinetic energy margin and power margin, and map the frequency regulation capability to the wind power participation factor Kw through the LOGISTIC function;

[0042] The energy storage management module is used to configure the energy storage system into a group containing multiple energy storage units, collect the state of charge (SOC) of each energy storage unit in real time, and calculate the normalized balance degree (B) of the energy storage group; dynamically adjust the droop coefficient (Kes) of the energy storage group based on the balance degree (B), and trigger the switching strategy of charging and discharging roles of each unit within the energy storage group when the SOC of any energy storage unit reaches a preset upper or lower limit.

[0043] The joint modeling module is used to construct a joint state-space model that includes the dynamics of wind power, energy storage, thermal power and controllable load, and uses the wind power participation coefficient Kw and droop coefficient Kes as model inputs or constraints.

[0044] The optimization scheduling module is used to solve the optimal control sequence within the preset prediction time domain by using model predictive control or quadratic programming methods based on the joint state space model and the wind power participation coefficient Kw and droop coefficient Kes obtained in real time, so as to minimize the comprehensive objective function within the prediction time domain.

[0045] The execution and communication module is used to send the control command of the first moment in the optimal control sequence obtained by solving to the corresponding wind turbine, energy storage, thermal power and controllable load actuators to implement real-time adjustment, and to collect the feedback status data of each actuator to complete closed-loop control.

[0046] Furthermore, the wind power regulation module is specifically used to obtain the kinetic energy margin. and the power margin The minimum value is used as an indicator of the wind turbine's available frequency regulation capability. and the wind power participation coefficient This is used to adjust the droop gain and active response upper limit of the fan when frequency deviation occurs, wherein the active response upper limit satisfies .

[0047] The technical solution provided by this invention has the following significant beneficial effects:

[0048] 1. Improved frequency regulation performance and system frequency stability: By calculating the kinetic and power margins of wind turbines in real time and generating adaptive wind power participation coefficients using the LOGISTIC mapping function, precise quantification and smooth adjustment of wind power frequency regulation capabilities are achieved. This enables wind power to participate in primary frequency regulation dynamically and optimally without exceeding its own safety constraints, complementing resources such as energy storage, thereby effectively reducing the maximum frequency deviation of the system, accelerating frequency recovery speed, and suppressing frequency oscillations.

[0049] 2. Effectively Extends the Lifespan of Energy Storage Systems: By grouping energy storage units for management and calculating a normalized balance based on the SOC of each unit within a group, the droop coefficient of the energy storage group is dynamically adjusted. Furthermore, when the SOC of a single unit exceeds its limit, a switching of charging and discharging roles is triggered. This mechanism proactively maintains the charge-discharge balance of each energy storage unit within the group, preventing overcharging or over-discharging of individual units. This significantly reduces the number of deep charge-discharge cycles, thereby delaying battery aging and improving the overall economy and lifespan of the energy storage system.

[0050] 3. Enhanced System Coordination Control and Engineering Practicality: By constructing a unified joint state-space model encompassing multiple types of frequency regulation resources and employing Model Predictive Control (MPC) or quadratic programming for rolling optimization, real-time coordination and optimal power allocation among wind power, energy storage, thermal power, and controllable loads are achieved. This method uses the aforementioned wind power participation coefficient and energy storage balance as key constraints or inputs, ensuring that the optimization results closely align with the actual operating state and safety boundaries of the equipment. Furthermore, by incorporating engineering protection logic such as dead zone, ramp-up limits, and commutation cooling, the reliability, safety, and good engineering feasibility of the control strategy are guaranteed. Attached Figure Description

[0051] Figure 1 This is a flowchart of a wind-storage joint frequency regulation method based on the wind power participation coefficient and the energy storage charge-discharge balance according to an embodiment of the present invention;

[0052] Figure 2 This is a schematic diagram of the fan curves (fan speed - power margin and kinetic energy margin) according to an embodiment of the present invention;

[0053] Figure 3 This is a schematic diagram of the variable droop coefficient of the dual energy storage system (DESS) according to an embodiment of the present invention;

[0054] Figure 4 This is a schematic diagram of the DESS droop coefficient adjustment factor in an embodiment of the present invention;

[0055] Figure 5 This is a schematic diagram of the flexible power distribution control framework of the wind storage system according to an embodiment of the present invention;

[0056] Figure 6 The LOGISTIC curve variation characteristics are shown in the embodiments of the present invention.

[0057] Figure 7 This is a speed-participation curve for an embodiment of the present invention;

[0058] Figure 8 The frequency response results of the system under different load disturbances in the embodiments of the present invention are shown.

[0059] Figure 9 The frequency adjustment results under different initial conditions in the embodiments of the present invention;

[0060] Figure 10 This is a schematic diagram of the SOC curve and balance change of DESS according to an embodiment of the present invention;

[0061] Figure 11 This is a schematic diagram comparing the embodiments of the present invention with the constant coefficient method;

[0062] Figure 12 Timing diagram of the working state of DESS in an embodiment of the present invention. Detailed Implementation

[0063] The present invention will now be further described based on preferred embodiments and with reference to the accompanying drawings.

[0064] In the description of the embodiments of the present invention, it should be noted that if there are indications of orientation or positional relationship such as "up", "down", "inner", "outer", etc., they are based on the orientation or positional relationship shown in the accompanying drawings, or are the orientation or positional relationship that is usually placed when the product of the embodiments of the present invention is used. They are only used to facilitate the description of the present invention and simplify the explanation, and do not constitute a limitation of the present invention.

