CN114942133B - Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox - Google Patents
Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox Download PDFInfo
- Publication number
- CN114942133B CN114942133B CN202210548982.2A CN202210548982A CN114942133B CN 114942133 B CN114942133 B CN 114942133B CN 202210548982 A CN202210548982 A CN 202210548982A CN 114942133 B CN114942133 B CN 114942133B
- Authority
- CN
- China
- Prior art keywords
- matrix
- pkmcq
- clustering
- spectrum
- fault
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/02—Gearings; Transmission mechanisms
- G01M13/021—Gearings
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/02—Gearings; Transmission mechanisms
- G01M13/028—Acoustic or vibration analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/70—Wind energy
- Y02E10/72—Wind turbines with rotation axis in wind direction
Abstract
The invention provides an optimal rank non-negative matrix decomposition-based early fault diagnosis method for a planetary gearbox. Firstly, acquiring a fault vibration signal of the planetary gearbox, and calculating an STFT spectrum of the vibration signal; then, inputting the original vibration signal frequency spectrum as K-means clusters, and obtaining a curve graph with a clustering quality index (pkmcq) as a vertical coordinate and a clustering number K as a horizontal coordinate; secondly, selecting a k value corresponding to a first inflection point in a k-pkmcq curve as an optimal rank of nonnegative matrix decomposition to decompose the STFT spectrum of the vibration signal to obtain a base matrix W containing the spectrum characteristics of the original signal; then, selecting a base vector with the maximum kurtosis value in the base matrix W as an optimal base vector to filter the original signal to obtain a filtered signal; finally, envelope demodulation is carried out on the filtered signals, characteristic frequency is extracted from an envelope spectrum, and the fault type of the gearbox is identified. The method can accurately diagnose the early weak fault of the planetary gear box.
Description
Technical Field
The invention belongs to the technical field of fault diagnosis and signal analysis of rotary machines, and relates to an early fault diagnosis method of a rotary machine, in particular to an early fault diagnosis method of a planetary gearbox based on optimal rank non-negative Matrix Factorization (non-negative Matrix Factorization), which can be used for fault diagnosis of gears and rolling bearing components in the planetary gearbox.
Background
Planetary gear transmission is widely used in the field of modern mechanical industry due to its advantages of small size, large transmission ratio, strong bearing capacity, etc. Considering the structural characteristics of the planetary gearbox, the planetary gearbox usually runs under severe working conditions such as heavy load, impact, pollution and the like, so that key parts such as gears and bearings are easy to break down, and the equipment is stopped even more seriously. Therefore, the method for monitoring the unfolding state of the planetary gear box, exploring an efficient fault diagnosis method, accurately extracting and identifying fault characteristics in the early stage of damage development and has important significance for ensuring the safe operation of equipment and reducing economic loss.
The operating conditions inside the planetary gearbox can be expressed by vibration information. The early failure characteristics of the gear box are weak, and the gear box is easily submerged in strong noise. The effective analysis and processing of the vibration signals containing the fault information of the gearbox are powerful ways to realize the accurate diagnosis of the early fault of the planetary gearbox. Meanwhile, due to the influence of a transmission path and strong environmental noise, the vibration signals of the planetary gearbox can be considered to be mixed by a plurality of signal sources, and the planetary gearbox has a complex frequency spectrum structure. In order to eliminate noise interference, characteristic information representing a gearbox fault mode is separated from an original vibration signal, and various signal analysis technologies such as wavelet transformation, empirical mode decomposition, ensemble empirical mode decomposition, a basis function method, multi-parameter joint estimation and the like are widely applied. However, these methods have their own drawbacks. Wavelet transformation requires predetermination of wavelet bases and the number of decomposition layers, and lacks adaptivity. Empirical mode decomposition, while well adaptive, can automatically decompose an input signal into a series of sums of natural mode functions and residuals. However, for gearbox vibration signals with weak characteristic information, modal aliasing is easily generated by decomposition. Although the ensemble empirical mode decomposition improves the modal aliasing problem of the decomposition to a certain extent, two input parameters of the intensity and the average times of the additional white noise are additionally introduced, and the operation amount is increased. Different from the method, the non-negative matrix decomposition is used as a matrix decomposition algorithm, the signal local characteristics can be effectively extracted through the clustering analysis of the non-negative matrix, and the algorithm is simple, has interpretability and clear physical significance, and is widely applied to the fields of signal processing, pattern recognition and the like. However, in the implementation process of the non-negative matrix factorization algorithm, the selection of the rank k is always a puzzling problem. For non-negative matrix factorization of mechanical fault signals, most of existing k selection means are selected based on experimental effects or artificial experiences after traversal factorization. The former is generally time consuming and the latter decomposition quality is not guaranteed. In addition, the precise screening of the components generated by the decomposition and how to extract the fault features based on the screened components also affect the application of the non-negative matrix decomposition in the aspect of weak signal feature extraction.
