Wind driven generator vibration signal processing method, computer equipment and storage medium

Abstract

The invention discloses a method for processing a vibration signal of a wind driven generator. The method comprises the following steps: acquiring the vibration signal of the wind driven generator; performing sample expansion on the vibration signal of the wind driven generator to obtain a plurality of sections of vibration signals; performing adaptive variational mode decomposition on each section of vibration signal to obtain an intrinsic mode component signal set of each section of vibration signal; and performing fast Fourier transform on each intrinsic mode component signal in the intrinsic mode component signal set to obtain a frequency spectrum of each intrinsic mode component signal so as to obtain a local oscillation mode component signal frequency spectrum set of each section of vibration signal. The invention provides a more advanced and effective solution for processing the vibration signal required by the fault diagnosis of the wind driven generator, and solves the technical problems of single feature and insufficient information extraction in the existing vibration signal processing method of the wind driven generator.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine fault diagnosis data processing, and in particular to a wind turbine vibration signal processing method, computer equipment and storage medium. Background Art

[0002] As the world pays more and more attention to sustainable development and environmental protection, wind energy, as an important new energy technology, has been widely used around the world. The energy conversion process of wind turbines involves the cooperation of multiple key rotating mechanical components, and the health status of these components is directly related to the overall operation status and power generation efficiency of wind turbines.

[0003] The blades, bearings, and gearboxes of wind turbines are the core components of wind power generation systems. They will suffer various damages under long-term operation or in harsh environments. Blades will have defects such as sand holes, depressions, bulges, and cracks due to external environments such as wind, sand, snow, and salt spray. Bearings are prone to wear and breakage due to poor lubrication, overload, and harsh environments. Gears in gearboxes will have defects such as broken teeth, pitting, bonding, and cracking under long-term operation. These defects can be reflected through vibration during the operation of wind turbines. Therefore, accurately processing and analyzing the vibration signals of wind turbines is of great significance for realizing intelligent monitoring and fault warning of wind turbines.

[0004] The processing of wind turbine vibration signals is based on a variety of principles and technologies, including signal preprocessing (such as filtering to reduce noise and downsampling to reduce data volume), time domain analysis (directly evaluating the periodicity, amplitude and other characteristics of the signal), frequency domain analysis (using Fourier transform and power spectrum analysis to reveal the signal spectrum characteristics) and time-frequency analysis (using short-time Fourier transform or wavelet transform and other technologies to integrate time domain and frequency domain information to obtain more comprehensive signal characteristics). The application of these principles and methods realizes the analysis of wind turbine vibration signals and provides an important basis for subsequent fault diagnosis.

[0005] However, traditional vibration signal processing methods, such as time domain analysis and frequency domain analysis, have limitations when processing nonlinear and multimodal vibration signals and cannot fully mine the effective information under superposition signals. Summary of the invention

[0006] In order to solve the technical problems existing in the above-mentioned prior art, the present invention proposes a wind turbine vibration signal processing method based on adaptive variational mode decomposition and Fourier transform, which can effectively separate signal components at different center frequencies to improve the accuracy and effectiveness of vibration signal processing.

[0007] According to one aspect of the present invention, a method for processing vibration signals of a wind turbine is provided, which comprises: acquiring a vibration signal of a wind turbine; performing sample expansion on the vibration signal of the wind turbine to obtain multiple vibration signals; performing adaptive variational modal decomposition on each vibration signal to obtain an intrinsic modal component signal set of each vibration signal; performing fast Fourier transform on each intrinsic modal component signal in the intrinsic modal component signal set to obtain a frequency spectrum of each intrinsic modal component signal, thereby obtaining a frequency spectrum set of the local oscillator modal component signal of each vibration signal.

[0008] In an example of a method for processing a vibration signal of a wind turbine provided in the above aspect, the method for sample expansion of a vibration signal of a wind turbine includes: collecting the vibration signal of the wind turbine at a set sampling frequency under a set wind turbine speed; intercepting the collected vibration signal according to a set overlap rate to obtain a plurality of vibration signals; wherein each intercepted vibration signal includes a vibration signal of at least one rotation period, and the rotation period is calculated based on the set sampling frequency and the wind turbine speed.

