A Collaborative Optimization Method for Wind Power Systems Considering Spatiotemporal Heterogeneity and Coupling Dependence

Abstract

This invention belongs to the field of wind power system control and dispatching technology. It proposes a collaborative optimization method for wind power systems that considers spatiotemporal heterogeneity and coupling dependence. First, this invention constructs a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power and grid-connected power electronic equipment, characterizing the spatiotemporal heterogeneity of frequency dynamic evolution after wide-area grid disturbances. Then, it establishes a coupling dependence model between new energy support characteristics and system strength, evaluates the maximum feasible region for new energy units to participate in frequency support, and quantifies the collaborative support capability of wind power participation in frequency regulation and thermal power start-up for regional frequency security. Finally, based on frequency security constraints, it constructs an integrated optimization framework for unit combination and frequency characteristics, collaboratively optimizing start-up methods and wind power frequency regulation parameters. The resulting dispatching scheme can fully release the support capacity of wind farms and energy storage systems, effectively reduce thermal power start-up, and at the same time ensure the regional frequency security of the system.

Description

Technical Field

[0001] This invention relates to the field of wind power system control and scheduling technology, and in particular to a collaborative optimization method for wind power systems that takes into account spatiotemporal heterogeneity and coupling dependence. Background Technology

[0002] With a high proportion of wind power continuously integrated into the grid, the high uncertainty of renewable energy output and the declining system inertia coexist. To ensure system frequency security, many thermal power units maintain low-output operation to provide system backup, severely restricting the renewable energy absorption capacity and exacerbating the contradiction between system economy and dynamic security. Against this backdrop, tapping into the active support potential of renewable energy has become an important way to alleviate the pressure on thermal power generation. Wind farms can actively participate in system frequency regulation through strategies such as rotor kinetic energy control, while grid-side energy storage can provide virtual inertia and frequency regulation support by adopting grid-based control. However, the spatial distribution of various power sources in the wide-area grid is highly uneven, and the power-frequency response speed and regulation characteristics of different power sources vary significantly. In particular, the inherent lag regulation characteristics of synchronous machines coexist with the fast response characteristics of grid-connected power electronic equipment, resulting in significant spatiotemporal heterogeneity in the dynamic frequency response of the system under heterogeneous power source collaborative networking conditions.

[0003] The frequency regulation characteristics of different wind farms significantly affect the system frequency dynamics, which in turn influence the feasible domain of wind farm support parameters through the physical constraints of wind turbine operation. Meanwhile, the start-up scheme of thermal power units determines the system's fundamental frequency strength. Improving the support capacity of renewable energy sources can reduce the thermal power capacity required to maintain frequency security, thus forming a coupled and interdependent relationship at the system level: unit combination, system strength, and renewable energy support capacity. Existing research largely focuses on unit combination schemes considering frequency security constraints, relying on the start-up of thermal power units to ensure frequency security, without considering the support capacity of renewable energy sources, resulting in a large number of thermal power units operating at low output as reserves; or it emphasizes evaluating the frequency regulation capacity that renewable energy can provide under a given operating mode, ignoring the interaction between the system operating mode and the feasible domain of renewable energy frequency regulation. Summary of the Invention

[0004] Therefore, one objective of this invention is to propose a collaborative optimization method for wind power systems that takes into account spatiotemporal heterogeneity and coupling dependence, in order to solve the problems mentioned in the background art and overcome the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides a method for collaborative optimization of wind power systems that considers spatiotemporal heterogeneity and coupling dependence, comprising: Construct a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment; Construct a coupled and interdependent model that considers the supporting characteristics of new energy sources and the strength of the system, and quantify the collaborative support capabilities of wind power participation in frequency regulation and thermal power generation for regional frequency security. Based on frequency security constraints, an integrated optimization framework is constructed to combine unit combinations and heterogeneous power supply frequency support characteristics, and to collaboratively optimize start-up methods and new energy frequency regulation parameters.

[0006] As a preferred option, the power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected / networked power electronic equipment includes: By introducing the potential node in the synchronous machine and the virtual potential node in the grid-type converter, the algebraic constraints are eliminated through network equivalent order reduction, while preserving the dynamic coupling relationship between the inertial source node and the new energy access node; Electromechanical transient model is used to describe the power frequency response characteristics of thermal power units, synchronous machine rotor dynamic model is used to describe grid-type energy storage converter, and first-order inertial element is used to describe the frequency regulation power response characteristics of grid-type new energy. The inertial response and primary frequency regulation process of the heterogeneous units after disturbance are modeled into a unified first-order matrix differential equation, and the node frequency expression is obtained by solving the differential equation through eigenvalue decomposition.

[0007] As a preferred approach, a coupled and dependent model based on frequency security constraints and considering the supporting characteristics of new energy sources and the strength of the system includes: To assess the frequency regulation support capability of a wind farm, considering the upper and lower limits of the wind turbine rotor speed and the rated overcurrent capacity constraints of the converter, frequency regulation power is allocated among the wind turbines based on the SOE index. The feasible region for evaluating the inertia-damping coefficient of grid-type energy storage is constrained by the rated power and state of charge of the energy storage. The starting status of thermal power units and the frequency regulation parameters of new energy sources are embedded as control variables into the power system node frequency model. The system frequency intensity changes under different combinations of thermal power start-up modes and new energy frequency regulation parameters are quantified. The frequency safety index is used as a constraint condition to inversely constrain the feasible domain of minimum thermal power start-up and new energy frequency regulation parameters.

[0008] As a preferred approach, the integrated optimization framework for unit combination and heterogeneous power supply frequency support characteristics includes: Construct an upper-level thermal-wind-storage coordinated unit combination optimization model with the goal of minimizing operating costs, and constrain the system power balance, line transmission capacity, thermal power unit operation, energy storage operation, and system reserve capacity. A time-varying maximum frequency regulation capability assessment model for the lower-level wind-storage system is constructed with the goal of maximizing the frequency regulation coefficient of new energy sources. The constraints include physical constraints and line power flow constraints. Construct regional frequency security constraints, and iteratively optimize thermal power plant start-up schemes and new energy frequency regulation parameters through frequency security verification. The frequency security verification is based on the region's maximum frequency change rate, maximum frequency difference, and quasi-steady-state frequency difference indicators.

