A fan blade and tower frequency coupling identification and active decoupling control method

Abstract

The application discloses a fan blade and tower frequency coupling identification and active decoupling control method, which can accurately identify the coupling between the blade through frequency and the blade natural frequency and the tower natural frequency under the variable speed condition; can distinguish forced coupling and modal coupling by the amplitude and phase characteristics of the cross-spectral density, the coherence function and the transfer function, and can exclude random interference such as turbulence; can change the coupling state and suppress vibration by adjusting the impeller speed, the variable pitch angle and enabling the notch filter according to the identification result; can realize the whole process automation and real-time, and is suitable for large variable speed and variable pitch fan; and can effectively solve the above-mentioned problems in the prior art and provide convenience for more users.

Description

Technical Field

[0001] This invention relates to the technical field of wind turbine control, and in particular to a method for frequency coupling identification and active decoupling control of wind turbine blades and towers. Background Technology

[0002] During wind turbine operation, there are two typical frequency domain coupling problems between the blades and the tower:

[0003] The blade passing frequency is coupled with the tower's natural frequency: Due to the tower's obstruction of airflow, the blades periodically pass in front of the tower, generating excitation forces with frequencies that are integer multiples of the impeller's rotational frequency, such as 1P, 2P, 3P, etc. When this passing frequency approaches the tower's natural frequency, it triggers forced resonance, causing the tower to vibrate significantly.

[0004] The natural frequencies of the blades and the tower are coupled: The blades themselves have bending / torsional modes, and their natural frequencies are affected by the centrifugal stiffening effect and increase with rotational speed. When this frequency is equal to or in an integer ratio with the natural frequency of the tower, such as 1:1, 1:2, or 2:1, structural mode coupling occurs, and energy can be transferred bidirectionally between the blades and the tower, inducing parametric resonance.

[0005] In existing technologies, forced coupling is often avoided by using speed-restricted zones; however, online identification methods are lacking for modal coupling, and most methods are only effective at constant speeds, failing to handle variable-speed operation. Furthermore, conventional spectral analysis struggles to distinguish between coupled vibrations and random turbulent excitations, and phase information is not fully utilized. Therefore, a complete technical method that can adapt to variable speeds, has high anti-interference capabilities, and can actively decouple is urgently needed. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a method for frequency coupling identification and active decoupling control of wind turbine blades and towers. Under variable speed conditions, it accurately identifies the coupling between the blade's passing frequency and the natural frequency of the blade and the natural frequency of the tower. By analyzing the amplitude and phase characteristics of the cross-spectral density, coherence function, and transfer function, it distinguishes between forced coupling and modal coupling, eliminating random interference such as turbulence. Based on the identification results, it actively controls and adjusts the impeller speed, pitch angle, and activates a notch filter to change the coupling state and suppress vibration. This achieves full automation and real-time control, and is applicable to large variable speed and pitch wind turbines. It effectively solves the problems existing in the current technology.

[0007] The objective of this invention is achieved through the following technical solution: a method for frequency coupling identification and active decoupling control of wind turbine blades and towers, comprising the following steps:

[0008] S1. Acquire the blade vibration signal, tower vibration signal, and instantaneous speed signal of the wind turbine;

[0009] S2. Perform preprocessing on the blade vibration signal and tower vibration signal, including detrending, bandpass filtering, and order tracking resampling;

[0010] S3. Perform a fast Fourier transform on the blade vibration signal, tower vibration signal and instantaneous rotation speed signal after the preprocessing in step S2, and further calculate the cross-order spectral density, self-order spectral density, amplitude squared coherence function and transfer function.

[0011] S4. Store the blade natural frequency curves and tower natural frequencies calibrated offline, and output the coupling type, coupling frequency, and calculated amplitude squared coherence value based on the cross-order spectral density, self-order spectral density, amplitude squared coherence function, transfer function, and preset discrimination rules.

[0012] S5. Execute the corresponding control action based on the determined coupling type;

[0013] S6. Calculate the real-time amplitude squared coherence value. When the real-time amplitude squared coherence value is less than the preset threshold, gradually restore the original speed of the wind turbine or turn off the adaptive notch filter integrated in the main controller of the wind turbine, exit the active control mode, and realize the identification and active decoupling control of the frequency coupling between the blade and the tower.

[0014] Furthermore, the method includes the following steps:

[0015] S7. Based on the blade vibration signal and tower vibration signal obtained in step S1, continuously monitor the peak acceleration at the top of the tower and the blade strain during the control process; if the blade vibration signal or tower vibration signal does not exceed the preset threshold, no operation is performed; if the blade vibration signal or tower vibration signal exceeds the preset threshold, immediately control the wind turbine to stop in an emergency feathering mode.

[0016] Furthermore, step S1 includes:

[0017] With synchronized sampling rate The blade vibration signal x(t), the tower vibration signal y(t), and the instantaneous impeller speed pulse n(t) are collected continuously for a duration T≥60 s; the instantaneous impeller speed pulse n(t) is converted into an instantaneous speed signal Ω(t).