[0065] Furthermore, to distinguish different modules or units, the terms "first" and "second" are used in this specification. These terms are only used to distinguish different modules or steps and do not indicate any limitation regarding order or importance. Those skilled in the art should understand that the above designations can be adjusted according to specific needs without affecting the scope of protection of this invention.

[0066] Unless otherwise expressly defined, the terms "set," "connect," "link," etc., used in this specification should be interpreted broadly, and may refer to fixed connections, detachable connections, integrally formed connections, indirect connections through an intermediate medium, or electrical connections or data connectivity between two components. Those skilled in the art can understand the specific meaning of each term in conjunction with the specific technical background.

[0067] This invention provides a wind-storage joint frequency regulation method based on wind power participation factor and energy storage charge-discharge balance. The system includes at least one wind power region and at least one energy storage subsystem. The wind power region includes multiple wind turbines, and the energy storage region includes one or more battery energy storage units. The system may further include other frequency regulation resources such as thermal power and controllable loads, and joint frequency regulation is achieved through a unified state-space model and frequency regulation control topology.

[0068] In some embodiments, such as Figure 1 As shown, the method includes the following steps:

[0069] Step 100: Establish a joint state-space model of wind power and energy storage, and construct a quantitative model of wind power regulation capacity and an energy storage balance model.

[0070] Step 200: Establish a wind-storage joint frequency regulation control topology based on the model, and determine the frequency regulation participation strategy, including participation coefficient, droop coefficient and priority rules.

[0071] Step 300: Generate adjustment commands for wind turbines and energy storage based on signal sampling and optimization results, and execute real-time grid frequency adjustment.

[0072] The specific implementation methods of steps 100, 200, and 300 are described in detail below.

[0073] <Step 100>

[0074] Multi-source power systems, under conditions of high wind power integration, exhibit complex characteristics of low inertia, strong coupling, and multi-timescale dynamics. Traditional LFC models are insufficient to describe the true capabilities of wind turbines and energy storage in primary frequency regulation. Therefore, this step requires the establishment of: a real-time quantitative model of wind power regulation capability to constrain the maximum participating power of wind turbines; a model of energy storage group balance and availability to describe the impact of SOC, SOH, and balance state on frequency regulation capability; and a joint state-space model to provide a unified prediction framework for subsequent real-time optimization and scheduling.

[0075] (1) Quantitative model of wind power regulation capacity:

[0076] Modern wind turbines mostly employ variable speed constant frequency structures. The frequency regulation capability of a wind turbine is primarily limited by two factors: the amount of kinetic energy that can be released from the rotor and the available power margin of the converter. To ensure the safety of the wind turbine during frequency regulation, this invention employs kinetic energy margin and power margin (such as...) Figure 2 As shown, the upper limit for wind turbine participation in frequency regulation is defined by taking the minimum value.

[0077] In one embodiment, the kinetic energy of the wind turbine rotor can be expressed as:

[0078]

[0079] in, This is the equivalent rotational inertia of the fan. ω is the rotor angular velocity.

[0080] The margin of kinetic energy that the wind turbine can safely release is denoted as:

[0081]

[0082] in, This is the minimum safe operating speed allowed for the wind turbine. Running below this value may cause aerodynamic instability of the blades or prevent normal grid connection.

[0083] When the shunting frequency of a wind turbine drops, it typically needs to increase its active power output, but the extent of this increase is limited by the converter capacity and the turbine's current operating point. In this invention, the adjustable power margin is defined as:

[0084]

[0085] in, This represents the maximum power output under the current wind speed conditions. This represents the current output of the wind turbine.

[0086] Based on safety principles, the maximum power adjustment that can participate in frequency modulation is defined as:

[0087]

[0088] This design ensures that the wind turbine will not exceed its mechanical safety or converter limits due to participation in frequency regulation.

[0089] To achieve smooth control and avoid abrupt changes, this invention employs an S-type logistic mapping function to map capability indicators. Mapped to frequency modulation participation factor:

[0090]

[0091] in, This is a sensitivity parameter used to control the steepness of the curve; The center point parameter is used to control the transition position from "low" to "high" in terms of participation level.

[0092] In various embodiments of the present invention, parameters and It can be determined offline based on the wind turbine model, wind speed prediction data, and frequency regulation requirements.

[0093] The LOGISTIC function can automatically reduce participation when the turbine capacity is low and increase response efficiency when the capacity is high, thus balancing safety and frequency regulation performance. Figure 6 The mapping characteristics of the participation coefficients under different parameters are shown.

[0094] (2) Energy Storage Grouping and Balance Model (DESS)

[0095] like Figure 10 and Figure 12 As shown, to improve the long-term lifespan and operating efficiency of energy storage, this invention preferably adopts a dual energy storage unit structure (DESS), where each energy storage group includes two energy storage units with the same or similar capacity (denoted as ES-1 and ES-2), which alternately undertake discharge and charge tasks during operation. To quantify the degree of imbalance within the group and adjust the control strategy accordingly, this invention proposes the following normalized balance definition and control logic.

[0096] Let the first The state of charge of each energy storage unit is (Expressed as a decimal or percentage), the allowable range for energy storage units is... The normalized charge ratio (dimensionless) is defined as:

[0097]

[0098] in .