[0009] Preferably, the power system node frequency model describes the dynamic process of frequency evolution of all nodes after a wide-area system disturbance. Potential nodes are introduced for thermal power units and grid-connected energy storage converters, and connected to the unit's grid connection point through generator transient reactance or grid-connected energy storage virtual impedance. The active power-voltage phase angle change relationship of all nodes is as follows: ; In this context, the subscripts G, C, and L of each variable represent the inertial source potential node, the wind farm grid-connected node, and the load node including the inertial source terminal node, respectively, with the following quantities: , , , These are the power change vectors for the aforementioned nodes. , , The phase angle changes of the nodes mentioned above are represented by the coefficient matrix on the right side of the equation. The system's Laplace matrix, whose block elements can be calculated from its constituent nodes using the following formula: ; in Let the initial voltage of bus n be denoted by the initial voltage value of the extended potential node, which is represented by the generator potential E. , and Let N be the initial phase angle difference, conductance, and susceptance of the branch nm.

[0010] As a preferred embodiment, the electromechanical transient model of the thermal power unit is as follows: ; ; Among them, H, F, D, and T are diagonal matrices composed of the unit's inertia coefficient, frequency regulation coefficient, reheat coefficient, damping coefficient, and reheat time constant, respectively.

[0011] As a preferred option, the SOE index is expressed as follows: ; in, and For minimum and maximum stored kinetic energy, The kinetic energy stored in each wind turbine blade, For the fan rotor speed, and Minimum and maximum fan rotor speed limits.

[0012] As a preferred option, the frequency regulation power is distributed among the wind turbines as follows: ; ; in, For wind farm frequency modulation power, Frequency regulation power supplied to the wind turbine.

[0013] As a preferred option, the optimization model for the upper-level thermal-wind-storage coordinated unit combination is as follows: ; The evaluation model for the time-varying maximum frequency regulation capability of the lower-level wind-storage system is as follows: ; in, These are the unit operation and start-up / shutdown costs, respectively. For energy storage operating costs, These are backup compensations for thermal power and energy storage, respectively. These refer to thermal power output and energy storage charging and discharging power, respectively. For the 0-1 variables of the start / stop flag of thermal power units, These are thermal power and energy storage reserve capacity, respectively.

[0014] As a preferred option, the frequency security constraints are as follows: ; Among them, if the current number Taiwan unit offline The unit was switched to the start-up state. ,at this time This indicates an improvement in the frequency index; if the first Taiwan unit online Flip it to the power off state. Then at this time This indicates that the frequency index deteriorated after the unit was shut down.

[0015] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: This invention first constructs a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment, characterizing the spatiotemporal heterogeneity of frequency dynamic evolution after a disturbance in a wide-area power grid. Then, it establishes a coupled and dependent model of new energy support characteristics and system strength, assesses the maximum feasible region for new energy units to participate in frequency support, and quantifies the collaborative support capability of wind power participation in frequency regulation and thermal power generation for regional frequency security. Finally, based on frequency security constraints, it constructs an integrated optimization framework for unit combination and frequency characteristics, collaboratively optimizing the start-up mode and wind power frequency regulation parameters. The resulting dispatch scheme can fully release the support capacity of wind farms and energy storage systems, effectively reduce thermal power generation, and simultaneously ensure regional frequency security of the system.

[0016] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0017] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the frequency response model of a heterogeneous unit according to an embodiment of the present invention; Figure 3 This is an optimized flowchart of an embodiment of the present invention; Figure 4 This is a schematic diagram of the system topology for a verification case of an embodiment of the present invention; Figure 5 This is a comparative schematic diagram of thermal power plant start-up schemes under different scenarios according to embodiments of the present invention; Figure 6 This is a waveform comparison diagram of system frequency dynamics under the minimum start-up mode of thermal power plants according to an embodiment of the present invention; Figure 7 This is a schematic diagram comparing the wind turbine's operational safety boundaries before and after optimization in an embodiment of the present invention; Figure 8 This is a schematic diagram comparing the thermal power plant start-up and new energy parameters before and after optimization in an embodiment of the present invention. Detailed Implementation

[0018] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0019] In wide-area power systems with a high proportion of wind power penetration, large-scale wind farms are connected to the grid across regions and at multiple nodes. Their grid-side converters mainly adopt grid-following control with additional active power frequency regulation strategies. Grid-based energy storage, as a new type of supporting unit connected to the system, possesses virtual inertia and damping characteristics similar to synchronous machines, and together with thermal power units, constitutes the system's inertial source. After a system disturbance, the frequency response modes of different types of power sources show significant differences, such as... Figure 2 As shown, after coupling with the wide-area power grid, the system frequency further exhibits significant spatial characteristics. Therefore, it is necessary to establish a wide-area system node frequency model for heterogeneous unit collaborative networking in order to characterize the spatiotemporal heterogeneity of the system frequency response after disturbance.

[0020] like Figure 1 As shown, an embodiment of the present invention provides a wind power system collaborative optimization method that considers spatiotemporal heterogeneity and coupling dependence, comprising S1-S3.

[0021] S1: Construct a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power and grid-connected / grid-connected power electronic equipment. This includes: introducing synchronous machine internal potential nodes and grid-connected converter virtual potential nodes; eliminating algebraic constraints through network equivalent order reduction; and preserving the dynamic coupling relationship between inertial source nodes and renewable energy access nodes. An electromechanical transient model is used to describe the power frequency response characteristics of thermal power units, a quasi-synchronous machine rotor dynamic model is used to describe grid-connected energy storage converters, and a first-order inertial element is used to describe the frequency regulation power response characteristics of grid-connected renewable energy. The inertial response and primary frequency regulation process of heterogeneous units after disturbance are modeled into a unified first-order matrix differential equation, and the differential equation is solved through eigenvalue decomposition to obtain the node frequency expression.