Disclosure of Invention
In order to solve the problems, the invention aims to provide an optimal rank non-negative matrix factorization-based early fault diagnosis method for the planetary gearbox. The method can get rid of dependence on priori knowledge such as a signal processing technology and diagnosis experience, and the like, overcomes the prior condition of the traditional nonnegative matrix decomposition algorithm on the decomposition rank, and solves the problems that the fault information of the planetary gear box is difficult to diagnose weakly and the like under the strong noise background.
The technical scheme of the invention is as follows:
an optimal rank non-negative matrix factorization-based early fault diagnosis method for a planetary gearbox comprises the following steps:
step S1: arranging an acceleration sensor right above a shell, in which the sensor is easy to place, of the planetary gear box, collecting a failure original vibration signal S (t) of the planetary gear box, and performing fast Fourier transform to obtain an original signal frequency spectrum S (f);
step S2: and taking the original signal spectrum S (f) of the gearbox as input, sequentially adding one to the initial clustering center number K =2, traversing and carrying out K-Means clustering to obtain a line graph with the clustering center number K as a horizontal coordinate and the spectrum K-Means clustering quality (pkmcq) as a vertical coordinate. The horizontal and vertical coordinates are dimensionless parameters;
wherein, the new index pkmcq is defined as the ratio of the intra-class average distance to the inter-class average distance of the signal spectrum K-means cluster:
in the formula, C i Denotes the ith cluster, num (C) i ) Number of samples representing ith cluster, c i Is C i Has a centroid of p is C i Sample point(s) in (c). pkmcq represents how good the clustering effect is. As the cluster number k gradually approaches the true cluster number from small to large, the rate of decrease of the pkmcq value also gradually decreases. When the number of clusters k exceeds the true number of clusters and gradually moves away from the true number of clusters, the rate of decrease in the pkmcq value shows a gentle tendency. And drawing a k-pkmcq curve by taking the clustering number k as an abscissa and the pkmcq value as an ordinate. Therefore, an obvious inflection point appears in the k-pkmcq curve for the first time, and the corresponding k value is the estimated clustering number;
and step S3: carrying out short-time Fourier transform on the original vibration signal s (t) to obtain a short-time Fourier transform spectrogram matrix V, and taking the short-time Fourier transform spectrogram matrix V as an input matrix of nonnegative matrix decomposition;
wherein, the spectrogram matrix V is obtained according to the following formula:
in the formula, γ (t) represents a short-time fourier transform window function. t denotes time and τ denotes time delay. The window function is displaced on a time axis along with the change of time t, and the Fourier transform of the truncated part of s (t) through the window function is calculated;
and step S4: taking the optimal clustering number k obtained according to the k-pkmcq curve as a nonnegative matrix to decompose the optimal rank, and guiding the nonnegative matrix decomposition of a spectrogram matrix V to obtain two nonnegative matrixes W = [ W = 1 ,w 2 ,…,w k ]∈R m×k And H = [ H ] 1 ,h 2 ,…,h k ] T ∈R k×n The specific decomposition process is as follows:
V m×n ≈W m×k ×H k×n (s.t.W≥0,H≥0)
in the formula, W is referred to as a base matrix, and H is a weight matrix.
