Online Identification and Evaluation Method and System for Virtual Primary Frequency Regulation of Wind and Solar Power Units

Abstract

This invention relates to the field of frequency control technology for new energy systems, and particularly to a method and system for online identification and evaluation of virtual primary frequency regulation for wind and solar turbine units. The method includes: determining the steady-state output power benchmark, maximum available power, and reserve capacity of the doubly-fed induction generator (DFIG) and photovoltaic (PV) units; collecting frequency data and active power data at the grid connection point and preprocessing them to obtain time-series data; dividing the time-series data into stages, and identifying the actual effective parameters corresponding to the virtual primary frequency regulation control components online during the steady-state frequency regulation stage; identifying the actual effective parameters corresponding to the virtual inertial control components online during the transient inertial stage; and evaluating the validity of the parameters based on the actual effective parameters, reserve capacity, and controller preset parameters to determine whether the DFIG and PV units have entered a power saturation state. This invention effectively solves the problem that existing new energy units easily enter a power saturation state under large disturbance conditions, causing theoretical control parameters to fail to reflect actual frequency regulation capabilities.

Description

Technical Field

[0001] This invention relates to the field of frequency control technology for new energy systems, and in particular to a method and system for online identification and evaluation of virtual primary frequency regulation of wind and solar turbine units. Background Technology

[0002] With the acceleration of the global energy transition, building a new power system dominated by new energy sources has become an inevitable trend. New energy units such as wind power and photovoltaic power, connected to the grid via power electronic converters, are replacing traditional synchronous generators on a large scale. However, this replacement has led to a significant decrease in the physical rotational inertia of the power system, weakening its ability to withstand disturbances. To address this challenge, grid connection guidelines in various countries require new energy units to have primary frequency regulation capabilities, that is, to simulate the frequency response characteristics of synchronous machines through control strategies.

[0003] Currently, most mainstream new energy generating units adopt grid-following control strategies, reserving a certain amount of standby capacity (i.e., load shedding) on ​​top of maximum power point tracking (MPPT), and introducing virtual inertial control (corresponding to the differential element) and virtual primary frequency regulation control (corresponding to the proportional element, i.e., droop control). However, unlike traditional thermal power units with relatively abundant fuel supply and a clearly defined regulation range, the frequency regulation capability of new energy units is strictly limited by fluctuations in their primary energy sources (wind speed, solar radiation) and converter capacity, exhibiting significant physical limiting characteristics. Specifically: there are capacity limitations. When a large disturbance occurs in the grid, causing a large frequency deviation, the target power increase calculated based on theoretical parameters may exceed the unit's current available standby capacity or reach the converter's power limit. At this time, the unit will enter the nonlinear saturation region, and the actual output power will be limited to the rated value or maximum available power, resulting in the actual effective equivalent primary frequency regulation coefficient being much smaller than the set value. Furthermore, existing evaluation methods have shortcomings. Most existing evaluation methods are based on offline simulation or assume that the unit is always in the linear region, lacking online evaluation methods for the accuracy of parameters under actual operating conditions. When disturbances occur, the dispatch center often cannot distinguish whether the controller parameter settings are effective or whether the unit is saturated, resulting in insufficient frequency regulation capability.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a method and system for online identification and evaluation of virtual primary frequency regulation of wind and solar power units, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for online identification and evaluation of virtual primary frequency regulation for wind and solar turbine units, the method comprising: Obtain the primary energy parameters and operating parameters of the doubly-fed wind turbine and the photovoltaic unit, and determine the steady-state output power reference, maximum available power and reserve capacity of the doubly-fed wind turbine and the photovoltaic unit; A frequency response control model including a virtual primary frequency modulation control component and a virtual inertial control component is constructed based on the steady-state output power reference. Frequency data and active power data of the grid connection points of the doubly fed wind turbine and the photovoltaic unit are collected and preprocessed to obtain time-series data; Based on the time-series data, the steady-state frequency modulation stage and the transient inertial stage are divided, and the actual effective parameters corresponding to the virtual primary frequency modulation control component are identified online during the steady-state frequency modulation stage. Based on the actual effective parameters corresponding to the virtual primary frequency modulation control component, the actual effective parameters corresponding to the virtual inertial control component are identified online during the transient inertial phase; The validity of the parameters is evaluated based on the actual effective parameters, the reserve capacity, and the controller preset parameters to determine whether the doubly fed wind turbine and the photovoltaic unit have entered a power saturation state.

[0007] Further, determining the operating parameters of the doubly-fed wind turbine and the photovoltaic unit includes: Based on the aerodynamic model of the doubly fed fan and the real-time wind speed, the optimal mechanical power of the doubly fed fan under maximum power point tracking (MPPT) conditions is calculated. The reduced-load operating power of the doubly-fed fan is determined based on the optimal mechanical power of the doubly-fed fan and the set reduced-load rate. The maximum available power of the photovoltaic unit is calculated based on real-time light intensity and ambient temperature. The steady-state output power reference of the photovoltaic unit is determined based on the maximum available power of the photovoltaic unit and the set load reduction rate; The reserve capacity reserved for frequency response by the doubly-fed wind turbine and the photovoltaic unit is determined based on the reduced-load operating power of the doubly-fed wind turbine and the steady-state output power of the photovoltaic unit.

[0008] Furthermore, a frequency response control model is constructed, including: An additional frequency control loop is introduced into the active power control circuit of the converter of the doubly fed wind turbine and the photovoltaic unit. The virtual primary frequency modulation control component is constructed based on the steady-state output power reference. The virtual inertial control components are constructed based on the rate of change of frequency; The steady-state output power reference, the virtual primary frequency control component, and the virtual inertial control component are superimposed to generate an active power reference command.

[0009] Furthermore, frequency data and active power data are collected and preprocessed, including: Real-time collection of frequency data and active power data at the grid connection points of the doubly-fed wind turbine and the photovoltaic unit; The frequency data and active power data are preprocessed using synchronous filtering to obtain a smooth sequence; The window length of the moving average filter is set to 200ms to eliminate noise and maintain the relative phase consistency between the frequency data and the active power data.

[0010] Furthermore, the steady-state frequency modulation stage and the transient inertial stage are divided into: Select a 2-second time window before the disturbance occurs; The median of the filtered frequency data within the time window is determined as the frequency reference value; The median of the filtered active power data within the time window is determined as the power reference value; Based on the time scale difference in the dynamic response of the doubly fed wind turbine and the photovoltaic unit, the post-disturbance process is divided into the steady-state frequency regulation stage and the transient inertial stage.

[0011] Furthermore, the actual effective parameters corresponding to the virtual primary frequency modulation control component are identified online, including: During the steady-state frequency modulation phase, the rapidly decaying inertial term is ignored, and a steady-state equation error model is established based on the frequency deviation and the active power increment. Construct a quadratic loss function based on the observation data sequences of system input and output; The actual effective parameters corresponding to the virtual primary frequency modulation control component are iteratively identified using the recursive least squares method with a forgetting factor. The actual effective parameters are updated online by updating the gain matrix, parameter estimates, and covariance matrix. The forgetting factor was set to 0.99 to obtain smooth steady-state parameter estimation results.