[0022] Specifically, for high-proportion wind power systems, a power system node frequency model is constructed that considers the differences in power frequency characteristics of synchronous, grid-connected, and grid-connected power electronic equipment. First, at the system network level, potential nodes within synchronous machines and virtual potential nodes of grid-connected converters are introduced. The algebraic constraints of other nodes are eliminated through network equivalent order reduction, while preserving the dynamic coupling relationship between inertial source nodes and renewable energy access nodes. For thermal power units, an electromechanical transient model incorporating rotor inertia, damping, and speed regulation is used to describe their power frequency response characteristics. For grid-connected energy storage converters, a quasi-synchronous machine rotor dynamic model based on virtual inertia and damping characteristics is constructed. For grid-connected renewable energy, a first-order inertial element is introduced to characterize its delayed frequency regulation power response characteristics. The inertial response and primary frequency regulation process of heterogeneous units after disturbance are modeled into a unified first-order matrix differential equation form. The differential equation is solved using eigenvalue decomposition to obtain an expression for the dynamic evolution of the unit frequency after disturbance. Furthermore, based on the improved frequency divider principle, the frequency of each node in the system is represented as a weighted combination of the frequency responses of various power supplies, thereby obtaining the frequency expressions of all nodes after system disturbance.

[0023] To accurately describe the frequency changes caused by the motion of the synchronous motor's electromechanical rotor and the virtual rotor motion of the grid-connected energy storage, this invention introduces internal potential nodes for thermal power units and grid-connected energy storage units (inertial source units) into the network model, and connects them to the grid connection point of the units through generator transient reactance or grid-connected energy storage virtual impedance. Further considering the frequency regulation power of the grid-connected wind farm during disturbances and other load nodes, the active power-voltage phase angle change relationship for all nodes in the entire system is as follows: (1) (2) In formula (1), the subscripts G, C, and L of each variable represent the inertial source potential node, the wind farm grid connection node, and the load node including the inertial source terminal node, respectively, and their quantities are respectively , , , These are the power change vectors for the aforementioned nodes. , , The phase angle changes of the nodes mentioned above are represented by the coefficient matrix on the right side of the equation. The system's Laplace matrix is ​​given by its block elements, which can be calculated from its constituent nodes according to formula (2), where... Let the initial voltage of bus n be denoted by the initial voltage value of the extended potential node, which is represented by the generator potential E. , and Let N be the initial phase angle difference, conductance, and susceptance of the branch nm.

[0024] From formula (1), taking the second and third rows of the matrix equation, we can obtain: (3) in Taking the derivative of both sides of equation (3) with respect to time t, we get: (4) In the formula , , These represent the frequency variations at the inertial source potential node, the wind farm grid connection node, and the load node, respectively. The system's nominal frequency, Let be the derivative of the wind farm power change with respect to time, when the disturbance is a power step disturbance. The derivative with respect to time can be considered as 0.

[0025] The network model is simplified using Kronos reduction, eliminating all load nodes and retaining only the inertial source potential node and the wind farm access node. From equation (1), taking the first row of the matrix equation yields: (5) (6) The new variable in formula (5) is shown in formula (6). Formula (5) indicates that the electromagnetic power imbalance borne by the inertial source unit after system disturbance mainly comes from: the power change caused by the change in the unit's power angle; the power change generated by the grid-connected wind farm participating in frequency regulation; and the power imbalance formed by the power disturbance of the system load node transmitted through the network.

[0026] For thermal power units, their rotor motion equations and frequency-modulated power responses can be expressed as: (7) (8) In the formula H, F, D, and T are diagonal matrices composed of the unit's inertia coefficient, frequency regulation coefficient, reheat coefficient, damping coefficient, and reheat time constant, respectively. As an inertial source with characteristics similar to a synchronous machine, the virtual rotor motion of grid-type energy storage is consistent with that of a synchronous machine and can be uniformly described by formula (7). However, grid-type energy storage participates in frequency regulation through converter control and does not have a reheat stage. To achieve a unified mathematical modeling form, this invention still uses formula (8) to describe the energy storage frequency regulation process, replacing the diagonal elements of the corresponding grid-type energy storage in matrix F with 0, and replacing the corresponding elements in matrix T with the energy storage response time constant, to reflect the rapid frequency regulation characteristics of grid-type energy storage that differ from those of thermal power units. Thus, under a unified inertial source modeling framework, by adjusting the elements in the parameter matrix, the differences in frequency regulation dynamic characteristics between thermal power units and grid-type energy storage are characterized, achieving unified modeling of the frequency regulation characteristics of the two types of inertial sources.

[0027] Frequency support for grid-connected wind farms is typically based on frequency measurements at the grid connection point, which are then used to generate frequency regulation power commands after a delay for tracking. The dynamic characteristics of its frequency regulation power response are modeled as a first-order inertial element, and its power change rate... for: (9) in Let be the diagonal matrix representing the time constant of the wind farm response. The diagonal matrix of the equivalent frequency modulation coefficients for the entire station is given by the above equation, which is then rearranged and transformed into a time-domain expression: (10) in, , Combining formulas (5)-(10), the system state variables can be summarized as follows: And expressed by a differential equation as: (11) (12) Considering that the frequency change is zero in the initial stage of the frequency disturbance, the expression for the rate of frequency change can be obtained by solving the differential equation: (13) in , For diagonal elements that are 1 A diagonal matrix. Diagonalizing matrix A and calculating the integral yields the expression for the frequency change of the inertial source unit: (14) in, Let X be a diagonal matrix consisting of all eigenvalues ​​of matrix A, and let X be the corresponding eigenvector matrix. Applying formula (14) to time Differentiation yields the expression for the frequency change rate of the inertial source unit: (15) in, Considering that the frequency regulation power of a wind farm will not change abruptly, the expression for its frequency regulation power change and its rate of change can be calculated in a similar manner as described above: (16) in Substituting equations (14) and (16) back into formula (4), we can obtain the expression for the frequency change of all nodes in the system. This constructs a wide-area power grid node frequency model under heterogeneous unit collaborative networking, which not only characterizes the regional frequency security features of the system but also provides a foundation for the subsequent optimization of the frequency regulation capability of regional wind farms.

[0028] S2: Construct a coupled and dependent model that considers the supporting characteristics of new energy sources and the strength of the system, and quantify the collaborative support capability of wind power participation in frequency regulation and thermal power generation for regional frequency security. This includes: assessing the frequency regulation support capability of wind farms, considering the upper and lower limits of wind turbine rotor speed and the rated overcurrent capacity constraints of converters, and allocating frequency regulation power among wind turbines based on the SOE index; assessing the feasible region of grid-type energy storage inertia-damping coefficient, constrained by the rated power of energy storage and state of charge; embedding the thermal power unit start-up status and new energy frequency regulation parameters as control variables into the power system node frequency model, quantifying the changes in system frequency strength under different combinations of thermal power start-up modes and new energy frequency regulation parameters, and using frequency security indicators as constraints to inversely constrain the feasible region of minimum thermal power start-up and new energy frequency regulation parameters.