[0099] Taking a two-unit group as an example, the balance within the group is defined. The difference between the normalized SOC of the two units:

[0100]

[0101] when When, it indicates that the two units are in equilibrium under the normalized SOC sense; when A higher value indicates that ES-1 has more power than ES-2, and the system should prioritize discharging ES-1 or reduce its charging priority; when... The opposite is true when (smaller or negative). For engineering implementation, priority should be given to... Normalization or limitation to ,For example This is a preferred value, indicating that a safety margin is still maintained under extreme conditions.

[0102] According to The ability of energy storage groups to participate in frequency regulation is dynamically adjusted, and an energy storage droop coefficient is introduced. A segmented mapping rule is preferred, as shown in the following example:

[0103] Set four key thresholds (in Middle area (For the ideal equilibrium region). Let the reference droop coefficient be... The adjustment coefficient is a constant. The segmentation is then defined as:

[0104]

[0105] In the left end area ( (Very small) The system increases charge droop or decreases discharge priority to avoid over-discharge; in the right-end region ( (Very large) The system improves discharge droop to provide discharge support; the middle region is the ideal working area, using a reference coefficient. .

[0106] Figure 3 and Figure 4 The balance zone and droop adjustment factor can be illustrated separately. A changing curve. Parameters Threshold Calibration can be performed during the project acceptance phase or based on historical operating data. The instruction manual should provide the optimal range and example values ​​(e.g., ...). ).

[0107] To avoid high-frequency charging and discharging switching caused by balance fluctuations, this invention introduces a cooling time during the commutation operation. With amplitude threshold. When any single unit Reaching the upper or lower limit ( or When this occurs, it triggers a swap of charging and discharging characters within the group, and forces a cooldown period after the swap. Internal interchange is prohibited; the cooling time can be set from several seconds to tens of seconds (preferably 10 s) to reduce mechanical and electrochemical losses. The interlock also includes a safety degradation strategy for communication failures: if the BMS / communication connection is lost, priority is given to ensuring system frequency stability and putting the energy storage into a safe mode.

[0108] (3) Battery Health State (SOH) Adaptive Threshold Model

[0109] Considering the decrease in usable capacity due to battery aging, this invention proposes a threshold adaptive strategy based on SOH to dynamically adjust the balance zone threshold, thereby avoiding the use of the same charge and discharge depth as new batteries during the battery aging period, thus slowing down the aging rate.

[0110] Let the normalized expression of SOH be: 1 represents the initial health state of the battery, which decreases over time. Define the initial threshold for a new battery. An example of the threshold adjusting linearly with SOH is as follows:

[0111]

[0112]

[0113] in, The threshold adjustment function is preferably in a linear or sublinear form, for example:

[0114]

[0115] Where, constant After calibration, it meets the requirements. That is, the initial threshold remains unchanged. When decrease This means narrowing the threshold and triggering protection actions earlier. An example is provided. More generally, It can be any monotonically increasing function and can be determined based on learning from historical degradation curves. This SOH adaptive mechanism is described as a preferred embodiment in the specification, and simulation results of parameter selection are given in the embodiment; see the Embodiments and Simulation sections.

[0116] (4) Wind-storage joint state-space model

[0117] To achieve real-time prediction and optimized allocation, this invention dynamically integrates wind power, energy storage, thermal power, and controllable loads into a unified state-space model. First, it presents a simplified continuous-time dynamic of the power system frequency, and then provides discretized state equations for use by the MPC (Multi-Process Control).

[0118] Under the equivalent single-machine model, the system frequency dynamics can be approximated as:

[0119]

[0120] in, Let be the system inertia constant (s); The damping coefficient is pu / Hz. The active power imbalance is defined as:

[0121]

[0122] in, For wind power output deviation, by Determined by the controller; Due to the deviation in energy storage output, SOC and commutation logic constraints; This refers to the output deviation of thermal power units. For controllable load response.

[0123] To construct a discrete-time state space that is easy for MPC to use, a state vector is selected. Includes, but is not limited to, the following quantities:

[0124]

[0125] Control input It can include wind turbine droop control variables, energy storage commands (discharge / charge power), thermal power regulation commands, and controllable load control variables. Disturbance vector. Includes sudden load increase Wind speed disturbances, etc.

[0126] Linearizing the nonlinear relationship near the operating point yields a linear approximation model, which can then be discretized in the time domain.

[0127] Discretization yields the commonly used prediction model form:

[0128]

[0129]

[0130] in, This is the state transition matrix, reflecting the coupling between inertia, damping, and subsystems;

[0131] The control input matrix describes the effect of each control action on the state. The perturbation transfer matrix; For measuring outputs (such as frequency, tie-line power, etc.).

[0132] Specific matrix items The parameters can be obtained analytically or through system identification based on the system scale, subsystem parameters, and linearized operating point. The embodiments in the specification will provide a parameter set for a benchmark small system (e.g., a 9 MW thermal power plant, a 6 MW wind power plant, or a 1.2 MWh DESS). Numerical examples and annotations are provided to facilitate the reader's reproduction of the simulation. The parameters and simulation settings for this small system can be found in the Examples section (figures and parameter tables) of this manual.