Step S5: calculating kurtosis value of each basis vector in the basis matrix W, and selecting the basis vector W with the largest kurtosis value o Taking the optimal component to perform the next processing;
step S6: according to the optimal basis vector w o Filtering the original signal spectrum S (f) to obtain a filtered signal spectrum S w (f) Then obtaining a filtered signal s by inverse fast Fourier transform w (t) of (d). The specific implementation process of the filtering is shown as follows:
step S7: for the filtered signal s w And (t) performing Hilbert transform demodulation processing, extracting characteristic information (local fault frequency, frequency conversion and combined frequency of the local fault frequency and the frequency conversion of the gear) reflecting the fault of the planetary gearbox from the envelope spectrum, and accurately judging the fault type.
The invention has the beneficial effects that:
(1) The method gives play to the advantages of a non-negative matrix factorization algorithm in the aspect of self-adaptive signal matrix factorization, and solves the problem of difficult early weak fault diagnosis of the planetary gearbox by effectively screening and processing the internal components of the matrix obtained by factorization.
(2) The invention provides a new estimation strategy for decomposing the optimal rank by a signal non-negative matrix, constructs a line graph of the number K of clustering centers and the clustering quality of a frequency spectrum K-Means, effectively overcomes the prior dependence of a non-negative matrix decomposition algorithm on the decomposed rank, and enables the non-negative matrix decomposition algorithm to be suitable for a strong noise background.
(3) The invention designs a novel self-adaptive filtering method, which is used for carrying out subsequent frequency domain filtering processing by self-adaptively selecting the optimal basis vector containing fault information in a basis matrix obtained by processing a non-negative matrix decomposition algorithm by means of a maximum kurtosis criterion. On one hand, the stability of the selection of the optimal basis vector is ensured, and on the other hand, the problem that the useful information is easy to filter out when the prior fault information is insufficient by the traditional filtering means is solved.
Drawings
FIG. 1 is a flow chart of an optimal rank non-negative matrix factorization-based planetary gearbox early fault diagnosis method provided by the invention;
FIG. 2 is a time domain waveform, a Fourier spectrum and an envelope spectrum of a sun gear tooth root crack fault vibration signal in an embodiment of the invention; in the figure, (a) is a time-domain waveform of the failure vibration signal, (b) is a fourier spectrum of the failure vibration signal, and (c) is an envelope spectrum of the failure vibration signal.
FIG. 3 is a line graph of the clustering center number K and the clustering quality pkmcq obtained by performing K-means clustering on a sun gear fault vibration signal frequency spectrum in the embodiment of the present invention.
FIG. 4 is a short-time Fourier transform spectrum of a sun gear fault vibration signal in an embodiment of the present invention.
FIG. 5 is a result diagram of adaptive filtering of sun gear fault vibration signals by using an optimal rank non-negative matrix factorization algorithm according to an embodiment of the present invention; in the figure, (a) is a time domain waveform of the filtered signal, (b) is a fourier spectrum of the filtered signal, and (c) is an envelope spectrum of the filtered signal.
Detailed Description
The following detailed description of the embodiments of the present invention is provided in connection with the accompanying drawings.
In the embodiment, fig. 1 shows a flowchart of the early fault diagnosis method for the planetary gearbox based on the optimal rank nonnegative matrix factorization, which includes the following steps:
the method comprises the following steps: an acceleration sensor is arranged right above a box body of the planetary gear box fault experiment table. The internal parameters of the planetary gearbox are shown in table 1. The output rotating speed of the motor is 600r/min (the rotating frequency is 10 Hz). And continuously acquiring the original vibration signal of the sun gear fault at a sampling frequency of 12800Hz, wherein the sampling time is 2s. To simulate a strong noise background, white Gaussian noise with a signal-to-noise ratio of-3 dB is added to the original vibration signal. And carrying out fast Fourier transform and Hilbert demodulation on the noisy vibration signal, and obtaining a frequency spectrum and an envelope spectrum. In fig. 2, (a), (b) and (c) are respectively time domain waveform, frequency spectrum and envelope spectrum of the sun gear fault vibration signal, the interference of noise makes (a) signal time domain waveform exhibit disorder, and (b) frequency spectrum information in fig. 2 is disordered. Meanwhile, no obvious frequency information indicating the local failure of the sun gear is observed in the envelope spectrum shown in fig. 2 (c).