[0018] Furthermore, step S2 includes:

[0019] The blade vibration signal x(t) and the tower vibration signal y(t) are preprocessed by detrending and bandpass filtering. The pass frequency range of the bandpass filter is set to 0.1~50 Hz to cover the main modes of the blade and the tower.

[0020] The key phase timescale is calculated based on the instantaneous impeller speed pulse n(t). The order tracking (COT) method is used to resample the blade vibration signal x(t) and the tower vibration signal y(t) into angle domain signals x(θ) and y(θ) to eliminate the influence of speed variation on frequency analysis. The angle domain signal is segmented at equal angular intervals per revolution to prepare for order spectrum analysis.

[0021] Furthermore, step S3 includes:

[0022] Perform FFT on the resampled x(θ) and y(θ) signals to obtain the order spectra X(O) and Y(O), where the order O = f / (Ω / 60) is dimensionless;

[0023] Calculate the mutual order spectral density ;

[0024] Calculate the self-order spectral density S xx (O) and S yy (O);

[0025] Calculate the squared amplitude coherence function ;

[0026] Calculate the transfer function H xy (O) = S xy (O) / S xx (O), and converts it into a frequency domain transfer function through the speed relationship;

[0027] in, For discrete frequency points, determined by the sampling rate The number of FFT points N determines the result: K represents the average frequency, which is used to segment long signals based on the Welch method. The larger the value, the smaller the variance, the smoother the spectral estimation, but the lower the frequency resolution. k represents the index segment, indicating the k-th segment of data, k=1,2,...,K. Let x(t) be the Fourier transform of the signal x(t) of the k-th blade segment, which is a complex number containing amplitude and phase, corresponding to the blade vibration; Let y(t) be the conjugate complex number of the Fourier transform of the signal y(t) of the k-th tower segment; This is an estimate of the cross-spectral density, also a complex number. Its real part represents the in-phase component, and its imaginary part represents the quadrature component. This value increases at the coupling frequency.

[0028] Furthermore, step S4 includes:

[0029] Establish calibration curves f of the blade's pass order and natural frequency as a function of rotational speed. b (Ω); When running online, based on the current average speed calculate:

[0030] Theoretical blade passing frequency f 1P = / 60, f 3P =3 / 60;

[0031] Theoretical blade natural frequency f b ( (This can be determined by looking up a table or by real-time identification using online modal analysis (OMA).)

[0032] Based on the cross-order spectral density, self-order spectral density, and frequency domain transfer function obtained in step S3, the coupling frequency is found. Or the corresponding order O c The amplitude squared coherence value is calculated based on the amplitude squared coherence function in step S3, satisfying the amplitude squared coherence value γ. xy 2 (f c )>γ th The preset discrimination rules are as follows:

[0033] (1) When At that time, the coupling type is determined to be 3P forced coupling, and the phase characteristic is |∣∠H. xy (f c ) |<60°;

[0034] (2) When and At that time, the coupling type is determined to be 1:1 modal coupling, and the phase characteristic is |∣∠H xy (f c ) |>75°;

[0035] (3) When or When n=2 or 3, the coupling type is determined to be subharmonic coupling, and the phase characteristics are unrestricted.

[0036] Where Δf is taken as half of the frequency resolution, This is the first-order natural frequency of the tower.

[0037] Furthermore, step S5 includes:

[0038] When the coupling type is 3P forced coupling, calculate the target rotational speed offset ΔΩ, and make the new... satisfy B is the tower resonance half-power bandwidth; by changing the generator torque setpoint, the speed is slowly adjusted to Ω. new =Ω+ΔΩ, the amplitude is limited within the preset safe operating range, and the pitch angle is increased by 1~2° simultaneously to reduce the aerodynamic excitation amplitude;

[0039] When the coupling type is 1:1 modal coupling, the calibration curve is used to calculate... Deviation Required speed change If the target speed deviation ΔΩ exceeds the allowable operating range, then large-scale speed regulation will be abandoned, and instead, the adaptive notch filter integrated into the wind turbine main controller will be activated, with the notch center frequency set to the current coupling frequency. The notch depth is automatically adjusted based on the coherence value. The quality factor Q=10; simultaneously, the high-frequency damping component in the generator torque fluctuation is increased through additional torque. Implementation, in which This is the speed signal at the top of the tower;

[0040] In all coupling cases, when the squared amplitude coherence value γ 2 (f c When the pitch is greater than 0.85, a resonance warning signal is sent to the main control system, and the pitch rate is temporarily increased to quickly get out of the dangerous speed.

[0041] Furthermore, step S6 includes:

[0042] Real-time calculation of amplitude squared coherence value γ 2 (f c ), when γ lasts for 30 seconds 2 (f c When all values ​​are below 0.3, gradually restore the original speed or turn off the notch filter to exit the active control mode.

[0043] A wind turbine blade and tower frequency coupling identification and active decoupling control system, used to implement the aforementioned wind turbine blade and tower frequency coupling identification and active decoupling control method, comprising:

[0044] The sensor unit is used to sense and output blade vibration signals, tower vibration signals, and instantaneous rotational speed signals;

[0045] The data acquisition and preprocessing unit is used to acquire blade vibration signals, tower vibration signals, and instantaneous rotational speed signals, and perform preprocessing such as detrending, bandpass filtering, and order tracking resampling.

[0046] The order tracking resampling unit calculates the order tracking algorithm based on the instantaneous rotation speed signal, converting the time domain signal into a uniform signal in the angle domain;

[0047] The cross-spectral analysis and coupling discrimination unit performs piecewise fast Fourier transform on the uniform signal in the angle domain, and further calculates the cross-order spectral density, the self-order spectral density, the amplitude squared coherence function, and the transfer function; it stores the offline calibrated blade natural frequency curves and tower natural frequencies, and outputs the coupling type, coupling frequency, and calculated amplitude squared coherence value according to the discrimination rules.

[0048] The active control unit calculates the speed deviation, pitch angle increment, and additional torque based on the coupling frequency, blade vibration signal, tower vibration signal, and instantaneous speed signal, and performs adaptive notch filtering to actively control the wind turbine.

[0049] The safety protection unit is used to independently monitor the peak acceleration at the top of the tower and the blade strain.

[0050] Furthermore, the sensor unit includes a blade vibration sensor, a tower vibration sensor, and a speed encoder; the blade vibration sensor is installed at the root or tip of the wind turbine blade to sense and output blade vibration signals; the tower vibration sensor is installed at the top of the wind turbine tower and communicates with the active control unit to sense and output tower vibration signals; the speed encoder is installed on the impeller shaft or generator input end of the wind turbine to sense impeller speed pulse signals, calculate and output instantaneous speed signals based on the impeller speed pulse signals.

[0051] The data acquisition and preprocessing unit includes a multi-channel synchronous acquisition card, an anti-aliasing filter, and a signal conditioning circuit that are connected in sequence. The multi-channel synchronous acquisition card is connected in communication with the sensor unit.

[0052] The order tracking resampling unit includes a digital signal processor, a field-programmable gate array, and an industrial computing module that are connected in sequence. The digital signal processor is also connected in communication with the data acquisition and preprocessing unit.

[0053] The active control unit is integrated in the wind turbine main controller and communicates with the cross-spectrum analysis and coupling discrimination unit. It includes a torque controller, a pitch controller, an adaptive notch filter, and an additional damping control module.

[0054] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0055] 1. Wide adaptability: The order tracking technology enables the invention to accurately lock the blade passing frequency and natural frequency under variable speed and variable speed operating conditions, overcoming the limitations of traditional spectrum analysis and eliminating the influence of speed fluctuations.

[0056] 2. Accurate identification: The joint cross-spectral density, coherence function and transfer function phase are used. The coherence function eliminates the influence of random noise such as turbulence, resulting in a low misjudgment rate. The phase characteristics ensure accurate differentiation of coupling types, and can clearly distinguish between forced coupling and modal coupling.

[0057] 3. Flexible control: It provides a variety of active means such as speed regulation, pitch angle adjustment, adaptive notch filtering, and additional damping control, which can be intelligently selected according to the coupling strength and operating boundaries.

[0058] 4. Safety and Economy: Independent safety protection units work in conjunction with active control to ensure unit safety under extreme operating conditions; no additional expensive hardware is required, and closed-loop operation can be achieved using only existing sensors and controllers; existing encoders are used to significantly reduce fatigue loads on the tower and blades, extending their service life; a safe exit mechanism is also provided to avoid over-adjustment.

[0059] 5. Easy to integrate: The sensor, data acquisition, analysis and control are integrated. The algorithm can be embedded in the wind turbine main controller or independent vibration monitoring module, and is compatible with existing wind power systems. Attached Figure Description

[0060] Figure 1 This is a flowchart of a method for frequency coupling identification and active decoupling control of wind turbine blades and towers.

[0061] Figure 2 The graph shows the natural frequency of the blade as a function of rotational speed and a schematic diagram of the natural frequency of the tower.

[0062] Figure 3 The order domain cross-spectral amplitude, coherence function, and transfer function phase diagram are obtained after order tracking processing.

[0063] Figure 4 This is an architecture diagram of the frequency coupling identification and active decoupling control system for wind turbine blades and towers.

[0064] Figure 5 A time-domain comparison of the vibration acceleration at the top of the tower before and after control. Detailed Implementation

[0065] The present invention will be further described below with reference to specific embodiments.

[0066] Example 1

[0067] See Figure 1 The image shows the wind turbine blade and tower frequency coupling identification and active decoupling control method provided in this embodiment, which includes the following steps:

[0068] S1. Acquire the blade vibration signal, tower vibration signal, and instantaneous speed signal of the wind turbine; at a synchronous sampling rate f s The blade vibration signal x(t), tower vibration signal y(t), and instantaneous impeller speed pulse n(t) are collected continuously for a duration T≥60 s. The instantaneous impeller speed pulse n(t) is converted into an instantaneous speed signal Ω(t), and the unit of the instantaneous speed signal Ω(t) is rpm.

[0069] S2. The blade vibration signal x(t) and tower vibration signal y(t) are preprocessed with detrending and bandpass filtering. The pass frequency range of the bandpass filter is set to 0.1~50 Hz to cover the main modes of the blade and tower. The key phase time scale is calculated based on the instantaneous impeller speed pulse n(t). The order tracking (COT) method is used to resample the blade vibration signal x(t) and tower vibration signal y(t) into angle domain signals x(θ) and y(θ) to eliminate the influence of speed variation on frequency analysis. The angle domain signal is segmented at equal angular intervals per revolution to prepare for order spectrum analysis.

[0070] S3. Perform Fast Fourier Transform on the blade vibration signal, tower vibration signal and instantaneous rotation speed signal after preprocessing in step S2, and further calculate the cross-order spectral density, self-order spectral density, amplitude squared coherence function and transfer function; perform FFT on the resampled x(θ) and y(θ) signals to obtain the order spectra X(O) and Y(O), where the order O = f / (Ω / 60) is dimensionless.

[0071] Calculate the mutual order spectral density ;

[0072] Calculate the self-order spectral density S xx (O) and S yy (O);

[0073] Calculate the squared amplitude coherence function ;

[0074] Calculate the transfer function H xy (O) = S xy (O) / S xx (O), and converted into a frequency domain transfer function through the speed relationship.

[0075] in, For discrete frequency points, determined by the sampling rate The number of FFT points N determines the result: K represents the average number of times the long signal is segmented based on the Welch method. The larger the value, the smaller the variance and the smoother the spectral estimation, but the frequency resolution will decrease. Usually, 8 to 32 segments are used with an overlap rate of 50%. k is the index segment, representing the k-th segment of data, k=1,2,...,K. Let x(t) be the Fourier transform of the signal x(t) of the k-th blade segment. It is a complex number containing amplitude and phase, corresponding to blade vibration, such as blade root bending moment and blade tip acceleration. Let y(t) be the conjugate complex number of the Fourier transform of the signal y(t) of the k-th tower segment. The conjugate operation (*) is to preserve the relative phase difference between the two signals when multiplying in the frequency domain. If the conjugate is not added, the phase information will be lost. This is a cross-spectral density estimate, also a complex number. Its real part represents the in-phase component (energy exchange), and its imaginary part represents the quadrature component (phase lag / lead). At the coupling frequency, this value increases significantly.

[0076] S4, see also Figure 2 As shown, the system stores the blade natural frequency curves and tower natural frequencies calibrated offline, and outputs the coupling type, coupling frequency, and calculated amplitude squared coherence value based on the cross-order spectral density, self-order spectral density, amplitude squared coherence function, transfer function, and preset discrimination rules.

[0077] Establish calibration curves f for the blade pass order (1, 2, 3, ...) and the blade natural frequency as a function of rotational speed. b (Ω); When running online, based on the current average speed calculate:

[0078] Theoretical blade passing frequency f 1P = / 60, f 3P =3 / 60;

[0079] Theoretical blade natural frequency f b ( (This can be determined by looking up a table or by real-time identification using online modal analysis (OMA).)

[0080] Based on the cross-order spectral density, self-order spectral density, and frequency domain transfer function obtained in step S3, the coupling frequency f is found. c Or the corresponding order O c The amplitude squared coherence value is calculated based on the amplitude squared coherence function in step S3, satisfying the amplitude squared coherence value γ. xy 2 (f c )>γ th γ th =0.7, the preset discrimination rules are as follows:

[0081] (1) When At that time, the coupling type is determined to be 3P forced coupling, and the phase characteristic is |∣∠H. xy (f c ) |<60°;

[0082] (2) When and At that time, the coupling type is determined to be 1:1 modal coupling, and the phase characteristic is |∣∠H xy (f c ) |>75°;

[0083] (3) When or When n=2,3, the coupling type is determined to be subharmonic coupling, and the phase characteristics are unrestricted.

[0084] Where Δf is taken as half of the frequency resolution, f t This is the first-order natural frequency of the tower.

[0085] S5. Execute the corresponding control action based on the determined coupling type;

[0086] When the coupling type is 3P forced coupling, calculate the target rotational speed offset ΔΩ, and make the new... satisfy B is the tower resonance half-power bandwidth; by changing the generator torque setpoint, the speed is slowly adjusted to Ω. new =Ω+ΔΩ, the amplitude is limited within the preset safe operating range. In this embodiment, the preset safe operating range is 0.7-1.1 times the rated speed. The pitch angle is increased by 1~2° simultaneously to reduce the aerodynamic excitation amplitude.