[0029] Specifically, all wind turbines in the wind farm operate under maximum power point tracking (MPPT) control during normal operation, determining the base power output based on wind speed. When frequency disturbances occur, the turbines switch from MPPT operation mode to frequency regulation mode based on rotor kinetic energy release. Changes in electromagnetic power during frequency regulation lead to an imbalance between electromagnetic and mechanical power, causing changes in turbine rotor speed. Operating constraints on the turbines during frequency regulation include upper and lower limits of rotor speed and rated overcurrent capacity constraints of the converter. Based on these speed and current constraints, the maximum rotor kinetic energy that a single turbine can release is calculated, and frequency regulation power is allocated at the wind farm level according to the proportion of available rotor kinetic energy for each turbine, thereby obtaining the feasible range of the wind farm's frequency regulation coefficient. Simultaneously, for grid-type energy storage systems, frequency changes are suppressed by setting virtual inertia, and steady-state frequency difference is adjusted through virtual damping. Under the constraints of the energy storage's rated power and state of charge, the feasible region of the grid-type energy storage inertia-damping coefficient is evaluated. The operating status of thermal power units and the frequency regulation support capability of new energy sources are used as control variables affecting the dynamics of system frequency. They are also embedded in the node frequency model to quantify the changes in system frequency intensity under different combinations of thermal power start-up modes and new energy frequency regulation parameters. Frequency safety indicators are used as constraints to inversely constrain the feasible domain of minimum thermal power start-up and new energy frequency regulation parameters, thus forming a dependent relationship in which the support capability of new energy sources is dynamically adjusted with changes in system intensity.

[0030] A fan extracts kinetic energy from airflow and converts it into mechanical energy; its power is expressed as: (17) in, Let be the mechanical power of the a-th wind turbine in the w-th wind farm. air density, For wind speed, is the wind energy capture factor, and is the tip speed ratio. and pitch angle Functions: (18) (19) in, This refers to the tip speed ratio of the fan blades. This refers to the fan speed. and For the wind turbine blade radius and blade pitch angle.

[0031] During normal operation, the wind turbine employs a maximum power point tracking (MPPT) strategy. Based on this, the turbine can participate in frequency support by rapidly releasing rotor kinetic energy; however, its support capability is constrained by the converter's overcurrent capacity and the turbine's speed limits. The mismatch between the turbine's mechanical and electromagnetic power will lead to changes in rotor speed, a process that can be represented by the following first-order model: (20) in, The mechanical rotational inertia of the turbine blades. , , These represent the mechanical torque, electromagnetic torque, and electromagnetic power of the fan, respectively. Discretizing this differential equation yields: ;(twenty one) in, The discrete time interval is used to discretize the primary frequency regulation process into a time series. When the system disturbance frequency changes, the wind turbine initiates a primary frequency regulation response after a delay. The wake effect causes uneven wind speed distribution within the wind farm. To ensure safe and full utilization of the support capacity of each wind turbine, the SOE index is typically used to achieve a reasonable parameter allocation among the wind turbines. The kinetic energy stored in each wind turbine blade... With rotor speed Proportional to the square: ;(twenty two) The moment of inertia of the fan is taken into account, considering the fan speed limit. and The SOE of the wind turbine is represented as: ;(twenty three) in and To minimize and maximize stored kinetic energy, when SOE approaches 0, it indicates that the corresponding fan's regulation capacity is almost exhausted. Therefore, to achieve synchronized adjustment, the optimal power allocation scheme should ensure that the unbalanced power of each fan is proportional to SOE, i.e.: ;(twenty four) Then the fan outputs electromagnetic power It can be represented as: (25) (26) in, The frequency regulation power of the wind field can be obtained from formula (16). The frequency difference after disturbance is related to the frequency regulation coefficient of the wind field. Influence, Frequency regulation power borne by the wind turbines. Through the wind turbine SOE allocation strategy, the power and energy balance of different wind turbines participating in frequency regulation within the wind farm can be ensured, and the support capacity of all wind turbines in the site can be fully mobilized.

[0032] Grid-based energy storage possesses virtual inertia and damping characteristics similar to synchronous machines, enabling it to spontaneously respond to frequency disturbances. Its frequency support capability is primarily constrained by the converter's overcurrent capacity and is also affected by its charging and discharging states. This is related to the frequency dynamics after the disturbance and the preset inertia coefficient. With frequency modulation coefficient Related. The frequency dynamics of the grid-connected energy storage power station after system disturbance are obtained according to formulas (14) and (15). and frequency change rate Then the energy storage frequency regulation power and output power It can be represented as: (27) (28) (29) in, Let the capacity of the s-th energy storage power station be... For the charging and discharging power of the sth energy storage unit, This represents the maximum energy storage discharge power. It is used for energy storage and discharge backup.

[0033] Based on the boundary modeling of the frequency regulation capabilities of wind farms and grid-connected energy storage, the support coefficients of both for the system frequency are no longer fixed parameters, but rather dynamically adjustable decision variables constrained by the operating state. The frequency regulation capability of a wind farm is primarily determined by its frequency regulation coefficient. Its support margin is limited by the wind turbine rotor speed and power operation under the wind speed level; grid energy storage, on the other hand, relies on its inertia. With frequency modulation parameters Participating in frequency support, the support capability boundary is affected by the charging and discharging state of energy storage. Furthermore, the feasible domains of frequency regulation parameters for different wind farms and grid-connected energy storage do not exist independently, but are interdependent coupled boundaries formed dynamically by the system frequency. Adjustment of any support resource parameter will change the system frequency evolution process, thereby reshaping their respective feasible domains.