TABLE 1 planetary gearbox internal parameters
Step two: and (4) taking the noisy fault signal frequency spectrum S (f) obtained in the first step as an input to perform K-means clustering processing, and calculating the clustering quality pkmcq under different clustering center numbers K to obtain a K-pkmcq line graph shown in figure 3. And selecting an abscissa value k =6 corresponding to the first inflection point of the line graph as an estimated non-negative matrix factorization optimal rank.
Step three: and carrying out short-time Fourier transform on the original vibration signal to obtain a time-frequency spectrogram as shown in figure 4. Noise is widely distributed in each frequency band of the signal, and the number of the frequency bands of the signal cannot be effectively distinguished in fig. 4. The empirical non-negative matrix factorization optimal rank estimation method proved to be ineffective.
Step four: and (4) carrying out non-negative matrix decomposition on the spectrogram matrix V obtained in the step three to obtain a component set containing a series of original signal characteristic information and a base matrix W.
Step five: calculating kurtosis value of each basis vector in the basis matrix W, and selecting the basis vector W with the largest kurtosis value o The next processing is performed as the best component.
Step six: according to the sensitive component w o The original signal spectrum S (f) shown in fig. 2 (b) is subjected to filtering processing to obtain a filtered signal spectrum S w (f) Then obtaining a filtered signal s by inverse fast Fourier transform w (t) of (d). The waveform and spectrum of the corresponding filtered signal are shown in fig. 5 (a) and (b).
Step seven: for the filtered signal s w (t) hilbert transform demodulation processing is performed, and in fig. 5, (c) shows an envelope spectrum after demodulation. Clear sun wheel local fault frequency f appears in envelope spectrum su Sun wheel rotation frequencyf r And a series of frequency doubling and combined frequencies f su -f r ，f su +f r ，2*f su ，3*f su ，4*f su ，5*f su ，6*f su ，4*f su -f r ，6*f su -f r . These are all favorable indicators of sun gear failure. The analysis result shows the effectiveness and feasibility of the planetary gearbox early fault diagnosis method based on the optimal rank nonnegative matrix decomposition.
Claims (1)
1. An optimal rank non-negative matrix factorization-based early fault diagnosis method for a planetary gearbox is characterized by comprising the following steps of:
step S1: arranging an acceleration sensor right above a shell on which a sensor is easy to place along a planetary gear box, collecting a fault original vibration signal S (t) of the planetary gear box, and performing fast Fourier transform to obtain an original signal frequency spectrum S (f);
step S2: taking the original signal frequency spectrum S (f) of the gearbox as input, sequentially adding one to the initial clustering center number K =2, traversing and carrying out K-Means clustering to obtain a broken line graph taking the clustering center number K as a horizontal coordinate and the frequency spectrum K-Means clustering quality pkmcq as a vertical coordinate; the horizontal and vertical coordinates are dimensionless parameters;
wherein the index pkmcq is defined as the ratio of the intra-class average distance to the inter-class average distance of the signal spectrum K-means cluster:
in the formula, C i Denotes the ith cluster, num (C) i ) Number of samples representing ith cluster, c i Is C i Has a centroid of p is C i The sample point of (1); pkmcq represents the quality of the clustering effect; as the clustering number k gradually approaches to the real clustering number from small to large, the decreasing rate of the pkmcq value is gradually decreased; when the clustering number k exceeds the real clustering number and gradually gets away from the real clustering number, the reduction rate of the pkmcq value presents a gentle trend; using the clustering number k as the abscissaDrawing a k-pkmcq curve by taking the pkmcq value as a vertical coordinate; therefore, an obvious inflection point appears in the k-pkmcq curve for the first time, and the corresponding k value is the estimated clustering number;
