CN111678691A - Gear fault detection method based on improved sparse decomposition algorithm - Google Patents
Gear fault detection method based on improved sparse decomposition algorithm Download PDFInfo
- Publication number
- CN111678691A CN111678691A CN202010732900.0A CN202010732900A CN111678691A CN 111678691 A CN111678691 A CN 111678691A CN 202010732900 A CN202010732900 A CN 202010732900A CN 111678691 A CN111678691 A CN 111678691A
- Authority
- CN
- China
- Prior art keywords
- signal
- wavelet
- reconstructed
- detected
- target
- 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.)
- Pending
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
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a gear fault detection method based on an improved sparse decomposition algorithm, which relates to the technical field of fault detection, and is characterized in that on the basis of traditional sparse reconstruction based on a parameter dictionary, signal preprocessing and optimal design on the parameter dictionary are added, the signal preprocessing is realized by combining dual-tree complex wavelet decomposition with the maximum kurtosis principle, the influence of noise on subsequent processing is greatly reduced, a complete dictionary is constructed by determining target characteristic parameters based on relevant filtering of Laplace wavelets, the redundancy of the dictionary can be effectively reduced, the designed dictionary is more similar to fault characteristics, and finally, the matching pursuit algorithm is combined to extract impact characteristics in a vibration signal to realize fault detection.
Description
Technical Field
The invention relates to the technical field of fault detection, in particular to a gear fault detection method based on an improved sparse decomposition algorithm.
Background
Gears are widely used in modern industrial equipment such as electric power, petroleum, transportation, and agriculture as one of indispensable parts of rotary machines. As the running conditions of the gears in the actual engineering are complex and changeable and inevitable faults occur, according to statistics, 80% of equipment faults in the traditional machinery are caused by the gears, and therefore, the effective fault diagnosis of the gears has very important significance for reducing economic loss and casualties.
At present, a diagnosis method based on vibration analysis is one of important ways for judging faults of rotary machines, time domain, frequency domain and time domain characteristics of vibration signals are important basis for fault diagnosis, but in practical application, because the influence of background noise and early fault characteristics are not obvious, fault diagnosis becomes very difficult, and effective fault identification is difficult to perform by using a common time-frequency analysis means. In recent years, sparse representation theory is applied to the field of fault diagnosis, feature extraction based on sparse decomposition becomes a new research hotspot, but the application of sparse decomposition in fault diagnosis still has many problems, because actual vibration signals are very complicated, a large data volume is required to ensure that complete fault information is acquired, which leads to a rapid increase of calculation amount, and in addition, a proper over-complete dictionary is constructed, which plays a crucial role in the effect of sparse reconstruction.
Disclosure of Invention
The invention provides a gear fault detection method based on an improved sparse decomposition algorithm aiming at the problems and the technical requirements, and the technical scheme of the invention is as follows:
a gear fault detection method based on an improved sparse decomposition algorithm is characterized by comprising the following steps:
collecting a vibration signal sample, carrying out dual-tree complex wavelet decomposition on the vibration signal sample to obtain a plurality of signal components, and selecting a target signal component containing most fault characteristic information from all the signal components according to a maximum kurtosis principle;
performing parameter identification on a target signal component based on relevant filtering of a Laplace wavelet to determine a target characteristic parameter;
generating an overcomplete dictionary matched with fault feature information by taking Laplacian wavelet atoms constructed by target feature parameters as a kernel function;
acquiring a vibration signal to be detected of a gear to be detected, and performing sparse reconstruction on the vibration signal to be detected by using an over-complete dictionary based on a matching tracking algorithm to obtain a reconstructed signal;
and demodulating and analyzing the reconstructed signal to realize fault detection of the gear to be detected.
The further technical scheme is that the method for identifying the parameters of the target signal components based on the relevant filtering of the Laplace wavelet to determine the target characteristic parameters comprises the following steps:
taking the Laplace wavelet as a kernel function, and forming a wavelet feature library by performing discrete construction on feature parameters of the Laplace wavelet;
calculating the correlation coefficient of each wavelet in the wavelet feature library and the target signal component;
and determining the characteristic parameter of the wavelet with the maximum correlation coefficient as a target characteristic parameter.