[0009] In an example of the method for processing vibration signals of a wind turbine provided in the above aspect, the overlap rate is 50%.

[0010] In an example of the wind turbine vibration signal processing method provided in the above aspect, the adaptive variational modal decomposition of each vibration signal is specifically to determine the optimal decomposition number k according to the correlation between the decomposed modes and the correlation between the modes and the original signal when solving the constrained variation problem composed of the following equations 1 and 2:

[0011]

[0012] Where min is the minimum value; ∑ is the sum function; i is one of the decomposition numbers, i = 1, 2, 3, ..., k; x sui is the ith intrinsic mode component signal; ω i is the center frequency of the i-th eigenmode component signal; ||·|| 2 is the Euclidean norm; is the partial differential; * represents convolution; δ is the Dirac function; j is the imaginary number symbol; π is the circumference of a circle; x s is the input vibration signal.

[0013] In an example of the wind turbine vibration signal processing method provided in the above aspect, the constrained variation problem composed of equations 1 and 2 is converted into an unconstrained variation problem by adding a quadratic penalty term and a Lagrange multiplier, which is expressed as the following equation 3,

[0014]

[0015] Among them, α is the quadratic penalty term and β is the Lagrange multiplier.

[0016] In an example of the wind turbine vibration signal processing method provided in the above aspect, a method for determining the optimal decomposition number k according to the correlation between the decomposed modes and the correlation between the modes and the original signal includes: step 1: starting from i=2; step 2: calculating x sui With x s Correlation coefficient and x sui-1 With x s Correlation coefficient And determine whether i is less than 3, where, when i < 3, proceed to step 3, and when i ≥ 3, proceed to step 4; Step 3: If Then the step ends and the current k=i-1 is determined, otherwise i=i+1, and returns to step 2; Step 4: Calculate x sui With x sui-1 Correlation coefficient and x sui-1 With x sui-2 Correlation coefficient like and The step ends, the current k=i-1 is determined, otherwise i=i+1, and returns to step 2.

[0017] In an example of the wind turbine vibration signal processing method provided in the above aspect, before performing fast Fourier transform, the wind turbine vibration signal processing method also includes: padding the intrinsic mode component signal set with zeros so that the length of each intrinsic mode component signal satisfies a power of 2.

[0018] In an example of the wind turbine vibration signal processing method provided in the above aspect, each intrinsic mode component signal in the intrinsic mode component signal set is subjected to a fast Fourier transform using the following equation 4:

[0019]

[0020] Where xf(m) is the data after fast Fourier transformation, m = 0, 1, ..., 2h-1, x suk (l) is the signal to be transformed, P is the total number of frequencies in the frequency sequence, l is the frequency value in the frequency sequence, and l=0 represents the DC component of the signal.

[0021] According to another aspect of the present invention, a computer device is provided, which includes a processor and a memory connected to the processor, wherein the memory stores program instructions; the processor is used to execute the program instructions stored in the memory to implement the wind turbine vibration signal processing method as described above.

[0022] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the storage medium stores program instructions, and when the program instructions are executed, the above-mentioned method for processing vibration signals of a wind turbine is implemented.

[0023] Beneficial effects: The present invention implements adaptive variational mode decomposition (VMD) and fast Fourier transform (FFT) technology for the vibration signal of the wind turbine, aiming to deeply mine the characteristic information hidden in the signal. Through this process, the spectrum of the intrinsic mode component signal can be extracted from the vibration signal, thereby forming a spectrum data set. Relying on the spectrum data set, feature extraction and feature selection operations can be further performed to extract and screen out key eigenvalues ​​as inputs for unsupervised classification algorithms and supervised classification algorithms. Combined with fault label information, a wind turbine fault diagnosis model based on vibration signals can be constructed. The present invention provides a more advanced and effective solution for the vibration signal processing required for wind turbine fault diagnosis, and solves the technical problems of single features and insufficient information extraction in existing wind turbine vibration signal processing methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0025] Figure 1 is a flow chart of a method for processing a vibration signal of a wind turbine according to an embodiment of the present invention;

[0026] Figure 2 is a time domain diagram of each intrinsic mode component signal obtained by adaptive variational mode decomposition of the vibration signal when k=4 according to an embodiment of the present invention;