and step S3: carrying out short-time Fourier transform on the original vibration signal s (t) to obtain a short-time Fourier transform spectrogram matrix V, and taking the short-time Fourier transform spectrogram matrix V as an input matrix of nonnegative matrix decomposition;
wherein, the spectrogram matrix V is obtained according to the following formula:
wherein γ (t) represents a short-time fourier transform window function; t represents time, τ represents time delay; the window function is displaced on a time axis along with the change of time t, and the Fourier transform of the truncated part of s (t) through the window function is calculated;
and step S4: taking the optimal clustering number k obtained according to the k-pkmcq curve as a nonnegative matrix decomposition optimal rank, and guiding nonnegative matrix decomposition of a spectrogram matrix V to obtain two matrixes W = [ W ] meeting nonnegativity 1 ,w 2 ,…,w k ]∈R m×k And H = [ H = 1 ,h 2 ,…,h k ] T ∈R k×n The specific decomposition process is as follows:
V m×n ≈W m×k ×H k×n (W≥0,H≥0)
in the formula, W is called a base matrix, and H is a weight matrix;
step S5: calculating the kurtosis value of each basis vector in the basis matrix W, and selecting the basis vector W with the largest kurtosis value o Taking the component as a sensitive component to be processed in the next step;
step S6: according to the sensitive component w o Filtering the original signal frequency spectrum S (f) to obtain a filtered signal frequency spectrum S w (f) Then obtaining a filtered signal s by inverse fast Fourier transform w (t); the specific implementation process of the filtering is shown as follows:
step S7: for the filtered signal s w And (t) carrying out Hilbert transform demodulation processing, extracting characteristic information reflecting the faults of the planetary gearbox from the envelope spectrum, wherein the characteristic information is the local fault frequency and the rotating frequency of the gear and the combined frequency of the local fault frequency and the rotating frequency of the gear, and accurately judging the fault type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210548982.2A CN114942133B (en) | 2022-05-20 | 2022-05-20 | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210548982.2A CN114942133B (en) | 2022-05-20 | 2022-05-20 | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114942133A CN114942133A (en) | 2022-08-26 |
| CN114942133B true CN114942133B (en) | 2023-04-14 |
Family
ID=82908367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210548982.2A Active CN114942133B (en) | 2022-05-20 | 2022-05-20 | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114942133B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103728135A (en) * | 2013-12-16 | 2014-04-16 | 西安交通大学 | Bearing fault feature extraction and diagnosis method of non-negative matrix factorization |
| CN107192553A (en) * | 2017-06-28 | 2017-09-22 | 石家庄铁道大学 | Gear-box combined failure diagnostic method based on blind source separating |
| CN109029977A (en) * | 2018-07-12 | 2018-12-18 | 福州大学 | A kind of epicyclic gearbox Incipient Fault Diagnosis method based on VMD-AMCKD |
| CN109632312A (en) * | 2019-01-22 | 2019-04-16 | 北京化工大学 | Bearing combined failure diagnostic method based on multiple constraint Algorithms of Non-Negative Matrix Factorization |
| JP2019067197A (en) * | 2017-10-02 | 2019-04-25 | 日本ユニシス株式会社 | Method for detecting trouble sign |
| CN109883693A (en) * | 2019-04-04 | 2019-06-14 | 西安交通大学 | Based on low-rank-sparse bayesian learning epicyclic gearbox method for diagnosing faults |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AT385359B (en) * | 1986-03-26 | 1988-03-25 | Asa Hydraulik Gmbh | DEVICE FOR CHECKING THE AXLE BEARINGS OF VEHICLE BICYCLES |
| US9727532B2 (en) * | 2008-04-25 | 2017-08-08 | Xerox Corporation | Clustering using non-negative matrix factorization on sparse graphs |
| US20210157312A1 (en) * | 2016-05-09 | 2021-05-27 | Strong Force Iot Portfolio 2016, Llc | Intelligent vibration digital twin systems and methods for industrial environments |