[0133] 4) Expression of frequency modulation dead zone and constraints

[0134] To avoid unnecessary mechanical and electrical energy consumption caused by frequent minor movements, a frequency modulation dead zone is introduced into the model. Concept: When At this time, some resources (such as controllable loads or low-priority energy storage) do not participate in frequency regulation, or their participation is reduced to zero by a certain scaling factor. This dead zone can be set separately for different resources (such as wind power dead zone is small, energy storage is next, and controllable load is the largest), and it is reflected in the constraint set in the form of inequalities (see the optimization constraint paragraph of step 200 / 300 in the following steps).

[0135] <Step 200>

[0136] Based on the wind power-storage joint state-space model given in step 100, this step establishes a control topology suitable for wind-storage coordinated primary frequency regulation, so as to incorporate wind turbine regulation capacity, energy storage balance, frequency deviation, tie-line power deviation, and priority strategy into a unified control framework. The control topology realizes the allocation of active power output from wind power, energy storage, and other resources through feedback channels between different modules, and can be expanded or tailored according to engineering needs.

[0137] In some embodiments, such as Figure 5 As shown, the wind-storage joint frequency regulation control topology of the present invention includes an input layer, a quantization and calculation layer, an optimization scheduling layer, and an output execution layer, specifically including the following parts:

[0138] (1) Input layer: real-time measurement and state estimation

[0139] The input layer is used to collect system operating status and pass the measurements to the model update and capability quantization module, including:

[0140] 1. System frequency deviation

[0141] 2. Wind turbine output power, rotor speed, and operating area

[0142] 3. Energy storage unit's SOC, SOH, current, voltage, and operating mode (discharge / charge)

[0143] 4. Regulation status between thermal power and other power sources

[0144] 5. Load-side disturbance prediction or real-time measurement

[0145] 6. Tie line power measurement (if the system has a multi-zone structure)

[0146] The input layer can be implemented based on SCADA, wind farm controllers, BMS, energy management systems, or PMU devices, and can be combined with filtering or state estimation algorithms to improve data stability. This input data will be used to: update the state-space model, calculate the wind power participation factor, calculate the energy storage balance and droop factor, adjust control priorities, and prepare the state vectors required for optimization solutions.

[0147] (2) Quantification and coefficient calculation layer: wind power participation coefficient and energy storage droop coefficient

[0148] The second layer of the control topology includes a capacity quantization and droop calculation module. Its core function is to input the adjustable capacity of wind turbines and energy storage into the optimizer in a parameterized manner, so that the system can automatically adjust the participation level of different resources during operation.

[0149] The results are calculated based on the kinetic energy margin and power margin in step 100. And generate participation coefficients through LOGISTIC mapping. During frequency modulation, if the actual frequency deviation exceeds the set dead zone... If the wind turbine participates in regulation, its command quantity will include the following restrictions in subsequent optimization constraints:

[0150]

[0151] when When the fan does not participate in frequency regulation, the controller will set the command to zero or shorten the response according to the preferred scaling strategy.

[0152] Energy storage droop coefficient Based on group balance Segmented or smooth mapping can be performed to reflect the preferential actions of energy storage in different equilibrium zones. For example, when When the energy level is too high, the system tends to instruct the stored energy to discharge in order to restore balance; when... When the energy level is low, the system tends to instruct the energy storage to charge in order to restore balance. The allowable range for energy storage regulation is defined by:

[0153]

[0154] However, the actual usable range is further reduced by the constraints of SOC and SOH, which will be explained in detail in step 300.

[0155] (3) Structure of multi-source joint control topology

[0156] like Figure 5 As shown, the joint control topology of the present invention can be summarized as consisting of the following modules:

[0157] 1. Wind power regulation module: Receives frequency deviation and participation coefficient, and generates wind turbine output deviation based on state-space model and instructions.

[0158] 2. Energy Storage Management Module: Calculates the charging and discharging power commands for energy storage based on the balance degree and droop coefficient, and executes commutation logic and interlocking strategies.

[0159] 3. Thermal power and controllable load module: used for coordinated regulation in multi-source frequency regulation scenarios.

[0160] 4. Joint Modeling Module: Updates the state predictions based on the model from step 100.

[0161] 5. Optimization Scheduling Module (MPC / QP): Performs objective function solving and constraint assignment, and issues optimal instructions.

[0162] 6. Feedback Execution Module: The actuator is responsible for writing the active power command of the wind turbine, the energy storage power command, and the thermal power regulation quantity into the equipment controller.

[0163] The above modules can be implemented either within the controller or in a distributed deployment across systems.

[0164] (4) Priority strategy design

[0165] In wind-storage coordinated frequency regulation, different resources have different costs, lifespan limitations, and response speeds. Therefore, this invention employs a hierarchical priority rule to ensure that, while meeting system frequency stability requirements, the burden on energy storage cycles is minimized and wind turbine safety is guaranteed.

[0166] In a preferred embodiment, the priority order adopted by the present invention is as follows:

[0167] The specific meanings are as follows: when the wind turbine has sufficient kinetic energy margin and power margin, the wind turbine will be given priority for regulation; when the wind turbine's regulation capability is insufficient or the frequency deviation is large, the energy storage will perform rapid compensation; and controllable loads will participate in regulation when both wind power and energy storage are close to their limits.

[0168] This priority can be configured as a parameter in the controller to adapt to different grid strategies. For example, the weight of wind turbines can be increased in scenarios where energy storage costs are high; the weight of energy storage can be increased in scenarios where wind turbine output fluctuates greatly.