[0027] Figure 3 is a time domain diagram of each intrinsic mode component signal obtained by adaptive variational mode decomposition of the vibration signal when k=5 according to an embodiment of the present invention;

[0028] Figure 4is a time domain diagram of each intrinsic mode component signal obtained by adaptive variational mode decomposition of the vibration signal when k=6 according to an embodiment of the present invention;

[0029] Figure 5 is a frequency spectrum diagram after fast Fourier transformation of each eigenmode component according to an embodiment of the present invention;

[0030] Figure 6 is a framework diagram of a computer device according to an embodiment of the present invention;

[0031] Figure 7 is a schematic diagram of a computer storage medium according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0034] Figure 1 is a flow chart of a method for processing a vibration signal of a wind turbine according to an embodiment of the present invention.

[0035] Reference Figure 1 In step S110, a vibration signal of the wind turbine is obtained.

[0036] Here, some suitable sensors (such as vibration sensors) may be respectively arranged on the blades, main bearings and gearbox of the wind turbine to obtain the vibration signal of the wind turbine.

[0037] In step S120, sample expansion is performed on the vibration signal of the wind turbine generator to obtain multiple vibration signals.

[0038] Specifically, first, the vibration signal X of the wind turbine is collected at a sampling frequency f when the wind turbine rotates at a speed n. The number of collected signal data points is L.

[0039] Secondly, according to the set overlap rate C o The collected vibration signal X is intercepted to obtain multiple vibration signals.

[0040] Here, each segment of the vibration signal x is intercepted s A vibration signal including at least one rotation period. The rotation period is obtained according to the wind turbine generator speed n. The rotation period T = 60 / n. The time unit of the collected wind turbine generator speed n is minutes, while the time unit of the rotation period T is seconds.

[0041] Furthermore, the number of signal data points of the vibration signal in each rotation axis T is in, The round-up symbol.

[0042] Since each segment of the intercepted vibration signal x s The vibration signal includes at least one complete rotation cycle. Therefore, each vibration signal x s The number of data points of the vibration signal included is 2P, that is, [120f / n]. In this way, while achieving sample expansion, it is ensured that each vibration signal x s With complete information.

[0043] In addition, when expanding the sample, if the number of signal data points contained in the last vibration signal (i.e., the last sample) is less than 2P, the last vibration signal is discarded. In this case, the number of samples after expansion (i.e., the number of vibration signals)

[0044] In a specific embodiment, the vibration signal X of the high-speed shaft of the wind turbine gearbox of a wind farm is obtained. At the rated speed of 1800 r / min, the meshing frequency between the planetary gear and the sun gear is 3.87 times the rotation frequency. The wind turbine speed n = 1687.5 r / min, the number of signal data points L = 8192, and the sampling frequency f = 12800 Hz. The number of signal data points per rotation cycle P = [60 × 12800 / 1687.5] = 456. Overlap rate C o = 50%, the vibration signal X is intercepted with a length of 2P for sample expansion, and the sample after expansion is x s The number is 17, that is, 17 vibration signals. Among them, the data of the vibration signal X before and after expansion are shown in Table 1. Table 1 shows the wind turbine vibration signal X to be processed and the expanded sample x s1 to x s17 .

[0045] Table 1

[0046]

[0047] In step S130, adaptive variational modal decomposition is performed on each vibration signal to obtain an intrinsic modal component signal set of each vibration signal.

[0048] Adaptive variational mode decomposition (VMD) is a method to calculate the optimal solution, adaptively calculate each mode, and achieve adaptive decomposition. Through the VMD method, complex vibration signals can be decomposed into several single modal components, which is convenient for subsequent processing and analysis.

[0049] Furthermore, for each vibration signal x s Perform adaptive variational mode decomposition to obtain each vibration signal x s The intrinsic mode component signal set X su =[x su1 , x su2 , x su3 , …, x suk ], X su ∈R k×2P , k is the optimal number of decompositions selected by adaptive variational mode decomposition.

[0050] In this embodiment, x in Table 1 s1 Perform adaptive variational mode decomposition and get x s1 The intrinsic mode component signal set X s1u =[x s1u1 , x s1u2 , x s1u3 , x s1u4 , x s1u5 ], X s1u ∈R 5×912 , at this time, the optimal number of decompositions selected by adaptive variational mode decomposition is k = 5. When k = 4, k = 5, k = 6, the optimal number of decompositions for x s1 and X s1u The time and amplitude are plotted, with the horizontal axis being time and the vertical axis being amplitude, and the vibration signal x is obtained. s1 And the intrinsic mode component signal set X s1u The time domain diagram of Figures 2 to 4 shown.

[0051] When k = 5, x s1 The data after adaptive variational mode decomposition is shown in Table 2 below.

[0052] Table 2

[0053]

[0054] Furthermore, the adaptive variational modal decomposition of each vibration signal is specifically to determine the optimal decomposition number k according to the correlation between the decomposed modes and the correlation between the mode and the original signal when solving the constrained variational problem composed of equations (1) and (2).

[0055]

[0056] Where min is the minimum value; ∑ is the sum function; i is one of the decomposition numbers, i = 1, 2, 3, ..., k; u i is the ith eigenmode component signal; ω i is the center frequency of the i-th eigenmode component signal; ||·|| 2 is the Euclidean norm; is the partial differential; * represents convolution; δ is the Dirac function; j is the imaginary number symbol; π is the circumference of a circle; x s is the input vibration signal.

[0057] Furthermore, the constrained variational problem composed of equations (1) and (2) is transformed into an unconstrained problem by adding quadratic penalty terms and Lagrange multipliers, and the alternating direction multiplier method is used to find the optimal solution.

[0058] The augmented Lagrangian equation with the addition of quadratic penalty terms and Lagrangian multipliers is shown in equation (3):

[0059]

[0060] Among them, α is the quadratic penalty term; β is the Lagrange multiplier.

[0061] Furthermore, the steps of determining the optimal number of decompositions k according to the correlation between the decomposed modes and the correlation between the modes and the original signal when solving the constrained variational problem are as follows:

[0062] Step 1) Start from i=2;

[0063] Step 2) Calculate x sui With x s Correlation coefficient and x sui-1 With x s Correlation coefficient And determine whether i is greater than 3; when i < 3, proceed to step 3), and when i ≥ 3, proceed to step 4);

[0064] Step 3) If Then the step ends and the current k=i-1 is determined, otherwise i=i+1, and returns to step 2);

[0065] Step 4) Calculate x sui With xsui-1 Correlation coefficient and x sui-1 With x sui-2 Correlation coefficient like and Then the step ends, determine the current k=i-1, otherwise i=i+1, and return to step 2).

[0066] Furthermore, the formula of the correlation coefficient is the following formula (4):

[0067]

[0068] Among them, Cov(A, B) is the covariance of variables A and B, Var(A) represents the variance of A, and Var(B) represents the variance of B. It should be understood that when calculating the correlation coefficient of specific variables, A and B can be replaced by the corresponding specific variables respectively.

[0069] In this embodiment, x in Table 1 s1 Adaptive variational mode decomposition is performed, and the optimal number of decompositions k is determined based on the correlation between the decomposed modes and the correlation between the modes and the original signal. When k is 1 to 7, x s1 and the eigenmode component signal set X s1u =[x s1u1 , x s1u2 , x s1u3 , …, x s1uk The correlation calculation results of each eigenmode component signal in ] are shown in Table 3 (the results are retained to 4 decimal places). When k is 1 to 7, the eigenmode component signal set X s1u =[x s1u1 , x s1u2 , x slu3 , …, x sluk ] s1uk With x s1uk-1 The correlation between x s1uk-1 With x s1uk-2 The correlation between The calculation results are shown in Table 4 (the results are rounded to 4 decimal places). According to the contents of Table 3 and Table 4, the optimal number of decompositions k=5 is selected by adaptive variational mode decomposition.

[0070] Table 3

[0071]

[0072] Table 4

[0073]

[0074] In step S140, a fast Fourier transform is performed on each eigenmode component signal in the eigenmode component signal set to obtain a frequency spectrum of each eigenmode component signal, thereby obtaining a frequency spectrum set of the local vibration mode component signals of each vibration signal segment.