[0012] Furthermore, the actual effective parameters corresponding to the virtual inertial control components are identified online, including: During the transient inertial phase, the frequency modulation component is stripped off according to the actual effective parameters corresponding to the virtual primary frequency modulation control component; Calculate the remaining inertial response components based on the results after stripping. Calculate the frequency change rate based on the frequency data; Under the condition of satisfying the physical validity constraint criterion, the recursive least squares method with forgetting factor is used to identify the actual effective parameters corresponding to the virtual inertial control components online. The forgetting factor was set to 0.95 to enhance the dynamic tracking capability of changes in inertial response.

[0013] Further, parameter validity assessment is conducted, including: The actual effective parameters corresponding to the virtual primary frequency modulation control component are compared with the preset parameters of the controller to calculate the accuracy evaluation index. Calculate the maximum increase in power output of the doubly-fed wind turbine and the photovoltaic unit under the current operating conditions; The theoretical power requirement is calculated based on the controller's preset parameters and steady-state frequency deviation. When the accuracy evaluation index is within the allowable error range and the theoretical power demand is not greater than the maximum achievable power, it is determined that the controller preset parameters are valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region. When the theoretical power demand exceeds the maximum power that can be increased, the doubly fed wind turbine and the photovoltaic unit are determined to have entered a power saturation state, and the accuracy evaluation index is used to characterize the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability.

[0014] A virtual primary frequency regulation online identification and evaluation system for wind and solar turbine units, the system comprising: The information acquisition module acquires the primary energy parameters and operating parameters of the doubly-fed wind turbine and photovoltaic unit, and determines the steady-state output power benchmark, maximum available power and reserve capacity of the doubly-fed wind turbine and photovoltaic unit; The model building module constructs a frequency response control model based on the steady-state output power benchmark, including virtual primary frequency modulation control components and virtual inertial control components. The grid connection processing module collects frequency data and active power data at the grid connection points of doubly-fed wind turbines and photovoltaic units, and performs preprocessing to obtain time-series data; The steady-state identification module divides the steady-state frequency modulation stage and the transient inertial stage based on time-series data, and identifies the actual effective parameters corresponding to the virtual primary frequency modulation control components online during the steady-state frequency modulation stage. The transient identification module identifies the actual effective parameters corresponding to the virtual primary frequency control component online during the transient inertial phase, based on the actual effective parameters corresponding to the virtual primary frequency control component. The saturation assessment module evaluates the validity of parameters based on actual effective parameters, standby capacity, and controller preset parameters to determine whether the doubly fed wind turbine and photovoltaic unit have entered a power saturation state.

[0015] Furthermore, the saturation evaluation module includes: The parameter comparison unit compares the actual effective parameters corresponding to the virtual primary frequency modulation control component with the preset parameters of the controller to calculate the accuracy evaluation index. The maximum power generation unit calculates the maximum power that can be increased for the doubly-fed wind turbine and photovoltaic unit under the current operating conditions; The theoretical calculation unit calculates the theoretical power demand based on the controller's preset parameters and steady-state frequency deviation; The linearity determination unit determines that the controller preset parameters are valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region when the accuracy evaluation index is within the allowable error range and the theoretical power demand is not greater than the maximum power that can be generated. The saturation determination unit determines that the doubly fed wind turbine and photovoltaic unit have entered the power saturation state when the theoretical power demand exceeds the maximum power that can be generated. The accuracy evaluation index is used to characterize the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability.

[0016] The technical solution of this invention can achieve the following technical effects: Based on the maximum available power and load shedding reserve capacity of doubly-fed wind turbines and photovoltaic units under current weather and operating conditions, a frequency response control model including virtual primary frequency regulation control components and virtual inertial control components is constructed. Then, based on the measured data of grid connection point frequency and active power, after synchronous preprocessing and phased division, the virtual primary frequency regulation parameters and virtual inertial parameters are identified online. Finally, the effectiveness of the parameters and the power saturation state are evaluated by combining the identification results, the controller preset parameters, and the unit's available adjustment capabilities. This effectively solves the problem that existing new energy units are prone to entering a power saturation state under large disturbance conditions, resulting in theoretical control parameters that cannot reflect the actual frequency regulation capability.

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the online identification and evaluation method for virtual primary frequency regulation of wind and solar turbine units; Figure 2 This is a schematic diagram of the grid-connected topology of a doubly-fed wind turbine. Figure 3 A schematic diagram of the maximum power tracking and rotor overspeed unloading operation curves of a doubly fed wind turbine at different wind speeds; Figure 4 This is a schematic diagram of virtual inertia control for a doubly-fed wind turbine. Figure 5 This is a schematic diagram of a fast frequency modulation control strategy for photovoltaic units. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] Example 1; like Figure 1 , Figure 2 and Figure 3 As shown, this application provides a method for online identification and evaluation of virtual primary frequency regulation for wind and solar power units. The method includes: S10: Obtain the primary energy parameters and operating parameters of the doubly-fed wind turbine and photovoltaic unit, and determine the steady-state output power benchmark, maximum available power and reserve capacity of the doubly-fed wind turbine and photovoltaic unit; S20: Construct a frequency response control model based on the steady-state output power reference, including virtual primary frequency modulation control components and virtual inertial control components; S30: Collect frequency data and active power data at the grid connection point of the doubly-fed wind turbine and photovoltaic unit, and perform preprocessing to obtain time series data; S40: Based on time-series data, the steady-state frequency modulation stage and the transient inertial stage are divided, and the actual effective parameters corresponding to the virtual primary frequency modulation control components are identified online during the steady-state frequency modulation stage; S50: Based on the actual effective parameters corresponding to the virtual primary frequency modulation control components, online identification of the actual effective parameters corresponding to the virtual inertial control components is performed during the transient inertial phase; S60: Based on the actual effective parameters, standby capacity and controller preset parameters, perform parameter validity assessment to determine whether the doubly fed wind turbine and photovoltaic unit have entered the power saturation state.