[0169] (5) Dead zone setting and engineering protection mechanism (constraints in control topology)

[0170] To improve system stability and reduce equipment wear, the following engineering protection mechanisms are introduced into the control topology:

[0171] 1. Frequency dead zone setting. When the frequency deviation is less than the dead zone corresponding to the resource, the resource will not participate in regulation. The dead zone value is set according to the preferred range, for example: 0.01 Hz for fans, 0.05 Hz for energy storage, and 0.1 Hz for controllable loads.

[0172] 2. Ramp-up rate limits. Ramp-up and ramp-down limits are set separately for active power commands from wind turbines, energy storage, and thermal power plants to ensure actuator safety.

[0173] 3. Communication Failure Degradation Strategy. When data from the BMS or wind farm controller is interrupted, the controller automatically switches to steady-state or limiting mode.

[0174] 4. Commutation Cooling Time Limit. The switching between charging and discharging roles within the energy storage unit must meet cooling time constraints to avoid losses caused by high-frequency switching.

[0175] These engineering protection measures are all reflected in the optimization constraints or instruction execution logic, and a more specific mathematical expression will be given in step 300.

[0176] (6) Data flow and instruction flow between topology layers

[0177] Combination Figure 4 The data stream and instruction stream can be described as follows:

[0178] Data Flow: Frequency Deviation → Capacity Quantization Module → State Prediction Module → Optimized Scheduling Module (SOC / SOH) → Balance Module → Energy Storage Regulation Module → Optimized Scheduling Module

[0179] Command Flow: Optimization Results → Wind Turbine Actuator Optimization Results → Energy Storage PCS Optimization Results → Thermal Power Actuator or Load Controller

[0180] The entire topology operates in a rolling manner, with a typical sampling period of 0.2 s, but can be selected between 0.05 and 1 s depending on engineering conditions.

[0181] <Step 300>

[0182] Based on the wind-storage joint control topology constructed in step 200, this step generates active power control commands for wind turbines, energy storage, thermal power, and controllable loads through a rolling optimization solution and constraint execution mechanism, and sends them to each execution terminal in real time, thereby achieving rapid stabilization of the system frequency.

[0183] (1) Constructing the optimization objective function (MPC framework)

[0184] This invention preferably employs Model Predictive Control (MPC) as the solver for joint frequency modulation, utilizing the discretized state-space equations obtained in step 100 for prediction. The objective function adopts a quadratic form structure:

[0185]

[0186] in, To predict the time domain length, for example, 10 to 20 steps; The frequency deviation weight is usually the largest; Weighting of wind power and energy storage output changes; Smooth the climbing rate weights to avoid drastic jumps; To maintain weight for energy storage balance; This is the terminal penalty weight, used to improve convergence.

[0187] The objective function embodies the following engineering principles: prioritize eliminating frequency deviation (maintain grid friendliness); minimize high-frequency cycling of energy storage (delay aging); avoid wind turbine speed dropping to near the safety boundary; maintain balance among energy storage units; and limit frequent load-side response actions.

[0188] The objective function can automatically adjust its weights based on the scenario. For example, in scenarios with a high proportion of wind power, increasing the weights... or To ensure user comfort when using wind power and energy storage, and to improve efficiency in scenarios where energy storage is expensive or lifespan is sensitive. and .

[0189] (2) Construction of constraint set

[0190] To ensure that wind turbines, energy storage, and other resources operate within safe limits, this invention explicitly incorporates constraints into the constraint set of the MPC. These constraints include:

[0191] (2.1) System dynamic constraint state equations

[0192]

[0193] Among them and As defined in step 100, it includes:

[0194] 1. Fan speed and output deviation

[0195] 2. SOC, SOH, and energy storage power

[0196] 3. System frequency deviation

[0197] 4. Other filtering and integration states

[0198] Dynamic constraints ensure that the instructions generated by the optimization solver are consistent with the actual reachable behavior of the system.

[0199] (2.2) Constraints on the regulating capacity of the fan

[0200] Power constraints:

[0201]

[0202] Upper limit of response constrained by kinetic energy margin:

[0203]

[0204] in and Calculated dynamically from step 100 / 200.

[0205] Lower speed limit protection:

[0206]

[0207] If the speed approaches the limit, the controller will automatically reduce the fan adjustment amount, see the priority mechanism in step 200 for details.

[0208] (2.3) Constraints on Energy Storage Operation

[0209] SOC range:

[0210]

[0211] Used to protect battery health.

[0212] Power constraints:

[0213]

[0214] Energy storage droop coefficient constraint:

[0215]

[0216] in It has been defined in step 100.

[0217] Reversing interlock and cooldown time:

[0218] If the last reversal time was ,but:

[0219]

[0220] This means that charging and discharging switching is prohibited during the cooling period. This constraint can be implemented by logic variables within the MPC or enforced by the outer control logic.

[0221] (2.4) Climbing rate constraint

[0222] Apply uniform ramp constraint to all resources:

[0223]

[0224]

[0225] This is used to prevent the actuator from moving too fast, which could lead to equipment safety risks.

[0226] (2.5) Dead Zone Constraint

[0227] If the resource corresponds to a dead zone ,but:

[0228]

[0229] Dead zones can be implemented using binary variables or constraint switching functions.

[0230] (3) Rolling solution mechanism

[0231] MPC employs a rolling time-domain optimization method:

[0232] 1. At any moment Update status based on measurement ;

[0233] 2. Solve the quadratic programming (QP) problem consisting of the above objective function and constraints;

[0234] 3. Obtaining several future moments ;

[0235] 4. Execute only the first step instruction. ;

[0236] 5. Time Measure again, update the state, and solve again.

[0237] This method can automatically adapt to changes in wind speed, sudden load changes, and changes in equipment operating status. Its stability and real-time performance can meet the frequency regulation requirements of the power system.

[0238] (4) Command output and device execution

[0239] The optimization results are implemented through the wind turbine controller, energy storage converter (PCS), and thermal power regulator.

[0240] (4.1) Wind turbine execution logic

[0241] The wind turbine actuator will send active bias commands. Switch to: Adjust the propeller angle (suitable for high-speed range), or adjust the speed droop (suitable for medium and low-speed range).

[0242] The command is automatically reduced when the speed approaches the lower limit.

[0243] (4.2) Energy storage execution logic

[0244] Energy storage system PCS receives power commands It also performs: DC bus voltage control, charge / discharge mode selection, interlocking and commutation protection, and SOC balance coordination among BMS members.

[0245] If the SOH is low, then the depth of cycling should be reduced first.

[0246] (4.3) Fault Degradation Strategy

[0247] If a communication anomaly, sensor error, or controller overflow occurs, the following safety strategy will be implemented: energy storage will be reduced to "standby mode" and maintained only within the SOC range; wind turbines will maintain maximum safe output and stop regulation; thermal power units will resume normal primary frequency regulation; the main control system will be interrupted to optimize the solution and enter steady-state protection logic.

[0248] (5) Action effect and system stability

[0249] Based on the above mechanism, the control commands generated by this invention can eliminate system frequency deviation within seconds and operate under different disturbances ( Figure 7 The energy storage balance remains stable under these conditions. Figure 3 ), wind turbine participation factor ( Figure 5 All of these are dynamically regulated in real time, thereby achieving: rapid frequency recovery, safe adjustment of wind turbines, efficient utilization of energy storage, delayed aging, and coordinated operation of multiple sources.

[0250] The following specific examples further illustrate the wind-storage combined frequency regulation method proposed in this invention. This embodiment is analyzed under a typical wind farm-energy storage system-grid structure, and uses... Figure 7 , Figure 8 The small-scale simulation system shown is for reference, demonstrating the actual operation of the invention under different disturbance conditions. To avoid conceptual confusion, the term "wind farm" refers to the wind power zone containing several variable-speed constant-frequency wind turbines; "energy storage system" refers to a dual energy storage group (DESS) containing two battery units; and "main grid" refers to the equivalent bus of the power system containing certain equivalent inertia and damping characteristics.

[0251] 1) Simulation system structure and parameters

[0252] This embodiment uses Figure 7 Based on the structure shown, the wind farm has a rated capacity of 6 MW, consisting of multiple fixed-pitch doubly-fed induction generators, equivalent to an adjustable wind power node. The energy storage system adopts a dual-battery structure with a capacity of 1.2 MWh, consisting of two battery cells with identical capacity, rated voltage, and rated power. Furthermore, the main grid is equivalent to a power system with a capacity much larger than the wind-storage system, and its frequency response characteristics are simulated using virtual inertia and damping coefficients.

[0253] The key parameters of the simulation system are listed in Table 1. The table includes indicators such as the wind turbine inertia constant, pitch adjustment time delay, minimum and maximum SOC of energy storage, initial SOH value, and maximum PCS power. These are typical values ​​and can be adjusted or calibrated according to different grid scales and equipment types.

[0254] Table 1. Main parameters of the simulation system

[0255]

[0256] 2) Capacity quantification and state calculation

[0257] At the start of the simulation, the system first quantifies the adjustability of wind power and energy storage based on actual operating conditions. For the wind power component, the kinetic energy margin is calculated based on the turbine speed and the turbine characteristic curve (…). Figure 3 )Depend on:

[0258]

[0259] Calculated, and obtained through, as Figure 6 The LOGISTIC function shown is mapped to the participation coefficient. When the system load suddenly increases, causing a frequency deviation, if the deviation exceeds the frequency dead zone preset by the fan, the fan will be adjusted according to this participation coefficient; if the instantaneous kinetic energy margin is insufficient, the participation coefficient will be automatically reduced to prevent the fan speed from falling to the safe boundary.

[0260] The energy storage section calculates the balance degree using the normalized difference of the SOC within the group. ,Right now:

[0261]

[0262] If battery aging is significant (e.g., SOH drops to around 0.85), the threshold range will automatically shrink through the SOH adaptive function, allowing the scheduler to reduce deep discharge behavior during optimization. This mechanism will be discussed in subsequent scenario analyses. Figure 9 Displayed in the form of [format].

[0263] 3) Frequency modulation event triggering and command generation

[0264] In this embodiment, the simulation system applies a load disturbance of 0.8 MW at t=3s. Figure 8 After the disturbance occurs, the frequency deviation rapidly increases from zero to approximately -0.18 Hz, exceeding the wind turbine dead zone. Under the MPC framework, the controller predicts the frequency evolution trend for the next few steps based on the state-space model, and simultaneously considers factors such as wind turbine kinetic energy margin, energy storage SOC state, balance, SOH degradation, and PCS power limitation to generate the optimal control sequence for the next 10 to 20 steps.

[0265] The control quantity at this time and It is not determined independently, but rather automatically calculated through multidimensional trade-offs of the objective function. For example, in the initial second-level response phase, wind power still maintains a high kinetic energy margin, therefore the participation coefficient... While maintaining a relatively high frequency range (e.g., 0.6–0.8), the regulator prioritizes using the droop capacity of the wind turbine to help the system recover frequency, while the energy storage compensates for the remaining deviation with a slight discharge of 0.1–0.2 MW during this phase. As the frequency decline slows and kinetic energy consumption decreases, the energy storage gradually increases its discharge power and stabilizes at around 0.6–0.8 MW within 10 seconds. This process is... Figure 8 The joint regulation trajectory is shown in the figure.

[0266] 4) Energy storage balance adjustment and aging protection

[0267] As frequency regulation continues, the State of Charge (SOC) of the two units in the energy storage system gradually deviates. However, because the controller constantly utilizes the balance... As part of the predicted state, the amount of energy stored is automatically distributed to prevent any single cell from over-discharging.

[0268] when Below When the output of ES-1 is reduced, the discharge of ES-2 will be increased. If the state of charge (SOH) is low, the system will automatically trigger the protection threshold in advance, making the depth of charge and discharge more conservative.

[0269] This process is in Figure 4 and Figure 9 The curves all reflect this, showing that energy storage maintains high consistency after 20-30 seconds of joint adjustment, thus significantly reducing the long-term aging rate.

[0270] 5) System frequency recovery process

[0271] In this embodiment, the system frequency recovery trajectory is as follows: Figure 8 As given in the text. From a time-domain perspective:

[0272] 0–3 s: The system is in a normal and stable state, with a frequency close to 50 Hz;

[0273] 3–4 s: Load disturbance causes a rapid drop in frequency;

[0274] 4–10 s: The fan participates in the regulation and contributes most of the initial response;

[0275] 10–30 s: The energy storage system gradually takes over the main regulation, making the frequency approach the rated value in a smooth manner;

[0276] After 30 seconds: the system returns to a stable state, the adjustment gradually decreases and eventually returns to zero.

[0277] This process demonstrates that the dynamic participation coefficient, balance control strategy, and MPC prediction mechanism proposed in this invention work together to enable wind-storage joint frequency regulation to be fast, smooth, and equipment-friendly, significantly outperforming traditional fixed-weight or prediction-free frequency regulation strategies. Figure 11 As shown.

[0278] This invention also provides a wind-storage joint frequency regulation device based on wind power participation factor and energy storage charge-discharge balance, for implementing the above method. The system includes:

[0279] The wind power regulation module is used to acquire the operating status data of each wind turbine in real time, including the turbine rotor speed, current output power and converter capacity limit, and calculate the kinetic energy margin and power margin of the wind turbine based on the operating status data; determine the available frequency regulation capability of the wind turbine based on the kinetic energy margin and power margin, and map the frequency regulation capability to the wind power participation factor Kw through the LOGISTIC function;

[0280] The energy storage management module is used to configure the energy storage system into a group containing multiple energy storage units, collect the state of charge (SOC) of each energy storage unit in real time, and calculate the normalized balance degree (B) of the energy storage group; dynamically adjust the droop coefficient (Kes) of the energy storage group based on the balance degree (B), and trigger the switching strategy of charging and discharging roles of each unit within the energy storage group when the SOC of any energy storage unit reaches a preset upper or lower limit.

[0281] The joint modeling module is used to construct a joint state-space model that includes the dynamics of wind power, energy storage, thermal power and controllable load, and uses the wind power participation coefficient Kw and droop coefficient Kes as model inputs or constraints.

[0282] The optimization scheduling module is used to solve the optimal control sequence within the preset prediction time domain by using model predictive control or quadratic programming methods based on the joint state space model and the wind power participation coefficient Kw and droop coefficient Kes obtained in real time, so as to minimize the comprehensive objective function within the prediction time domain.

[0283] The execution and communication module is used to send the control command of the first moment in the optimal control sequence obtained by solving to the corresponding wind turbine, energy storage, thermal power and controllable load actuators to implement real-time adjustment, and to collect the feedback status data of each actuator to complete closed-loop control.

[0284] The specific embodiments of the present invention have been described in detail above. For those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principle of the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims

1. A wind-storage joint frequency regulation method based on wind power participation coefficient and energy storage charge-discharge balance, used for real-time coordinated control of wind turbines, energy storage, and other frequency regulation resources in a power system to maintain system frequency stability, characterized in that... Includes the following steps: Real-time acquisition of operating status data for each wind turbine, including turbine rotor speed, current output power and converter capacity limit, and calculation of the kinetic energy margin and power margin of the wind turbine based on the operating status data; The available frequency regulation capability of the wind turbine is determined based on the kinetic energy margin and power margin, and the frequency regulation capability is mapped to the wind power participation factor using the LOGISTIC function. ; The energy storage system is configured as a group containing multiple energy storage units. The state of charge (SOC) of each energy storage unit is collected in real time, and the normalized balance of the energy storage group is calculated. Based on the aforementioned balance Dynamically adjust the droop coefficient of this energy storage group And when the SOC of any energy storage unit is detected to reach the preset upper or lower limit, the strategy of switching the charging and discharging roles of each unit within the energy storage group is triggered. Construct a joint state-space model that includes the dynamics of wind power, energy storage, thermal power, and controllable loads, and use the wind power participation coefficient as the basis for this model. and droop coefficient As model input or constraint; Based on the aforementioned joint state-space model and the real-time acquired wind power participation coefficient and droop coefficient Within the preset prediction time domain, the optimal control sequence is solved using model predictive control or quadratic programming methods to minimize the comprehensive objective function within the prediction time domain. The control command at the first moment of the optimal control sequence obtained by the solution is sent to the corresponding actuators of wind turbines, energy storage, thermal power and controllable loads to implement real-time adjustment, and the feedback status data of each actuator is collected to complete the closed-loop control.

