CN-121996931-A - Virtual inertia identification method suitable for doubly-fed wind farm

CN121996931ACN 121996931 ACN121996931 ACN 121996931ACN-121996931-A

Abstract

The invention discloses a virtual inertia identification method suitable for a doubly-fed wind power plant, which comprises the steps of constructing a second-order transfer function and a virtual moment of inertia expression of grid-connected power and grid frequency of the doubly-fed wind power plant, collecting data, selecting a stable active power sequence and a grid angular frequency sequence as identification data, denoising, selecting quasi-steady state data to calculate a sagging control coefficient, discretizing the transfer function, identifying the coefficient of the discretized transfer function based on a regularized least square method, restoring the second-order transfer function based on the identified parameter, and calculating virtual moment of inertia by utilizing the coefficient of the second-order transfer function and the virtual moment of inertia expression. The method has the advantages that no phase delay exists in the process of identifying the virtual inertia of the doubly-fed wind power plant, the peak characteristic of inertia response is reserved, the inertia identification error caused by parameter coupling in the traditional method is solved, the phenomenon of severe oscillation of fitting parameters is prevented by introducing regularization factors, and the identification precision of the virtual inertia is remarkably improved.

Inventors

TANG WENZUO
ZHOU JIAXIN
YU YUE
LI BO
LIU YONGCHAO

Assignees

国网重庆市电力公司经济技术研究院

Dates

Publication Date: 20260508
Application Date: 20260129

Claims (3)

1. The virtual inertia identification method suitable for the doubly-fed wind farm is characterized by comprising the following steps of: 1) Aiming at a doubly-fed wind power plant in a power system, constructing a second-order transfer function of grid-connected power and grid frequency of the doubly-fed wind power plant: (1) (2) Wherein: is the variation of the grid-connected power, Is the variation of the angular frequency of the power grid, Is the rotational inertia of the doubly-fed asynchronous wind power generator, Is the equivalent time constant of the active power control loop of the doubly-fed asynchronous wind power generator, Is the damping coefficient of the doubly-fed asynchronous wind power generator, Is the pneumatic damping coefficient of the doubly-fed asynchronous wind power generator, Is a virtual coefficient of inertia which is the coefficient of inertia, In order for the droop control coefficient to be a factor, Is the pole pair number of the doubly-fed asynchronous wind power generator, Is the angular velocity of the machine at steady state, Is the active power of the stator in a steady state, For the grid-tie point steady state angular frequency, Tracking the coefficient for maximum power; s is the Laplace operator; b) The doubly-fed wind power plant is approximately equivalent to a virtual doubly-fed asynchronous wind power generator, and a virtual rotational inertia expression of the doubly-fed wind power plant is constructed: (3) in the formula, The pole pair number of other generators except the doubly-fed asynchronous wind power generator in the power system is the pole pair number; 2) The method comprises the steps of applying load disturbance to a power system, then collecting active power sequences and power grid angular frequency sequences at grid connection points of a doubly-fed wind power plant, collecting real-time wind speed data at each doubly-fed asynchronous wind power generator in the doubly-fed wind power plant, and calculating a full-field wind speed average value at each sampling moment according to the collected real-time wind speed data to obtain a full-field wind speed average value sequence of the doubly-fed wind power plant; 3) Judging whether the acquired wind speed average value sequence is a stable sequence or not, and selecting an active power sequence and a power grid angular frequency sequence which correspond to the acquisition time of the stable sequence as identification data; 4) Denoising the active power sequence and the power grid angular frequency sequence which are taken as identification data; 5) Selecting quasi-steady state data in the later period of frequency disturbance from the denoised active power sequence and the power grid angular frequency sequence to calculate a droop control coefficient: (4) Wherein, the Representing the sag control coefficient, Is the average value of the quasi-steady-state active power variation, Is the average value of the angular frequency variation of the quasi-steady-state power grid, t represents a sampling time; (5) (6) Wherein, the Is the active power acquired at the time t in the selected active power quasi-steady state data sequence, Is the first sampling value in the active power data sequence after original denoising; (7) (8) Wherein, the Is the power grid angular frequency acquired at the time t in the selected power grid angular frequency quasi-steady state data sequence, Is the first sampling value in the grid angular frequency data sequence after the original denoising; Wherein, the The lengths of the active power quasi-steady state data sequence and the grid angular frequency quasi-steady state data sequence are respectively; then subtracting steady-state components from active power sequence data obtained by denoising in the step 4) to obtain a dynamic power sequence: (9) Wherein, the A sequence of dynamic powers is represented and, Representing the original denoised active power sequence, For the sequence of the angular frequency variation of the power grid, Is a steady-state component; 6) Transfer function is to Discretizing to obtain: (10) in the formula, Is a discrete transform operator; Least square form of power grid angular frequency sequence data obtained by denoising treatment in the measure 4) and dynamic power sequence structure obtained in the measure 5): (11) Wherein: To be the kth sample value in the dynamic power sequence, T is the sampling period; Is the first sample value in the dynamic power sequence; As the parameter to be identified, , Representing the power grid angular frequency sequence data obtained by the denoising treatment in the step 4); the objective function of constructing the least squares is: (12) Wherein, the The requirements are as follows: (13) Wherein: n is the sequence length; ; Solving to obtain : (14) Introducing regularization factors Constructing a cost function, and obtaining the cost function by adopting a ridge regression algorithm to the formula (13): (15) Wherein I is an identity matrix, and the parameters are obtained by solving the formula (15) ; 7) Identifying the parameters obtained in step 6) Carry-in transfer function The following formula is used: (16) Will be Restoring to a continuous domain transfer function form shown in the formula (1) to obtain coefficients in the formula (1) And calculating by using the formula (3) to obtain the virtual moment of inertia of the doubly-fed wind power plant contributing to the power system.
2. The method for identifying virtual inertia suitable for a doubly-fed wind farm according to claim 1, wherein in step 3), the method for determining whether the collected wind speed average sequence is a stationary sequence comprises: Calculating the average value of all the samples in the full-field wind speed average value sequence, subtracting the average value of all the samples in the full-field wind speed average value sequence by using each sample value in the wind speed average value sequence, comparing the absolute value of the difference with a set screening threshold value, and if the absolute value of each difference is smaller than or equal to the screening threshold value, the wind speed average value sequence is a stable sequence, otherwise, the wind speed average value sequence is an unstable sequence.
3. The virtual inertia identification method suitable for the doubly-fed wind power plant according to claim 1, wherein the step 4) is characterized in that a Savitzky-Golay filter is adopted to conduct denoising processing on an active power sequence and a grid angular frequency sequence which are used as identification data.

Description

Virtual inertia identification method suitable for doubly-fed wind farm Technical Field The invention relates to the technical field of power system control and parameter identification, in particular to a method for identifying virtual inertia of a doubly-fed wind power plant. Background With the improvement of permeability of renewable energy sources such as wind power, the moment of inertia of a power system is gradually reduced, and the frequency stability is challenging. Doubly fed asynchronous wind generators (DFIGs) typically support grid frequencies through additional frequency control, doubly fed wind farms are a power generation base of numerous doubly fed wind generators. In order to evaluate the frequency supporting capability of the system, it is important to accurately identify the equivalent inertia constant of the doubly-fed wind farm. The prior art mainly has the following problems when inertia identification is carried out: 1) Noise sensitivity-inertia identification depends on the frequency change rate, the power grid frequency signal usually contains measurement noise, and direct difference can amplify high-frequency noise, so that the identification result is severely oscillated. 2) Filtering delay-in order to suppress noise, the prior art often employs a low pass filter such as butterworth, but this introduces a significant phase delay. Resulting in a power change at time t in the identification equationCorresponding but isTime of dayThis phase delay can disrupt the physical causality of the inertia response, causing large recognition errors. 3) Parameter coupling conventional least squares tend to minimize global errors. Because the inertia response only exists in the transient process, and the droop control dominates the steady-state process, if the integral fitting is performed indiscriminately, the algorithm always sacrifices the accuracy of transient inertia parameters for the purpose of migrating steady-state accuracy, and the accuracy of the identified inertia value is low. And the traditional least square method can not even calculate and fit when facing the data matrix irreversibility. Disclosure of Invention In view of the above, the present invention provides a virtual inertia identification method suitable for a doubly-fed wind farm, so as to solve the problems of large differential signal noise, high filtering delay and inaccurate identification caused by steady-state parameter coupling in the inertia identification of the existing doubly-fed wind farm. The virtual inertia identification method suitable for the doubly-fed wind power plant comprises the following steps: 1) Aiming at a doubly-fed wind power plant in a power system, constructing a second-order transfer function of grid-connected power and grid frequency of the doubly-fed wind power plant: (1) (2) Wherein: is the variation of the grid-connected power, Is the variation of the angular frequency of the power grid,Is the rotational inertia of the doubly-fed asynchronous wind power generator,Is the equivalent time constant of the active power control loop of the doubly-fed asynchronous wind power generator,Is the damping coefficient of the doubly-fed asynchronous wind power generator,Is the pneumatic damping coefficient of the doubly-fed asynchronous wind power generator,Is a virtual coefficient of inertia which is the coefficient of inertia,In order for the droop control coefficient to be a factor,Is the pole pair number of the doubly-fed asynchronous wind power generator,Is the angular velocity of the machine at steady state,Is the active power of the stator in a steady state,For the grid-tie point steady state angular frequency,Tracking the coefficient for maximum power; s is the Laplace operator; b) The doubly-fed wind power plant is approximately equivalent to a virtual doubly-fed asynchronous wind power generator, and a virtual rotational inertia expression of the doubly-fed wind power plant is constructed: (3) in the formula, The pole pair number of other generators except the doubly-fed asynchronous wind power generator in the power system is the pole pair number; 2) The method comprises the steps of applying load disturbance to a power system, then collecting active power sequences and power grid angular frequency sequences at grid connection points of a doubly-fed wind power plant, collecting real-time wind speed data at each doubly-fed asynchronous wind power generator in the doubly-fed wind power plant, and calculating a full-field wind speed average value at each sampling moment according to the collected real-time wind speed data to obtain a full-field wind speed average value sequence of the doubly-fed wind power plant; 3) Judging whether the acquired wind speed average value sequence is a stable sequence or not, and selecting an active power sequence and a power grid angular frequency sequence which correspond to the acquisition time of the stable sequence as identification data; 4) Denoising the a