CN115855509A - Data-driven train bearing fault diagnosis method - Google Patents
Data-driven train bearing fault diagnosis method Download PDFInfo
- Publication number
- CN115855509A CN115855509A CN202310167675.4A CN202310167675A CN115855509A CN 115855509 A CN115855509 A CN 115855509A CN 202310167675 A CN202310167675 A CN 202310167675A CN 115855509 A CN115855509 A CN 115855509A
- Authority
- CN
- China
- Prior art keywords
- bearing
- frequency domain
- value
- determining
- lossless
- 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.)
- Granted
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 44
- 238000000034 method Methods 0.000 title claims abstract description 34
- 230000036541 health Effects 0.000 claims abstract description 23
- 239000013598 vector Substances 0.000 claims abstract description 21
- 238000009825 accumulation Methods 0.000 claims abstract description 13
- 230000001186 cumulative effect Effects 0.000 claims description 41
- 230000006378 damage Effects 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 10
- 208000027418 Wounds and injury Diseases 0.000 claims description 4
- 208000014674 injury Diseases 0.000 claims description 4
- 230000017105 transposition Effects 0.000 claims description 4
- 230000003862 health status Effects 0.000 claims description 3
- 238000002405 diagnostic procedure Methods 0.000 claims 3
- 238000004364 calculation method Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 3
- 230000007246 mechanism Effects 0.000 abstract description 3
- 238000007781 pre-processing Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 238000013499 data model Methods 0.000 description 2
- 230000006735 deficit Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- 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
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a data-driven train bearing fault diagnosis method, which comprises the following steps: acquiring an acoustic signal of a bearing to be tested; determining a frequency domain accumulation value according to the acoustic signal; determining a normalized logarithmic Bayes factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model; wherein the nonparametric probabilistic regression model is established based on the correlation vector machine and the acoustic signals of the lossless bearing; and determining the diagnosis result of the bearing to be detected according to the normalized logarithmic Bayesian factor. The diagnosis method provided by the invention has the advantages that the data-driven sparse model about the accumulated frequency domain value is established through the correlation vector machine, the model is simple, the calculation is rapid, and the health state of the bearing can be monitored in real time. The frequency domain accumulated value is proposed for modeling for the first time, the mechanism is simple, and the effect is directly obvious.
Description
Technical Field
The invention relates to the technical field of fault diagnosis, in particular to a data-driven train bearing fault diagnosis method.
Background
Since the speed is much faster than that of a general train, the safety of a high-speed railway is required to be higher. The rolling bearing is used as a key component of a high-speed railway train and plays an important role in the safe operation of the train. An important part for supporting the mechanical rotating body is a portion which is extremely likely to fail. The method is limited by installation and operation conditions of various trains, acceleration or sensors cannot be installed at all parts with bearing components, and fault diagnosis of the train bearings is difficult in the prior art.
Accordingly, there is a need for improvements and developments in the art.
Disclosure of Invention
The invention aims to solve the technical problem that the fault diagnosis of the train bearing driven by data is difficult in the prior art.
The technical scheme adopted by the invention for solving the technical problem is as follows:
a data-driven train bearing fault diagnosis method comprises the following steps:
acquiring an acoustic signal of a bearing to be tested;
determining a frequency domain cumulative value according to the acoustic signal;
determining a normalized logarithmic Bayes factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model; wherein the nonparametric probabilistic regression model is established based on the correlation vector machine and the acoustic signals of the lossless bearing;
and determining the diagnosis result of the bearing to be detected according to the normalized logarithm Bayesian factor.
The method for diagnosing the fault of the train bearing driven by the data, wherein the step of determining the frequency domain accumulated value according to the acoustic signal comprises the following steps:
carrying out Fourier transform on the acoustic signal to obtain a frequency domain value of the acoustic signal;
and determining a frequency domain accumulated value according to the frequency domain value.
The data-driven train bearing fault diagnosis method comprises the following steps of:
wherein ,
representnFrequency domain cumulative value of several frequency points, <' >>
Representing the second in the acoustic signaliThe frequency domain amplitude, Σ, for each frequency bin represents the summation.