[0034] Secondly, the start-up status of thermal power units A two-way coupling relationship is formed between thermal power unit operating capacity and renewable energy frequency support capability. The operating scale of thermal power units directly determines the system frequency intensity. Increasing the operating capacity of thermal power units improves the system inertia and frequency regulation capability, and reduces the frequency deviation and rate of change after disturbance, thereby expanding the feasible domain of renewable energy frequency regulation parameters under given physical constraints. Conversely, improving renewable energy frequency regulation parameters can reduce the minimum thermal power unit operating requirement under frequency safety constraints. Thus, thermal power operating capacity, system frequency intensity, and renewable energy frequency regulation parameters form an interdependent coupling closed loop through the frequency dynamic response process.

[0035] Based on the above analysis, the controllable variable of the system frequency intensity can be expressed as: By embedding Q into the node frequency model and combining it with the station support capabilities, the spatiotemporal heterogeneity of system frequency and the dependence of frequency regulation parameters are integrated. Referring to Section 1, a time-varying state equation coefficient matrix is ​​constructed based on unit start-up mode, frequency regulation parameters, and system power flow. , And calculate the system frequency: (30) At this point, the supporting capacity of each wind farm and grid-connected energy storage power station is not only coupled with the system strength, but also coupled with the capacity of each station, which can be expressed as: ; (31-1) (31-2) (31-3) (31-4) (31-5) ; (32-1) (32-2) (32-3) (32-4) in, and These represent variables that influence the supporting capacity of wind farms and grid-connected energy storage. and As the corresponding physical boundary, the dependence and coupling relationship between the two is achieved through a function. This indicates that the function is implicit in the modeling of formulas (1)-(30). This represents the sum of the frequency regulation capabilities of all wind farms in the system except for the w-th wind farm. This represents the sum of the frequency regulation and inertia coefficients of all energy storage devices in the system except for the s-th energy storage power station. This is the wind speed vector of the wind turbine. For system power disturbances, disturbances of different locations and magnitudes will significantly affect the frequency response characteristics.

[0036] The aforementioned coupling relationship exhibits high-dimensional nonlinearity and is difficult to model explicitly. However, by incorporating the entire process of the field-grid frequency response under disturbance scenarios into a unified optimization model, a coordination mechanism among various wind farms can be effectively realized, thereby achieving a simultaneous improvement in the system's dynamic frequency performance and the collaborative support capability of wind farms.

[0037] S3: Based on frequency security constraints, construct an integrated optimization framework for unit combination and heterogeneous power source frequency support characteristics, and collaboratively optimize start-up methods and renewable energy frequency regulation parameters. This includes: constructing an upper-level thermal-wind-storage coordinated unit combination optimization model with the goal of minimizing operating costs, including system power balance constraints, line transmission capacity constraints, thermal power unit operation constraints, energy storage operation constraints, and system reserve capacity constraints; constructing a lower-level wind-storage time-varying maximum frequency regulation capability assessment model with the goal of maximizing renewable energy frequency regulation coefficients, including physical constraints and line power flow constraints; and constructing regional frequency security constraints, iteratively optimizing thermal power start-up schemes and renewable energy frequency regulation parameters through frequency security verification, where frequency security verification is based on the region's maximum frequency change rate, maximum frequency difference, and quasi-steady-state frequency difference indicators.

[0038] Specifically, based on the construction of a node frequency model and a new energy support capability dependency model, a thermal power unit combination model is first constructed within the scheduling cycle, with the goal of minimizing operating costs. Constraints include power balance, unit start-up and shutdown states, minimum output, ramp-up capability, and reserve capacity constraints. This yields the initial thermal power start-up scheme and the basic energy storage operation scheme. Subsequently, under this start-up scheme, based on wind speed prediction and the corresponding wind turbine operating states, the configurable frequency regulation coefficients for each wind farm are calculated. The virtual inertia and frequency regulation parameters of the grid-type energy storage are determined based on the energy storage operation scheme. These new energy frequency regulation parameters are then substituted into the node frequency model to perform dynamic frequency calculations for preset disturbance scenarios, constructing regional frequency safety constraints and incorporating them into the thermal power start-up optimization decision. If the frequency indicators do not meet the safety threshold, the thermal power start-up scheme is adjusted according to the impact of unit start-up and shutdown state changes on the frequency response, and the new energy frequency regulation parameters are re-evaluated. By repeatedly executing the unit combination solution, frequency regulation parameter configuration, and frequency safety verification process, a thermal power start-up scheme and new energy frequency regulation parameter combination that meets both frequency safety and new energy operation safety constraints are obtained.

[0039] For the thermal-wind-storage coordinated unit combination optimization model, the first step is to establish a unit combination model that considers the basic operating constraints of the system to determine the thermal power start-up scheme. .

[0040] (33) This model aims to minimize the system's operating cost. In the formula, These are the unit operation and start-up / shutdown costs, respectively. For energy storage operating costs, These are backup compensations for thermal power and energy storage, respectively. These refer to thermal power output and energy storage charging and discharging power, respectively. For the 0-1 variables of the start / stop flag of thermal power units, These are thermal power and energy storage reserve capacity, respectively.

[0041] The constraints include: ; 1) System power balance constraints: (34-1) in, These represent the wind farm output and the nodal load, respectively. These are collections of thermal power units, energy storage power stations, wind farms, and load nodes.

[0042] 2) Line transmission capacity constraints: (34-2) (34-3) in, The transmission power and maximum transmission power of branch l. Power injection for node k, Let be the power generation transfer distribution factor of node k to branch l.

[0043] 3) Operating constraints of thermal power units: (35-1) (35-2) (35-3) (35-4) (35-5) (35-6) (35-7) in, For the start-up and shutdown status of thermal power units, 0-1 variables For the maximum and minimum limit output of thermal power, For thermal power ramp-up power, These are the minimum start-up and shutdown times for thermal power plants.

[0044] 4) Energy storage operation constraints: (36-1) (36-2) (36-3) (36-4) (36-5) (36-6) (36-7) in, Let be the stored energy at time t and the maximum and minimum values, respectively. This refers to the charge / discharge efficiency.

[0045] 5) System standby capacity constraints: (37) in, For the system's N-1 spare capacity, For wind power uncertainty reserve ratio, this constraint is used to ensure that the total system reserve capacity is not lower than the required minimum reserve level.