[0075] Specifically, for the eigenmode component signal set X su Each eigenmode component signal x in suk Before performing fast Fourier transform, in order to increase the frequency resolution or avoid spectrum aliasing, the intrinsic mode component signal set X su Zero padding is performed so that the signal length of each eigenmode component satisfies the power of 2. When padding zero, the zero padding matrix O is designed according to the power of 2 closest to the signal length. k×D , so that X su ∈R k×(2P+D) , (2P+D)∈2 N After fast Fourier transform, each eigenmode component signal x is obtained suk The spectrum of xf suk , and then transform them in sequence to obtain the intrinsic modal component signal spectrum data set X of each vibration signal that can be used for further feature extraction f =[xf su1 , xf su2 , xf su3 ,…,xf suk ],

[0076] In this embodiment, the eigenmode component signal set X is firstly s1u =[x s1u1 , x s1u2 , x s1u3 , x s1u4 , x s1u5 ], X s1u ∈R 5×912 Fill with zeros, the zero-filling matrix is ​​O k×(1024-912) , and obtain the new intrinsic mode component signal set X′ s1u =[x′ s1u1 , x′ s1u2 , x′ s1u3 , x′ s1u4 , x′ s1u5 ],X′ s1u ∈R 5×1024 Then, for the eigenmode component signal set X′ s1u =[x′ s1u1 , x′ s1u2 , x′ s1u3 , x′ s1u4 , x′ s1u5 ],X′ s1u ∈R 5×1024Each eigenmode component in is transformed by fast Fourier transform to form the eigenmode component signal spectrum data set X f =[xf s1u1 , xf s1u2 , xf s1u3 , xf s1u4 , xf s1u5 ], X f ∈R 5×512 .

[0077] Among them, the zero-filling matrix O k×(1024-912) Specifically:

[0078]

[0079] According to the frequency and amplitude of each point of each intrinsic mode component, the frequency and amplitude are plotted, with the horizontal axis being the frequency and the vertical axis being the amplitude, to obtain the spectrum diagram of each intrinsic mode component after fast Fourier transformation, such as Figure 5 As shown in the figure. Through adaptive variational mode decomposition and fast Fourier transform, the implicit characteristic information in the vibration signal is deeply excavated, such as xf s1u1 The peak values ​​with amplitudes greater than 0.02g RMS are included at 100±12.5hz, 187.5±12.5hz, and 300±12.5hz. The first high-amplitude frequency is about 3.87 times the rotation frequency of 25.83hz. It can be judged that the signal has a fault of poor meshing between the planetary gear and the sun gear. Combined with the intrinsic mode component signal spectrum data set X f and the eigenmode component signal set X′ s1u Further performing feature extraction and feature selection operations can solve the technical problems of single features and insufficient information extraction in existing wind turbine vibration signal processing methods.

[0080] Among them, for the intrinsic mode component signal set X′ s1u The data after fast Fourier transform of the eigenmode components are shown in Table 5 below.

[0081] Table 5

[0082]

[0083] To implement the wind turbine vibration signal processing method in the above embodiment, the present application also provides a computer device 300, see Figure 6 The computer device 300 of the embodiment of the present application includes a processor 31, a memory 32, an input and output device 33 and a bus 34.

[0084] The processor 31 , the memory 32 , and the input / output device 33 are respectively connected to the bus 34 . The memory 32 stores program data. The processor 31 is used to execute the program data to implement the wind turbine vibration signal processing method described in the above embodiment.

[0085] In the embodiment of the present application, the processor 31 may also be referred to as a CPU (Central Processing Unit). The processor 31 may be an integrated voltage control system chip with signal processing capabilities. The processor 31 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated voltage control system (ASIC), a field programmable gate array (FPGA), or other programmable logic devices, discrete gates or transistor logic devices, or discrete hardware components. A general-purpose processor may be a microprocessor or the processor 31 may also be any conventional processor, etc.

[0086] This application also provides a computer storage medium, please continue to refer to Figure 7 , Figure 7 It is a structural diagram of an embodiment of a computer storage medium provided in the present application. The computer storage medium 40 stores program data 41. When the program data 41 is executed by the processor, it is used to implement the wind turbine vibration signal processing method of the above embodiment.

[0087] When the embodiments of the present application are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (processor) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, RandomAccess Memory), a disk or an optical disk.

[0088] The above description is only an implementation method of the present application, and does not limit the patent scope of the present application. Equivalent structures or equivalent process changes made by utilizing the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for processing vibration signals of a wind turbine generator, characterized in that: The wind turbine generator vibration signal processing method comprises: Obtain vibration signals of wind turbines; Perform sample expansion on the vibration signal of the wind turbine to obtain multiple vibration signals; Performing adaptive variational modal decomposition on each vibration signal to obtain an intrinsic modal component signal set of each vibration signal; Performing a fast Fourier transform on each eigenmode component signal in the eigenmode component signal set to obtain the frequency spectrum of each eigenmode component signal, thereby obtaining the frequency spectrum set of the local oscillation mode component signal of each vibration signal.

2. The method for processing wind turbine vibration signals according to claim 1, characterized in that: The method for performing sample expansion on the vibration signal of the wind turbine generator comprises: Collecting the vibration signal of the wind turbine generator at a set sampling frequency at a set wind turbine generator speed; The collected vibration signal is intercepted according to the set overlap rate to obtain multiple vibration signals; Each segment of the intercepted vibration signal includes a vibration signal of at least one rotation period, and the rotation period is calculated according to a set sampling frequency and a rotation speed of the wind turbine generator.

3. The method for processing wind turbine vibration signals according to claim 2, characterized in that: The overlap ratio is 50%.

4. The method for processing wind turbine vibration signals according to claim 1, characterized in that: The adaptive variational modal decomposition of each vibration signal is specifically to determine the optimal decomposition number k according to the correlation between the decomposed modes and the correlation between the modes and the original signal when solving the constrained variation problem composed of the following equations 1 and 2. Where min is the minimum value; ∑ is the sum function; i is one of the decomposition numbers, i = 1, 2, 3, ..., k; x sui is the ith eigenmode component signal; ω i is the center frequency of the i-th eigenmode component signal; ||·||2 is the Euclidean norm; is the partial differential; * represents convolution; δ is the Dirac function; j is the imaginary number symbol; π is the circumference of a circle; x s is the input vibration signal.

5. The method for processing wind turbine vibration signals according to claim 4, characterized in that: By adding the quadratic penalty term and the Lagrange multiplier, the constrained variational problem composed of equations 1 and 2 is transformed into an unconstrained variational problem, which is expressed as the following equation 3, Among them, α is the quadratic penalty term and β is the Lagrange multiplier.

6. The method for processing wind turbine vibration signals according to claim 5, characterized in that: Methods for determining the optimal decomposition number k based on the correlation between the decomposed modes and the correlation between the modes and the original signal include: Step 1: Start from i=2; Step 2: Calculate x sui With x s Correlation coefficient and x sui-1 With x s Correlation coefficient And determine whether i is less than 3, where when i<3, proceed to step 3, and when i≥3, proceed to step 4; Step 3: If Then the step ends and the current k=i-1 is determined, otherwise i=i+1, and returns to step 2; Step 4: Calculate x sui With x sui-1 Correlation coefficient and x sui-1 With x sui-2 Correlation coefficient like and The step ends, the current k=i-1 is determined, otherwise i=i+1, and returns to step 2.

7. The method for processing wind turbine vibration signals according to claim 1, characterized in that: Before performing the fast Fourier transform, the wind turbine vibration signal processing method further includes: padding the intrinsic mode component signal set with zeros so that the length of each intrinsic mode component signal satisfies a power of 2.

8. The method for processing wind turbine vibration signals according to claim 1, characterized in that: The following formula 4 is used to perform a fast Fourier transform on each eigenmode component signal in the eigenmode component signal set: Where xf(m) is the data after fast Fourier transformation, m = 0, 1, ..., 2h-1, x suk (l) is the signal to be transformed, h is half of the total number of frequencies in the frequency sequence, l is the frequency value in the frequency sequence, and l=0 represents the DC component of the signal.

9. A computer device, characterized in that: The computer device comprises a processor and a memory connected to the processor, wherein: The memory stores program instructions; The processor is used to execute the program instructions stored in the memory to implement the wind turbine vibration signal processing method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The storage medium stores program instructions, and when the program instructions are executed, the method for processing vibration signals of a wind turbine generator according to any one of claims 1 to 8 is implemented.

Images