CN109100144B - Automobile hub bearing fault feature extraction method based on optimal quality factor selection - Google Patents

Automobile hub bearing fault feature extraction method based on optimal quality factor selection Download PDF

Info

Publication number
CN109100144B
CN109100144B CN201810864229.8A CN201810864229A CN109100144B CN 109100144 B CN109100144 B CN 109100144B CN 201810864229 A CN201810864229 A CN 201810864229A CN 109100144 B CN109100144 B CN 109100144B
Authority
CN
China
Prior art keywords
quality factor
low
resonance
automobile hub
hub bearing
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
Application number
CN201810864229.8A
Other languages
Chinese (zh)
Other versions
CN109100144A (en
Inventor
李仲兴
周庄
薛红涛
江洪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangyin Intellectual Property Operation Co., Ltd
Original Assignee
Jiangsu University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu University filed Critical Jiangsu University
Priority to CN201810864229.8A priority Critical patent/CN109100144B/en
Publication of CN109100144A publication Critical patent/CN109100144A/en
Application granted granted Critical
Publication of CN109100144B publication Critical patent/CN109100144B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction

Landscapes

  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • General Physics & Mathematics (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)

Abstract

The invention discloses an automobile hub bearing fault feature extraction method based on optimal quality factor selection, which comprises the steps of firstly collecting vibration signals, initializing resonance sparse decomposition parameters, then obtaining an optimal quality factor by utilizing a successive optimization algorithm and taking an RSK index as a target function, and finally carrying out envelope analysis on low resonance components obtained by carrying out resonance sparse decomposition under the optimal quality factor of the signals to obtain an envelope spectrum, thereby effectively extracting fault features; the method solves the problems that in the traditional resonance sparse decomposition method, the quality factor is large in randomness due to manual selection, uncertainty is caused, and an ideal decomposition effect is difficult to obtain, can adaptively select the optimal quality factor, and can effectively extract the fault characteristics of the automobile hub bearing under the noise of intermittent strong interference.