[0023] Specifically, firstly, the available power boundaries are established based on the current primary energy parameters and operating parameters. For doubly-fed wind turbines, the optimal mechanical power under maximum power point tracking is determined using aerodynamic relationships and real-time wind speed. Furthermore, an overspeed load reduction strategy is employed to operate the turbine at a suboptimal power point, reserving reserve capacity for active power adjustment while sacrificing some steady-state output power. For photovoltaic (PV) units, the current maximum available power is estimated by combining real-time irradiance and ambient temperature. Then, a steady-state output power benchmark is determined according to a preset load reduction rate, similarly reserving reserve capacity for frequency support. This process does not simply use fixed rated values ​​as the basis for frequency regulation capability, but rather considers wind speed, irradiance, and temperature... The changes in generating capacity caused by the degree of stress are explicitly incorporated into the subsequent parameter evaluation basis, ensuring from the source that the subsequent identification results are consistent with the actual physical capacity of the unit. Next, an additional frequency control loop is introduced into the active power control circuit of the converter for both doubly-fed induction generators (DFIGs) and photovoltaic (PV) units. This allows the total active power reference command to be formed by superimposing the steady-state output power reference, the virtual primary frequency regulation control component, and the virtual inertial control component. The virtual primary frequency regulation control component provides continuous active power support based on frequency deviation, while the virtual inertial control component rapidly releases support power in the early stages of disturbance based on the frequency change rate. For DFIGs, virtual inertial support can preferably be achieved using rotor energy storage; for PV units, virtual inertial support can preferably be achieved using DC-side energy storage. By utilizing energy storage or reserved power, this frequency response control structure enables wind turbines and photovoltaic units to respond quickly in the initial stage of frequency drops and then continuously follow, thus forming a complete frequency support process. Simultaneously, it provides a unified control basis for subsequent identification of virtual primary frequency regulation parameters and virtual inertial parameters. During the online parameter identification stage, it is preferable to collect frequency and active power data from the grid connection points of the doubly-fed induction generator (DFIG) and photovoltaic units in real time. A synchronous moving average filtering method is preferred for preprocessing the two types of data, with a preferred filter window length of 200ms and a sampling period of 2ms. This setting aims to suppress high-frequency noise while maintaining the relative phase between the frequency and power signals. To mitigate phase mismatch and avoid identification errors, a further optimization is made to select a 2-second time window before the disturbance occurs. The median of the filtered frequency and active power data within this window is then used as the frequency and power reference values, respectively. This is because the median is less sensitive to outliers than the mean and is more suitable for actual noisy measurement scenarios. Subsequently, based on the differences in the dynamic response of wind and solar turbines over time, the disturbance process is divided into a steady-state frequency regulation stage and a transient inertial stage. The steady-state frequency regulation stage is preferably selected when the inertial component has decayed and the primary frequency regulation response remains. The transient inertial stage is preferably selected during the initial time period after the disturbance occurs, such as 0.5 seconds to 1 second after the disturbance.Within a 5-second interval; during the steady-state frequency modulation phase, it is preferable to ignore the rapidly decaying inertial component and establish a steady-state identification relationship based solely on the frequency deviation and active power increment. A recursive least squares algorithm with a forgetting factor is used for online identification, preferably with a forgetting factor close to 0.99 to ensure the algorithm has long memory and is insensitive to new noise, thus obtaining smooth and stable virtual primary frequency modulation actual effective parameters. Subsequently, during the transient inertial phase, the frequency modulation component is stripped using the previously identified virtual primary frequency modulation actual effective parameters to obtain the remaining inertial response component. This is combined with the frequency change rate to construct a second-stage identification relationship, and the actual effective parameters corresponding to the virtual inertial control component are identified online. In this stage, a smaller forgetting factor, such as 0.95, is preferred to enhance the tracking ability for rapid dynamic changes. Furthermore, it is preferable to update the inertial parameters only when the physical validity constraint criterion is met to avoid introducing errors from invalid data. This allows for a phased decoupled identification of the virtual primary frequency modulation parameters and virtual inertial parameters, compared to... The overall identification method, which does not distinguish between time periods, can more clearly differentiate the contributions of steady-state proportional support and transient inertia support, improving identification accuracy under complex disturbances and measurement noise conditions. Finally, based on the actual effective parameters obtained from online identification, the pre-determined reserve capacity, and the controller's preset parameters, parameter validity assessment and power saturation determination are performed. Preferably, the actual effective parameters of the virtual primary frequency regulation are compared with the controller's preset parameters to form an accuracy assessment index. Furthermore, the maximum achievable additional power output of the unit under the current operating conditions and the theoretical power demand obtained based on the controller's preset parameters and the steady-state frequency deviation are calculated. When the accuracy assessment index is within the allowable error range and the theoretical power demand is not greater than the maximum achievable additional power output, the controller parameters can be determined to be effective and the unit is operating in the linear region. When the theoretical power demand exceeds the maximum achievable additional power output, the unit is determined to have entered the power saturation region. At this time, the equivalent parameters obtained from identification will be significantly lower than the preset parameters and can be used to quantify the ratio of the unit's actual frequency regulation capability to its theoretical frequency regulation capability.

[0024] The technical solution of this invention constructs a frequency response control model containing virtual primary frequency regulation control components and virtual inertial control components based on the maximum available power and load reduction reserve capacity of doubly-fed wind turbines and photovoltaic units under current meteorological and operating conditions. Then, based on the measured data of grid connection point frequency and active power, the virtual primary frequency regulation parameters and virtual inertial parameters are identified online after synchronous preprocessing and phased division. Finally, the effectiveness of the parameters and the power saturation state are evaluated by combining the identification results, the controller preset parameters, and the unit's available adjustment capabilities. This effectively solves the problem that existing new energy units are prone to entering a power saturation state under large disturbance conditions, resulting in theoretical control parameters that cannot reflect the actual frequency regulation capabilities.

[0025] Furthermore, determining the operating parameters of the doubly-fed wind turbine and photovoltaic unit includes: Based on the aerodynamic model and real-time wind speed calculation of the doubly fed wind turbine, the optimal mechanical power of the doubly fed wind turbine under maximum power point tracking (MPPT) conditions is calculated. The reduced-load operating power of the doubly-fed fan is determined based on the optimal mechanical power and the set reduced-load rate of the doubly-fed fan. The maximum available power of the photovoltaic unit is calculated based on real-time light intensity and ambient temperature. The steady-state output power benchmark of the photovoltaic unit is determined based on the maximum available power of the photovoltaic unit and the set load shedding rate. The reserve capacity for frequency response of the doubly-fed induction generator (DFIG) and the photovoltaic (PV) unit is determined based on the reduced-load operating power of the DFIG and the steady-state output power of the PV unit.

[0026] As a preferred embodiment of the above embodiments, for a doubly-fed induction generator (DFIG), it is preferable to first characterize the process of the rotor capturing mechanical power from the incoming airflow based on an aerodynamic model, and then determine the maximum power point tracking (MPPT) operating state under the current wind conditions by combining real-time wind speed, rotor swept area, rotor radius, rotor mechanical angular velocity, and wind energy utilization coefficient. In the operating region where the wind speed is lower than the rated wind speed, the blade pitch angle is preferably kept constant. The control objective is to adjust the generator torque to maintain the tip speed ratio near the optimal tip speed ratio, thereby keeping the wind energy utilization coefficient at its maximum value. This yields the optimal mechanical power of the DFIG under maximum power point tracking. The advantages of this approach are... Therefore, the optimal mechanical power is not a static setpoint, but changes in real time with wind speed and rotational speed, truly reflecting the upper limit of wind energy that the turbine can capture. Based on this, to ensure the doubly-fed induction generator (DFIG) has the power margin required for primary frequency regulation, an overspeed load shedding strategy is preferred. This shifts the turbine from the optimal power point corresponding to maximum power tracking to the suboptimal power point. Specifically, this can be achieved by adjusting the generator's electromagnetic torque to appropriately increase the rotor speed, thereby causing the blade tip speed ratio to deviate from the optimal value and the wind energy utilization coefficient to decrease. Ultimately, the output power is reduced to the load shedding operating power, and the difference between the optimal mechanical power and the load shedding operating power forms the reserved reserve capacity. For example, such as... Figure 4As shown, at a reference wind speed of 11.4 m / s, the output power of a doubly-fed induction generator (DFIG) under maximum power point tracking (MPPT) is 5 MW. When the load shedding rate is set to 0.85, the load shedding operating power is 4.25 MW, with a corresponding reserve capacity of 0.75 MW. The turbine reserve is obtained directly through real-time power limits and load shedding operating power, rather than estimating the reserve afterward, which is more conducive to subsequent judgment of whether the unit has entered the saturation zone due to large disturbances. Similarly, for photovoltaic (PV) units, the preferred approach is not to simply use the current output power as the adjustable frequency reference, but to first estimate the maximum available power under the current operating conditions by combining real-time irradiance and ambient temperature. Specifically, the preferred approach is to use the rated maximum power at the reference irradiance and reference temperature as a basis, correct the photocurrent through real-time irradiance, and then correct the maximum power point voltage through the influence of temperature on the open-circuit voltage and the maximum power point voltage, thereby determining the maximum power output under the current environment. Since changes in irradiance have a more significant impact on current, and temperature... The effect of temperature on voltage is more significant, so this calculation method can more accurately reflect the upper limit of the actual available power of the photovoltaic array under different weather and temperature conditions. On this basis, the steady-state output power benchmark of the photovoltaic unit is determined according to the set load shedding rate, so that the photovoltaic unit retains a certain proportion of reserve power when the frequency is undisturbed, which is used to quickly increase the output when the frequency drops. In other words, the steady-state output power benchmark is preferably determined directly by the product of the current maximum available power and the load shedding rate, rather than by a fixed reduction of the rated value. This can avoid overestimating the frequency regulation potential of the photovoltaic unit under low irradiance or high temperature conditions. Furthermore, after the load shedding operating power has been obtained on the wind turbine side and the steady-state output power benchmark has been obtained on the photovoltaic side, it is preferable to calculate the adjustable power margin of the two types of units relative to their respective current maximum available power, and use this margin as the reserve capacity as the physical constraint basis for subsequent virtual primary frequency regulation control and parameter effectiveness evaluation.

[0027] Furthermore, such as Figure 5 As shown, a frequency response control model is constructed, including: An additional frequency control loop is introduced into the active power control circuit of the converter in doubly fed wind turbines and photovoltaic units; A virtual primary frequency modulation control component is constructed based on a steady-state output power reference. Virtual inertial control components are constructed based on the rate of change of frequency; The steady-state output power reference, the virtual primary frequency control component, and the virtual inertial control component are superimposed to generate the active power reference command.

[0028] As a preferred embodiment of the above, the frequency response control model is not an isolated addition to the original grid-connected control of the new energy unit, but rather preferably directly built into the active power control loop of the converter of each of the doubly-fed wind turbine and photovoltaic unit, so that the unit has an inertial response and primary frequency regulation response similar to that of a synchronous generator while maintaining its original grid-connected operation capability. The frequency response control model is composed of a steady-state reference power, a virtual droop control component, and a virtual inertial control component, and its function is to simulate the inertia and primary frequency regulation characteristics of a synchronous generator. Preferably, the virtual primary frequency regulation control component is based on the grid connection point. The construction of frequency deviation between real-time and rated frequency essentially converts the frequency deviation into a proportional active power regulation quantity to undertake the task of eliminating steady-state frequency deviation. In actual implementation, when the grid frequency is detected to be lower than the rated value, the virtual primary frequency regulation control component outputs a positive regulation command, causing doubly-fed induction generators to call up reserved reserve capacity and photovoltaic units to increase the output power by utilizing the active power margin retained after load reduction. When the frequency recovers or exceeds the rated value, the regulation quantity is reduced accordingly, thus forming a negative feedback regulation process. This control component based on steady-state frequency difference preferably plays a continuous support role, addressing the grid's... The issue of frequency deviation still needing correction in the later stages of disturbances; virtual droop control is used to simulate the characteristics of synchronous motor governors and to eliminate steady-state frequency deviation. When the frequency drops, wind turbines increase output using reserved reserve capacity, and photovoltaic units increase output according to the same principle. In conjunction with this, the virtual inertial control component is preferably constructed based on the frequency change rate. Its function is not to handle steady-state deviations, but to simulate the dynamic process of synchronous generator rotors releasing or absorbing kinetic energy when the frequency changes rapidly in the early stages of grid disturbances, thereby suppressing the frequency change rate and shortening the response start-up delay. For doubly-fed induction generators, this control component can preferably be constructed using... Rapidly adjusting the electromagnetic power of the converter converts the rotor kinetic energy stored in the wind turbine, gearbox, and generator rotor into short-term active power support, thus enabling it to take effect before droop control during the rapid frequency drop phase. For photovoltaic units, although they lack physical rotational inertia, virtual inertia control is introduced to simulate the dynamic inertial process of the synchronous machine rotor, and rapid support is achieved by utilizing the photovoltaic DC-side energy storage characteristics or reserved power. Therefore, it is preferable to use the same RoCoF trigger-instantaneous support control logic for both wind turbines and photovoltaic units, only the actual energy sources correspond to rotor kinetic energy and photovoltaic-side callable power, respectively.More preferably, in the formation of the total active power reference command, the virtual inertial control component and the virtual primary frequency regulation control component are not mutually exclusive, but rather have a cooperative relationship with different time scales and functional divisions. That is, in the initial stage after the frequency disturbance occurs, the control system first causes the virtual inertial control component to act rapidly according to the frequency change rate, prioritizing the provision of rapid but usually short-duration instantaneous support. Subsequently, as the frequency deviation gradually appears and enters a relatively slow stage, the virtual primary frequency regulation control component then undertakes the task of continuous support. Therefore, when the two are superimposed with the steady-state output power reference, they can form an active power response covering the entire process from the initial disturbance stage to the steady-state recovery stage. For example, in a doubly-fed induction generator (DFIG) wind turbine embodiment, the steady-state output power reference after load reduction can be determined first based on the current wind speed, and this value is used as the basic active power command when the frequency is undisturbed. When the grid connection frequency drops rapidly, the control loop first forms a virtual inertial control component based on the frequency change rate, causing the turbine to release rotor kinetic energy briefly to increase output. Subsequently, a virtual primary frequency regulation control component is formed based on the persistent frequency deviation, further utilizing reserved standby capacity to maintain support. In a photovoltaic (PV) unit embodiment, the steady-state output power reference after load reduction can be determined first based on real-time irradiance and ambient temperature, and the final active power reference command is generated according to the same superposition logic when there is a frequency disturbance.