The further technical scheme is that the correlation coefficient of each wavelet in the wavelet feature library and the target signal component is calculated, and the calculation comprises the following steps of for each wavelet in the wavelet feature library:
according to the formula
Calculating a correlation coefficient of the wavelet with the target signal component, wherein krFor correlation coefficient,. phiγ(t) is wavelet, and x (t) is target signal component.
The further technical scheme is that an overcomplete dictionary matched with fault feature information is generated by taking Laplace wavelet atoms constructed by target feature parameters as a kernel function, and the overcomplete dictionary comprises the following steps:
and taking Laplace wavelet atoms constructed by the target characteristic parameters as a kernel function, and translating the kernel function on a signal time course to generate an over-complete dictionary.
The further technical scheme is that sparse reconstruction is carried out on the vibration signal to be detected by using the over-complete dictionary based on the matching pursuit algorithm to obtain a reconstructed signal, and the sparse reconstruction method comprises the following steps:
and sequentially selecting atoms in the over-complete dictionary as local optimal atoms according to the maximum inner product principle, and performing linear addition on the selected local optimal atoms to obtain a reconstructed signal.
The further technical scheme is that atoms in the overcomplete dictionary are sequentially selected as local optimal atoms according to the inner product maximum principle, and the selected local optimal atoms are subjected to linear addition to obtain a reconstructed signal, and the method comprises the following steps:
using the vibration signal to be detected as a signal to be reconstructed of the first iteration;
in the ith iteration, calculating the inner product of the signal to be reconstructed and each atom in the overcomplete dictionary, and determining the atom corresponding to the maximum value of the inner product as a local optimal atom;
performing linear addition on the local optimal atoms selected by the ith iteration to obtain an ith iteration result, wherein each iteration result comprises a reconstruction signal and a residual signal;
if i is less than m, making i equal to i +1, taking the residual signal in the ith iteration result as the signal to be reconstructed, and executing the step of calculating the inner product of the signal to be reconstructed and each atom in the overcomplete dictionary again;
and when i is m, taking the reconstructed signal in the m-th iteration result as the reconstructed signal of the vibration signal to be detected.
The further technical scheme is that the demodulation analysis of the reconstructed signal comprises the following steps: and carrying out frequency domain analysis and envelope analysis on the reconstructed signal.
The further technical scheme is that the fault detection of the gear to be detected is realized by demodulating and analyzing the reconstructed signal, and the fault detection method comprises the following steps:
if a modulation side frequency band taking the meshing frequency as the center appears in the reconstructed signal during frequency domain analysis and/or a frequency doubling phenomenon of frequency conversion appears in the reconstructed signal during envelope analysis, determining that the gear to be detected has a fault;
wherein the frequency is converted
Mesh frequency of fm＝z·frN is the rotating speed of the shaft where the gear to be detected is located, and z is the number of teeth of the gear to be detected.
The beneficial technical effects of the invention are as follows:
the application discloses a gear fault detection method based on an improved sparse decomposition algorithm, which is characterized in that on the basis of traditional sparse reconstruction based on a parameter dictionary, signal preprocessing and optimization design of the parameter dictionary are added, and a matching pursuit algorithm is combined to extract impact characteristics in a vibration signal. The influence of noise on subsequent processing is greatly reduced by preprocessing the signals, the dictionary is designed to be more similar to fault characteristics by the optimized design of dictionary parameters, the redundancy of the dictionary is effectively reduced by the constructed over-complete dictionary, and the calculation speed of sparse representation is improved. The method can effectively reduce the influence of noise, can realize the effective separation of complex signals, has the characteristics of translation invariance and modal aliasing resistance, is applied to the processing of gear signals, can separate fault impact components, modulation components and noise in the signals, hardly introduces false components, can improve the calculation efficiency of sparse representation, and realizes effective fault diagnosis.
Drawings
FIG. 1 is a schematic flow diagram of a gear fault detection method of the present application.
Fig. 2 is a diagram illustrating the result of a dual-tree complex wavelet decomposition of a signal in an example of the present application.