Description

Automobile hub bearing fault feature extraction method based on optimal quality factor selection
Technical Field
The invention belongs to the field of automobile hub bearing fault diagnosis, and particularly relates to an automobile hub bearing fault feature extraction method based on optimal quality factor selection of resonance sparse decomposition.
Background
The automobile hub bearing is one of important parts for automobile transmission and bearing, bears axial load and radial load, and the performance of the automobile hub bearing can directly influence the driving safety of an automobile and the riding comfort of passengers. Because the driving working conditions of the automobile are complex and changeable, the hub bearing is often in a working environment with high load, frequent speed change and load change, mechanical faults such as local abrasion and the like are easily induced, further, the hub is damaged, and the direction of the automobile is out of control in the driving process in serious cases. The main reason for the failure of the hub bearing is that the inner ring, the outer ring and the rolling bodies are damaged, so that abnormal vibration response is generated in the operation process. When the hub bearing rotates to pass through a damage position, a periodic transient impact acting force can be generated in the vibration signal, and the frequency corresponding to the impact is the fault characteristic frequency. Therefore, the fault diagnosis of the automobile hub bearing is realized, and the method has important significance for the life and property safety of passengers and the road traffic safety.
At present, the commonly used bearing fault feature extraction methods mainly comprise Fourier transform, wavelet transform, statistical filtering and the like. However, the operation condition of the automobile hub bearing is complex and changeable, and the fault characteristics are difficult to be completely extracted by the traditional methods such as Fourier transform under the condition of intermittent strong interference noise. To better implement bearing fault diagnosis, Selessnick, IW, university in New York, 2011, proposed a resonance sparse decomposition and adjustable quality factor wavelet transform in a paper entitled "sparse representation using the tunable Q-factor wavelet transform" published by Conference Conference on Wavelets and Spectrity XIV. However, the quality factor is set by people for experience, and the randomness is too large, so that the fault feature extraction cannot be accurately realized. In the document CN201210515071.6, a bearing fault diagnosis method based on a composite Q-factor basis algorithm is proposed, in which the normal vibration component and the fault impact component of the bearing are matched with the high-Q and low-Q factor bases, respectively. However, the high-Q and low-Q factor bases are too random to achieve optimal decomposition by manual selection through human experience. For example, in the document CN201610916137.0, a wind turbine gearbox fault diagnosis method based on the adaptive resonance sparse decomposition theory is proposed, and the quality factor and the scale coefficient of the resonance sparse decomposition are simultaneously optimized by using a genetic algorithm to obtain an optimal parameter matrix, so as to realize the resonance sparse decomposition of a fault signal. However, the fitness function of the genetic algorithm cannot accurately evaluate the signal impact component and the normal component, and the selection of genetic algebra and population number of the genetic algorithm causes the calculation to be too complicated.
Disclosure of Invention
Aiming at the problem that the selection randomness of the quality factors of the conventional resonance sparse decomposition is high, the invention provides a method for extracting the fault characteristics of an automobile hub bearing based on the selection of the optimal quality factors of the resonance sparse decomposition.
The invention is realized by the following technical scheme: the method comprises the following steps:
the method comprises the following steps: acquiring a vibration signal x of an automobile hub bearing;
step two: setting initial resonance sparse decomposition parameters, high quality factor Q h3, redundancy rhNumber of decomposition layers J3h30; low quality factor Ql1, redundancy rlNumber of decomposition layers J3l＝11；