[0029] Furthermore, frequency data and active power data are collected and preprocessed, including: Real-time acquisition of frequency and active power data at grid connection points of doubly-fed wind turbines and photovoltaic units; Synchronous filtering is used to preprocess frequency data and active power data to obtain a smooth sequence; The window length of the moving average filter is set to 200ms to eliminate noise and maintain the relative phase consistency between frequency data and active power data.

[0030] As a preferred embodiment of the above embodiments, to ensure that the frequency response data used for subsequent online identification can accurately reflect the dynamic characteristics of grid disturbances caused by doubly-fed induction generators (DFIGs) and photovoltaic (PV) units, and to avoid introducing additional identification errors due to measurement noise and signal misalignment, it is preferable to adopt a synchronous acquisition and synchronous preprocessing method for the frequency data and active power data at the grid connection point. Specifically, the electrical quantities at the grid connection point of the DFIGs and PV units are preferably sampled continuously in real time, and the frequency time series and active power time series are calculated from the sampled data. The sampling period can preferably be set to 2ms to fully preserve the rapid frequency changes and power response details at the initial stage of the disturbance. Due to practical... Grid-connected measurement signals typically contain high-frequency noise, and subsequent parameter identification involves calculations of frequency deviation, power increment, and frequency change rate. If the raw data is used directly, the noise will be further amplified during differential and recursive identification. Therefore, it is preferable to perform filtering before the data enters the identification process. Ideally, synchronous filtering should be used to process both types of data simultaneously, ensuring they undergo the same filtering delay and smoothing process. This suppresses noise while maintaining their relative phase consistency, avoiding relative phase shifts between the frequency and power signals caused by inconsistent processing links. Preferably, synchronous filtering can be implemented using a moving average filter. The window length is set to 200ms. This setting effectively reduces high-frequency disturbances in grid-connected measurements, smoothing the frequency and active power sequences, while avoiding excessive stretching of the response characteristics, thus preserving key dynamic information suitable for subsequent window division and parameter identification. Although the two signal sequences experience a consistent delay after using this moving average filter, their relative phase relationship remains unchanged because the frequency and active power data use the exact same window length and synchronous update mechanism. Therefore, the frequency deviation, power increment, and frequency change rate constructed based on the filtered data can more accurately represent... The unit response status at the same moment is crucial for subsequent phased identification of virtual primary frequency regulation parameters and virtual inertial parameters. For example, when frequency disturbances occur at the grid connection point of wind and solar power plants, if the original sampled signal is directly identified, small measurement noise may cause drastic fluctuations in the rate of frequency change, leading to distortion in the inertial response estimation. However, by using synchronous moving average filtering with a sampling period of 2ms and a window length of 200ms, the frequency curve and active power curve can be significantly denoised while maintaining relative synchronization, making the subsequently extracted dynamic features more stable and reliable, thereby improving the accuracy and robustness of the online identification results.

[0031] Furthermore, the steady-state frequency modulation stage and the transient inertial stage are divided into: Select a 2-second time window before the disturbance occurs; The median of the filtered frequency data within the time window is determined as the frequency reference value; The median of the filtered active power data within the time window is determined as the power reference value; Based on the difference in time scale between the dynamic response of doubly-fed wind turbines and photovoltaic units, the post-disturbance process is divided into a steady-state frequency regulation stage and a transient inertial stage.

[0032] As a preferred embodiment of the above, after completing the synchronous filtering of the grid connection point frequency data and active power data, the steady-state baseline before the disturbance is first determined, and then the response process after the disturbance is divided into stages according to the time scale difference. Specifically, a time window of 2 seconds can be selected before the disturbance occurs, and the median of the filtered frequency data and active power data within the time window is calculated respectively. The results are used as the frequency baseline value and the power baseline value. This is because the median is less susceptible to occasional outliers and local sampling jitter than the mean, and can more stably characterize the pre-disturbance state. The reference level for both the wind turbine and the power grid is in a quasi-steady state, thus providing a reliable starting point for constructing the frequency deviation and active power increment sequence. After obtaining the frequency and power reference values, the post-disturbance process is further divided into a steady-state frequency regulation stage and a transient inertial stage based on the difference in the dynamic response time scale of the doubly-fed induction generator (DFIG) and photovoltaic (PV) units after frequency disturbances. The transient inertial stage preferably selects the initial period after the disturbance to characterize the rapid response process of the units to the frequency change rate. For example, a time interval of 0.5s to 1.5s after the disturbance can be selected to cover the participation of wind and solar units in inertia. The main support period; the steady-state frequency regulation stage is preferably selected when the inertial component has decayed after the disturbance and the frequency response has gradually entered a relatively stable period. This ensures that the active power change during this stage is mainly caused by the virtual primary frequency regulation, rather than being dominated by transient inertial support. The key to this approach is not to treat the entire disturbance process as a single response for overall identification, but rather to first establish a unified reference by selecting a pre-disturbance benchmark, and then strictly separate the steady-state frequency regulation component and the transient inertial component in time through a two-stage division. This reduces the mutual coupling of the two types of control actions in parameter identification and improves the subsequent understanding of the virtual primary frequency regulation. The accuracy and interpretability of secondary frequency regulation parameters and virtual inertial parameters identification; as an example, when a frequency drop disturbance occurs at the grid connection point of a wind and solar turbine at a certain moment, the filtered median frequency and median power within the first 2-second window can be used as a benchmark. Then, the initial stage of the disturbance is classified into the transient inertial stage to characterize rapid inertial support, and the stable response range after the frequency change slows down is classified into the steady-state frequency regulation stage to characterize continuous active power support. In this way, the identification model established later can be calculated for different physical mechanisms, avoiding parameter deviations caused by mixing power responses of different time scales.

[0033] Furthermore, online identification of the actual effective parameters corresponding to the virtual primary frequency control component includes: During the steady-state frequency regulation phase, the rapidly decaying inertial term is ignored, and a steady-state equation error model is established based on the frequency deviation and the active power increment. Construct a quadratic loss function based on the observation data sequences of system input and output; The recursive least squares method with a forgetting factor is used to iteratively identify the actual effective parameters corresponding to the virtual primary frequency modulation control components. The actual effective parameters are updated online by updating the gain matrix, parameter estimates, and covariance matrix. The forgetting factor was set to 0.99 to obtain smooth steady-state parameter estimation results.

[0034] As a preferred embodiment of the above, after selecting the reference value before the disturbance, the deviation of the filtered real-time frequency from the frequency reference value is used as the input, and the increment of the filtered active power from the power reference value is used as the observed output. During the steady-state frequency regulation phase, the rapidly decaying inertial term is ignored, and a steady-state equation error model is established, so that the parameter to be identified directly corresponds to the actual effective parameter of the virtual primary frequency regulation control component. The advantage of this approach is that there is a clear correspondence between the identified object and the physical control action, which can more accurately reflect the true equivalent frequency regulation capability of the unit under the effect of steady-state frequency difference. In terms of the solution method, a recursive least squares method with a forgetting factor is preferred. The actual effective parameters are identified iteratively online. Specifically, a quadratic loss function can be constructed based on the observation data sequence of the system input and output, and the parameter estimates are gradually converged by continuously minimizing this loss function. Compared with one-time offline least squares estimation, the advantage of using a recursive form is that it does not need to save all historical data, but can continuously correct the parameter estimates as new sampling points arrive, which is more suitable for online monitoring and real-time tracking in the grid-connected operation scenario of wind and solar turbine units. Furthermore, to achieve continuous online updates of parameters, it is preferable to update the gain matrix, parameter estimates, and covariance matrix sequentially at each sampling time, where the gain matrix is ​​used to measure the current new observation data. The impact of parameter correction is considered. The parameter estimate is used to characterize the latest identification result of the droop parameter at the current moment, and the covariance matrix is ​​used to characterize the dispersion of the parameter estimation error and reflect the algorithm's sensitivity to subsequent new data. Through the recursive update of the above three, continuous tracking of the frequency regulation coefficient of wind and solar turbines can be achieved while ensuring the numerical stability of the algorithm. In the selection of the forgetting factor, it is preferable to set it to 0.99, so that the algorithm has a longer memory length in the steady-state frequency regulation stage, thereby reducing the impact of instantaneous measurement noise and occasional fluctuations on the identification result and obtaining a smoother and more stable steady-state parameter estimate. In other words, a larger forgetting factor allows new sampling points to participate in the parameter estimation. The method corrects the values ​​but does not cause drastic fluctuations in the estimated values, making it more suitable for identifying virtual primary frequency regulation parameters in the steady-state phase. For example, during the grid-connected operation of wind and solar turbines, when the grid experiences disturbances and enters a relatively stable frequency regulation phase, the filtered frequency deviation and active power increment sequences can be continuously read. The recursive least squares identification process with forgetting factor described above is used to output the estimated values ​​of the virtual primary frequency regulation parameters in real time. As the estimated values ​​gradually converge, they can be used as the actual effective parameters corresponding to the virtual primary frequency regulation control components for frequency regulation component stripping in subsequent transient inertial parameter identification, as well as for the final parameter validity assessment and power saturation state determination.

[0035] Furthermore, online identification of the actual effective parameters corresponding to the virtual inertial control components includes: During the transient inertial phase, the frequency modulation component is stripped off based on the actual effective parameters corresponding to the virtual primary frequency modulation control component; Calculate the remaining inertial response components based on the results after stripping. Calculate the rate of frequency change based on the frequency data; Under the condition of satisfying the physical validity constraint criterion, the recursive least squares method with forgetting factor is used to identify the actual effective parameters corresponding to the virtual inertial control components online. The forgetting factor was set to 0.95 to enhance the dynamic tracking capability of changes in inertial response.

[0036] As a preferred embodiment of the above, during the transient inertia stage, based on the obtained actual effective parameters of the virtual primary frequency regulation and the frequency deviation at the corresponding moment, the theoretical droop adjustment power at that moment is first calculated, and then subtracted from the filtered actual active power increment to obtain the residual inertial response component. Since this residual component mainly originates from the transient support for rapid frequency changes, it is more suitable as the observation object of the virtual inertial control component. On this basis, the frequency change rate is further calculated based on the frequency data, and the frequency change rate is used as the input and the residual inertial response component as the observation output to construct the second-stage recursive identification relationship. Due to transient inertia... This stage corresponds to the rapid dynamic process in the initial stage of the disturbance. The rate of frequency change and inertial support power typically change faster and for shorter durations. Therefore, it is preferable to still use the recursive least squares method with a forgetting factor for online identification. However, unlike the steady-state frequency tuning stage, this stage emphasizes the ability to track rapid changes rather than smoothness. To ensure the physical rationality of the identification results, it is preferable to introduce a physical validity constraint criterion before parameter updates. Only when transient samples simultaneously meet the preset validity conditions will the corresponding data be used to update the actual effective parameters corresponding to the virtual inertial control components. This avoids issues when the rate of frequency change is extremely small, the inertial response components are not obvious, or the samples are incomplete. In representative cases, parameter updates are mistakenly triggered, thereby improving the stability and reliability of identification. Preferably, the forgetting factor of the recursive least squares method with a forgetting factor is set to 0.95 in this stage. A smaller forgetting factor means that the algorithm is more sensitive to newly entered transient data and can correct parameter estimates more quickly, thereby enhancing the dynamic tracking ability of the rapidly changing inertial response. This setting, combined with the strategy of using a larger forgetting factor to obtain smooth estimates in the steady-state stage, allows the two-stage identification to adapt to control characteristics at different time scales. As an example, when the power grid experiences a rapid frequency drop disturbance, the doubly-fed induction generator can first release the rotor. The kinetic and photovoltaic units can first utilize the virtual inertial support capability. During this short period, if the inertial parameters of the total active response are directly estimated, the droop adjustment component caused by frequency deviation will still be present, which may easily lead to overestimation or underestimation of the virtual inertial parameters. In this embodiment, the droop component is first removed using the identified actual effective parameters of the virtual primary frequency regulation, and then the relationship between the remaining inertial response components and the rate of frequency change is recursively estimated. This can more accurately obtain the actual effective parameters corresponding to the virtual inertial control components and provide a basis for subsequent judgment on whether the wind and solar units respond effectively according to the preset inertial support capability in the early stage of disturbance.

[0037] Furthermore, the parameter validity assessment includes: The actual effective parameters corresponding to the virtual primary frequency modulation control component are compared with the preset parameters of the controller to calculate the accuracy evaluation index. Calculate the maximum increase in power generation for the doubly-fed wind turbine and photovoltaic unit under the current operating conditions; The theoretical power requirement is calculated based on the controller's preset parameters and steady-state frequency deviation. When the accuracy assessment index is within the allowable error range and the theoretical power demand is not greater than the maximum power that can be generated, the controller preset parameters are deemed to be valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region. When the theoretical power demand exceeds the maximum power that can be increased, the doubly fed wind turbine and photovoltaic unit are determined to have entered a power saturation state, and the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability is characterized by an accuracy evaluation index.

[0038] As a preferred embodiment of the above, the parameter validity assessment is not merely a static comparison between the parameter values ​​obtained from online identification and the controller's preset parameters. Instead, it further incorporates the unit's physical adjustability under current operating conditions into a unified judgment process. This distinguishes between two different scenarios: the control parameter settings themselves are valid, but the unit cannot fully execute them due to physical capacity limitations; and the control parameter settings are inconsistent with the actual response. Specifically, the actual effective parameters corresponding to the virtual primary frequency regulation control component are first compared with the controller's preset parameters to calculate an accuracy assessment index. This accuracy assessment index is preferably used to characterize the degree of closeness between the actual effective frequency regulation capability and the theoretically set frequency regulation capability. Based on this, the maximum increaseable power is determined according to the maximum available power and steady-state output power benchmark of the doubly-fed wind turbine and photovoltaic unit under current operating conditions. This is the active power margin that the unit can continue to increase under current wind speed, sunlight, and temperature conditions. Simultaneously, the theoretical power demand is calculated by combining the controller's preset parameters and the steady-state frequency deviation, thereby establishing a correspondence between the identification results, theoretical commands, and physical boundaries. When the accuracy assessment index is within the allowable error range and the theoretical power demand is not greater than the maximum increaseable power, a judgment can be made. When the preset parameters of the frequency controller are valid and the unit operates in the linear region under the current disturbance, the actual output of the unit can change according to the preset control law, and the online identification results are basically consistent with the preset values. However, when the theoretical power demand exceeds the maximum available additional power, it indicates that although the unit is still attempting to increase power according to the preset parameters, it is limited by the current available reserve capacity or the upper limit of the rated power, and the actual output can no longer increase linearly. The unit enters a power saturation state, and the actual effective parameters identified at this time will be significantly smaller than the preset parameters. The accuracy evaluation index can then quantitatively characterize the actual frequency regulation capability of the unit. The ratio of theoretical frequency regulation capability provides a direct basis for the dispatching side to determine whether the unit truly possesses the set frequency support capability. For example, in a scenario where the rated power of the doubly-fed induction generator (DFIG) is 5MW and the load reduction rate is 0.85, the unit's reserved reserve capacity is approximately 0.75MW. When a minor disturbance occurs in the system, the theoretically required additional power does not exceed this reserve capacity. The error between the actual effective parameters of the virtual primary frequency regulation obtained online and the preset parameters of the controller is within the allowable range, indicating that the parameters are valid and the unit is operating in the linear region. However, when a major disturbance occurs in the system, the theoretically required additional power exceeds 0.75MW.For a 75MW unit, with output limited by a 5MW upper limit, the actual power increment stops increasing after reaching the maximum available power. At this point, the calculated accuracy assessment index deviates significantly from the allowable range, indicating the unit has entered power saturation. This index reflects that the actual frequency regulation contribution is only a portion of the theoretical expectation. The same judgment logic applies to photovoltaic units under both small and large disturbance scenarios: when the theoretical power demand does not exceed the current available reserve, the parameter assessment result is consistent with the preset value; however, when the theoretical power demand exceeds the available reserve, the actual output becomes saturated due to the maximum available power limit, and the accuracy assessment index decreases accordingly. This achieves a unified assessment of the parameter effectiveness and saturation state of wind and solar turbine units.

[0039] Example 2; Based on the same inventive concept as the online identification and evaluation method for virtual primary frequency regulation of wind and solar turbine units in the foregoing embodiments, the present invention also provides an online identification and evaluation system for virtual primary frequency regulation of wind and solar turbine units, the system comprising: The information acquisition module acquires the primary energy parameters and operating parameters of the doubly-fed wind turbine and photovoltaic unit, and determines the steady-state output power benchmark, maximum available power and reserve capacity of the doubly-fed wind turbine and photovoltaic unit; The model building module constructs a frequency response control model based on the steady-state output power benchmark, including virtual primary frequency modulation control components and virtual inertial control components. The grid connection processing module collects frequency data and active power data at the grid connection points of doubly-fed wind turbines and photovoltaic units, and performs preprocessing to obtain time-series data; The steady-state identification module divides the steady-state frequency modulation stage and the transient inertial stage based on time-series data, and identifies the actual effective parameters corresponding to the virtual primary frequency modulation control components online during the steady-state frequency modulation stage. The transient identification module identifies the actual effective parameters corresponding to the virtual primary frequency control component online during the transient inertial phase, based on the actual effective parameters corresponding to the virtual primary frequency control component. The saturation assessment module evaluates the validity of parameters based on actual effective parameters, standby capacity, and controller preset parameters to determine whether the doubly fed wind turbine and photovoltaic unit have entered a power saturation state.

[0040] The adjustment system described above in this invention can effectively realize the online identification and evaluation method of virtual primary frequency regulation for wind and solar turbine units. The technical effects it can achieve are as described in the above embodiments, and will not be repeated here.

[0041] Furthermore, the saturation evaluation module includes: The parameter comparison unit compares the actual effective parameters corresponding to the virtual primary frequency modulation control component with the preset parameters of the controller to calculate the accuracy evaluation index. The maximum power generation unit calculates the maximum power that can be increased for the doubly-fed wind turbine and photovoltaic unit under the current operating conditions; The theoretical calculation unit calculates the theoretical power demand based on the controller's preset parameters and steady-state frequency deviation; The linearity determination unit determines that the controller preset parameters are valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region when the accuracy evaluation index is within the allowable error range and the theoretical power demand is not greater than the maximum power that can be generated. The saturation determination unit determines that the doubly fed wind turbine and photovoltaic unit have entered the power saturation state when the theoretical power demand exceeds the maximum power that can be generated. The accuracy evaluation index is used to characterize the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability.

[0042] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0043] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for online identification and evaluation of virtual primary frequency regulation of wind turbine generators, characterized in that, The method includes: Obtain the primary energy parameters and operating parameters of the doubly-fed wind turbine and the photovoltaic unit, and determine the steady-state output power reference, maximum available power and reserve capacity of the doubly-fed wind turbine and the photovoltaic unit; A frequency response control model including a virtual primary frequency modulation control component and a virtual inertial control component is constructed based on the steady-state output power reference. Frequency data and active power data of the grid connection points of the doubly fed wind turbine and the photovoltaic unit are collected and preprocessed to obtain time-series data; Based on the time-series data, the steady-state frequency modulation stage and the transient inertial stage are divided, and the actual effective parameters corresponding to the virtual primary frequency modulation control component are identified online during the steady-state frequency modulation stage. Based on the actual effective parameters corresponding to the virtual primary frequency modulation control component, the actual effective parameters corresponding to the virtual inertial control component are identified online during the transient inertial phase; The validity of the parameters is evaluated based on the actual effective parameters, the reserve capacity, and the controller preset parameters to determine whether the doubly fed wind turbine and the photovoltaic unit have entered a power saturation state.