Fig. 3 is a diagram illustrating the result of correlation filtering in an example of the present application.
Fig. 4 is a graph showing a correlation coefficient with frequency and viscous damping at τ of 0.034s in the example of the present application.
Fig. 5 is a reconstructed signal of a signal obtained in an example of the present application.
Fig. 6 is a residual signal of a signal obtained in an example of the present application.
Fig. 7 is a schematic diagram of frequency domain characteristics of a reconstructed signal in an example of the present application.
Fig. 8 is a diagram of an envelope spectrum in an example of the present application.
Detailed Description
The following further describes the embodiments of the present invention with reference to the drawings.
The application discloses a gear fault detection method based on an improved sparse decomposition algorithm, please refer to a flow chart shown in fig. 1, and the method comprises the following steps:
and step S1, collecting a vibration signal sample, carrying out dual-tree complex wavelet decomposition on the vibration signal sample to obtain a plurality of signal components, and selecting a target signal component containing the most fault characteristic information from all the signal components according to the maximum kurtosis principle.
Firstly, performing dual-tree complex wavelet decomposition on a vibration signal sample, wherein the dual-tree complex wavelet decomposition is different from the traditional discrete wavelet transformation, the dual-tree complex wavelet transformation adopts a pair of wavelet filters which satisfy approximate analysis relation to perform the discrete wavelet transformation on the signal, and the obtained coefficients are real part wavelet coefficients respectively
With imaginary wavelet coefficients
Corresponding scale factor
Where x (t) is a vibration signal sample, #h(t) is the real tree wavelet,. psigAnd (t) is an imaginary tree wavelet, J is the decomposition layer number, and J is 1,2, … and J. The components of the corresponding layer can be reconstructed by calculation according to the wavelet coefficients and the scale coefficients:
a plurality of components can be obtained after dual-tree complex wavelet decomposition, then the kurtosis of each component is calculated, the kurtosis is numerical statistics for representing the deviation degree of a random variable from normal distribution, and is very sensitive to an impact signal, the more obvious the impact is, the larger the absolute value of the kurtosis is, and the existence of the impact signal is one of important characteristics of gear failure, so that the kurtosis can be used as a factor for measuring gear failure, and the component with most failure information is selected by utilizing the sensitivity of the kurtosis to the impact signal. The kurtosis is calculated as follows:
e () represents the mathematical expectation, x represents the components, μ represents the mean of each component, and σ represents the variance of each component.
And step S2, performing parameter identification on the target signal component based on the relevant filtering of the Laplace wavelet to determine the target characteristic parameter. The laplacian wavelet is a single-side vibration attenuation signal, and the expression is as follows:
the characteristic parameters of the laplacian wavelet include a characteristic frequency f, a viscous damping zeta and a time parameter tau, and the characteristic parameters determine the characteristics of the wavelet. A is normalized by waveletTransformation factor, WsThe width of the tight support interval of the wavelet.
A wavelet feature library can be constructed by taking the Laplace wavelet as a kernel function and dispersing the feature parameters f, zeta and tau.
Computing individual wavelets psi in a wavelet feature libraryγ(t) a correlation coefficient with the target signal component x (t) by the formula
The correlation coefficient krThe size of (a) reflects the degree of similarity between the component and the atom, and therefore the characteristic parameter of the wavelet atom corresponding to the maximum value of the correlation coefficient is considered as the modal parameter reflecting the component, and therefore the characteristic parameter of the wavelet with the maximum correlation coefficient is determined as the target characteristic parameter, which is used for the construction of the overcomplete dictionary.
Step S3, an overcomplete dictionary matched with fault feature information is generated by taking Laplace wavelet atoms constructed by target feature parameters as a kernel function, the motion characteristics of a rotary machine are applied, impact signals with obvious periodicity are generated when local faults occur, and the feature parameters of each impact are the same, so that the overcomplete dictionary is generated by translating the kernel function on a signal time history. The constructed overcomplete dictionary effectively reduces the redundancy of the dictionary, improves the calculation speed of sparse representation, and simultaneously considers the similarity of the dictionary and signals.
