Method and device for monitoring wind power main shaft bearing based on acoustic emission technology

Through the online fault monitoring method of acoustic emission technology, the problem of difficult detection of early faults of low-speed heavy-load bearings of wind power spindle bearings is solved, and early detection and accurate diagnosis of faults is achieved, and the operating efficiency of wind turbines is improved.

CN119064009APending Publication Date: 2024-12-03BEIJING TONGTAI HENGSHENG TECH CO LTD +1
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410095087.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-01-23
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

When monitoring wind power spindle bearing failures, the early failures of low-speed heavy-load bearings are difficult to detect, and the fault diagnosis is inaccurate, resulting in increased maintenance costs.

Method used

The online fault status monitoring method based on acoustic emission technology is adopted to obtain the acoustic emission signals of wind power spindle bearings, and frequency domain feature extraction, Bartlett inspection, factor analysis and entropy value analysis are carried out to achieve early detection and accurate diagnosis of faults.

Benefits of technology

It realizes early detection and effective early warning of wind power spindle bearing failures, improves the operating efficiency of wind turbine units, and reduces maintenance costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119064009A_ABST
    Figure CN119064009A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a method, device and equipment for monitoring a wind power main shaft bearing based on an acoustic emission technology and a computer readable storage medium. The method comprises the following steps: acquiring an acoustic emission signal of a wind power main shaft bearing; frequency domain feature extraction is carried out on the acoustic emission signals, and a first fault diagnosis feature set is constructed; carrying out Bartlett test on the first fault diagnosis feature set, and if a test result meets a preset condition, carrying out factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set; and entropy analysis is carried out on the second fault diagnosis feature set, whether the current wind power main shaft bearing has a fault is determined, and monitoring of the current electric main shaft bearing is completed. In this way, online fault state monitoring and analysis are conducted on the main shaft bearing of the wind generating set through the acoustic emission technology, early discovery and effective early warning of faults of the main shaft bearing are achieved, and the overall benefits of operation of the wind generating set are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present application relate to the field of wind turbine operation and maintenance management, and in particular to methods, devices, equipment and computer-readable storage media for monitoring wind turbine main shaft bearings based on acoustic emission technology. Background Art

[0002] In recent years, with the adjustment of economic structure, wind energy as a green energy has received more and more attention, which has promoted the rapid development of wind power equipment. However, statistics on the small-power wind turbines that have been in operation show that due to the harsh living environment of wind turbine main shaft bearings, failures and shutdown accidents often occur. Therefore, research on fault diagnosis of wind turbine main shaft bearings is of great significance to reduce the risks that affect the safe and stable operation of wind turbines.

[0003] The inspection process of wind turbine main shaft bearings is cumbersome and complicated. It requires manual spectrum recognition, which consumes a lot of labor costs and expert experience. It is often difficult for ordinary quality inspectors to fully extract fault information from the spectrum image. At present, the common method for online fault status monitoring and analysis of main shaft bearings is the vibration method. The vibration method is to collect the vibration signal generated by the equipment during operation through the vibration sensor installed on the equipment under test, and then perform spectrum analysis on the vibration signal to extract the fault characteristic frequency and determine whether the equipment has a fault and the type of fault. However, the vibration method has the following shortcomings for fault status monitoring of low-speed and heavy-loaded bearings:

[0004] (1) The impact of early defects in low-speed heavy-load bearings at low speeds is not enough to cause the equipment under test to generate sufficient vibration signals for detection, making it difficult to detect early faults;

[0005] (2) Low-speed heavy-load bearings have low speeds, their fault characteristic frequencies are extremely close, and the vibration analysis algorithm is not easy to accurately distinguish the fault characteristics, resulting in inaccurate fault diagnosis;

[0006] (3) When detecting and analyzing spindle bearing failures, the extent of bearing damage may have exceeded the planned time required for preventive maintenance, resulting in increased maintenance costs. Summary of the invention

[0007] According to an embodiment of the present application, a solution for monitoring wind turbine main shaft bearings based on acoustic emission technology is provided, which utilizes acoustic emission technology to perform online fault status monitoring and analysis on the main shaft bearings of wind turbine generator sets, thereby achieving early detection and effective warning of main shaft bearing faults and improving the overall efficiency of wind turbine generator set operation.