Step three: successively optimizing high-quality factors and low-quality factors to obtain the best high resonance component xh *And an optimum low resonance component xl *；
The successive optimization method comprises the following implementation steps:
step A: maintaining a high quality factor QhConstant 3, low quality factor Qlp1+0.1(p-1), p being an argument for a low quality factor, p being 1;
and B: by a high quality factor QhAnd a low quality factor QlpRunning resonance sparse decomposition to obtain high-quality factor QhTransformed basis functions DhAnd a low quality factor QlpTransformed basis functions DlpConstructing an optimized objective function
Figure GDA0002387310610000021
whp、wlpRespectively representing basis functions DhAnd DlpThe transform coefficients of (a); lambda [ alpha ]hAnd λlIs a regularization parameter; iterative calculation is carried out on the target function by adopting a splitting and amplifying Lagrange contraction algorithm, and the high resonance component and the low resonance component are respectively obtained as follows:
Figure GDA0002387310610000022
whp *、wlp *are respectively L (w)hp,wlp) A high resonance transformation coefficient and a low resonance transformation coefficient corresponding to the minimum;
and C: high resonance component xhpAnd a low resonance component xlpSubstituting the objective function
Figure GDA0002387310610000023
μ and σ are each xlpMean and standard deviation of；
Step D: judging whether p is greater than 20, if not, adding 1 to p and returning to the step B; if yes, continuing the step E:
step E: finding RSKpMinimum value of (3) and corresponding QlpThen Q is corresponded tolpFor the best low quality factor Ql *；
Step F: maintaining an optimally low quality factor Ql *Constant, high quality factor QhqQ is 3+0.1(q-1), q is an argument for a high quality factor, q is 1;
step G: by a high quality factor QhqAnd an optimal low quality factor Ql *Running resonance sparse decomposition to obtain high resonance component xhqAnd a low resonance component xlq；
Step H: high resonance component xhqAnd a low resonance component xlqSubstituting the objective function
Figure GDA0002387310610000031
μ and σ are each xlqMean and standard deviation of;
step I: judging whether q is larger than 20, if not, adding 1 to q and returning to the step G, and if so, continuing the step J;
step J: finding RSKqMinimum value of (3) and corresponding QhqWill correspond to QhqTo achieve the best high quality factor Qh *；
Step K: with an optimum high quality factor Qh *And an optimal low quality factor Ql *Running resonance sparse decomposition to obtain optimal high resonance component xh *And a low resonance component xl *；
Step four: for the best low resonance component xl *Performing envelope demodulation to obtain envelope spectrum xb；
Step five: extracting envelope spectra xbCharacteristic frequency f inoAnd frequency doubling to obtain the fault characteristics of the automobile hub bearing.
The invention has the beneficial effects that:
1. according to the method, the RSK target function is constructed, the RSK index is taken as the target function, the optimal high-quality factor and the optimal low-quality factor are obtained through a successive optimization algorithm, the problems that the quality factors are large in randomness due to manual selection, uncertainty is caused, and an ideal decomposition effect is difficult to obtain in the traditional resonance sparse decomposition method are solved, the optimal quality factors can be selected in a self-adaptive mode, and therefore the automobile hub bearing fault characteristics are extracted.
2. According to the invention, the fault signal of the automobile hub bearing with the optimal quality factor is decomposed, so that the fault characteristics of the automobile hub bearing under the noise of intermittent strong interference can be effectively extracted.
Drawings
Fig. 1 is a hardware structure diagram adopted by the method for selecting and extracting the automobile hub bearing fault characteristics based on the optimal quality factors.
FIG. 2 is a flow chart of the present invention;
FIG. 3 is a detailed flow chart of the successive optimization algorithm of FIG. 2;
FIG. 4 is an original waveform of a fault of an outer ring of a bearing of an automobile hub in the embodiment;