And step S4, acquiring a vibration signal to be detected of the gear to be detected, and performing sparse reconstruction on the vibration signal to be detected by utilizing the over-complete dictionary based on the matching tracking algorithm to obtain a reconstructed signal, so as to extract an impact component. The matching pursuit is a greedy algorithm for local optimization, atoms in an overcomplete dictionary are sequentially selected as local optimal atoms according to the inner product maximization principle, and a reconstructed signal is obtained by linearly adding the selected local optimal atoms. Let H be Hilbert space, overcomplete dictionary
Assume an arbitrary signal f ∈ H, based on a matching pursuit algorithmThe process of solving for its sparse representation is as follows: firstly, using a vibration signal to be detected as a signal to be reconstructed of a first iteration, calculating an inner product of the signal to be reconstructed and each atom in an overcomplete dictionary in the ith iteration, determining an atom corresponding to the maximum value of the inner product as a local optimal atom, and determining the counts of the atoms in the first iteration<f,gγ0>|＝sup|<f,gγ>|，gγIs any one atom, gγ0Is the locally optimal atom determined after the first iteration. Linearly adding the local optimal atoms selected by the ith iteration to obtain an ith iteration result, wherein each iteration result comprises a reconstruction signal and a residual signal, and the first iteration result can be expressed as
R is the projection of the signal on the atom, i.e. the reconstruction thereof1f is a residual signal. If i<And m, setting i to be i +1, taking a residual signal in the ith iteration result as a signal to be reconstructed, and performing the step of calculating the inner product of the signal to be reconstructed and each atom in the overcomplete dictionary again, wherein when i is m, after m iterations, a signal f is expressed as
And taking the reconstructed signal in the mth iteration result as the reconstructed signal of the vibration signal to be detected.
And step S5, demodulating and analyzing the reconstructed signal to realize fault detection of the gear to be detected. According to the fault mechanism of the gear, when the gear has a local fault, such as pitting corrosion, tooth breakage and the like, the energy distribution of the vibration signal is changed, periodic pulse impact occurs, and then the signal modulation phenomenon is caused. When demodulation analysis is carried out, frequency domain analysis is carried out by combining a spectrogram, and Hilbert envelope analysis is carried out by combining an envelope map. When demodulation analysis is carried out, if a modulation side frequency band taking meshing frequency as the center appears when the reconstruction signal is subjected to frequency domain analysis, and/or the reconstruction signal is subjected to frequency conversion frequency doubling phenomenon when envelope analysis is carried out, determining that the reconstruction signal to be detected is to be detectedThere is a failure of the gear. Wherein the frequency is converted
Mesh frequency of fm＝z·frN is the rotating speed of the shaft where the gear to be detected is located, and z is the number of teeth of the gear to be detected.
The method of the present application is further illustrated by the following example:
firstly, fault data are collected, a rotor test bed is built, the rotating speed is set to be 600r/min, the sampling frequency is 10000Hz, and a gear broken tooth fault vibration signal is collected. The 4-layer dual-tree complex wavelet decomposition is performed on the acquired signals, and the decomposition results are shown in fig. 2. When the dual-tree complex wavelet is decomposed into the third and fourth layers (D3 and D4), obvious impact attenuation components appear in the components, and the component D4 with the largest kurtosis value contains most fault information by combining the criterion of the largest kurtosis value. And selecting a D4 component as a function to be identified of the correlation filtering to identify the parameter.
Performing relevant filtering based on Laplace wavelet on the components, and identifying characteristic parameters of the components; and (3) combining a signal time domain, a signal frequency domain and a theoretical rule, respectively setting frequency, damping and time parameters, and constructing a Laplace wavelet feature library:
F＝{320:1:370}
Z＝{0.01:0.001:0.1,0.2:0；1:0.9}
T＝{0:dt:0.1}
where dt is the sampling interval. As shown in fig. 3, it is understood from the graph of the result of the correlation filtering that the correlation coefficient Kr reaches the maximum value of 0.755 when the time parameter τ is 0.034s, which indicates that the atom is most similar to the signal. Fig. 4 shows the relationship between the correlation coefficient and the frequency and the viscous damping under the time parameter, and it can be seen from the graph that the frequency and the damping corresponding to the maximum correlation coefficient are respectively 355Hz and 0.016.