2. The method as described in claim 1, characterized in that, The kinetic energy margin of the wind turbine Defined as: ； in, This represents the current rotor kinetic energy of the wind turbine. For rotational inertia, The rotor speed is and These are the maximum and minimum available kinetic energies calculated based on the safe upper and lower limits of the fan speed.

3. The method as described in claim 1, characterized in that, The power margin of the wind turbine Defined as: ； in, This is the current output power. , This represents the upper and lower power limits of the wind turbine under the constraints of the converter and control.

4. The method as described in claim 1, characterized in that, The wind power participation coefficient Determined by the following formula: ； in, , These are real-valued parameters used for modulation sensitivity and threshold.

5. The method as described in claim 1, characterized in that, The energy storage group balance The calculation is as follows: ； in For energy storage units The state of charge is used to achieve a normalized measure of balance; for groups of more than two units, a [specific term] can be defined. This is the variance or range index of the normalized SOC for each unit.

6. The method as described in claim 1, characterized in that, Based on the balance Dynamically adjust the droop coefficient of this energy storage group Specifically, this includes: the energy storage droop coefficient. According to balance Dynamic adjustments are made using segmented or continuous mapping relationships; when Less than a certain lower threshold When this occurs, energy storage groups prioritize entering charging mode or increase the charging droop coefficient; when Greater than a certain upper limit threshold When this occurs, the energy storage group prioritizes entering discharge mode or increases the discharge droop coefficient; when When in the middle range, energy storage participates in frequency regulation with a conventional droop coefficient.

7. The method as described in claim 1, characterized in that, The state variables of the joint state-space model include system frequency, wind turbine output power deviation, energy storage output power deviation, energy storage SOC, thermal power unit output power deviation, and tie-line power deviation+, and the discrete-time state equation is obtained through linearization / discretization: ； It is the state variable at time k+1; It is the state variable at time k; It is the input variable at time k; It is the perturbation variable at time k; It is a state variable matrix; It is the input variable matrix; It is the perturbation variable matrix.

8. The method as described in claim 1, characterized in that, The objective function of the MPC includes a frequency deviation squared term, a tie-line power deviation term, and a control input / variation cost term, denoted as: ； It is the objective function; It is a time index variable; N is the total duration of the data being monitored; yes Frequency deviation at any given moment; yes Line power deviation at any given time; yes Deviation of input variables at any given time; It is the weighting coefficient of the frequency deviation term; It is the weighting coefficient for the line power deviation term; These are the weighting coefficients for the input variable deviation term; Furthermore, the optimization process is subject to constraints such as active power upper and lower limits, ramp rate constraints, energy storage SOC constraints, and frequency regulation dead zone constraints. For prediction in the time domain, the range is 10 to 50; These are weighting coefficients, which are used in simulation / engineering calibration to balance frequency performance with control costs.

9. A wind-storage joint frequency regulation device based on wind power participation factor and energy storage charge-discharge balance, used to implement the method according to any one of claims 1 to 8, characterized in that, The system includes: The wind power regulation module is used to acquire the operating status data of each wind turbine in real time, including the turbine rotor speed, current output power and converter capacity limit, and calculate the kinetic energy margin and power margin of the wind turbine based on the operating status data; determine the available frequency regulation capability of the wind turbine based on the kinetic energy margin and power margin, and map the frequency regulation capability to the wind power participation factor Kw through the LOGISTIC function; The energy storage management module is used to configure the energy storage system into a group containing multiple energy storage units, collect the state of charge (SOC) of each energy storage unit in real time, and calculate the normalized balance degree (B) of the energy storage group; dynamically adjust the droop coefficient (Kes) of the energy storage group based on the balance degree (B), and trigger the switching strategy of charging and discharging roles of each unit within the energy storage group when the SOC of any energy storage unit reaches a preset upper or lower limit. The joint modeling module is used to construct a joint state-space model that includes the dynamics of wind power, energy storage, thermal power and controllable load, and uses the wind power participation coefficient Kw and droop coefficient Kes as model inputs or constraints. The optimization scheduling module is used to solve the optimal control sequence within the preset prediction time domain by using model predictive control or quadratic programming methods based on the joint state space model and the wind power participation coefficient Kw and droop coefficient Kes obtained in real time, so as to minimize the comprehensive objective function within the prediction time domain. The execution and communication module is used to send the control command of the first moment in the optimal control sequence obtained by solving to the corresponding wind turbine, energy storage, thermal power and controllable load actuators to implement real-time adjustment, and to collect the feedback status data of each actuator to complete closed-loop control.

10. The apparatus as claimed in claim 9, characterized in that, The wind power regulation module is specifically used to obtain the kinetic energy margin. and the power margin The minimum value is used as an indicator of the wind turbine's available frequency regulation capability. and the wind power participation coefficient This is used to adjust the droop gain and active response upper limit of the fan when frequency deviation occurs, wherein the active response upper limit satisfies .

Images