The method for diagnosing the fault of the train bearing driven by the data, wherein the nonparametric probabilistic regression model comprises the following steps: the distribution model of the health hypothesis and the distribution model of the damage hypothesis are obtained by establishing the following steps:
acquiring acoustic signals of a plurality of lossless bearings;
determining a frequency domain cumulative value of the lossless bearing according to the acoustic signal of the lossless bearing;
determining a distribution model of a health hypothesis according to frequency domain accumulated values of a plurality of lossless bearings based on a correlation vector machine;
determining a distribution model of the damage hypothesis based on the distribution model of the health hypothesis.
The data-driven train bearing fault diagnosis method comprises the following steps of:
the distribution model of the damage hypothesis is:
wherein ,
Indicates a health hypothesis>
Represents a normal distribution, is>
Represents the mean value of the frequency-domain cumulative value of the lossless bearing, < > is >>
Standard deviation, representing a frequency domain cumulative value for a lossless bearing>
Indicates a damage hypothesis, is>
Representing an offset factor determined from a standard deviation of the frequency domain cumulative values of the lossless bearing.
The data-driven train bearing fault diagnosis method comprises the following steps of:
wherein ,
represents a normalized logarithmic Bayes factor,. Sup.>
Represents the number of frequency data points, <' > based on the number of the data points>
Represents the transposition of the vector>
Representing a frequency domain cumulative value>
And the mean value of the frequency-domain cumulative value of the lossless bearing->
The difference between, representing the frequency domain cumulative value->
And the mean value of the frequency-domain cumulative value of the lossless bearing->
Based on the offset factor>
The difference between them.
The data-driven train bearing fault diagnosis method comprises the following steps of: health status and injury status; determining a diagnosis result of the bearing to be detected according to the normalized logarithm Bayesian factor, wherein the step of determining the diagnosis result comprises the following steps:
when the normalized logarithm Bayes factor is a positive value, the bearing to be tested is in a healthy state;
and when the normalized logarithmic Bayes factor is a negative value, the bearing to be detected is in a damaged state.
A data driven train bearing fault diagnostic system, comprising:
the acquisition module is used for acquiring an acoustic signal of the bearing to be detected;
the frequency domain accumulation module is used for determining a frequency domain accumulation value according to the acoustic signal;
the factor module is used for determining a normalized logarithmic Bayesian factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model;
and the diagnosis module is used for determining the diagnosis result of the bearing to be detected according to the normalized logarithmic Bayesian factor.
A computer device comprising a memory storing a computer program and a processor, wherein the processor implements the steps of the method as claimed in any one of the above when executing the computer program.
A computer-readable storage medium, on which a computer program is stored, wherein the computer program, when being executed by a processor, realizes the steps of the method as set forth in any of the above.
Has the advantages that: the diagnosis method provided by the invention has the advantages that the data-driven sparse model about the accumulated frequency domain value is established through the correlation vector machine, the model is simple, the calculation is rapid, and the health state of the bearing can be monitored in real time. The frequency domain accumulated value is proposed for modeling for the first time, the mechanism is simple, and the effect is directly obvious.
Drawings
Fig. 1 is a time chart of an acoustic signal of a microphone pickup bearing in an embodiment of the present invention.
Fig. 2 is a time domain diagram of a preprocessed acoustic signal in an embodiment of the invention.
Fig. 3 is a frequency domain plot of a preprocessed acoustic signal in an embodiment of the invention.
Fig. 4 is a diagram of a plurality of frequency domain cumulative value sample sets in an embodiment of the present invention.
FIG. 5 is a correlation vector machine model in an embodiment of the invention.
FIG. 6 is a normalized Bayesian factor graph for a bearing in a healthy state in an embodiment of the present invention.
FIG. 7 is a normalized Bayesian factor graph of a bearing under a damaged condition in an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1 to 7, the present invention provides some embodiments of a data-driven train bearing fault diagnosis method.
The data-driven train bearing fault diagnosis method provided by the embodiment of the invention comprises the following steps of:
and S100, acquiring an acoustic signal of the bearing to be measured.
Specifically, the original acoustic signal of the bearing under test collected by the microphone is shown in fig. 1. The acquired original acoustic signal contains a plurality of noise components irrelevant to the bearing, so the original acoustic signal is required to be preprocessed firstly, a common bearing preprocessing method such as an autoregressive model obtains discrete signals such as a shaft or a gear, the discrete signals are removed from the original signal to serve as a first preprocessing step, envelope analysis preprocessing is carried out in a second step, although fault diagnosis can be carried out primarily after the envelope analysis preprocessing, whether an envelope frequency spectrum is matched with fault frequency needs to be observed, the fault diagnosis intuitiveness is not obvious under certain working conditions, and the bearing damage degree cannot be quantized. The acoustic signal of the bearing to be tested is obtained by preprocessing the original acoustic signal of the bearing to be tested, as shown in fig. 2.
And S200, determining a frequency domain accumulation value according to the acoustic signal.
Specifically, a frequency domain cumulative value of the bearing to be tested is determined according to an acoustic signal of the bearing to be tested, the frequency domain amplitude trend of the acoustic signal is large, smoothness is poor, and direct modeling is difficult. For this reason, the acoustic signals are converted into frequency domain accumulated values, and the consistency of data trends is increased. The frequency domain cumulative value becomes a monotonically increasing curve, which is easy to model.
Step S200 specifically includes:
step S210, carrying out Fourier transform on the acoustic signal to obtain a frequency domain value of the acoustic signal.
And step S220, determining a frequency domain accumulation value according to the frequency domain value.
Specifically, a Fourier transform is adopted to transform the acoustic signal of the bearing to be tested into the frequency domain value of the acoustic signal. And then obtaining a frequency domain accumulated value according to the frequency domain value. As shown in fig. 3, after fourier transform, the frequency domain values of the acoustic signals are obtained, and the range of the frequency points is 0-1000Hz. And accumulating the amplitudes of all the frequency points to obtain a frequency domain accumulated value. As shown in fig. 4, the frequency domain cumulative value gradually increases as the frequency of the frequency point increases, and modeling is easy since the frequency domain cumulative value becomes a monotonously increasing curve.
The frequency domain cumulative value is:
wherein ,
to representnFrequency-domain cumulative value for frequency points>
Representing the second in the acoustic signaliThe frequency domain amplitude, Σ, for each frequency bin represents the summation.
Step S300, based on a nonparametric probability regression model, determining a normalized logarithmic Bayes factor according to the frequency domain cumulative value; wherein the nonparametric probabilistic regression model is established based on the correlation vector machine and the acoustic signals of the lossless bearing.
Specifically, a nonparametric probability regression model is established through acoustic signals of the lossless bearing, and a normalized logarithmic Bayes factor of the bearing to be tested is obtained according to the frequency domain accumulated value of the bearing to be tested based on the nonparametric probability regression model.
The normalized logarithm Bayesian factor is as follows:
wherein ,
represents a normalized logarithmic Bayes factor,. Sup.>
Represents the number of frequency data points, <' > based on the number of the data points>
Represents the transposition of the vector>
Standard deviation ^ representing the frequency domain cumulative value of a lossless bearing>
Represents a frequency-domain cumulative value->
And the mean value of the frequency-domain cumulative value of the lossless bearing->
The difference between, representing the frequency domain cumulative value->
And the mean value of the frequency-domain cumulative value of the lossless bearing->
Based on the offset factor>
The difference between them. Because the nonparametric probability regression model is established based on the relevance vector machine, only a few frequency data points can be deduced to be relevance vectors, so that the model expression is very simplified, and the calculation of the normalized logarithm Bayesian factor is simpler.
After the nonparametric probability regression model is established, the mean value of the frequency domain accumulated values of the lossless bearing can be obtained
Based on a deviation factor>
And the standard deviation of the frequency-domain cumulative value of the lossless bearing->
A normalized logarithmic bayesian factor can then be calculated.
And S400, determining a diagnosis result of the bearing to be detected according to the normalized logarithmic Bayesian factor.
Specifically, after the normalized logarithmic Bayes factor is obtained, the state of the bearing to be tested is judged according to the normalized logarithmic Bayes factor, and a diagnosis result is obtained. The diagnosis result comprises: health status and injury status; when the bearing to be tested is in a healthy state, the bearing to be tested is not damaged; when the bearing to be detected is in a damaged state, the bearing to be detected is damaged, and the damaged degree of the bearing to be detected can be quantized according to the magnitude of the normalized logarithmic Bayes factor.
Step S400 specifically includes:
and S410, when the normalized logarithmic Bayes factor is a positive value, the bearing to be tested is in a healthy state.
And step S420, when the normalized logarithm Bayes factor is a negative value, the bearing to be detected is in a damage state.
Specifically, when the normalized logarithmic Bayes factor is a positive value, the bearing to be tested is in a healthy state; and when the normalized logarithmic Bayes factor is a negative value, the bearing to be detected is in a damaged state.
The nonparametric probabilistic regression model includes: the distribution model of the health hypothesis and the distribution model of the damage hypothesis are obtained by establishing the following steps:
step A100, acoustic signals of a plurality of lossless bearings are obtained.
Step A200, determining a frequency domain accumulated value of the lossless bearing according to the acoustic signal of the lossless bearing.
And A300, determining a distribution model of the health hypothesis according to the frequency domain accumulated values of a plurality of lossless bearings based on a correlation vector machine.
And A400, determining a distribution model of the damage hypothesis according to the distribution model of the health hypothesis.
Specifically, steps a100 and a200 are similar to steps S100 and S200, respectively, and in steps a100 and a200, the original acoustic signals of the lossless bearing are collected by the microphone and are preprocessed to obtain the acoustic signals of the lossless bearing. Then, according to the acoustic signals of the lossless bearings, frequency domain accumulated values of the lossless bearings are determined. Specifically, fourier transform is carried out on the acoustic signal of the lossless bearing to obtain a frequency domain value of the acoustic signal of the lossless bearing; and determining a frequency domain accumulated value of the lossless bearing according to the frequency domain value of the acoustic signal of the lossless bearing. And acquiring acoustic signals of a plurality of lossless bearings, wherein the acoustic signals can be the acoustic signals of the same lossless bearing or the acoustic signals of different lossless bearings. Finally, the frequency domain accumulated values of the plurality of lossless bearings can be obtained, as shown in fig. 4, a plurality of monotonically increasing curve samples are obtained, the curves represent the trend of the frequency domain accumulated values in the healthy state, and the frequency domain accumulated values include the frequency domain information of the bearings in the healthy state. By means of the data, a complex physical model can be abandoned, and a simple data model can be directly established. It can be seen that the frequency domain cumulative value samples in the same state are not fixed values, but obey a certain distribution, and therefore need to be regarded as random variables and subjected to probability modeling.
And a non-parametric probability model is established by selecting a relevant vector machine, so that the method has the advantages of repeated exertion of the sparse characteristics of the data model, simple model parameters and robustness. Under the correlation vector machine framework, the above-mentioned frequency domain cumulative value sample set can be represented by the following mathematical model:
wherein ,
represents an expectation of a sample set regression model>
Then it is at each frequency data point pick>
Radial basis kernel function as mean:
wherein ,
represents the width of the kernel, which is a kind of hyper-parameter. And/or>
Representing the weight of each kernel function. />
Representing noise, which is used to characterize the uncertainty of the model. Most of the->
The weight of (c) will approach zero and thus be ignored by the model, and only a small number of data points will be derived as correlation vectors, thus making the model representation very simplified.
，
Where T denotes transposition.
Fig. 5 shows the trained correlation vector machine model, and it can be found that the weight of the kernel function of only 6 data points is not zero, so that the expression of the model is robust and concise.