And generating an over-complete dictionary for sparse reconstruction based on target characteristic parameters tau, f, 355Hz and ζ, 0.016, wherein the fault characteristic extraction based on sparse decomposition is not used for accurately reconstructing an original signal but used for accurately extracting fault characteristic components in the signal, which is different from the traditional sparse representation. And terminating the matching pursuit algorithm by controlling the iteration times. Fig. 5 shows a reconstructed signal, and fig. 6 shows a residual signal after reconstruction.
And performing FFT and envelope spectrum demodulation on the reconstructed signal, and further performing precise diagnosis on the gear. Fig. 7 shows the frequency domain characteristics of the reconstructed signal, and the frequency domain diagram has significant modulation sidebands, whose center frequency is about 4 times the meshing frequency, and the sidebands are separated by the transition frequency, and the envelope spectrum in fig. 8 also has significant one-time multiplication and two-time multiplication of the transition frequency.
What has been described above is only a preferred embodiment of the present application, and the present invention is not limited to the above embodiment. It is to be understood that other modifications and variations directly derivable or suggested by those skilled in the art without departing from the spirit and concept of the present invention are to be considered as included within the scope of the present invention.
Claims (8)
1. A gear fault detection method based on an improved sparse decomposition algorithm is characterized by comprising the following steps:
collecting a vibration signal sample, carrying out dual-tree complex wavelet decomposition on the vibration signal sample to obtain a plurality of signal components, and selecting a target signal component containing most fault characteristic information from all the signal components according to a maximum kurtosis principle;
performing parameter identification on the target signal component based on the relevant filtering of the Laplace wavelet to determine a target characteristic parameter;
generating an overcomplete dictionary matched with fault feature information by taking Laplacian wavelet atoms constructed by the target feature parameters as a kernel function;
acquiring a vibration signal to be detected of a gear to be detected, and performing sparse reconstruction on the vibration signal to be detected by using the over-complete dictionary based on a matching pursuit algorithm to obtain a reconstructed signal;
and demodulating and analyzing the reconstructed signal to realize fault detection of the gear to be detected.
2. The method of claim 1, wherein the performing of the parameter identification on the target signal component based on the laplacian wavelet correlation filtering to determine the target feature parameter comprises:
taking a Laplace wavelet as a kernel function, and forming a wavelet feature library by performing discrete construction on feature parameters of the Laplace wavelet;
calculating the correlation coefficient of each wavelet in the wavelet feature library and the target signal component;
and determining the characteristic parameter of the wavelet with the maximum correlation coefficient as the target characteristic parameter.
3. The method of claim 2, wherein the computing the correlation coefficient of each wavelet in the wavelet feature library with the target signal component comprises, for each wavelet in the wavelet feature library:
according to the formula
Calculating a correlation coefficient of the wavelet with the target signal component, wherein krFor said correlation coefficient, #γ(t) is the wavelet, and x (t) is the target signal component.
4. The method of claim 1, wherein the generating an overcomplete dictionary matched with fault feature information by using the Laplacian wavelet atoms constructed by the target feature parameters as a kernel function comprises:
and taking Laplacian wavelet atoms constructed by the target characteristic parameters as a kernel function, and translating the kernel function on a signal time course to generate the overcomplete dictionary.
5. The method according to claim 1, wherein the sparse reconstruction of the vibration signal to be detected by using the overcomplete dictionary based on the matching pursuit algorithm to obtain a reconstructed signal comprises:
and sequentially selecting atoms in the overcomplete dictionary as local optimal atoms according to the maximum inner product principle, and performing linear addition on the selected local optimal atoms to obtain the reconstruction signal.