-
2022
- 2022-05-20 CN CN202210548982.2A patent/CN114942133B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103728135A (en) * | 2013-12-16 | 2014-04-16 | 西安交通大学 | Bearing fault feature extraction and diagnosis method of non-negative matrix factorization |
| CN107192553A (en) * | 2017-06-28 | 2017-09-22 | 石家庄铁道大学 | Gear-box combined failure diagnostic method based on blind source separating |
| JP2019067197A (en) * | 2017-10-02 | 2019-04-25 | 日本ユニシス株式会社 | Method for detecting trouble sign |
| CN109029977A (en) * | 2018-07-12 | 2018-12-18 | 福州大学 | A kind of epicyclic gearbox Incipient Fault Diagnosis method based on VMD-AMCKD |
| CN109632312A (en) * | 2019-01-22 | 2019-04-16 | 北京化工大学 | Bearing combined failure diagnostic method based on multiple constraint Algorithms of Non-Negative Matrix Factorization |
| CN109883693A (en) * | 2019-04-04 | 2019-06-14 | 西安交通大学 | Based on low-rank-sparse bayesian learning epicyclic gearbox method for diagnosing faults |
Non-Patent Citations (2)
| Title |
|---|
| Nonnegative factorization of spectrogram for local damage detection of belt conveyor gearboxes;Jacek Wodecki等;《International Federation of Automatic Control》;20171231;4714-4718 * |
| 基于 k 值优化 VMD 的滚动轴承故障诊断方法;王奉涛等;《振动 、测试与诊断》;20180630;540-547 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114942133A (en) | 2022-08-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Early fault diagnosis for planetary gearbox based on adaptive parameter optimized VMD and singular kurtosis difference spectrum | |
| Su et al. | Rolling element bearing faults diagnosis based on optimal Morlet wavelet filter and autocorrelation enhancement | |
| CN111089726B (en) | Rolling bearing fault diagnosis method based on optimal dimension singular spectrum decomposition | |
| CN110657985B (en) | Gearbox fault diagnosis method and system based on singular value spectrum manifold analysis | |
| CN109883706B (en) | Method for extracting local damage weak fault features of rolling bearing | |
| CN102937522B (en) | Composite fault diagnosis method and system of gear case | |
| CN110806315B (en) | Gearbox composite fault diagnosis method based on inversion editing | |
| CN108801630B (en) | Gear fault diagnosis method for single-channel blind source separation | |
| Shao et al. | Multi-fault feature extraction and diagnosis of gear transmission system using time-frequency analysis and wavelet threshold de-noising based on EMD | |
| CN113375939B (en) | Mechanical part fault diagnosis method based on SVD and VMD | |
| He et al. | A joint adaptive wavelet filter and morphological signal processing method for weak mechanical impulse extraction | |
| Chen et al. | Fault feature extraction and diagnosis of gearbox based on EEMD and deep briefs network | |
| CN110795843A (en) | Method and device for identifying faults of rolling bearing | |
| Xu et al. | An adaptive spectrum segmentation method to optimize empirical wavelet transform for rolling bearings fault diagnosis | |
| CN113390631A (en) | Fault diagnosis method for gearbox of diesel engine | |
| Xiao et al. | Research on fault feature extraction method of rolling bearing based on NMD and wavelet threshold denoising | |
| CN107966287B (en) | Weak fault feature extraction method for self-adaptive electromechanical equipment | |
| Li et al. | Development of a morphological convolution operator for bearing fault detection | |
| CN114964769A (en) | Wind power gear box vibration signal fault diagnosis method | |
| CN114942133B (en) | Optimal rank non-negative matrix factorization-based early fault diagnosis method for planetary gearbox | |
| CN113033304B (en) | Multi-resonance-band amplitude demodulation analysis method for overcoming frequency domain overlapping interference | |
| CN111582248B (en) | SVD-based gearbox signal noise reduction method | |
| CN110044619B (en) | Multi-fault feature identification method based on sparse multi-cycle group lasso | |
| Du et al. | Translation invariance-based deep learning for rotating machinery diagnosis | |
| CN115434872A (en) | Wind turbine generator gearbox composite fault diagnosis method based on AVMD and improved RSSD |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |