Wind turbine generator time-varying vibration signal noise reduction method and device based on adaptive chirp mode decomposition and parameter optimization

Abstract

The invention discloses a wind turbine generator time-varying vibration signal noise reduction method and device based on adaptive chirp mode decomposition and parameter optimization, and relates to the field of signal noise reduction, the method comprises the following steps: constructing an adaptive chirp mode decomposition model, and carrying out parameter initialization; extracting a component dominant chirp mode in the vibration signal of the wind turbine generator by using an adaptive chirp mode decomposition model; performing signal reconstruction on the plurality of component dominant chirp modes to obtain a noise reduction signal; calculating the time-frequency kurtosis of the noise reduction signal, and optimizing a bandwidth coefficient and a smoothing coefficient by using a Bayesian optimization algorithm according to the time-frequency kurtosis; and carrying out loop iteration on the above steps until the maximum number of iterations is reached, and outputting the optimal bandwidth coefficient, smoothing coefficient and noise reduction signal. The accuracy of chirp mode component extraction is improved, the noise suppression level of the time-varying vibration signal of the wind turbine generator is further improved, and the fault diagnosis and state early warning level is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of signal denoising, in particular to a wind turbine time-varying vibration signal denoising method and device based on adaptive chirplet modal decomposition and parameter optimization. BACKGROUND

[0002] In the transmission chain of a wind power system, the gear box as a key component of energy transmission has long been subjected to the influence of dynamic load caused by time-varying wind speed, and its operating state is directly related to the safety and reliability of the wind turbine. Under the time-varying speed condition, the vibration signal of the gear box presents significant non-stationary characteristics, which aggravates the vibration signal modulation and forms a complex coupling with the environmental noise, resulting in nonlinear distortion of the characteristic frequency in the time-frequency domain. This signal spectrum smearing problem caused by speed fluctuation poses a serious challenge to the traditional vibration analysis method based on the steady-state assumption.

[0003] In summary, the current wind turbine time-varying vibration signal is low, which hinders the fault diagnosis and state warning of the wind turbine. SUMMARY

[0004] The purpose of the present application is to provide a wind turbine time-varying vibration signal denoising method based on adaptive chirplet modal decomposition and parameter optimization, which comprises the following steps: S1. obtaining the vibration signal of the wind turbine; S2. constructing an adaptive chirplet modal decomposition model and initializing the target parameters of the model; the target parameters at least include: bandwidth coefficient a, smoothing coefficient b and maximum iteration number; S3. using the adaptive chirplet modal decomposition model to extract the component dominant chirplet mode in the vibration signal of the wind turbine; S4. after signal reconstruction of multiple component dominant chirplet modes, a denoising signal is obtained; S5. calculating the time-frequency kurtosis K TF of the denoising signal according to the time-frequency kurtosis K TF optimizing the bandwidth coefficient a and the smoothing coefficient b by using the Bayesian optimization algorithm; S6. repeating steps S2-S5 until the maximum iteration number is reached, and outputting the optimal bandwidth coefficient a, the optimal smoothing coefficient b and the optimal denoising signal.

[0005] In the second aspect, the present application provides a computer device, comprising: a memory, a processor to store a computer program on the memory and run the computer program on the processor, and the processor executes the computer program to realize the wind turbine time-varying vibration signal denoising method based on adaptive chirplet modal decomposition and parameter optimization described above.

[0006] According to the specific embodiments provided by the present application, the following technical effects are disclosed:

[0007] The application firstly collects the vibration signal of the wind turbine, constructs an adaptive chirp modal decomposition model and initializes target parameters such as a bandwidth coefficient, a smoothing coefficient and a maximum iteration number; the model is used to extract a component dominant chirp mode in the vibration signal, the dominant chirp mode is reconstructed into a noise reduction signal, the time-frequency kurtosis index of the noise reduction signal is calculated, and the bandwidth coefficient and the smoothing coefficient are dynamically optimized based on the index through a Bayesian optimization algorithm; the parameters are continuously updated through multiple iterations, and finally the optimal bandwidth coefficient, the optimal smoothing coefficient and the optimal noise reduction signal are output. The method significantly improves the extraction accuracy of the chirp modal component, effectively enhances the noise suppression capability of the time-varying vibration signal, and thus improves the fault diagnosis accuracy and state early warning reliability of the wind turbine. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technical solutions, the drawings needed to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0009] Figure 1 A wind turbine time-varying vibration signal noise reduction method based on adaptive chirp modal decomposition and parameter optimization provided by an embodiment of the present application Figure 1 .

[0010] Figure 2 A wind turbine time-varying vibration signal diagram provided by an embodiment of the present application; wherein, Figure 2 (a) a wind turbine time-varying vibration signal time domain diagram; Figure 2 (b) a wind turbine time-varying vibration signal FFT frequency spectrum; Figure 2 (c) a wind turbine time-varying vibration signal STFT time-frequency diagram.

[0011] Figure 3 An initial adaptive chirp modal decomposition model provided by an embodiment of the present application extracts different components.

[0012] Figure 4 A noise signal diagram after initial parameter reconstruction provided by an embodiment of the present application.

[0013] Figure 5 An adaptive chirp modal decomposition model provided by an embodiment of the present application extracts different components after parameter optimization.

[0014] Figure 6 A noise reduction result diagram of the adaptive chirp modal decomposition model after parameter optimization provided by an embodiment of the present application.

[0015] Figure 7A wind turbine time-varying vibration signal denoising method based on adaptive chirp mode decomposition and parameter optimization is provided Figure 2 .

[0016] Figure 8 A structural diagram of a computer device is provided for the embodiments of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0018] The performance of the traditional signal noise method is limited in that: first, the signal stationarity requirement is relatively high. For example, the EMD method relies on the search of local extreme values in the time domain of the signal, and when the signal frequency fluctuates dramatically (such as the time-varying speed working condition), the extraction error of the local extreme value points will exacerbate the modal aliasing. The fixed basis function of WT (Wavelet transform) cannot adapt to the frequency change. Second, the limitation of the narrowband signal assumption of the method itself. For example, VMD (Variational mode decomposition) and EWT (Empirical wavelet transform) assume that the signal components are distributed in narrow bands on the fast Fourier transform spectrum, and separate the modes through a pre-set filter group.

[0019] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0020] Embodiment 1, as shown in Figure 1 and Figure 7 The present embodiment provides a wind turbine time-varying vibration signal denoising method based on adaptive chirp mode decomposition and parameter optimization, which comprises:

[0021] S1. Obtain the vibration signal of the wind turbine.

[0022] S2. Construct an adaptive chirp mode decomposition model (ACMD), and initialize the target parameters of the model; the target parameters at least include: bandwidth coefficient α, smoothing coefficient β and maximum iteration number.

[0023] S3. Extract the dominant chirp mode of the wind turbine vibration signal using the adaptive chirp mode decomposition model.

[0024] Further, the step S3 specifically comprises:

[0025] S31. Construct an optimization problem based on the wind turbine vibration signal.

[0026] S32. Discretize the optimization problem to obtain a discretized optimization problem.

[0027] Further, the expression of the optimization problem is as follows:

[0028]

[0029] a i = [a i (t0), …, a i (t N-1 )] T .

[0030] b i = [b i (t0), …, b i (t N-1 )] T .

[0031] In the formula, a i (t) and b i (t) are two demodulation signals of the signal component i at t time; and are the signal smoothness of the two demodulation signals of the signal component i at t time; f i (t) is the instantaneous frequency of the signal component i at t time; is the total energy of the signal component at t time x(t) after removing the energy x i (t) of the signal component i at t time; and α is a bandwidth coefficient, which is used to represent a weighting coefficient.

[0032] S33. Solve the discretized optimization problem to obtain the demodulation signal and the instantaneous frequency.

[0033] Further, the expression of the discretized optimization problem is as follows:

[0034]

[0035] In the formula, Θ is a diagonal matrix wherein Ω is a second-order difference matrix; u i is the time-domain signal of the i-th component; f i is the instantaneous frequency of the signal; and G iGini coefficient of the signal component i.

[0036] Optionally, the calculation formula of the Gini coefficient is as follows:

[0037]

[0038] In the formula, diag[·] represents a diagonal matrix.

[0039] Further, the calculation formula of the instantaneous frequency is as follows:

[0040]

[0041] In the formula, f i j+1 is the instantaneous frequency of the signal component i at the j+1th iteration of t; f i j is the total instantaneous frequency of the signal component i at the jth iteration of t0-N-1; f is the instantaneous frequency of the signal component i at the jth iteration of t; β is a smoothing coefficient; Ω is a second-order difference matrix of size (N-2)×N; N is the total number of time points; T is a transpose; and I is a unit matrix. is the total frequency increment of the signal component i at the jth iteration of t0-N-1; f is the instantaneous frequency increment of the signal component i at the jth iteration of t.