6. The method according to claim 5, wherein the sequentially selecting atoms in the overcomplete dictionary as locally optimal atoms according to the inner product maximization principle, and obtaining the reconstructed signal by linearly adding the selected locally optimal atoms comprises:
taking the vibration signal to be detected as a signal to be reconstructed of the first iteration;
in the ith iteration, calculating the inner product of the signal to be reconstructed and each atom in the overcomplete dictionary, and determining the atom corresponding to the maximum value of the inner product as a local optimal atom;
performing linear addition on the local optimal atoms selected by the ith iteration to obtain an ith iteration result, wherein each iteration result comprises a reconstruction signal and a residual signal;
if i is less than m, making i equal to i +1, taking a residual signal in the ith iteration result as the signal to be reconstructed, and executing the step of calculating the inner product of the signal to be reconstructed and each atom in the overcomplete dictionary again;
and when i is m, taking the reconstructed signal in the m-th iteration result as the reconstructed signal of the vibration signal to be detected.
7. The method of claim 1, wherein the performing demodulation analysis on the reconstructed signal comprises: and carrying out frequency domain analysis and envelope analysis on the reconstructed signal.
8. The method according to claim 7, wherein the demodulating and analyzing the reconstructed signal realizes fault detection of the gear to be detected, and comprises:
if a modulation side frequency band taking the meshing frequency as the center appears in the reconstructed signal during frequency domain analysis, and/or a frequency doubling phenomenon of frequency conversion appears in the reconstructed signal during envelope analysis, determining that the gear to be detected has a fault;
wherein the frequency is converted
Mesh frequency of fm＝z·frN is the rotating speed of the shaft where the gear to be detected is located, and z is the number of teeth of the gear to be detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010732900.0A CN111678691A (en) | 2020-07-27 | 2020-07-27 | Gear fault detection method based on improved sparse decomposition algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010732900.0A CN111678691A (en) | 2020-07-27 | 2020-07-27 | Gear fault detection method based on improved sparse decomposition algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111678691A true CN111678691A (en) | 2020-09-18 |
Family
ID=72458013
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010732900.0A Pending CN111678691A (en) | 2020-07-27 | 2020-07-27 | Gear fault detection method based on improved sparse decomposition algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111678691A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112906632A (en) * | 2021-03-18 | 2021-06-04 | 中北大学 | Automatic extraction method of highly adaptive time delay target signal |
| CN114136604A (en) * | 2021-11-30 | 2022-03-04 | 三门核电有限公司 | Rotary equipment fault diagnosis method and system based on improved sparse dictionary |
| CN115407162A (en) * | 2022-08-31 | 2022-11-29 | 国网山东省电力公司郯城县供电公司 | Low-current ground fault line selection method and system based on sparse representation |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101275702B1 (en) * | 2012-02-15 | 2013-06-17 | 목포대학교산학협력단 | Method and apparatus for fault diagnosis and fault detection of gearbox using vibration signal |
| CN105241666A (en) * | 2015-09-21 | 2016-01-13 | 华南理工大学 | Rolling bearing fault feature extraction method based on signal sparse representation theory |
| CN109839271A (en) * | 2018-12-29 | 2019-06-04 | 昆明理工大学 | A kind of bearing fault characteristics extracting method based on match tracing Corresponding Sparse Algorithm |
| CN110907162A (en) * | 2019-12-13 | 2020-03-24 | 北京天泽智云科技有限公司 | Rotating machinery fault feature extraction method without tachometer under variable rotating speed |
| CN111222289A (en) * | 2020-01-10 | 2020-06-02 | 燕山大学 | Data-model-driven K-SVD-based rolling bearing impact fault diagnosis method |
-
2020
- 2020-07-27 CN CN202010732900.0A patent/CN111678691A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101275702B1 (en) * | 2012-02-15 | 2013-06-17 | 목포대학교산학협력단 | Method and apparatus for fault diagnosis and fault detection of gearbox using vibration signal |
| CN105241666A (en) * | 2015-09-21 | 2016-01-13 | 华南理工大学 | Rolling bearing fault feature extraction method based on signal sparse representation theory |