[0008] In a first aspect of the present application, a method for monitoring a wind turbine main shaft bearing based on acoustic emission technology is provided. The method comprises:

[0009] Obtain the acoustic emission signal of the wind turbine main shaft bearing;

[0010] Performing frequency domain feature extraction on the acoustic emission signal to construct a first fault diagnosis feature set;

[0011] Performing a Bartlett test on the fault diagnosis feature set, and if the test result meets a preset condition, performing a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set;

[0012] An entropy analysis is performed on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing, thereby completing the monitoring of the current wind turbine main shaft bearing.

[0013] Furthermore, it also includes:

[0014] The bearing acoustic emission signal is denoised to obtain a clean signal.

[0015] Furthermore, the denoising process is performed on the acoustic emission signal to obtain a clean signal, including:

[0016] The optimal wavelet coefficients are determined by the following formula:

[0017] x(t)=∑ i ciψ i (t)

[0018]

[0019] Among them, ci is the wavelet coefficient;

[0020] ψ i (t) is the wavelet basis function;

[0021] f1, f2, f3, f4, and f5 are the resonant frequencies of the three components;

[0022] t1, t2, t3, t4, and t5 are the time points at which the three component signals occur respectively;

[0023] σ1, σ2, σ3, σ4, and σ5 are the standard deviations of the three components;

[0024] According to the optimal wavelet coefficients, the acoustic emission signal is denoised to obtain a clean signal.

[0025] Furthermore, the denoising process of the acoustic emission signal to obtain a clean signal also includes:

[0026] Decomposing the acoustic emission information based on a preset number of decomposition modes;

[0027] The components with kurtosis values ​​greater than the preset threshold in the decomposition results are retained to suppress the noise of the acoustic emission signal and obtain a clean signal.

[0028] Furthermore, the kurtosis value is determined by:

[0029]

[0030] Among them, KU is the kurtosis value;

[0031] σ is the standard deviation.

[0032] Furthermore, the extracting frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set includes:

[0033] Based on the clean signal, construct a high-dimensional matrix;

[0034] Decomposing the high-dimensional matrix to obtain multiple singular values;

[0035] Based on the multiple singular values, a first fault diagnosis feature set is constructed.

[0036] Furthermore, decomposing the high-dimensional matrix to obtain a plurality of singular values ​​includes:

[0037]

[0038] Among them, Σ is a non-negative diagonal matrix;

[0039] Matrices U and V are the left and right singular matrices of matrix A, respectively.

[0040] In a second aspect of the present application, a device for monitoring a wind turbine main shaft bearing based on acoustic emission technology is provided. The device comprises:

[0041] An acquisition module, used for acquiring acoustic emission signals of wind turbine main shaft bearings;

[0042] An extraction module, used for extracting frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set;

[0043] A first analysis module is used to perform a Bartlett test on the fault diagnosis feature set, and if the test result meets a preset condition, perform a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set;

[0044] The second analysis module is used to perform entropy analysis on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing and complete the monitoring of the current electric main shaft bearing.

[0045] In a third aspect of the present application, an electronic device is provided, which includes a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the program, the method described above is implemented.

[0046] In a fourth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the method according to the first aspect of the present application is implemented.

[0047] The method for monitoring wind turbine main shaft bearings based on acoustic emission technology provided in the embodiment of the present application obtains the acoustic emission signal of the wind turbine main shaft bearing; extracts the frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set; performs a Bartlett test on the fault diagnosis feature set, and if the test result meets the preset conditions, performs factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set; performs entropy analysis on the second fault diagnosis feature set to determine the current wind turbine main shaft bearing fault type, completes the monitoring of the current electric main shaft bearing, realizes the early detection, accurate diagnosis and effective warning of wind turbine main shaft bearing faults, and improves the overall efficiency of wind turbine generator set operation.

[0048] It should be understood that the contents described in the Summary of the Invention are not intended to limit the key or important features of the embodiments of the present application, nor are they intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, wherein:

[0050] Figure 1 A flowchart of a method for monitoring a wind turbine main shaft bearing based on acoustic emission technology according to an embodiment of the present application;

[0051] Figure 2 A block diagram of a device for monitoring a wind turbine main shaft bearing based on acoustic emission technology according to an embodiment of the present application;

[0052] Figure 3 It is a schematic diagram of the structure of a terminal device or a server suitable for implementing the embodiments of the present application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0054] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0055] Noun analysis:

[0056] Acoustic emission technology is a non-destructive testing technology that detects and analyzes elastic wave signals generated by microscopic deformation or damage inside the material.

[0057] Acoustic emission technology has the following advantages:

[0058] (1) The acoustic emission signal is directly generated by the internal defects of the material and has nothing to do with external interference. Therefore, it can more sensitively reflect the internal state of the material;

[0059] (2) The acoustic emission signal has a higher frequency range, so it can more clearly distinguish different types and degrees of defects;

[0060] (3) Acoustic emission signals have strong positioning capabilities, so they can determine the location and scope of defects more accurately.

[0061] Acoustic emission technology has obvious advantages in detecting and discovering early faults of low-speed heavy-load bearings. Therefore, modern signal processing methods are used to extract the characteristics of acoustic emission signals, and then the acoustic emission monitoring technology is combined with nonlinear signal and information processing technologies such as neural network expert systems to establish a wind turbine main shaft bearing fault feature information database with a large amount of knowledge, realize intelligent classification and diagnosis of bearing working conditions, and achieve the purpose of real-time intelligent monitoring.

[0062] Figure 1 A flow chart of a method for monitoring a wind turbine main shaft bearing based on acoustic emission technology according to an embodiment of the present disclosure is shown. The method comprises:

[0063] S110, acquiring an acoustic emission signal of a wind turbine main shaft bearing.

[0064] In some embodiments, the acoustic emission signal may be acquired by an acoustic emission sensor pre-installed on the wind turbine main shaft bearing.

[0065] S120, extracting frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set.

[0066] During the operation of the wind turbine main shaft bearing, various factors may cause damage or failure, leading to the occurrence of faults. Once a fault occurs, it will cause serious safety hazards. In the early stage of the wind turbine main shaft bearing failure, there is a lot of noise interference in the signal collection, which makes it difficult to extract fault features and cannot detect the fault as early as possible. Therefore, the noise reduction problem of the early fault signal of the rolling bearing is the key to fault diagnosis. Therefore, in the present disclosure, it is necessary to give priority to the noise reduction of the acoustic emission signal obtained in step S110.

[0067] In some embodiments, the acoustic emission signal can be de-noised by:

[0068] On the basis of the traditional denoising by wavelet threshold, a new method for selecting an improved wavelet threshold function is proposed to make up for the shortcomings of the traditional wavelet threshold. The wavelet threshold disclosed in the present invention can adapt to various environments by selecting different values ​​of the adjustment parameter a. The value range of the adjustment parameter a is usually between 0-1. The value of a can be adjusted according to the needs of different vibration signals, and four threshold functions of hard threshold, a=0.3, a=0.5, and a=0.7 are used for noise reduction processing. The value of a can be adjusted according to the needs of different vibration signals to achieve mutual conversion between the new threshold function and the traditional hard threshold function. Threshold processing is to operate on the wavelet coefficients, and process them according to different threshold functions.

[0069] Specifically, for each wavelet coefficient ci, the processing rules are as follows:

[0070] (1) If the absolute value of |ci| is less than the set threshold, ci is set to 0.

[0071] (2) If the absolute value of |ci| is greater than or equal to the set threshold, the original value of ci is retained.

[0072] The specific value of the threshold depends on the selected threshold function, which may be a hard threshold or a threshold adjusted by setting parameter a to 0.3, 0.5, or 0.7.

[0073] Furthermore, the following multiple exponentially decaying signal mixing method is used to simulate the acoustic emission signal, and the noise reduction effects of the above thresholds are analyzed respectively. The acoustic emission signal simulation model is:

[0074] x(t)=∑ i ciψ i (t)

[0075]

[0076] Among them, ci is the wavelet coefficient;

[0077] ψ i (t) is the wavelet basis function;

[0078] f1, f2, f3, f4, and f5 are the resonant frequencies (kHz) of the three components respectively; the acoustic emission signal usually includes three orthogonal directional components, each of which can be represented by one or more exponentially decaying sinusoidal signals. In the model, the x-direction component is represented by an exponentially decaying sinusoidal signal, the y-direction component is represented by two exponentially decaying sinusoidal signals, and the z-direction component is represented by two exponentially decaying sinusoidal signals, so there are five parameters in total;

[0079] t1, t2, t3, t4, and t5 are the time points (ms) at which the three component signals occur;

[0080] σ1, σ2, σ3, σ4, and σ5 are the standard deviations of the three components;

[0081] The three components in the above formula refer to the three oscillation components of the signal x(t) with different frequencies, which can be represented by three different wavelet basis functions. Each wavelet basis function has a resonant frequency and an occurrence time point.

[0082] Specifically, the first component is an oscillating component of frequency, which occurs at a point in time and can be represented by a wavelet basis function;

[0083] The second component is an oscillating component of frequency, which occurs at a point in time and can be represented by a wavelet basis function;

[0084] The third component is an oscillating component of frequency, which occurs at a point in time and can be represented by a wavelet basis function.

[0085] The optimal wavelet coefficients are determined through the above model, and the acoustic emission signal is denoised according to the optimal wavelet coefficients to obtain a clean signal.

[0086] In some embodiments, depending on the application scenario, after the above-mentioned denoising process, the following method can be further used for secondary denoising (secondary denoising of the clean signal), or the following method (denoising based on VMD waveform reconstruction with kurtosis factor correction) can be used alone for separate denoising.

[0087] Specifically, the VMD algorithm is an adaptive, non-recursive modal decomposition method suitable for nonlinear and non-stationary signal processing. The algorithm decomposes the acoustic emission signal into K intrinsic mode components (IMF) from low frequency to high frequency, which can satisfy the sum of the modal components equal to the original input signal and the sum of the estimated signal bandwidth of each mode is minimized. When using VMD to decompose the signal, it is necessary to preset the number of decomposition modes K. If the K value is too large, useless components will be decomposed and modal aliasing will occur. If the K value is too small, the decomposition will be insufficient and some useful information of the signal will be lost. Therefore, a reasonable value of K can improve the decomposition effect of VMD.

[0088] The kurtosis factor is usually used to indicate the flatness of a waveform and to describe the distribution of a variable. The kurtosis of a normal distribution is equal to the kurtosis threshold. When the kurtosis is less than the kurtosis threshold, the distribution curve will be "flatter", and when the kurtosis is greater than the kurtosis threshold, the distribution curve will be "steeper". The kurtosis threshold can be determined according to the actual application scenario, preferably 3.

[0089] The kurtosis coefficient is sensitive to early bearing failures. Once a bearing fails, the K value increases. That is, when the rolling surface of the bearing is gradually damaged, periodic large-amplitude pulses will be generated in the vibration signal. Since the numerator in the kurtosis coefficient expression is the fourth power of x, and the denominator is the square of the mean square value, that is, the denominator is an average, which will cause the numerator to increase faster than the denominator, and K will rise quickly, reflecting that a fault has occurred. The calculation of kurtosis can be expressed as:

[0090]

[0091] Among them, KU is the kurtosis value;

[0092] σ is the standard deviation;

[0093] In the present disclosure, the K value is first initialized, and the K value is gradually increased in a step size so that the center frequency can maintain a certain distance. After multiple experiments, the K value is preferably 8 (4 is used as the initial K value). After VMD decomposition, the components whose kurtosis value is greater than the kurtosis threshold are retained for reconstruction to complete noise suppression.

[0094] In some embodiments, after VMD denoising, that is, after the signal is decomposed by the VMD algorithm, a high-dimensional intrinsic mode function IMFs feature matrix can be obtained. The feature matrix has a large amount of data. If it is directly used as a fault feature vector for fault diagnosis, it will lead to problems such as low fault diagnosis accuracy and poor robustness. Therefore, after completing the decomposition of the acoustic emission signal, it is necessary to perform dimensionality reduction processing on the extracted high-dimensional feature matrix to improve the fault diagnosis effect.

[0095] Specifically, in the present disclosure, a singular value dimensionality reduction method can be used to process the acoustic emission signal after VMD denoising to obtain singular value features and centroid frequency features obtained by Hilbert transform and centroid frequency calculation to construct a first fault diagnosis feature set; the first fault diagnosis feature set includes the singular value features and centroid frequency features.

[0096] Singular value decomposition is a commonly used high-dimensional data dimensionality reduction method with good decorrelation characteristics. Singular values ​​are inherent characteristics of a matrix and can measure the stability of a matrix. If the elements in the matrix change, the singular values ​​will also change accordingly. The IMFs obtained from each VMD decomposition are combined into a high-dimensional matrix. Using the singular value decomposition method, the singular values ​​of the matrix are calculated, and singular values ​​of different sizes will be obtained under different bearing conditions. Therefore, it is feasible to describe the bearing state with singular value features. For any m×n-order matrix A, there are orthogonal matrices U and V:

[0097]

[0098] Where Σ is a non-negative diagonal matrix (the singular value matrix of A), Σ=diag(σ1,σ2,…,σ r );

[0099] Matrices U and V are the left singular matrix and right singular matrix of matrix A respectively;

[0100] The value of r is the smaller value of m and n, and satisfies σ1≥σ2≥…≥σ r ≥0, that is, the singular values ​​are arranged in descending order. r means the rank of matrix A, that is, the number of the largest linearly independent column vectors (or row vectors) of A. The rank of matrix A reflects the dimension of A, that is, the dimension of the linear space that A can represent. The larger the rank, the greater the amount of information in A and the higher the stability of A.

[0101] Hilbert transform can be used to perform envelope demodulation on the denoised acoustic emission signal, extract the low-frequency signal related to the fault from the denoised acoustic emission signal (high-frequency modulation signal), obtain the envelope signal containing the fault information, and perform spectrum analysis on the envelope to obtain the envelope spectrum. That is, the centroid frequency feature is obtained by Hilbert transform and centroid frequency calculation. The principle of envelope demodulation is: use Hilbert transform to generate the analytical signal of the signal, and then calculate the envelope of the signal, that is, the amplitude of the signal.

[0102] Specifically, the steps for obtaining the envelope by Hilbert are as follows:

[0103] Perform Hilbert transform on the acoustic emission signal xj(t) to obtain hj(t):

[0104]

[0105] And construct the parsing signal:

[0106]

[0107] Where xj(t) is the time domain waveform of the jth acoustic emission signal;

[0108] t is the time variable;

[0109] hj(t) is the Hilbert transform of the jth acoustic emission signal, which is a complex analysis signal and can be used to obtain the envelope and phase of the signal;

[0110] aj(t) is the analytical signal of the jth acoustic emission signal, which is a complex signal composed of the original signal and its Hilbert transform, and can be expressed as the product of amplitude and phase;

[0111] ej(t) is the envelope signal amplitude of the jth acoustic emission signal, which is the modulus of the analytical signal and reflects the amplitude change of the signal.

[0112] Take the modulus of the analytical signal and demodulate the envelope signal amplitude:

[0113]

[0114] Simple spectrum analysis cannot derive the fault frequency due to signal modulation. After envelope demodulation, the envelope spectrum of the faulty bearing shows different frequency distributions. At the same time, the frequency domain parameter centroid frequency can characterize the frequency center distribution of the signal. Therefore, the centroid frequency FC of the acoustic emission signal envelope can be used as a feature of fault diagnosis. The centroid frequency calculation is:

[0115]

[0116] in,

[0117]

[0118] FC is the centroid frequency of the acoustic emission signal envelope, which is a frequency domain parameter that characterizes the frequency center distribution of the signal and can be used as one of the features for fault diagnosis.

[0119] N is the number of sampling points of the signal, which determines the length and resolution of the signal.

[0120] f s is the signal sampling frequency (Hz).

[0121] S130, performing a Bartlett test on the fault diagnosis feature set, and if the test result meets a preset condition, performing a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set.

[0122] Among them, factor analysis is a data simplification technique that explores the basic structure of observed data by studying the internal dependencies between many variables and using a few hypothetical variables (factors) to represent the original data. Factors can reflect the main information of many original variables. The main purpose of factor analysis is to describe some more basic latent variables hidden in a set of measured variables that cannot be directly measured.

[0123] The process of factor analysis includes:

[0124] Factor analysis (exploratory factor analysis) is used to explore how many factors (variables) the analysis items (quantitative data) should be divided into. For example, how many aspects should 20 scale items be divided into? Users can set the number of factors by themselves. If not, the system will set the number of factors based on the eigenvalue greater than 1 as the judgment standard. In particular, when factor analysis is performed for factor condensation, it usually goes through multiple repeated cycles, deletes unreasonable items, and repeats multiple cycles to finally obtain reasonable results.

[0125] The specific steps include:

[0126] (1) Determine whether to conduct factor analysis, the judgment standard is that the KMO value is greater than 0.6;

[0127] (2) Determine the correspondence between factors and items.

[0128] Determine the correspondence between factors and items: Assume that there are 3 expected factors (variables) and 10 analysis items; the intersection of factors and items gives a total of 30 numbers, which are called "factor loading coefficients (the factor loading coefficient value indicates the degree of correlation between the analysis item and the factor); for each factor (variable), there are 10 corresponding factor loading coefficients", and for each analysis item, there are 3 "factor loading coefficient values" (such as 0.765, -0.066, 0.093). Select the value (0.765) with an absolute value greater than 0.4 among the 3 numbers. If it corresponds to factor 1, it means that this item should be classified under factor 1.

[0129] Deleting unreasonable items usually includes three situations:

[0130] Category 1:

[0131] If the commonality (common factor variance) value of the analysis item is less than 0.4, the corresponding analysis item should be deleted;

[0132] Category 2:

[0133] If the absolute values ​​of the "factor loading coefficient" corresponding to a certain analysis item are all less than 0.4, then this analysis item needs to be deleted:

[0134] Category 3:

[0135] If there is a serious deviation between the correspondence between an analysis item and a factor (often referred to as "putting the cart before the horse"), the analysis item needs to be deleted.

[0136] In some embodiments, it is first necessary to perform a Bartlett test on the first fault diagnosis feature set, and if the test result meets the preset conditions, factor analysis is performed on the first fault diagnosis feature set. That is, a Bartlett test is performed on the first fault diagnosis feature set, and if there is a certain correlation between the variables in the first fault diagnosis feature set (there is a linear relationship between the variables directly), factor analysis is performed on the fault diagnosis feature set.

[0137] Among them, the Bartlett test is a statistical method used to test whether the covariance matrix between the observed variables is an identity matrix (there is no linear relationship between the variables). The null hypothesis of this test is that the covariance matrix between the variables is an identity matrix, that is, there is no linear relationship between the variables.

[0138] The basic idea of ​​the Bartlett test is to determine whether there is a correlation between variables by the difference between the observed sample covariance matrix and the identity matrix. If the difference is significant, the null hypothesis can be rejected, indicating that there is some form of correlation between the variables.

[0139] In the context of factor analysis, the Bartlett test is often used to confirm whether factor analysis is appropriate. If the result of the Bartlett test is significant, it means that there is a certain correlation between the variables, and it may make sense to conduct factor analysis at this time. If the result of the Bartlett test is not significant, it means that the linear relationship between the variables is weak, and factor analysis may not be applicable. The p-value of the Bartlett test is used to determine the probability of whether the variances of the samples are equal. If the p-value is very small, it means that the difference between the observed data and the null hypothesis (equal variances) is very large, then we have reason to reject the null hypothesis and believe that the variances of the samples are not equal. Conversely, if the p-value is very large, it means that the difference between the observed data and the null hypothesis (equal variances) is very small, then we have no reason to reject the null hypothesis and believe that the variances of the samples are equal.

[0140] Furthermore, if the test result meets the preset conditions, the first fault diagnosis feature set is subjected to factor analysis in the above manner to obtain the hidden variables corresponding to the first fault diagnosis feature set, and the hidden variables are evaluated and optimized to obtain the second fault diagnosis feature set. Among them, the hidden variables can be optimized by deleting the original variables, that is, some unreasonable original variables can be deleted according to the commonality, factor loading coefficient and the correspondence between factors and items; the hidden variables can be optimized by adjusting the number of factors and / or the combination of factors, that is, the number of factors or the combination of factors can be adjusted according to the explanatory power of the factors, the distinguishability of the factors and the naming of the factors; the hidden variables can also be optimized by multiple cycles.

[0141] Specifically, factor analysis is used to perform one-dimensional classification on the data in the first fault diagnosis feature set, and specific analysis is performed based on the obtained KMO and Bartlett values:

[0142] Analyze KMO value: if this value is higher than the first threshold, it means that it is very suitable for factor analysis; the first threshold is preferably 0.8; if this value is between the first threshold and the second threshold, it means that it is more suitable for factor analysis; the second threshold is preferably 0.7; if this value is between the second threshold and the third threshold, it means that factor analysis can be performed; the third threshold is preferably 0.6; if this value is less than the third threshold, it means that it is not suitable for factor analysis; if the p value corresponding to the Bartlett test is less than the fourth threshold, it means that it is suitable for factor analysis; the fourth threshold is preferably 0.05. Then, the weights of the classified factors are obtained according to the column of the variance explanation rate after rotation in the variance explanation rate table and the last accumulated value. The last accumulated value represents the sum of the variance explanation rates after rotation of all factors, that is, the total percentage of the variance of the original variables that all factors can explain. The closer the last accumulated value is to 100%, the better the effect of factor analysis.

[0143] Variance Explained Table

[0144]

[0145]

[0146] Components or factors in the table: Indicates the number of the extracted principal component or common factor;

[0147] Initial eigenvalue: represents the eigenvalue value of each principal component or factor before rotation, that is, the variance of each principal component or factor; the larger the eigenvalue, the more information the principal component or factor contains;

[0148] Initial variance explanation rate: It indicates the percentage of the initial characteristic root of each principal component or factor to the sum of the variances of all variables, that is, the explanatory power of each principal component or factor;

[0149] Initial cumulative variance explanation rate: It means the sum of the initial variance explanation rates of all principal components or factors from the first principal component or factor to the current principal component or factor, that is, the cumulative explanatory power;

[0150] Rotated characteristic root: represents the characteristic root value of each principal component or factor after rotation, that is, the variance of each principal component or factor. The purpose of rotation is to minimize the correlation between principal components or factors, thereby improving the explanatory effect;

[0151] Variance explanation rate after rotation: It indicates the percentage of the characteristic root of each principal component or factor after rotation to the sum of the variances of all variables, that is, the explanatory power of each principal component or factor;

[0152] Cumulative variance explanation rate after rotation: It indicates the sum of the variance explanation rates of all principal components or factors after rotation, starting from the first principal component or factor to the current principal component or factor, that is, the cumulative explanatory power.

[0153] S140, performing entropy analysis on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing, and completing the monitoring of the current wind turbine main shaft bearing.

[0154] In some embodiments, the multiple factors obtained after the factor analysis are analyzed separately using the SPSS entropy method. That is, the second fault diagnosis feature set is subjected to entropy analysis using the SPSS entropy method. Each factor will obtain a summary table of weight calculation results using the entropy method, in which the weight coefficient is the weight corresponding to the sub-factor contained in a factor. Repeating this operation for each factor will obtain the weights of all required sub-factors.

[0155] Finally, the one-dimensional index (factor) and the two-dimensional index (sub-factor) are weighted and calculated. The specific formula is as follows:

[0156]

[0157] Finally, an upper and lower limit of a domain value of this set of data is obtained, and whether there is a fault can be judged based on the range of the upper and lower limits. That is, if it exceeds the upper and lower limits, a fault exists and the monitoring of the current electric spindle bearing is completed.

[0158] According to the embodiments of the present disclosure, the following technical effects are achieved:

[0159] The online fault status monitoring and analysis of the main shaft bearings of wind turbines is carried out through acoustic emission technology, which realizes the early detection and effective early warning of main shaft bearing faults and improves the overall efficiency of wind turbine operation.

[0160] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0161] The above is an introduction to the method embodiment. The following is a further explanation of the scheme described in this application through an apparatus embodiment.

[0162] Figure 2 FIG. 1 is a block diagram of a device for monitoring a wind turbine main shaft bearing based on acoustic emission technology according to an embodiment of the present application, as shown in FIG. Figure 2 Shown include:

[0163] An acquisition module 210 is used to acquire an acoustic emission signal of a wind turbine main shaft bearing;

[0164] An extraction module 220 is used to extract frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set;

[0165] A first analysis module 230 is used to perform a Bartlett test on the fault diagnosis feature set, and if the test result meets a preset condition, perform a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set;

[0166] The second analysis module 240 is used to perform entropy analysis on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing and complete the monitoring of the current wind turbine main shaft bearing.

[0167] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0168] Figure 3 A schematic diagram of the structure of a terminal device or server suitable for implementing an embodiment of the present application is shown.

[0169] like Figure 3As shown, the terminal device or server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 302 or the program loaded from the storage part 308 into the random access memory (RAM) 303. In the RAM 303, various programs and data required for the operation of the terminal device or server are also stored. CPU 301, ROM 302 and RAM 303 are connected to each other through a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0170] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed, so that a computer program read therefrom is installed into the storage section 308 as needed.

[0171] In particular, according to an embodiment of the present application, the above method flow steps can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a machine-readable medium, and the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, the above-mentioned functions defined in the system of the present application are executed.

[0172] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0173] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of the code, and the aforementioned module, program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0174] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0175] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the method described in the present application.

[0176] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of application involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned application concept. For example, the above features are replaced with (but not limited to) technical features with similar functions applied in the present application.

Claims

1. A method for monitoring wind turbine main shaft bearings based on acoustic emission technology. It is characterized in that include: Obtain the acoustic emission signal of the wind turbine main shaft bearing; Performing frequency domain feature extraction on the acoustic emission signal to construct a first fault diagnosis feature set; Performing a Bartlett test on the first fault diagnosis feature set, and if the test result meets a preset condition, performing a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set; An entropy analysis is performed on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing, thereby completing the monitoring of the current wind turbine main shaft bearing.

2. The method according to claim 1, It is characterized in that Also includes: The bearing acoustic emission signal is denoised to obtain a clean signal.

3. The method according to claim 2, It is characterized in that The denoising process of the acoustic emission signal to obtain a clean signal comprises: The optimal wavelet coefficients are determined by the following formula: x(t)=∑ i c i ψ i (t) Among them, ci is the wavelet coefficient; ψ i (t) is the wavelet basis function; f 1 、f 2 、f 3 、f 4 、f 5 are the resonant frequencies of the three components respectively; t 1 ,t 2 ,t 3 ,t 4 ,t 5 are the time points at which the three component signals occur respectively; σ 1 , σ 2 , σ 3 , σ 4 , σ 5 are the standard deviations of the three components respectively; According to the optimal wavelet coefficients, the acoustic emission signal is denoised to obtain a clean signal.

4. The method according to claim 3, It is characterized in that The denoising process of the acoustic emission signal to obtain a clean signal further comprises: Decomposing the acoustic emission information based on a preset number of decomposition modes; The components with kurtosis values ​​greater than the preset threshold in the decomposition results are retained to suppress the noise of the acoustic emission signal and obtain a clean signal.

5. The method according to claim 4, It is characterized in that The kurtosis value is determined as follows: Among them, KU is the kurtosis value; σ is the standard deviation.

6. The method according to claim 5, It is characterized in that The extracting frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set comprises: Based on the clean signal, construct a high-dimensional matrix; Decomposing the high-dimensional matrix to obtain multiple singular values; Based on the multiple singular values, a first fault diagnosis feature set is constructed.

7. The method according to claim 6, It is characterized in that Decomposing the high-dimensional matrix to obtain multiple singular values ​​includes: Among them, Σ is a non-negative diagonal matrix; Matrices U and V are the left singular matrix and right singular matrix of matrix A, respectively.

8. A device for monitoring wind turbine main shaft bearings based on acoustic emission technology. It is characterized in that include: An acquisition module, used for acquiring acoustic emission signals of wind turbine main shaft bearings; An extraction module, used for extracting frequency domain features of the acoustic emission signal to construct a first fault diagnosis feature set; A first analysis module is used to perform a Bartlett test on the fault diagnosis feature set, and if the test result meets a preset condition, perform a factor analysis on the fault diagnosis feature set to obtain a second fault diagnosis feature set; The second analysis module is used to perform entropy analysis on the second fault diagnosis feature set to determine whether there is a fault in the current wind turbine main shaft bearing and complete the monitoring of the current electric main shaft bearing.

9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory. It is characterized in that When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Citation Information

Patent Citations

  • Wind turbine generator running state online evaluation method

    CN112580940A

  • Early fault diagnosis method for wind power main shaft bearing

    CN115165368A

  • Active area distribution network line abnormal line loss diagnosis method based on factor analysis

    CN115640950A

  • Bearing fault diagnosis method and system based on digital signal processing

    CN117250004A