Experiments show that the frequency domain accumulated value of the acoustic signals of the damaged bearing can deviate from a relevant vector machine model established in a lossless state, and the bearing abnormity can be judged according to the deviation. To quantify the offset, we use a bayesian factor. Suppose that two states exist for a bearing, namely a health hypothesis
And the injury hypothesis>
. A distribution model of health assumptions is shown in FIG. 5 and is labeled @>
. And the impairment hypothesis->
Is set to the mean of the health hypothesis plus the offset factor @>
Distribution model of impairment hypotheses, scored >>
。/>
Represents a normal distribution, is>
Represents the mean value of the frequency-domain cumulative value of the lossless bearing, < > is >>
A standard deviation representing a frequency domain cumulative value for a lossless bearing, the offset factor being determined from the standard deviation of the frequency domain cumulative value for the lossless bearing. It is recommended that the offset factor is defined as three times the standard deviation (` based on `) of the healthy model>
) This is also the range of the model 95% confidence interval.
After the frequency domain accumulated value of the bearing to be detected with unknown health state exists, the deviation degree of the bearing to be detected with the health state regression model can be judged according to the Bayesian factor, and whether the bearing is abnormal or damaged is judged.
wherein ,
and->
. By taking the logarithm, anDivided by each->
Number of data points containedNObtaining a normalized logarithm Bayesian factor:
after the logarithm condition is taken, when the normalized logarithm Bayes factor is positive, the new data are more consistent with the distribution of the health hypothesis, and the bearing state is normal, namely in a healthy state; otherwise, the new data is more consistent with the distribution of the damage hypothesis, and the bearing state is abnormal, namely in a damaged state. When the normalized logarithmic Bayes factor is negative, different values of the normalized logarithmic Bayes factor represent different damage degrees, and the smaller the value, the more the data collected from the bearing deviates from the normal state, the larger the damage is. Fig. 6 and 7 show the health and damage of the bearing, respectively, and count the values of the bayesian factors obtained from the frequency domain accumulation values of the conversion of several acoustic signals, wherein 10 samples of the health state are not samples for establishing the model, but are newly collected. It can be seen that the value of the normalized bayes factor in the healthy state is positive, while the values of the 15 samples in the damaged state are negative. The potential of this method in identifying bearing damage can thus be seen.
Compared with the traditional method based on acceleration or other sensors, the diagnosis method provided by the invention has the advantages that the data driving sparse model related to the accumulated frequency domain value is established through the relevant vector machine, the model is simple, the calculation is rapid, and the health state of the bearing can be monitored in real time. The frequency domain accumulated value is proposed for modeling for the first time, the mechanism is simple, and the effect is directly obvious. The acoustic signals can be collected through the simple microphone, the microphone array can be simply installed and detached at any time according to needs, and the microphone array is high in efficiency, low in cost and convenient to assemble and disassemble.
Based on the data-driven train bearing fault diagnosis method described in any of the above embodiments, the present invention further provides a preferred embodiment of a data-driven train bearing fault diagnosis system:
the data-driven train bearing fault diagnosis system of the invention comprises:
the acquisition module is used for acquiring an acoustic signal of the bearing to be detected;
the frequency domain accumulation module is used for determining a frequency domain accumulation value according to the acoustic signal;
the factor module is used for determining a normalized logarithmic Bayesian factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model;
and the diagnosis module is used for determining the diagnosis result of the bearing to be detected according to the normalized logarithmic Bayesian factor.
The frequency domain accumulation module is specifically used for performing Fourier transform on the acoustic signal to obtain a frequency domain value of the acoustic signal; and determining a frequency domain accumulation value according to the frequency domain value.
The diagnosis module is specifically used for judging that the bearing to be detected is in a healthy state when the normalized logarithmic Bayesian factor is a positive value; and when the normalized logarithm Bayesian factor is a negative value, the bearing to be detected is in a damage state.
Based on the data-driven train bearing fault diagnosis method described in any one of the above embodiments, the present invention further provides an embodiment of a computer device:
the computer device of the present invention comprises a memory storing a computer program and a processor implementing the steps of the method according to any one of the above embodiments when the processor executes the computer program.