| CN109839271A (en) * | 2018-12-29 | 2019-06-04 | 昆明理工大学 | A kind of bearing fault characteristics extracting method based on match tracing Corresponding Sparse Algorithm |
| CN110907162A (en) * | 2019-12-13 | 2020-03-24 | 北京天泽智云科技有限公司 | Rotating machinery fault feature extraction method without tachometer under variable rotating speed |
| CN111222289A (en) * | 2020-01-10 | 2020-06-02 | 燕山大学 | Data-model-driven K-SVD-based rolling bearing impact fault diagnosis method |
Non-Patent Citations (2)
| Title |
|---|
| 罗春艳: "小波字典下齿轮箱复合故障瞬态特征的稀疏表示研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
| 胥永刚 等: "基于双树复小波包变换能量泄漏特性分析的齿轮故障诊断", 《农业工程学报》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112906632A (en) * | 2021-03-18 | 2021-06-04 | 中北大学 | Automatic extraction method of highly adaptive time delay target signal |
| CN112906632B (en) * | 2021-03-18 | 2022-11-11 | 中北大学 | Automatic extraction method of highly adaptive time delay target signal |
| CN114136604A (en) * | 2021-11-30 | 2022-03-04 | 三门核电有限公司 | Rotary equipment fault diagnosis method and system based on improved sparse dictionary |
| CN114136604B (en) * | 2021-11-30 | 2023-11-28 | 三门核电有限公司 | Rotary equipment fault diagnosis method and system based on improved sparse dictionary |
| CN115407162A (en) * | 2022-08-31 | 2022-11-29 | 国网山东省电力公司郯城县供电公司 | Low-current ground fault line selection method and system based on sparse representation |
| CN115407162B (en) * | 2022-08-31 | 2024-01-23 | 国网山东省电力公司郯城县供电公司 | Low-current ground fault line selection method and system based on sparse representation |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Ding et al. | Detection of faults in rotating machinery using periodic time-frequency sparsity | |
| Fan et al. | Sparse representation of transients in wavelet basis and its application in gearbox fault feature extraction | |
| Chen et al. | Compressed sensing based on dictionary learning for extracting impulse components | |
| CN111678691A (en) | Gear fault detection method based on improved sparse decomposition algorithm | |
| Huang et al. | Multi-source fidelity sparse representation via convex optimization for gearbox compound fault diagnosis | |
| Cai et al. | Sparsity-enabled signal decomposition using tunable Q-factor wavelet transform for fault feature extraction of gearbox | |
| Feng et al. | Complex signal analysis for planetary gearbox fault diagnosis via shift invariant dictionary learning | |
| Zhang et al. | Time-varying singular value decomposition for periodic transient identification in bearing fault diagnosis | |
| CN108106830A (en) | A kind of Variable Speed Rotating Machinery method for diagnosing faults based on time-frequency spectrum segmentation | |
| CN104156585B (en) | A kind of doubledictionary match tracing bearing fault degree evaluation method based on many index feature | |
| CN104807534B (en) | Equipment eigentone self study recognition methods based on on-line vibration data | |
| CN111413089A (en) | Gear fault diagnosis method based on combination of VMD entropy method and VPMCD | |
| CN113176092B (en) | Motor bearing fault diagnosis method based on data fusion and improved experience wavelet transform | |
| Dong et al. | A repeated single-channel mechanical signal blind separation method based on morphological filtering and singular value decomposition | |
| Zhang et al. | Sparsity-aware tight frame learning with adaptive subspace recognition for multiple fault diagnosis | |
| CN112284727B (en) | Rotary machine fault diagnosis method based on convolution maximum and minimum penalty algorithm | |
| Li et al. | Transient feature self-enhancement via shift-invariant manifold sparse learning for rolling bearing health diagnosis | |
| CN107966287B (en) | Weak fault feature extraction method for self-adaptive electromechanical equipment | |
| CN112733612A (en) | Cross-domain rotating machinery fault diagnosis model establishing method and application thereof | |
| CN113865872A (en) | Bearing fault diagnosis method based on wavelet packet reconstruction imaging and CNN | |
| CN114136604B (en) | Rotary equipment fault diagnosis method and system based on improved sparse dictionary | |
| CN116086807A (en) | Fault judging method for rolling bearing and related product | |
| CN114441174A (en) | Diagnosis method, system, equipment and medium for composite fault of rolling bearing | |
| Yu et al. | A clustering K-SVD-based sparse representation method for rolling bearing fault diagnosis | |
| CN115130500A (en) | Motor fault diagnosis system and method based on SDAE algorithm |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200918 |