[0042] Further, the calculation formula of the frequency increment is as follows:

[0043]

[0044] In the formula, diag[·] represents a diagonal matrix. is the instantaneous frequency increment of the signal component i at the jth iteration of t; f and are two demodulation signals of the signal component i at the jth iteration of t.

[0045] S34. Looping the steps S31-S33, calculating the signal residual energy value, and constructing a component dominant chirp mode based on all the extracted demodulation signals and instantaneous frequencies, iterating until the signal residual energy value is less than the residual energy threshold, stopping the iteration, and obtaining a plurality of component dominant chirp modes.

[0046] Further, the calculation formula of the signal residual energy value is as follows:

[0047]

[0048] In the formula, E is the signal residual energy value; s i is the estimation result of the ith component, and δ is an energy threshold.

[0049] Optionally, the residual energy threshold (cycle termination indicator) is calculated as follows:

[0050]

[0051] wherein, represents the error between the signal component i at the jth and (j-1)th iteration; represents the estimation result of the signal component i at the jth iteration of the jth iteration; represents the estimation result of the signal component i at the (j-1)th iteration of the jth iteration.

[0052] In actual application, the iteration is stopped when the signal residual energy value is less than the preset convergence accuracy by repeating the above iteration steps, that is, a target mode is extracted from the original signal. Until the signal residual energy ratio is less than the set threshold δ, that is, all signal components in the original signal have been extracted, the iteration is stopped, and the cycle is ended.

[0053] S5. Calculate the time-frequency kurtosis K TF of the denoised signal, and according to the time-frequency kurtosis K TF , optimize the bandwidth coefficient α and the smoothing coefficient β by using the Bayesian optimization algorithm (BOA).

[0054] Further, the time-frequency kurtosis K TF of the denoised signal is calculated, and according to the time-frequency kurtosis K TF , the bandwidth coefficient α and the smoothing coefficient β are optimized by using the Bayesian optimization algorithm, which specifically includes:

[0055] S51. According to the impact characteristics of the time domain and the frequency domain, the time-frequency kurtosis K TF of the denoised signal is calculated.

[0056] Further, the time-frequency kurtosis K TF includes the time domain kurtosis K T and the frequency domain kurtosis K F ; and the calculation formula of the time-frequency kurtosis K TF is as follows:

[0057] K TF = K T · K F .

[0058]

[0059] wherein, K T is the time domain kurtosis; KF is the kurtosis of frequency domain; N represents the total number of time points; x(t) is the vibration amplitude value corresponding to time point t; is the mean value of x(t); F(k) represents the amplitude value corresponding to frequency point k; M represents the total number of frequency points; represents the mean value of the spectrum amplitude.

[0060] S52. Take the time-frequency kurtosis K TF The objective function of the Bayesian optimization algorithm is established to maximize the target.

[0061] S53. Use the Bayesian optimization algorithm to iteratively search for the optimal bandwidth coefficient a and the smoothing coefficient b.

[0062] S6. Recursively iterate steps S2-S5 until the maximum iteration number is reached, and output the optimal bandwidth coefficient a, the optimal smoothing coefficient b, and the optimal denoising signal.

[0063] To verify the effectiveness of the above method, the application performs test verification on the simulated wind turbine gearbox intermediate stage motor side bearing fault data. The known fault type is bearing inner ring cracking. The signal sampling frequency is 1000 Hz, the sampling time is 5 s, and the signal-to-noise ratio SNR=10 dB. The time domain graph, FFT spectrum and STFT time-frequency graph are as shown in Figure 2 .

[0064] Step 1: Initialize the adaptive chirp modal decomposition model parameters, and set the optimization range and maximum optimization algebra of the bandwidth coefficient a and the smoothing coefficient b.

[0065] For this signal, the adaptive chirp modal decomposition model parameter residual energy threshold is set to 0.4, the optimization range of a and b is set to [10 -14 , 10 -4 ], the maximum iteration number is 20, and the maximum optimization time is 14 h.

[0066] Step 2: Use the adaptive chirp modal decomposition model method (initial parameters) to extract a plurality of CMs in the wind turbine signal.

[0067] 1) For a non-stationary signal x(t), an optimization problem is constructed.

[0068] 2) Discretize the above optimization problem.

[0069] 3) Update the demodulation signal.

[0070] 4) Calculate the frequency increment.

[0071] 5) Estimate the instantaneous frequency.

[0072] 6) Repeat the above iteration steps when When the convergence accuracy is less than the preset convergence accuracy, the iteration is stopped, and a single CM is extracted.

[0073] 7) Repeat the above iterative extraction process until the signal residual energy ratio is less than the set threshold, stop iteration, and end the loop.

[0074] Among them, the adaptive chirp modal decomposition model can extract 10 CM components for the signal, as shown in Figure 3 .

[0075] Step 3: Add the CMs obtained above to obtain a denoised signal.

[0076] Add the 10 CM components above to obtain a denoised signal, as shown in Figure 4 .

[0077] Step 4: Take the time-frequency kurtosis KTF of the above denoised signal as the objective function, and cycle steps 2 and 3 to optimize the bandwidth coefficient α and the smoothing coefficient β of the adaptive chirp modal decomposition model until the set maximum iteration number is reached, and the optimal α and β are obtained.

[0078] The Bayesian optimization algorithm optimization process is shown in Table 1. The observed maximum value of the objective function appears in the 14th generation, and the value is 8.3271. The observed optimal feasible point is α = 5.8441e -09 , β = 1.4729e -11 .

[0079] Table 1 Bayesian optimization algorithm parameter optimization table

[0080]

[0081] Step 5: Bring the final bandwidth coefficient α and smoothing coefficient β into the adaptive chirp modal decomposition model, and repeat steps 2 and 3 to obtain the final denoised signal.

[0082] The component extraction result of the adaptive chirp modal decomposition model is shown in Figure 5 . The reconstruction result, i.e. the denoising result, is shown in Figure 6 .

[0083] The signal denoising result is quantitatively evaluated by calculating the signal-to-noise ratio (SNR), the root mean square error (RMSE), and the correlation coefficient (CC). The calculation results are shown in Table 2.

[0084] Table 2 Quantitative evaluation of signal denoising results

[0085] SNR RMSE CC 10.5417 0.2447 0.9956

[0086] The technical effects of the present application are as follows:

[0087] The present application obtains an original vibration signal and initializes target parameters of an adaptive chirplet modal decomposition model; then uses the model to extract a plurality of dominant chirplet modes in the signal to reconstruct a denoised signal; further, the bandwidth coefficient and the smoothing coefficient are iteratively optimized by a Bayesian optimization algorithm to maximize the time-frequency kurtosis of the denoised signal; finally, the optimal parameters and the denoising result are output after reaching the maximum number of iterations. The present application uses the Bayesian optimization algorithm to match the optimal bandwidth coefficient and the smoothing coefficient in the adaptive chirplet modal decomposition, improves the accuracy of the chirplet modal component extraction, and further improves the noise suppression level of the wind turbine time-varying vibration signal, and improves the fault diagnosis and state warning level. Through testing, for the SNR = 10dB simulated wind turbine intermediate stage motor side bearing vibration signal, the present application can improve the signal-to-noise ratio by more than 10dB, and the root mean square error and the correlation coefficient of the original signal are 0.2447 and 0.9956, respectively.

[0088] In embodiment 2, the present application also provides a computer device, which can be a server or a terminal, and an internal structure diagram thereof can be as shown in Figure 8 The computer device includes a processor, a memory, an input / output interface (I / O), and a communication interface. The processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is used to store processing data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through network connection. The computer program is executed by the processor to implement the above-mentioned methods.

[0089] Those skilled in the art can understand that Figure 8 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0090] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0091] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist, it should be considered as the scope of the present application.

[0092] The principles and implementation modes of the present application are described by using specific examples in this paper, and the above-mentioned embodiments are only used to help understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. A method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization, characterized in that, The method includes: S1. Acquire vibration signals from the wind turbine generator; S2. Construct an adaptive chirped mode decomposition model and initialize the objective parameters of the model; the objective parameters include at least: bandwidth coefficient α, smoothness coefficient β, and maximum number of iterations; S3. Using an adaptive chirped mode decomposition model, extract the dominant chirped modes from the vibration signal of the wind turbine. S4. After reconstructing the signals of the multiple dominant chirped modes, a denoised signal is obtained; S5. Calculate the time-frequency kurtosis K of the denoised signal. TF According to the time-frequency kurtosis K TF The bandwidth coefficient α and the smoothness coefficient β are optimized using a Bayesian optimization algorithm. S6. Repeat steps S2-S5 until the maximum number of iterations is reached, and output the optimal bandwidth coefficient α, the optimal smoothness coefficient β, and the optimal noise reduction signal.

2. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 1, characterized in that, Using an adaptive chirped mode decomposition model, the dominant chirped modes in the vibration signal of a wind turbine are extracted, specifically including: S31. Optimization problem based on vibration signal of wind turbine unit; S32. Discretize the optimization problem to obtain the discretized optimization problem; S33. Solve the discretized optimization problem to obtain the demodulated signal and instantaneous frequency; S34. Iterate through steps S31-S33, calculate the signal residual energy value, and construct the component-dominated chirped mode based on all extracted demodulated signals and instantaneous frequencies. Stop iterating when the signal residual energy value is less than the residual energy threshold, and obtain multiple component-dominated chirped modes.

3. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 2, characterized in that, The expression for the optimization problem is as follows: In the formula, a i (t) and b i (t) represent the two demodulated signals of signal component i at time t; and f represents the signal smoothness of the two demodulated signals of signal component i at time t; i (t) represents the instantaneous frequency of signal component i at time t; Let x(t) be the total energy of the signal components at time t, after removing the energy of signal component i at time t. i The remaining energy after (t); α is the bandwidth coefficient, and α is used to characterize the weighting coefficient.

4. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 2, characterized in that, The expression for the discretized optimization problem is as follows: In the formula, Θ is a diagonal matrix. Where Ω is a second-order difference matrix; u i f is the time-domain signal of the i-th component; i G is the instantaneous frequency of the signal. i Let be the Gini coefficient of signal component i.

5. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 2, characterized in that, The formula for calculating the instantaneous frequency is as follows: In the formula, Let be the instantaneous frequency of signal component i in the (j+1)th iteration; Let be the total instantaneous frequency of signal component i during the time interval t0 to N-1 of the j-th iteration; Let be the instantaneous frequency of signal component i at time t in the j-th iteration; β be the smoothing coefficient; Ω be a second-order difference matrix of size (N-2)×N; N be the total number of time points; T be the transpose. I is a single identity matrix; Let i be the total frequency increment of signal component i during the time interval t0 to N-1 of the j-th iteration; Let be the instantaneous frequency increment of signal component i at time t during the j-th iteration.

6. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 2, characterized in that, The formula for calculating the frequency increment is as follows: In the formula, Let be the instantaneous frequency increment of signal component i at time t in the j-th iteration; and These are the two demodulated signals of signal component i at time t in the j-th iteration.

7. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 1, characterized in that, The formula for calculating the signal residual energy value is as follows: In the formula, E is the signal residual energy value; s i δ represents the estimated result of the i-th component, where δ is the energy threshold.

8. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 1, characterized in that, Calculate the time-frequency kurtosis K of the denoised signal TF According to the time-frequency kurtosis K TF The bandwidth coefficient α and the smoothing coefficient β are optimized using a Bayesian optimization algorithm, specifically including: S51. Calculate the time-frequency kurtosis K of the noise-reduced signal based on the impulse characteristics in the time and frequency domains. TF ; S52. Using time-frequency kurtosis K TF To maximize the objective function, we establish the objective function of the Bayesian optimization algorithm. S53. Using the Bayesian optimization algorithm, iteratively search for the optimal bandwidth coefficient α and smoothness coefficient β.

9. The method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization according to claim 1, characterized in that, The time-frequency kurtosis K TF Including: temporal kurtosis K T and frequency domain kurtosis K F The time-frequency kurtosis K TF The calculation formula is as follows: K TF =K T ·K F ; Among them, K T K represents the time-domain kurtosis. F Here, represents the frequency domain kurtosis; N represents the total number of time points; x(t) is the vibration amplitude corresponding to time t; Let x(t) be the mean of x(t); F(k) represent the amplitude corresponding to frequency point k; M represents the total number of frequency points. This represents the mean of the spectral amplitude.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method for denoising time-varying vibration signals of wind turbine generators based on adaptive chirped mode decomposition and parameter optimization as described in any one of claims 1-9.

Images