Based on the data-driven train bearing fault diagnosis method described in any one of the above embodiments, the present invention further provides an embodiment of a computer-readable storage medium:
the computer-readable storage medium of the present invention has stored thereon a computer program which, when being executed by a processor, carries out the steps of the method according to any one of the above-mentioned embodiments.
It will be understood that the invention is not limited to the examples described above, but that modifications and variations will occur to those skilled in the art in light of the above teachings, and that all such modifications and variations are considered to be within the scope of the invention as defined by the appended claims.
Claims (10)
1. A data-driven train bearing fault diagnosis method is characterized by comprising the following steps:
acquiring an acoustic signal of a bearing to be tested;
determining a frequency domain cumulative value according to the acoustic signal;
determining a normalized logarithmic Bayes factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model; wherein the nonparametric probabilistic regression model is established based on the correlation vector machine and the acoustic signals of the lossless bearing;
and determining the diagnosis result of the bearing to be detected according to the normalized logarithmic Bayesian factor.
2. The data driven train bearing fault diagnostic method of claim 1, wherein the determining a frequency domain cumulative value from the acoustic signal comprises:
carrying out Fourier transform on the acoustic signal to obtain a frequency domain value of the acoustic signal;
and determining a frequency domain accumulation value according to the frequency domain value.
3. The data-driven train bearing fault diagnosis method according to claim 2, wherein the frequency domain cumulative value is:
wherein ,
to representnFrequency domain cumulative value of several frequency points, <' >>
Representing the second in the acoustic signaliThe frequency domain amplitude, Σ, for each frequency bin represents the summation.
4. The data driven train bearing fault diagnosis method according to any one of claims 1 to 3, wherein the non-parametric probabilistic regression model comprises: the distribution model of the health hypothesis and the distribution model of the damage hypothesis are obtained by establishing the following steps:
acquiring acoustic signals of a plurality of lossless bearings;
determining a frequency domain cumulative value of the lossless bearing according to the acoustic signal of the lossless bearing;
determining a distribution model of a health hypothesis according to frequency domain accumulated values of a plurality of lossless bearings based on a correlation vector machine;
determining a distribution model of the damage hypothesis based on the distribution model of the health hypothesis.
5. The data driven train bearing fault diagnostic method of claim 4, wherein the distribution model of the health hypothesis is:
the distribution model of the damage hypothesis is:
wherein ,
indicates a health hypothesis>
Represents a normal distribution, is>
Represents the mean of the frequency domain cumulative values for the lossless bearing,
standard deviation representing a frequency domain cumulative value for a lossless bearing>
Indicates a damage hypothesis, is>
Representing an offset factor determined from a standard deviation of the frequency domain cumulative values of the lossless bearing.
6. The data-driven train bearing fault diagnosis method according to claim 5, wherein the normalized log Bayesian factor is:
wherein ,
represents a normalized logarithmic Bayes factor,. Sup.>
Represents the number of frequency data points, <' > based on the number of the data points>
Represents the transposition of the vector>
Represents a frequency-domain cumulative value->
And the mean value of the frequency-domain cumulative value of the lossless bearing->
The difference between->
Represents a frequency-domain cumulative value->
And the mean value of the frequency-domain cumulative value of the lossless bearing->
Based on the offset factor>
The difference between them.
7. The data driven train bearing fault diagnostic method of claim 1, wherein the diagnostic result comprises: health status and injury status; the determining the diagnosis result of the bearing to be detected according to the normalized logarithm Bayesian factor comprises the following steps:
when the normalized logarithm Bayes factor is a positive value, the bearing to be tested is in a healthy state;
and when the normalized logarithm Bayesian factor is a negative value, the bearing to be detected is in a damage state.