[0087] When the coupling type is 1:1 modal coupling, the calibration curve is used to calculate... Deviation Required speed change If the target speed deviation ΔΩ exceeds the allowable operating range, then large-scale speed regulation will be abandoned, and instead, the adaptive notch filter integrated into the wind turbine main controller will be activated, with the notch center frequency set to the current coupling frequency. The notch depth is automatically adjusted based on the coherence value. The quality factor Q=10; simultaneously, the high-frequency damping component in the generator torque fluctuation is increased through additional torque. Implementation, in which This is the speed signal at the top of the tower;

[0088] In all coupling cases, when the squared amplitude coherence value γ 2 (f c When the pitch is greater than 0.85, a resonance warning signal is sent to the main control system, and the pitch rate is temporarily increased to quickly get out of the dangerous speed.

[0089] S6. Real-time calculation of amplitude squared coherence value γ 2 (f c ), when γ lasts for 30 seconds 2 (f c When all values ​​are below 0.3, the original speed of the wind turbine is gradually restored or the adaptive notch filter integrated in the main controller of the wind turbine is turned off, exiting the active control mode, and realizing the identification and active decoupling control of the frequency coupling between the blade and the tower.

[0090] S7. Based on the blade vibration signal and tower vibration signal obtained in step S1, continuously monitor the peak acceleration at the top of the tower and the blade strain during the control process; if the blade vibration signal or tower vibration signal does not exceed the preset threshold, no operation is performed; if the blade vibration signal or tower vibration signal exceeds the preset threshold, immediately control the wind turbine to stop in an emergency feathering mode.

[0091] Example 2

[0092] See Figure 4 As shown, this embodiment discloses a wind turbine blade and tower frequency coupling identification and active decoupling control system, used to implement the wind turbine blade and tower frequency coupling identification and active decoupling control method described in Embodiment 1, including:

[0093] The sensor unit is used to sense and output blade vibration signals, tower vibration signals, and instantaneous speed signals. The sensor unit includes a blade vibration sensor, a tower vibration sensor, and a speed encoder. The blade vibration sensor is installed at the root or tip of the wind turbine blade to sense and output blade vibration signals. The tower vibration sensor is installed at the top of the wind turbine tower and communicates with the active control unit to sense and output tower vibration signals. The speed encoder is installed on the impeller shaft or generator input end of the wind turbine to sense impeller speed pulse signals and calculate and output instantaneous speed signals based on the impeller speed pulse signals.

[0094] The data acquisition and preprocessing unit is used to acquire blade vibration signals, tower vibration signals, and instantaneous rotational speed signals, and perform preprocessing such as detrending, bandpass filtering, and order tracking resampling. The data acquisition and preprocessing unit includes a multi-channel synchronous acquisition card, an anti-aliasing filter, and a signal conditioning circuit that are connected in sequence. The multi-channel synchronous acquisition card is connected in communication with the sensor unit.

[0095] The order tracking resampling unit calculates the order tracking algorithm based on the instantaneous rotation speed signal and converts the time domain signal into a uniform signal in the angle domain. The order tracking resampling unit includes a digital signal processor, a field-programmable gate array, and an industrial computing module that are connected in sequence. The digital signal processor is connected in communication with the data acquisition and preprocessing unit.

[0096] The cross-spectral analysis and coupling discrimination unit performs piecewise fast Fourier transform on the uniform signal in the angle domain, and further calculates the cross-order spectral density, the self-order spectral density, the amplitude square coherence function, and the transfer function; it stores the offline calibrated blade natural frequency curves and tower natural frequencies, and outputs the coupling type, coupling frequency, and calculated amplitude square coherence value according to the discrimination rules.

[0097] The active control unit calculates the speed deviation, pitch angle increment, and additional torque based on the coupling frequency, blade vibration signal, tower vibration signal, and instantaneous speed signal, and performs adaptive notch filtering to actively control the wind turbine. The active control unit is integrated into the wind turbine main controller and communicates with the cross-spectrum analysis and coupling discrimination unit. It includes a torque controller, pitch controller, adaptive notch filter, and additional damping control module.

[0098] The safety protection unit is used to independently monitor the peak acceleration at the top of the tower and the blade strain.

[0099] Example 3

[0100] This embodiment discloses a wind turbine blade and tower frequency coupling identification and active decoupling control method based on the method described in Embodiment 1, which is applied to the identification and control of a 2.5 MW variable speed and pitch wind turbine where 3P forced coupling and blade natural frequency mode coupling exist simultaneously.

[0101] The parameters of the 2.5 MW variable speed and variable pitch wind turbine are as follows:

[0102] Rated speed: 14 rpm; Number of blades: 3; First natural frequency of the tower =0.95 Hz.

[0103] Blade first-order flapping natural frequency calibration: 0.80 Hz at 5 rpm; 0.92 Hz at 10 rpm; 0.96 Hz at 12 rpm; 1.00 Hz at 14 rpm (centrifugal stiffening).

[0104] 3P frequency range: 0.25~0.70 Hz (corresponding to 5~14 rpm).