2. The wind-PV hybrid generator virtual primary frequency regulation online identification and evaluation method according to claim 1, characterized in that, Determining the operating parameters of the doubly-fed wind turbine and the photovoltaic unit includes: Based on the aerodynamic model of the doubly fed fan and the real-time wind speed, the optimal mechanical power of the doubly fed fan under maximum power point tracking (MPPT) conditions is calculated. The reduced-load operating power of the doubly-fed fan is determined based on the optimal mechanical power of the doubly-fed fan and the set reduced-load rate. The maximum available power of the photovoltaic unit is calculated based on real-time light intensity and ambient temperature. The steady-state output power reference of the photovoltaic unit is determined based on the maximum available power of the photovoltaic unit and the set load reduction rate; The reserve capacity reserved for frequency response by the doubly-fed wind turbine and the photovoltaic unit is determined based on the reduced-load operating power of the doubly-fed wind turbine and the steady-state output power of the photovoltaic unit.

3. The wind-PV hybrid generator virtual primary frequency regulation online identification and evaluation method according to claim 1, characterized in that, Constructing a frequency response control model includes: An additional frequency control loop is introduced into the active power control circuit of the converter of the doubly fed wind turbine and the photovoltaic unit. The virtual primary frequency modulation control component is constructed based on the steady-state output power reference. The virtual inertial control components are constructed based on the rate of change of frequency; The steady-state output power reference, the virtual primary frequency control component, and the virtual inertial control component are superimposed to generate an active power reference command.

4. The wind-PV hybrid generator virtual primary frequency regulation online identification and evaluation method according to claim 1, characterized in that, Collect frequency data and active power data and perform preprocessing, including: Real-time collection of frequency data and active power data at the grid connection points of the doubly-fed wind turbine and the photovoltaic unit; The frequency data and active power data are preprocessed using synchronous filtering to obtain a smooth sequence; The window length of the moving average filter is set to 200ms to eliminate noise and maintain the relative phase consistency between the frequency data and the active power data.

5. The wind-PV hybrid generator virtual primary frequency regulation online identification and evaluation method according to claim 4, characterized in that, The process is divided into a steady-state frequency modulation stage and a transient inertial stage, including: Select a 2-second time window before the disturbance occurs; The median of the filtered frequency data within the time window is determined as the frequency reference value; The median of the filtered active power data within the time window is determined as the power reference value; Based on the time scale difference in the dynamic response of the doubly fed wind turbine and the photovoltaic unit, the post-disturbance process is divided into the steady-state frequency regulation stage and the transient inertial stage.

6. The wind-PV hybrid generator virtual primary frequency regulation online identification and evaluation method according to claim 1, characterized in that, Online identification of the actual effective parameters corresponding to the virtual primary frequency modulation control component includes: During the steady-state frequency modulation phase, the rapidly decaying inertial term is ignored, and a steady-state equation error model is established based on the frequency deviation and the active power increment. Construct a quadratic loss function based on the observation data sequences of system input and output; The actual effective parameters corresponding to the virtual primary frequency modulation control component are iteratively identified using the recursive least squares method with a forgetting factor. The actual effective parameters are updated online by updating the gain matrix, parameter estimates, and covariance matrix. The forgetting factor was set to 0.99 to obtain smooth steady-state parameter estimation results.

7. The method for online identification and evaluation of virtual primary frequency regulation of wind and solar turbine units according to claim 1, characterized in that, Online identification of the actual effective parameters corresponding to the virtual inertial control components includes: During the transient inertial phase, the frequency modulation component is stripped off according to the actual effective parameters corresponding to the virtual primary frequency modulation control component; Calculate the remaining inertial response components based on the results after stripping. Calculate the frequency change rate based on the frequency data; Under the condition of satisfying the physical validity constraint criterion, the recursive least squares method with forgetting factor is used to identify the actual effective parameters corresponding to the virtual inertial control components online. The forgetting factor was set to 0.95 to enhance the dynamic tracking capability of changes in inertial response.

8. The method for online identification and evaluation of virtual primary frequency regulation of wind and solar turbine units according to claim 1, characterized in that, Conduct parameter validity assessment, including: The actual effective parameters corresponding to the virtual primary frequency modulation control component are compared with the preset parameters of the controller to calculate the accuracy evaluation index. Calculate the maximum increase in power output of the doubly-fed wind turbine and the photovoltaic unit under the current operating conditions; The theoretical power requirement is calculated based on the controller's preset parameters and steady-state frequency deviation. When the accuracy evaluation index is within the allowable error range and the theoretical power demand is not greater than the maximum achievable power, it is determined that the controller preset parameters are valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region. When the theoretical power demand exceeds the maximum power that can be increased, the doubly fed wind turbine and the photovoltaic unit are determined to have entered a power saturation state, and the accuracy evaluation index is used to characterize the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability.

9. A virtual primary frequency regulation online identification and evaluation system for wind and solar turbine units, characterized in that, The system includes: The information acquisition module acquires the primary energy parameters and operating parameters of the doubly-fed wind turbine and photovoltaic unit, and determines the steady-state output power benchmark, maximum available power and reserve capacity of the doubly-fed wind turbine and photovoltaic unit; The model building module constructs a frequency response control model based on the steady-state output power benchmark, including virtual primary frequency modulation control components and virtual inertial control components. The grid connection processing module collects frequency data and active power data at the grid connection points of doubly-fed wind turbines and photovoltaic units, and performs preprocessing to obtain time-series data; The steady-state identification module divides the steady-state frequency modulation stage and the transient inertial stage based on time-series data, and identifies the actual effective parameters corresponding to the virtual primary frequency modulation control components online during the steady-state frequency modulation stage. The transient identification module identifies the actual effective parameters corresponding to the virtual primary frequency control component online during the transient inertial phase, based on the actual effective parameters corresponding to the virtual primary frequency control component. The saturation assessment module evaluates the validity of parameters based on actual effective parameters, standby capacity, and controller preset parameters to determine whether the doubly fed wind turbine and photovoltaic unit have entered a power saturation state.

10. The online identification and evaluation system for virtual primary frequency regulation of wind and solar power units according to claim 9, characterized in that, The saturation evaluation module includes: The parameter comparison unit compares the actual effective parameters corresponding to the virtual primary frequency modulation control component with the preset parameters of the controller to calculate the accuracy evaluation index. The maximum power generation unit calculates the maximum power that can be increased for the doubly-fed wind turbine and photovoltaic unit under the current operating conditions; The theoretical calculation unit calculates the theoretical power demand based on the controller's preset parameters and steady-state frequency deviation; The linearity determination unit determines that the controller preset parameters are valid and the doubly fed wind turbine and photovoltaic unit are operating in the linear region when the accuracy evaluation index is within the allowable error range and the theoretical power demand is not greater than the maximum power that can be generated. The saturation determination unit determines that the doubly fed wind turbine and photovoltaic unit have entered the power saturation state when the theoretical power demand exceeds the maximum power that can be generated. The accuracy evaluation index is used to characterize the ratio of the actual frequency regulation capability to the theoretical frequency regulation capability.

Images