8. A data driven train bearing fault diagnostic system, comprising:
the acquisition module is used for acquiring an acoustic signal of the bearing to be detected;
the frequency domain accumulation module is used for determining a frequency domain accumulation value according to the acoustic signal;
the factor module is used for determining a normalized logarithmic Bayesian factor according to the frequency domain accumulated value based on a nonparametric probabilistic regression model;
and the diagnosis module is used for determining the diagnosis result of the bearing to be detected according to the normalized logarithm Bayesian factor.
9. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310167675.4A CN115855509B (en) | 2023-02-27 | 2023-02-27 | Data-driven train bearing fault diagnosis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310167675.4A CN115855509B (en) | 2023-02-27 | 2023-02-27 | Data-driven train bearing fault diagnosis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115855509A true CN115855509A (en) | 2023-03-28 |
| CN115855509B CN115855509B (en) | 2023-06-16 |
Family
ID=85659044
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310167675.4A Active CN115855509B (en) | 2023-02-27 | 2023-02-27 | Data-driven train bearing fault diagnosis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115855509B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2402731A1 (en) * | 2010-07-01 | 2012-01-04 | Schaeffler Technologies AG & Co. KG | Method for training a system for classifying a roller bearing state, method for classifying a roller bearing state and system for classifying a roller bearing state |
| CN204666087U (en) * | 2015-04-02 | 2015-09-23 | 香港理工大学深圳研究院 | Based on novel vibrating measuring table and the proving installation of accurate zero stiffness theory |
| US20170003352A1 (en) * | 2014-01-03 | 2017-01-05 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method, device and system for estimating the state of health of a battery in an electric or hybrid vehicle during operation thereof, and method for creating model for estimation of said type |
| CN107677472A (en) * | 2017-08-25 | 2018-02-09 | 中国矿业大学 | The bearing state noise diagnostics algorithm that network-oriented Variable Selection merges with Characteristic Entropy |
| CN109829402A (en) * | 2019-01-21 | 2019-05-31 | 福州大学 | Different operating condition lower bearing degree of injury diagnostic methods based on GS-SVM |
| CN112229633A (en) * | 2020-09-03 | 2021-01-15 | 中国长江三峡集团有限公司福建分公司 | Fan bearing fault diagnosis method based on multivariate feature fusion |
| US10896384B1 (en) * | 2017-04-28 | 2021-01-19 | Microsoft Technology Licensing, Llc | Modification of base distance representation using dynamic objective |
| CN112528564A (en) * | 2020-12-08 | 2021-03-19 | 中国矿业大学 | Bridge structure damage identification method based on sparse Bayesian learning |
| CN113959719A (en) * | 2021-09-30 | 2022-01-21 | 桂林电子科技大学 | SOT-SVM-based variable working condition rolling bearing fault diagnosis method |
| US20220283576A1 (en) * | 2021-03-03 | 2022-09-08 | Aktiebolaget Skf | Automatic diagnosis method, system and storage medium for equipment |
| CN115221930A (en) * | 2022-09-20 | 2022-10-21 | 苏州鸿哲智能科技有限公司 | Fault diagnosis method for rolling bearing |
-
2023
- 2023-02-27 CN CN202310167675.4A patent/CN115855509B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2402731A1 (en) * | 2010-07-01 | 2012-01-04 | Schaeffler Technologies AG & Co. KG | Method for training a system for classifying a roller bearing state, method for classifying a roller bearing state and system for classifying a roller bearing state |
| US20170003352A1 (en) * | 2014-01-03 | 2017-01-05 | Commissariat A L'energie Atomique Et Aux Energies Alternatives | Method, device and system for estimating the state of health of a battery in an electric or hybrid vehicle during operation thereof, and method for creating model for estimation of said type |
| CN204666087U (en) * | 2015-04-02 | 2015-09-23 | 香港理工大学深圳研究院 | Based on novel vibrating measuring table and the proving installation of accurate zero stiffness theory |
| US10896384B1 (en) * | 2017-04-28 | 2021-01-19 | Microsoft Technology Licensing, Llc | Modification of base distance representation using dynamic objective |