[0046] The aforementioned unit combination model is a mixed-integer linear programming problem, which can be solved directly using existing commercial solvers, thereby obtaining the system unit combination scheme. For the assessment model and solution method of wind-storage time-varying maximum frequency regulation capability: based on the current unit combination scheme. Under the physical boundary constraints of energy storage power stations and wind farms, the maximum frequency regulation coefficient of each wind farm and grid-connected energy storage power station is used as the objective to evaluate the maximum frequency regulation capability of new energy under the current system intensity. The objective function can be expressed as: (38) The constraints include: 1) Physical constraints of frequency regulation process in wind farms and energy storage power stations: Equations (30)-(32); 2) Frequency-modulated quasi-steady-state line power flow constraints: (39-1) (39-2) Equation (39) represents the line transmission power constraint during the quasi-steady-state stage of primary frequency modulation. This is the primary frequency regulation exit time for the wind farm. This layer of the model optimizes the maximum frequency support coefficient for each wind farm under the current unit combination scheme by embedding a dependent coupling model of energy storage and wind farm support capabilities.

[0047] Since this process involves a multivariable, strongly nonlinear continuous parameter coupling optimization problem, this invention employs a Bayesian optimization method based on the Kriging surrogate model for solution. In this process, the operating state of the thermal power unit and the system topology remain fixed; only the continuous variable optimization of the frequency regulation parameters of the wind farm and energy storage power station is involved. This allows the method to maintain optimal results while possessing high computational efficiency. The optimization variables in this model are: The constraints (30)-(32) and (39) are treated as penalty terms. At this point, the objective function (38) is equivalent to: (40) By pre-calculating a given set of samples ,in Represents the set of observation points, observation points This is a vector composed of the frequency regulation coefficients of each station. This represents the set of values ​​corresponding to the objective function. Based on the sample set, the Kriging model can model a Gaussian distribution and predict new observations. Mean and variance: (41) in, and express The mean and variance of the sample set can be obtained from the covariance function of the sample set. Bayesian optimization uses this information to guide the search for optimal parameters and utilizes the expected improvement function. Determine new observation points : (42) in, and It is the cumulative sum probability density function of the standard normal distribution. New observation points are evaluated by solving for the maximum value of the EI function. Then calculate the true target value at the new point. and the historical minimum value Compare and update the optimal value. New observation point and its function value. The Kriging model will be continuously updated by adding samples to the sample set. The optimization process will iterate until the EI converges to the preset threshold, and then, under the gradual guidance, the optimal solution will be approximated, realizing the nonlinear optimization evaluation of the maximum frequency regulation coefficient of the station in each time period.

[0048] For the regional frequency security constraint embedding and the integrated optimization architecture of unit combination-frequency regulation characteristics, based on the current unit combination scheme and the maximum frequency regulation parameters of the wind farm Maximum supporting parameters for grid-connected energy storage First, a system-wide frequency security check is performed. If the check passes, the current result is the minimum power-on mode; otherwise, iteration continues. This invention uses the maximum frequency change rate, maximum frequency difference, and quasi-steady-state frequency difference of the region as check indicators to quantitatively describe the most severe frequency degradation.

[0049] (43) Based on the current unit start-stop status For different regional load disturbance scenarios, the impact of unit state changes on frequency safety indicators is evaluated by reversing the start-stop state of the units within the disturbance area using a numerical difference method, and a start-stop reversal sensitivity function is defined. Based on this, a frequency-safe cutting plane is constructed: (44) in, To divide the system into regional indexes, Defined as: under the condition that the states of the other units remain unchanged, the function is expressed as the sum of the operating state index and the shutdown state index of the i-th unit; therefore, the value of this function is always less than 0. This function is used to construct the system's linear secant plane frequency safety constraint, specifically expressed as: (45) If the current number is Taiwan unit offline The unit was switched to the start-up state. ,at this time This indicates an improvement in the frequency index. Conversely, if the first... Taiwan unit online Flip it to the power off state. Then at this time This indicates that the frequency index deteriorated after the unit was shut down.

[0050] This constraint characterizes the marginal impact of unit start-up and shutdown decisions on frequency safety margin. If the current frequency index does not meet the safety boundary, the optimization model increases the frequency safety margin by putting previously shut-down units that contribute significantly to frequency support into operation. When the frequency safety constraint is met, the model reduces start-up costs by shutting down units that contribute less to frequency support, gradually approaching the frequency safety boundary. This constraint maintains linear compatibility with the unit combination model and does not change the solution structure of other models, thus enabling complex regional frequency safety constraints to be effectively embedded in the unit combination model framework and achieving the minimum start-up solution through multiple iterations.

[0051] Based on equations (33)-(42), a unified scheduling optimization framework is constructed based on the dependence of wind-storage support capacity and system strength evolution, and the optimization process is optimized. Figure 3As shown, the start-up mode of thermal power plants and the frequency regulation parameters of new energy sources are optimized in a coordinated manner, thereby achieving the safety guarantee of steady-state operation of the system and dynamic regional frequency under the constraints of source-grid coordination.

[0052] Specifically, the optimization process mainly includes the following steps: 1) First, obtain the 24-hour load and wind power forecast data, network topology and unit parameters for the next day. Ignore frequency security constraints in the first iteration and solve for the initial unit combination with the minimum start-up cost; 2) Subsequently, based on wind speed forecast and energy storage charging and discharging status, evaluate the maximum frequency support coefficient of wind farms and grid-type energy storage in each time period under the current start-up mode; 3) Verify the regional frequency security, calculate the limit frequency security index of the system after regional disturbance. If the index meets the security boundary, determine the final unit combination and frequency regulation parameters. Otherwise, generate regional frequency security constraints and embed them into the unit combination model. Through the iterative process of "unit combination - support capability assessment - frequency verification and constraint embedding", continuously correct the feasible region until the results converge, and realize the coordinated optimization of start-up mode and new energy frequency regulation parameters.

[0053] This invention first constructs a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment, characterizing the spatiotemporal heterogeneity of frequency dynamic evolution after a disturbance in a wide-area power grid. Then, it establishes a coupled and dependent model of new energy support characteristics and system strength, assesses the maximum feasible region for new energy units to participate in frequency support, and quantifies the collaborative support capability of wind power participation in frequency regulation and thermal power generation for regional frequency security. Finally, based on frequency security constraints, it constructs an integrated optimization framework for unit combination and frequency characteristics, collaboratively optimizing the start-up mode and wind power frequency regulation parameters. The resulting dispatch scheme can fully release the support capacity of wind farms and energy storage systems, effectively reduce thermal power generation, and simultaneously ensure regional frequency security of the system.