[0105] At 12 rpm, = 0.6 Hz, and They do not overlap; but =0.96 Hz and If the Hz value is very close to 0.95 Hz, then modal coupling exists.

[0106] At 14 rpm, = 0.7 Hz, still lower than However, the design goal is to demonstrate forced coupling. Assuming that the other wind turbine tower has a lower frequency, a scenario is added for a complete demonstration: the wind turbine is running at 19 rpm (over-rated, only allowed for a short time). = 0.95 Hz, equal to Forced coupling occurs.

[0107] The identification and control of the two types of coupling are described below. In practical applications, this control method can handle both simultaneously.

[0108] Scenario 1: 1:1 modal coupling between blade natural frequency and tower frequency (12 rpm)

[0109] (1) Signal acquisition:

[0110] a. Sampling frequency 200 Hz, continuous sampling for 120 seconds.

[0111] b. The speed encoder measured an average speed of 12.1 rpm, with a fluctuation of ±0.2 rpm.

[0112] c. Blade root bending moment signal, tower top acceleration signal.

[0113] (2) Order tracking and cross-spectral analysis:

[0114] a. The angle domain signal is obtained by resampling the rotational speed pulse to 1024 points per revolution.

[0115] b. Calculate the order spectrum and find the cross-spectral peak at order O = 0.95 / (12.1 / 60) = 4.71.

[0116] c. Convert the order to frequency. =4.71×(12.1 / 60)=0.95 Hz.

[0117] d. Coherence function γ 2 (0.95Hz)=0.93, transfer function phase -87∘.

[0118] e. From the table (12.1 rpm) = 0.96 Hz, and =0.95 Hz, deviation less than 0.02 Hz.

[0119] f. The condition is determined to be blade natural frequency-tower mode coupling.

[0120] (3) Active control, see Figure 3 As shown:

[0121] in, Figure 3 (a) is the cross-spectral amplitude spectrum of the order domain. A significant peak appears at order 3, corresponding to 3P forced coupling. Another peak appears at order 5, corresponding to the modal coupling between the blade natural frequency and the tower natural frequency. Figure 3 (b) is the squared amplitude coherence function γ 2 At the coupling order, the coherence value is close to 1, indicating that the two signals are highly linearly correlated at this order. Figure 3(c) is the phase spectrum of the transfer function. The phase at the forced coupling point is about -30° and the phase at the modal coupling point is about -88°, which reflects the resonance characteristics.

[0122] Calibration curve slope ≈(0.96-0.92) / (12-10)=0.02 Hz / rpm.

[0123] Objective: To make Moving to a frequency beyond 0.85 Hz or 1.05 Hz results in a half-power bandwidth of 0.04 Hz and a safety interval of 0.12 Hz. Choosing the descent direction requires Δf = -0.11 Hz, corresponding to ΔΩ = -0.11 / 0.02 = -5.5 rpm.

[0124] The new speed = 12.1 - 5.5 = 6.6 rpm, which is still within the allowable range of 5~14 rpm.

[0125] Control execution, see Figure 1 As shown: The generator torque is gradually reduced, and the speed drops to 6.6 rpm within 45 seconds.

[0126] Simultaneously, an adaptive notch filter is activated with a center frequency of 0.95 Hz and a depth of -9.3 dB (γ). 2 =0.93).

[0127] Real-time coherent monitoring: When the speed drops to 9 rpm, fb≈0.87 Hz, and the coherence drops to 0.5; at 6.6 rpm, the coherence is 0.21.

[0128] Stop adjusting and maintain 6.6 rpm for 10 minutes. The RMS vibration at the top of the column decreased from 0.32 g to 0.09 g. See [link to relevant documentation]. Figure 5 As shown.

[0129] Scenario 2: 3P forced coupling (19 rpm, short-term operation)

[0130] (1) Identification:

[0131] a. Rotational speed 19 rpm = 0.95 Hz.

[0132] b. Cross-spectral analysis at 0.95 Hz γ 2 =0.88γ 2 =0.88, phase angle -32° -32°.

[0133] c. Natural frequency of the blade (19 rpm) is 1.08 Hz, which is unrelated to 0.95 Hz.

[0134] d. Determined to be 3P forced coupling.

[0135] (2) Active control, see Figure 1 As shown:

[0136] a. Calculate the target rotational speed: make 3P deviate from f t It is 0.12 Hz, which is the new = 0.83 Hz or 1.07 Hz. Select downward adjustment: New = 0.83 Hz → New speed = 0.83×60 / 3 = 16.6 rpm.

[0137] b. Adjust the torque to reduce the speed to 16.6 rpm in 30 seconds.

[0138] c. Simultaneously increase the pitch angle by 1.5° to reduce aerodynamic excitation.

[0139] d. The coherence function drops to 0.28, and the vibration decreases by 68%.

[0140] Example 4

[0141] This embodiment discloses a non-transitory computer-readable medium storing instructions that, when executed by a processor, perform the steps of the wind turbine blade and tower frequency coupling identification and active decoupling control method according to Embodiment 1.

[0142] In this embodiment, the non-transitory computer-readable medium can be a disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), USB flash drive, portable hard drive, etc.

[0143] Example 5

[0144] This embodiment discloses a computing device, including a processor and a memory for storing processor-executable programs. When the processor executes the program stored in the memory, it implements the wind turbine blade and tower frequency coupling identification and active decoupling control method described in Embodiment 1.

[0145] The computing device described in this embodiment may be a desktop computer, laptop computer, smartphone, PDA handheld terminal, tablet computer, programmable logic controller (PLC), or other terminal device with processor function.

[0146] The above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Therefore, any changes made in accordance with the shape and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for frequency coupling identification and active decoupling control of wind turbine blades and towers, characterized in that, Includes the following steps: S1. Acquire the blade vibration signal, tower vibration signal, and instantaneous speed signal of the wind turbine; S2. Perform preprocessing on the blade vibration signal and tower vibration signal, including detrending, bandpass filtering, and order tracking resampling; S3. Perform a fast Fourier transform on the blade vibration signal, tower vibration signal and instantaneous rotation speed signal after the preprocessing in step S2, and further calculate the cross-order spectral density, self-order spectral density, amplitude squared coherence function and transfer function. S4. Store the blade natural frequency curves and tower natural frequencies calibrated offline, and output the coupling type, coupling frequency, and calculated amplitude squared coherence value based on the cross-order spectral density, self-order spectral density, amplitude squared coherence function, transfer function, and preset discrimination rules. S5. Execute the corresponding control action based on the determined coupling type; S6. Calculate the real-time amplitude squared coherence value. When the real-time amplitude squared coherence value is less than the preset threshold, gradually restore the original speed of the wind turbine or turn off the adaptive notch filter integrated in the main controller of the wind turbine, exit the active control mode, and realize the identification and active decoupling control of the frequency coupling between the blade and the tower.

2. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 1, characterized in that, Includes the following steps: S7. Based on the blade vibration signal and tower vibration signal obtained in step S1, continuously monitor the peak acceleration at the top of the tower and the blade strain during the control process; if the blade vibration signal or tower vibration signal does not exceed the preset threshold, no operation is performed. If the blade vibration signal or tower vibration signal exceeds the preset threshold, the wind turbine will be immediately controlled to shut down in an emergency feathering mode.

3. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 1, characterized in that, Step S1 includes: With synchronized sampling rate The blade vibration signal x(t), the tower vibration signal y(t), and the instantaneous impeller speed pulse n(t) are collected continuously for a duration T≥60 s; the instantaneous impeller speed pulse n(t) is converted into an instantaneous speed signal Ω(t).

4. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 3, characterized in that, Step S2 includes: The blade vibration signal x(t) and the tower vibration signal y(t) are preprocessed by detrending and bandpass filtering. The pass frequency range of the bandpass filter is set to 0.1~50 Hz to cover the main modes of the blade and the tower. The key phase timescale is calculated based on the instantaneous impeller speed pulse n(t). The order tracking (COT) method is used to resample the blade vibration signal x(t) and the tower vibration signal y(t) into angle domain signals x(θ) and y(θ) to eliminate the influence of speed variation on frequency analysis. The angle domain signal is segmented at equal angular intervals per revolution to prepare for order spectrum analysis.

5. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 4, characterized in that, Step S3 includes: Perform FFT on the resampled x(θ) and y(θ) signals to obtain the order spectra X(O) and Y(O), where the order O = f / (Ω / 60) is dimensionless; Calculate the mutual order spectral density ; Calculate the self-order spectral density S xx (O) and S yy (O); Calculate the squared amplitude coherence function ; Calculate the transfer function H xy (O) = S xy (O) / S xx (O), and converts it into a frequency domain transfer function through the speed relationship; in, For discrete frequency points, determined by the sampling rate The number of FFT points N determines the result: K represents the average frequency, which is used to segment long signals based on the Welch method. The larger the value, the smaller the variance, the smoother the spectral estimation, but the lower the frequency resolution. k represents the index segment, indicating the k-th segment of data, k=1,2,...,K. Let x(t) be the Fourier transform of the signal x(t) of the k-th blade segment, which is a complex number containing amplitude and phase, corresponding to the blade vibration; Let y(t) be the conjugate complex number of the Fourier transform of the signal y(t) of the k-th tower segment; This is an estimate of the cross-spectral density, also a complex number. Its real part represents the in-phase component, and its imaginary part represents the quadrature component. This value increases at the coupling frequency.

6. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 5, characterized in that, Step S4 includes: Establish calibration curves f of the blade's pass order and natural frequency as a function of rotational speed. b (Ω); When running online, based on the current average speed calculate: Theoretical blade passing frequency f 1P = / 60, f 3P =3 / 60; Theoretical blade natural frequency f b ( (This can be determined by looking up a table or by real-time identification using online modal analysis (OMA).) Based on the cross-order spectral density, self-order spectral density, and frequency domain transfer function obtained in step S3, the coupling frequency is found. Or the corresponding order O c The amplitude squared coherence value is calculated based on the amplitude squared coherence function in step S3, satisfying the amplitude squared coherence value γ. xy 2 (f c )>γ th The preset discrimination rules are as follows: (1) When At that time, the coupling type is determined to be 3P forced coupling, and the phase characteristic is |∣∠H. xy (f c ) |<60°; (2) When and At that time, the coupling type is determined to be 1:1 modal coupling, and the phase characteristic is |∣∠H xy (f c ) |>75°; (3) When or When n=2 or 3, the coupling type is determined to be subharmonic coupling, and the phase characteristics are unrestricted. Where Δf is taken as half of the frequency resolution, This is the first-order natural frequency of the tower.

7. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 6, characterized in that, Step S5 includes: When the coupling type is 3P forced coupling, calculate the target rotational speed offset ΔΩ, and make the new... satisfy B is the tower resonance half-power bandwidth; by changing the generator torque setpoint, the speed is slowly adjusted to Ω. new =Ω+ΔΩ, the amplitude is limited within the preset safe operating range, and the pitch angle is increased by 1~2° simultaneously to reduce the aerodynamic excitation amplitude; When the coupling type is 1:1 modal coupling, the calibration curve is used to calculate... Deviation Required speed change If the target speed deviation ΔΩ exceeds the allowable operating range, then large-scale speed regulation will be abandoned, and instead, the adaptive notch filter integrated into the wind turbine main controller will be activated, with the notch center frequency set to the current coupling frequency. The notch depth is automatically adjusted based on the coherence value. The quality factor Q=10; simultaneously, the high-frequency damping component in the generator torque fluctuation is increased through additional torque. Implementation, in which This is the speed signal at the top of the tower; In all coupling cases, when the squared amplitude coherence value γ 2 (f c When the pitch is greater than 0.85, a resonance warning signal is sent to the main control system, and the pitch rate is temporarily increased to quickly get out of the dangerous speed.

8. The method for frequency coupling identification and active decoupling control of wind turbine blades and towers according to claim 7, characterized in that, Step S6 includes: Real-time calculation of amplitude squared coherence value γ 2 (f c ), when γ lasts for 30 seconds 2 (f c When all values ​​are below 0.3, gradually restore the original speed or turn off the notch filter to exit the active control mode.

9. A frequency coupling identification and active decoupling control system for wind turbine blades and towers, characterized in that, The method for implementing the frequency coupling identification and active decoupling control of wind turbine blades and towers as described in any one of claims 1-8 includes: The sensor unit is used to sense and output blade vibration signals, tower vibration signals, and instantaneous rotational speed signals; The data acquisition and preprocessing unit is used to acquire blade vibration signals, tower vibration signals, and instantaneous rotational speed signals, and perform preprocessing such as detrending, bandpass filtering, and order tracking resampling. The order tracking resampling unit calculates the order tracking algorithm based on the instantaneous rotation speed signal, converting the time domain signal into a uniform signal in the angle domain; The cross-spectral analysis and coupling discrimination unit performs piecewise fast Fourier transform on the uniform signal in the angle domain, and further calculates the cross-order spectral density, the self-order spectral density, the amplitude squared coherence function, and the transfer function; it stores the offline calibrated blade natural frequency curves and tower natural frequencies, and outputs the coupling type, coupling frequency, and calculated amplitude squared coherence value according to the discrimination rules. The active control unit calculates the speed deviation, pitch angle increment, and additional torque based on the coupling frequency, blade vibration signal, tower vibration signal, and instantaneous speed signal, and performs adaptive notch filtering to actively control the wind turbine. The safety protection unit is used to independently monitor the peak acceleration at the top of the tower and the blade strain.

10. The wind turbine blade and tower frequency coupling identification and active decoupling control system according to claim 9, characterized in that: The sensor unit includes a blade vibration sensor, a tower vibration sensor, and a speed encoder. The blade vibration sensor is installed at the root or tip of the wind turbine blade to sense and output blade vibration signals. The tower vibration sensor is installed at the top of the wind turbine tower and communicates with the active control unit to sense and output tower vibration signals. The speed encoder is installed on the impeller shaft or generator input end of the wind turbine to sense impeller speed pulse signals, calculate and output instantaneous speed signals based on the impeller speed pulse signals. The data acquisition and preprocessing unit includes a multi-channel synchronous acquisition card, an anti-aliasing filter, and a signal conditioning circuit that are connected in sequence. The multi-channel synchronous acquisition card is connected in communication with the sensor unit. The order tracking resampling unit includes a digital signal processor, a field-programmable gate array, and an industrial computing module that are connected in sequence. The digital signal processor is also connected in communication with the data acquisition and preprocessing unit. The active control unit is integrated in the wind turbine main controller and communicates with the cross-spectrum analysis and coupling discrimination unit. It includes a torque controller, a pitch controller, an adaptive notch filter, and an additional damping control module.

Images