| CN107677472A (en) * | 2017-08-25 | 2018-02-09 | 中国矿业大学 | The bearing state noise diagnostics algorithm that network-oriented Variable Selection merges with Characteristic Entropy |
| CN109829402A (en) * | 2019-01-21 | 2019-05-31 | 福州大学 | Different operating condition lower bearing degree of injury diagnostic methods based on GS-SVM |
| CN112229633A (en) * | 2020-09-03 | 2021-01-15 | 中国长江三峡集团有限公司福建分公司 | Fan bearing fault diagnosis method based on multivariate feature fusion |
| CN112528564A (en) * | 2020-12-08 | 2021-03-19 | 中国矿业大学 | Bridge structure damage identification method based on sparse Bayesian learning |
| US20220283576A1 (en) * | 2021-03-03 | 2022-09-08 | Aktiebolaget Skf | Automatic diagnosis method, system and storage medium for equipment |
| CN113959719A (en) * | 2021-09-30 | 2022-01-21 | 桂林电子科技大学 | SOT-SVM-based variable working condition rolling bearing fault diagnosis method |
| CN115221930A (en) * | 2022-09-20 | 2022-10-21 | 苏州鸿哲智能科技有限公司 | Fault diagnosis method for rolling bearing |
Non-Patent Citations (1)
| Title |
|---|
| 毛文贵等: ""遗传智能采样技术的贝叶斯理论识别滑动轴承一 转子系统不平衡量"", 《振动工程学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115855509B (en) | 2023-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Intelligent fault diagnosis of roller bearings with multivariable ensemble-based incremental support vector machine | |
| Li et al. | Fault diagnosis of rotating machinery with a novel statistical feature extraction and evaluation method | |
| Brkovic et al. | Early fault detection and diagnosis in bearings for more efficient operation of rotating machinery | |
| Yan et al. | An efficient approach to machine health diagnosis based on harmonic wavelet packet transform | |
| Purushotham et al. | Multi-fault diagnosis of rolling bearing elements using wavelet analysis and hidden Markov model based fault recognition | |
| Jiang et al. | Study on Hankel matrix-based SVD and its application in rolling element bearing fault diagnosis | |
| CN108760266B (en) | The virtual degeneration index building method of mechanical key component based on learning distance metric | |
| Ma et al. | GNAR-GARCH model and its application in feature extraction for rolling bearing fault diagnosis | |
| Omoregbee et al. | Fault detection in roller bearing operating at low speed and varying loads using Bayesian robust new hidden Markov model | |
| Attoui et al. | Novel machinery monitoring strategy based on time–frequency domain similarity measurement with limited labeled data | |
| Lu et al. | CEEMD-assisted bearing degradation assessment using tight clustering | |
| Susilo et al. | Fault diagnosis of roller bearing using parameter evaluation technique and multi-class support vector machine | |
| CN111076934A (en) | Method for diagnosing potential fault of bearing based on S transformation | |
| JP7188143B2 (en) | Anomaly sign detection system, anomaly sign detection method | |
| Liu et al. | Transfer learning with inception ResNet-based model for rolling bearing fault diagnosis | |
| Sadoughi et al. | A physics-based deep learning approach for fault diagnosis of rotating machinery | |
| Zaeri et al. | Artificial neural network based fault diagnostics of rolling element bearings using continuous wavelet transform | |
| Gundewar et al. | Rolling element bearing fault diagnosis using supervised learning methods-artificial neural network and discriminant classifier | |
| CN115855509A (en) | Data-driven train bearing fault diagnosis method | |
| CN108051213A (en) | A kind of online fault detection method of rolling bearing | |
| CN114186593B (en) | Bearing fault identification method, device, equipment and storage medium under noise condition | |
| CN115860708A (en) | Ship equipment state monitoring and fault diagnosis method | |
| Kang et al. | Gearbox fault prognosis based on CHMM and SVM | |
| CN116451105A (en) | Wind turbine generator system gearbox fault diagnosis method based on ICEEMDAN-MPE-RF and SVM | |
| Kang et al. | A class incremental learning approach based on autoencoder without manual feature extraction for rail vehicle fault diagnosis |
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 |