[0054] In one embodiment, the invention is verified using a modified IEEE 39-node system, transforming the eastern section of the system into a large-scale centralized wind power grid-connected area, such as... Figure 4 As shown. The thermal power units at nodes 34, 35, 36, and 38 are replaced with wind farms of equal capacity. The total installed capacity of thermal power in the system is 4700MW, and the installed capacity of wind power is approximately 2700MW. The peak and valley loads are 5700MW and 3800MW respectively. A frequency safety boundary is set. The corresponding frequencies are 1Hz / s, 0.5Hz, and 0.25Hz. To verify the improvement effect of the method of the present invention on the low-output start-up standby of thermal power units, this embodiment sets up multiple sets of operating conditions as shown in Table 1 for comparison: Table 1 Operating Condition Settings Thermal power plant start-up methods in scenarios 1-4 are as follows Figure 5As shown, the costs are 2.51 million, 2.24 million, 2.22 million, and 2.28 million yuan, respectively, and the standby rates of thermal power units are 37.9%, 14.7%, 13.3%, and 16.5%, respectively. From the perspective of energy balance and relocation, energy storage can perform peak shaving through flexible charging and discharging, improve the wind power absorption rate, and provide backup for the system, reducing the standby rate of thermal power units. At the same time, the start-up and shutdown decisions of thermal power units are not only affected by operating costs, but also directly related to the regional frequency security level. Due to the instantaneous inertia response of the inertia source provided by the grid-connected wind farm, Area 3, the concentrated wind power area, is a low inertia area of ​​the system. This area faces the problem of a sharp increase in RoCoF after disturbance. Therefore, the spatial distribution and inertia size of thermal power start-up and grid-connected energy storage will directly determine the maximum RoCoF of the nodes. In Scenario 1, G4 is a frequency security-critical unit in Area 3 and cannot be shut down at any time. After grid-connected energy storage is connected to the grid, it can alleviate the problem of RoCoF shortage in the region. When the grid-connected energy storage has sufficient capacity and can provide more inertia support, G4 can even be shut down for part of the time (Scenario 2, t=24h).

[0055] Based on the results of scenarios 2 and 5, a comparative analysis was conducted on the system frequency dynamic safety boundary and the wind turbine operation safety boundary before and after optimization, such as... Figure 6 and Figure 7 As shown. Figure 6 The minimum number of thermal power units started in two scenarios (t=24h) was selected, and a uniform disturbance was applied to Area3 with the disturbance size equal to the maximum unit capacity. The frequency dynamics of the two scenarios were compared. After optimization, the system frequency safety indicators of Scenario 2 were all limited to within the safety threshold. Figure 6 To define the safe operating boundaries of the wind turbine within WF2 under both operating conditions, during low wind speed periods, both the turbine rotor speed and mechanical output power are at relatively low levels. The turbine mainly participates in frequency regulation by releasing rotor kinetic energy, and its frequency support capability is primarily constrained by the lower limit of rotor speed. During medium to high wind speed periods, the turbine rotor kinetic energy reserve increases significantly, and the frequency regulation capability becomes constrained by the converter's overcurrent capability. If parameter optimization is not performed and fixed coefficients or proportional allocation is used, physical operating boundary exceedance issues are likely to occur in different wind speed ranges.

[0056] Based on the results of scenarios 2 and 6, a comparative analysis is conducted on the interdependent coupling relationship between thermal power plant start-up and renewable energy frequency regulation capabilities through system strength and frequency dynamics. Figure 8(a) and (b) show the total frequency regulation support provided by wind farms and grid-connected energy storage during all time periods throughout the day in scenarios 2 and 6, respectively, and the lowest system frequency after disturbance. The support capacity of grid-connected energy storage is significantly affected by the charging and discharging state. In scenario 2, during the morning and evening off-peak periods, it has a larger discharge margin when in a charging state, thus allowing for the allocation of a larger frequency regulation coefficient. The frequency regulation capacity of wind farms is mainly constrained by wind speed levels and unit operating status. It is worth noting that although the number of thermal power units in operation is higher during peak load periods (12–18h) than during morning and evening off-peak periods, through synergistic optimization of the support parameters of wind farms and grid-connected energy storage, under the same disturbance conditions, the maximum frequency difference or maximum frequency change rate is actually smaller during some off-peak periods. This result indicates that when the support capacity of wind farms and grid-connected energy storage is fully released and the improvement in system strength reaches a certain level, their short-term frequency support effect can, to some extent, surpass that of thermal power unit operation. In contrast, Scenario 6 does not incorporate wind power and energy storage support capabilities into frequency security constraints. To meet frequency security requirements, a high level of thermal power generation must be maintained, requiring units G1–G6 to operate 24 / 7. Although the system strength is higher under this operating mode, wind farms and grid-connected energy storage can be configured with larger support coefficients, and the maximum frequency difference under the same disturbance can be reduced. However, this comes at the cost of a significant increase in thermal power generation. In this scenario, the thermal power reserve rate reaches as high as 38%, and the total system startup cost is 3.2 million yuan.

[0057] Based on the above comparisons, in systems with a high proportion of renewable energy penetration, relying solely on thermal power units for frequency backup will significantly increase the scale of power generation and operating costs. However, by characterizing the interdependent coupling relationship between thermal power generation, system frequency intensity, and wind power support capacity, and by coordinating and optimizing thermal power generation methods and wind power frequency regulation parameters within the physical operating boundaries of renewable energy power plants, it is possible to effectively improve system economy and clean energy absorption capacity while ensuring frequency security.

[0058] This invention first constructs a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment, characterizing the spatiotemporal heterogeneity of frequency dynamic evolution after a disturbance in a wide-area power grid. Then, it establishes a coupled and dependent model of new energy support characteristics and system strength, assesses the maximum feasible region for new energy units to participate in frequency support, and quantifies the collaborative support capability of wind power participation in frequency regulation and thermal power generation for regional frequency security. Finally, based on frequency security constraints, it constructs an integrated optimization framework for unit combination and frequency characteristics, collaboratively optimizing the start-up mode and wind power frequency regulation parameters. The resulting dispatch scheme can fully release the support capacity of wind farms and energy storage systems, effectively reduce thermal power generation, and simultaneously ensure regional frequency security of the system.

[0059] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0060] It will be readily understood by those skilled in the art that this invention includes any combination of the inventive description and specific embodiments outlined in the foregoing specification, as well as the various parts shown in the accompanying drawings. Due to space limitations and for the sake of brevity, not all of these combinations have been described in detail. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

[0061] Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention without departing from the principles and spirit of the invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A collaborative optimization method for wind power systems that considers spatiotemporal heterogeneity and coupling dependence, characterized in that, include: Construct a power system node frequency model that takes into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment; Construct a coupled and interdependent model that considers the supporting characteristics of new energy sources and the strength of the system, and quantify the collaborative support capabilities of wind power participation in frequency regulation and thermal power generation for regional frequency security. Based on frequency security constraints, an integrated optimization framework is constructed to combine unit combinations and heterogeneous power supply frequency support characteristics, and to collaboratively optimize start-up methods and new energy frequency regulation parameters.

2. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 1, characterized in that, The construction of power system node frequency models that take into account the differences in power frequency characteristics between thermal power plants and grid-connected power electronic equipment includes: By introducing the potential node in the synchronous machine and the virtual potential node in the grid-type converter, the algebraic constraints are eliminated through network equivalent order reduction, while preserving the dynamic coupling relationship between the inertial source node and the new energy access node; Electromechanical transient model is used to describe the power frequency response characteristics of thermal power units, synchronous machine rotor dynamic model is used to describe grid-type energy storage converter, and first-order inertial element is used to describe the frequency regulation power response characteristics of grid-type new energy. The inertial response and primary frequency regulation process of the heterogeneous units after disturbance are modeled into a unified first-order matrix differential equation, and the node frequency expression is obtained by solving the differential equation through eigenvalue decomposition.

3. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 1, characterized in that, The coupling and dependence model based on frequency security constraints and the characteristics of new energy support and system strength includes: To assess the frequency regulation support capability of a wind farm, considering the upper and lower limits of the wind turbine rotor speed and the rated overcurrent capacity constraints of the converter, frequency regulation power is allocated among the wind turbines based on the SOE index. The feasible region for evaluating the inertia-damping coefficient of grid-type energy storage is constrained by the rated power and state of charge of the energy storage. The starting status of thermal power units and the frequency regulation parameters of new energy sources are embedded as control variables into the power system node frequency model. The system frequency intensity changes under different combinations of thermal power start-up modes and new energy frequency regulation parameters are quantified. The frequency safety index is used as a constraint condition to inversely constrain the feasible domain of minimum thermal power start-up and new energy frequency regulation parameters.

4. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 1, characterized in that, The framework for integrating and optimizing the frequency support characteristics of unit combination and heterogeneous power sources includes: Construct an upper-level thermal-wind-storage coordinated unit combination optimization model with the goal of minimizing operating costs, and constrain the system power balance, line transmission capacity, thermal power unit operation, energy storage operation, and system reserve capacity. A time-varying maximum frequency regulation capability assessment model for the lower-level wind-storage system is constructed with the goal of maximizing the frequency regulation coefficient of new energy sources. The constraints include physical constraints and line power flow constraints. Construct regional frequency security constraints, and iteratively optimize thermal power plant start-up schemes and new energy frequency regulation parameters through frequency security verification. The frequency security verification is based on the region's maximum frequency change rate, maximum frequency difference, and quasi-steady-state frequency difference indicators.

5. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 2, characterized in that, The power system node frequency model describes the dynamic process of frequency evolution of all nodes after a wide-area system disturbance. Potential nodes are introduced for thermal power units and grid-connected energy storage converters, connected to the unit's grid connection point through generator transient reactance or grid-connected energy storage virtual impedance. The active power-voltage phase angle relationship for all nodes is as follows: ； In this context, the subscripts G, C, and L of each variable represent the inertial source potential node, the wind farm grid-connected node, and the load node including the inertial source terminal node, respectively, with the following quantities: , , , These are the power change vectors for the aforementioned nodes. , , The phase angle changes of the nodes mentioned above are represented by the coefficient matrix on the right side of the equation. The system's Laplace matrix, whose block elements can be calculated from its constituent nodes using the following formula: ； in Let the initial voltage of bus n be denoted by the initial voltage value of the extended potential node, which is represented by the generator potential E. , and Let N be the initial phase angle difference, conductance, and susceptance of the branch nm.

6. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 2, characterized in that, The electromechanical transient model of the thermal power unit is as follows: ； ； Among them, H, F, D, and T are diagonal matrices composed of the unit's inertia coefficient, frequency regulation coefficient, reheat coefficient, damping coefficient, and reheat time constant, respectively.

7. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 3, characterized in that, The SOE index is expressed as follows: ； in, and For minimum and maximum stored kinetic energy, The kinetic energy stored in each wind turbine blade, For the fan rotor speed, and Minimum and maximum fan rotor speed limits.

8. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 7, characterized in that, The frequency regulation power is distributed among the fans as follows: ； ； in, For wind farm frequency modulation power, Frequency regulation power supplied to the wind turbine.

9. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 4, characterized in that, The optimization model for the upper-level thermal-wind-storage coordinated unit combination is as follows: ； The evaluation model for the time-varying maximum frequency regulation capability of the lower-level wind-storage system is as follows: ； in, These are the unit operation and start-up / shutdown costs, respectively. For energy storage operating costs, These are backup compensations for thermal power and energy storage, respectively. These refer to thermal power output and energy storage charging and discharging power, respectively. For the 0-1 variables of the start / stop flag of thermal power units, These are thermal power and energy storage reserve capacity, respectively.

10. The wind power system collaborative optimization method considering spatiotemporal heterogeneity and coupling dependence as described in claim 4, characterized in that, Frequency security constraints are as follows: ； Among them, if the current number Taiwan unit offline The unit was switched to the start-up state. ,at this time This indicates an improvement in the frequency index; if the first Taiwan unit online Flip it to the power off state. Then at this time This indicates that the frequency index deteriorated after the unit was shut down.

Images