FIG. 5 is a waveform of a high resonance component of a fault of an outer ring of a bearing of an automobile hub in the embodiment;
FIG. 6 is a waveform of a low resonance component of a fault of an outer ring of a bearing of an automobile hub in the embodiment;
FIG. 7 is an envelope spectrum of a low resonance component of a fault of an outer ring of a bearing of an automobile hub in the embodiment.
Detailed Description
As shown in fig. 1, a signal acquisition module, a data processing module and a result display module are installed on an automobile. The signal acquisition module is an acceleration sensor and is connected with the data processing module. The acceleration sensor is arranged on a bearing seat of the automobile and used for collecting vibration signals x in the vertical direction and outputting the vibration signals x to the data processing module. The input end of the data processing module is connected with the output end of the signal acquisition module, and the optimal high and low quality factors Q are obtained by running a successive optimization algorithmh *And Ql *Then inputting the best quality factor Qh *And Ql *To carry outObtaining the best high resonance component x by resonance sparse decompositionh *And a low resonance component xl *. For the best low resonance component xl *Envelope demodulation to obtain envelope spectrum xb. Finally, reading the characteristic frequency f in the envelope spectrumoAnd frequency multiplication thereof, if the characteristic frequency foEqual to or in multiple relation with the failure passing frequency of the outer ring of the bearing, i.e. BPFOoNamely the extracted fault characteristic frequency, outputting a signal to a result display module. The input of the result display module is connected with the output of the signal acquisition module, and the fault is displayed through a screen.
As shown in FIG. 1, the method firstly collects vibration signals, then initializes resonance sparse decomposition parameters, and then obtains the best quality factor by using a successive optimization algorithm and taking RSK index as an objective function. And finally, carrying out envelope analysis on a low resonance component obtained by carrying out resonance sparse decomposition under the signal optimal quality factor to obtain an envelope spectrum, thereby effectively extracting fault characteristics. The method comprises the following specific steps:
the method comprises the following steps: the signal acquisition module acquires vibration signals x and sampling frequency f of the automobile hub bearing through the acceleration sensors100kHz, 0.5s of sampling time t and 50000 of sampling points N.
Step two: setting initial resonance sparse decomposition parameters, high quality factor Q h3, redundancy rhNumber of decomposition layers J3h30; low quality factor Ql1, redundancy rlNumber of decomposition layers J3l＝11。
The automobile hub bearing outer ring fault impact response is a periodic impact signal which should be decomposed into low resonance components as much as possible. Background noise is a periodic harmonic signal that should be decomposed into high-resonance components as much as possible. The high and low resonance components differ by differences in the resonance properties, which are defined by the quality factor Q. The larger the quality factor Q, the higher the resonance property of the signal; conversely, the smaller the quality factor Q, the lower the resonance properties of the signal. The quality factor for the high resonance component is typically between 3-10 and the quality factor for the low resonance component is typically between 1-3. Pass through pairThe automobile hub bearing vibration signal is analyzed, and the morphological characteristics and Q of the automobile hub bearing vibration signal h3 and QlThe morphology was most similar when 1. The redundancy r is the over-sampling rate of the signal filtering, which typically takes a value of 3. Number of decomposition layers Jh、JlAre respectively represented by the maximum value JhmaxMinimum value JlmaxDetermining:
1≤Jh≤Jhmax,1≤Jl≤Jlmax，
Figure GDA0002387310610000041
Figure GDA0002387310610000042
in the formula, N is the number of signal sampling points,
Figure GDA0002387310610000043
to round the symbol down.
Step three: adopting a quality factor successive optimization algorithm to successively optimize high-quality factors and low-quality factors, wherein the quality factor successive optimization algorithm takes an RSK (ratio of smoothening and Kurtosis) index as an objective function to obtain the optimal high-resonance component xh *And a low resonance component xl *. The successive optimization algorithm is described in detail in fig. 3.
Step four: data processing module for optimal low resonance component x of signall *Envelope demodulation is performed (a specific method is described in the document entitled "harmonic-ratio-based envelope demodulation band determination method" under the notice No. CN 104819766B) to obtain an envelope spectrum xb。
Step five: data processing module extracts envelope spectrum xbCharacteristic frequency f inoAnd frequency doubling thereof. If the characteristic frequency foEqual to or multiplied by the failure passing frequency of the bearing outer ring, i.e. BPFO, foThe automobile hub bearing fault detection method is characterized by automobile hub bearing fault.
The fault passing frequency of the outer ring of the bearing, namely the BPFO formula is as follows:
Figure GDA0002387310610000051
wherein n is the number of rolling elements of the bearing, frAnd d is the bearing rotation frequency, d is the diameter of a bearing rolling element, S is the bearing pitch diameter, and α is the bearing contact angle.
The flow of the high quality factor and low quality factor successive optimization algorithm is shown in fig. 3.
The method comprises the following steps: maintaining a high quality factor QhConstant 3, low quality factor QlThe values are as follows:
Qlp＝1+0.1(p-1)，
where p is an argument of a low quality factor, and the initial value of p is 1, i.e., p is 1 at the time of first optimization.
Step two: using QhAnd QlpRunning resonance sparse decomposition to obtain high resonance component xhpAnd a low resonance component xlp. And the resonance sparse decomposition method carries out resonance sparse decomposition on the hub bearing fault vibration signal through the resonance sparse decomposition parameters. (the specific method is described in Selesnick, IW, university of New York, 2011 in the paper entitled "Sparse signal representation using the tunable Q-factor wall transform" published by Conference Congression on wavets and spark XIV). First, the acquired vibration signal x is input to the resonance sparse decomposition parameter Q corresponding to the high resonance componenth＝3，rh＝3，JhObtaining a high quality factor Q for a filter bank determined 30hTransformed basis functions Dh. Similarly, the acquired vibration signal x is input to the resonance sparse decomposition parameter Q corresponding to the low resonance componentlp＝1+0.1(p-1)，rl＝3，JlObtaining a low quality factor Q for a filter bank determined 11lpTransformed basis functions Dlp. Then, a basis function D is inputhAnd Dlp. The optimization objective function is constructed as follows:
Figure GDA0002387310610000052
in the formula, whp、wlpRespectively representing basis functions DhAnd DlpThe transform coefficients of (a); lambda [ alpha ]hAnd λlTo regularize the parameters whose values depend on the energy of the high and low resonance components, this embodiment sets them to λh＝0.6，λl＝0.4。
And (4) performing iterative calculation on the target function by adopting a split augmented Lagrange shrinkage algorithm in the paper. w is ahp *、wlp *Are respectively L (w)hp,wlp) And if the high resonance transformation coefficient and the low resonance transformation coefficient correspond to the minimum, the obtained high resonance component and the low resonance component are respectively as follows:
Figure GDA0002387310610000061
step three: high resonance component xhpAnd a low resonance component xlpIts substitution into the target function RSKp，RSKpThe functional expression is:
Figure GDA0002387310610000062
where μ and σ are the signals x, respectivelylpMean and standard deviation of.
Step four: and judging whether p is greater than 20, if not, determining that p is p +1, and returning to the step two to continue the operation. If yes, the next step is continued to be operated.
Step five: finding RSKpMinimum value of (3) and corresponding QlpIt is recorded as the best low quality factor Ql *。
Step six: maintaining an optimally low quality factor Ql *Constant, high quality factor QhThe values of (A) are as follows:
Qhq＝3+0.1(q-1)。
where q is an argument of the high quality factor, and the initial value of q is 1, i.e., the first time q is 1.
Step seven: using QhqAnd Ql *Running resonance sparse decomposition to obtain high resonance component xhqAnd a low resonance component xlq. And the resonance sparse decomposition method carries out resonance sparse decomposition on the hub bearing fault vibration signal through the resonance sparse decomposition parameters. First, the acquired vibration signal x is input to the resonance sparse decomposition parameter Q corresponding to the high resonance componenthq＝3+0.1(q-1)，rh＝3，JhObtaining a high quality factor Q for a filter bank determined 30hqTransformed basis functions Dhq. Similarly, the acquired vibration signal x is input to the resonance sparse decomposition parameter Q corresponding to the low resonance componentlp＝Ql *，rl＝3，JlObtaining a low quality factor Q for a filter bank determined 11l *Transformed basis functions Dl *. Then, a basis function D is inputhqAnd Dl *. The optimization objective function is constructed as follows:
Figure GDA0002387310610000063
in the formula, whq、wlqRespectively representing basis functions DhAnd DlpThe transform coefficients of (a); lambda [ alpha ]hAnd λlTo regularize the parameters whose values depend on the energy of the high and low resonance components, this embodiment sets them to λh＝0.6，λl＝0.4。
And (4) performing iterative calculation on the target function by adopting a split augmented Lagrange shrinkage algorithm in the paper. w is ahq *、wlq *For the high resonance transformation coefficient and the low resonance transformation coefficient corresponding to the minimum L, the obtained high resonance component and low resonance component are respectively:
Figure GDA0002387310610000064
step eight: high resonance component xhqAnd a low resonance component xlqIts substitution into the target function RSKq，RSKqThe functional expression is:
Figure GDA0002387310610000071
where μ and σ are the signals x, respectivelylqMean and standard deviation of.
Step nine: and judging whether q is greater than 20, if so, changing q to q +1, and returning to the step seven to continue the operation. If yes, the next step is continued to be operated.
Step ten: finding RSKqMinimum value of (3) and corresponding QhqIt is recorded as the best high quality factor Qh *。
Step eleven: using Qh *And Ql *Running resonance sparse decomposition to obtain optimal high resonance component xh *And a low resonance component xl *。
Examples
First, the automobile hub bearing shown in table 1 was selected. Then, a groove having a width of 0.3mm and a depth of 0.05mm was cut in the bearing outer ring. And finally, mounting the acceleration sensor on a bearing seat to obtain a vibration signal x in the vertical direction.
TABLE 1 parameters of hub bearings
Figure GDA0002387310610000072
The method comprises the following steps: the information acquisition module acquires a vibration signal x of the automobile hub bearing through the acceleration sensor, and the waveform of the vibration signal x is shown in fig. 4. Sampling frequency fs100kHz, 0.5s of sampling time t and 50000 of sampling points N.
Step two: resonance sparse decomposition parameter initialization module sets initial resonance sparse decomposition parameter Qh＝3，rh＝3，Jh＝30；Ql＝1，rl＝3，Jl＝11。
Step three: quality factor successive optimization for high and low quality factor QhAnd QlOptimizing and finding the best quality factorIs Qh *＝3.8，Ql *2.2. Reuse of best quality factor Qh *And Ql *Carrying out resonance sparse decomposition to obtain the optimal high resonance component xh *And a low resonance component xl *. As shown in figures 5 and 6, respectively.
Step five: envelope demodulation analysis module for optimal low-resonance component x of signall *Performing envelope demodulation to obtain envelope spectrum xbAs shown in fig. 7.
Step six: the fault feature extraction module reads the feature frequency f in the envelope spectrumoAnd frequency doubling thereof. From FIG. 7, fo36Hz and its frequency multiplication of 2 to 7 can be found. Calculated BPFO is 36 Hz. Because f isoAnd the failure characteristics of the automobile hub bearing can be extracted as BPFO. The embodiment results show that the fault characteristics of the automobile hub bearing can be extracted.

Claims (5)

1. An automobile hub bearing fault feature extraction method based on optimal quality factor selection is characterized by comprising the following steps:
the method comprises the following steps: acquiring a vibration signal x of an automobile hub bearing;
step two: setting initial resonance sparse decomposition parameters, high quality factor Qh3, redundancy rhNumber of decomposition layers J3h30; low quality factor Ql1, redundancy rlNumber of decomposition layers J3l＝11；
Step three: successively optimizing high-quality factors and low-quality factors to obtain the best high resonance component xh *And an optimum low resonance component xl *；
The successive optimization method comprises the following implementation steps:
step A: maintaining a high quality factor QhConstant 3, low quality factor Qlp1+0.1(p-1), p being an argument for a low quality factor, p being 1;
and B: by a high quality factor QhAnd a low quality factor QlpRunning resonance sparse decomposition to obtain high-quality factor QhTransformed basis functions DhAnd a low quality factor QlpTransformed basis functions DlpConstructing an optimized objective function
Figure FDA0002387310600000011
whp、wlpRespectively representing basis functions DhAnd DlpThe transform coefficients of (a); lambda [ alpha ]hAnd λlIs a regularization parameter; iterative calculation is carried out on the target function by adopting a splitting and amplifying Lagrange contraction algorithm, and the high resonance component and the low resonance component are respectively obtained as follows:
Figure FDA0002387310600000012
whp *、wlp *are respectively L (w)hp,wlp) A high resonance transformation coefficient and a low resonance transformation coefficient corresponding to the minimum;
and C: high resonance component xhpAnd a low resonance component xlpSubstituting the objective function
Figure FDA0002387310600000013
μ and σ are each xlpMean and standard deviation of;
step D: judging whether p is greater than 20, if not, adding 1 to p and returning to the step B; if yes, continuing the step E:
step E: finding RSKpMinimum value of (3) and corresponding QlpThen Q is corresponded tolpFor the best low quality factor Ql *；
Step F: maintaining an optimally low quality factor Ql *Constant, high quality factor QhqQ is 3+0.1(q-1), q is an argument for a high quality factor, q is 1;
step G: by a high quality factor QhqAnd an optimal low quality factor Ql *Running resonance sparse decomposition to obtain high resonance component xhqAnd a low resonance component xlq；
Step H: high resonance component xhqAnd a low resonance component xlqSubstituting the objective function
Figure FDA0002387310600000014
μ and σ are each xlqMean and standard deviation of;
step I: judging whether q is larger than 20, if not, adding 1 to q and returning to the step G, and if so, continuing the step J;
step J: finding RSKqMinimum value of (3) and corresponding QhqWill correspond to QhqTo achieve the best high quality factor Qh *；
Step K: with an optimum high quality factor Qh *And an optimal low quality factor Ql *Running resonance sparse decomposition to obtain optimal high resonance component xh *And a low resonance component xl *；
Step four: for the best low resonance component xl *Performing envelope demodulation to obtain envelope spectrum xb；
Step five: extracting envelope spectra xbCharacteristic frequency f inoAnd frequency doubling to obtain the fault characteristics of the automobile hub bearing.
2. The method for extracting the automobile hub bearing fault features selected based on the optimal quality factors as claimed in claim 1, wherein the method comprises the following steps: in the first step, the sampling frequency f of the vibration signal x of the automobile hub bearings100kHz, 0.5s of high sampling time t and 50000 of sampling points N.
3. The method for extracting the automobile hub bearing fault features selected based on the optimal quality factors as claimed in claim 1, wherein the method comprises the following steps: in the second step, the number of decomposition layers Jh、JlAre respectively represented by the maximum value JhmaxMinimum value JlmaxDetermining: j is not less than 1h≤Jhmax,1≤Jl≤Jlma，
Figure FDA0002387310600000021
Figure FDA0002387310600000022
N is the number of sampling points of the automobile hub bearing vibration signal x,
Figure FDA0002387310600000023
to round the symbol down.
4. The method for extracting the automobile hub bearing fault features selected based on the optimal quality factors as claimed in claim 1, wherein the method comprises the following steps: in step five, if the characteristic frequency foEqual to or multiplied by the fault passing frequency of the outer ring of the bearing, foThe automobile hub bearing fault characteristic is obtained; bearing outer ring fault passing frequency
Figure FDA0002387310600000024
n is the number of rolling elements of the bearing, frAnd d is the bearing rotation frequency, d is the diameter of a bearing rolling element, S is the bearing pitch diameter, and α is the bearing contact angle.
5. The method for extracting the automobile hub bearing fault features selected based on the optimal quality factors as claimed in claim 1, wherein the method comprises the following steps: in step B, the vibration signal x is input into the resonance sparse decomposition parameter Q corresponding to the high resonance componenth＝3，rh＝3，JhObtaining a high quality factor Q for a filter bank determined 30hTransformed basis functions Dh(ii) a Inputting the collected vibration signal x into a resonance sparse decomposition parameter Q corresponding to a low resonance componentlp＝1+0.1(p-1)，rl＝3，JlObtaining a low quality factor Q for a filter bank determined 11lpTransformed basis functions Dlp。
CN201810864229.8A 2018-08-01 2018-08-01 Automobile hub bearing fault feature extraction method based on optimal quality factor selection Active CN109100144B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810864229.8A CN109100144B (en) 2018-08-01 2018-08-01 Automobile hub bearing fault feature extraction method based on optimal quality factor selection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810864229.8A CN109100144B (en) 2018-08-01 2018-08-01 Automobile hub bearing fault feature extraction method based on optimal quality factor selection

Publications (2)

Publication Number Publication Date
CN109100144A CN109100144A (en) 2018-12-28
CN109100144B true CN109100144B (en) 2020-06-09

Family

ID=64848320

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810864229.8A Active CN109100144B (en) 2018-08-01 2018-08-01 Automobile hub bearing fault feature extraction method based on optimal quality factor selection

Country Status (1)

Country Link
CN (1) CN109100144B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109655267B (en) * 2019-01-07 2020-12-18 江苏大学 Method for extracting fault features of automobile hub bearing
CN111693279B (en) * 2020-05-22 2021-06-01 电子科技大学 Mechanical fault diagnosis method based on MPGA parametric resonance sparse decomposition
CN112362343A (en) * 2020-10-28 2021-02-12 华南理工大学 Distributed fault feature extraction method for gearbox under variable rotating speed based on frequency modulation dictionary
CN114580480A (en) * 2022-03-09 2022-06-03 南京工业大学 Fault diagnosis method for spindle box of numerical control high-speed gear milling machine

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102998119A (en) * 2012-12-04 2013-03-27 北京工业大学 Bearing fault diagnosis method based on composite Q-factor base algorithm
CN105841803A (en) * 2016-03-15 2016-08-10 大连理工大学 Processing vibration signal decomposition method based on quality factor minimization
CN106441871A (en) * 2016-10-20 2017-02-22 哈尔滨工业大学 Wind power gearbox fault diagnosis method based on self-adaptive resonance sparse decomposition theory

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102998119A (en) * 2012-12-04 2013-03-27 北京工业大学 Bearing fault diagnosis method based on composite Q-factor base algorithm
CN105841803A (en) * 2016-03-15 2016-08-10 大连理工大学 Processing vibration signal decomposition method based on quality factor minimization
CN106441871A (en) * 2016-10-20 2017-02-22 哈尔滨工业大学 Wind power gearbox fault diagnosis method based on self-adaptive resonance sparse decomposition theory

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于最优信号共振稀疏分解的旋转机械故障诊断方法研究;张顶成;《CNKI中国优秀硕士学位论文全文数据库(电子期刊)》;20170228(第2期);正文第8-49页 *

Also Published As

Publication number Publication date
CN109100144A (en) 2018-12-28

Similar Documents

Publication Publication Date Title
CN109100144B (en) Automobile hub bearing fault feature extraction method based on optimal quality factor selection
Qin et al. M-band flexible wavelet transform and its application to the fault diagnosis of planetary gear transmission systems
CN111397896B (en) Fault diagnosis method and system for rotary machine and storage medium
CN109827776B (en) Bearing fault detection method and system
CN110595780B (en) Bearing fault identification method based on vibration gray level image and convolution neural network
Zhou et al. Mechanical equipment fault diagnosis based on redundant second generation wavelet packet transform
CN109102005A (en) Small sample deep learning method based on shallow Model knowledge migration
CN107608936B (en) Method for extracting compound fault characteristics of planetary gear box
Liu et al. Acoustic emission analysis for wind turbine blade bearing fault detection under time-varying low-speed and heavy blade load conditions
CN109813547B (en) Rotary machine local fault diagnosis method based on sparse decomposition optimization algorithm
CN109145727A (en) A kind of bearing fault characteristics extracting method based on VMD parameter optimization
CN113935467B (en) DAS (data acquisition system) well exploration data noise suppression method based on iterative multi-scale attention network
CN107525671B (en) Method for separating and identifying compound fault characteristics of transmission chain of wind turbine generator
CN107689059B (en) Method and device for identifying abnormal variable pitch of wind generating set
Zhang et al. Multi-model ensemble deep learning method for intelligent fault diagnosis with high-dimensional samples
CN109682600A (en) A kind of improvement variation mode decomposition diagnostic method for Main Shaft Bearing of Engine fault diagnosis
CN110046476A (en) The ternary two of rolling bearing fault is into the sparse diagnostic method of Fractal Wavelet
Chen et al. Fault identification method for planetary gear based on DT-CWT threshold denoising and LE
CN114112400A (en) Mechanical bearing fault diagnosis method based on multi-angle information fusion
CN112945546B (en) Precise diagnosis method for complex faults of gearbox
Zhang et al. A novel compound fault diagnosis method using intrinsic component filtering
CN109540523A (en) A kind of ship centrifugal pump Fault Diagnosis of Roller Bearings
CN106500991A (en) Based on the bearing fault signal characteristic extracting methods that self-adapting multi-dimension AVGH is converted
Zhang et al. Fault diagnosis of rotating machinery based on time-frequency image feature extraction
Wang et al. Multidimensional blind deconvolution method based on cross-sparse filtering for weak 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
TR01 Transfer of patent right

Effective date of registration: 20210406

Address after: No. 159, Chengjiang Middle Road, Jiangyin City, Wuxi City, Jiangsu Province

Patentee after: Jiangyin Intellectual Property Operation Co., Ltd

Address before: Zhenjiang City, Jiangsu Province, 212013 Jingkou District Road No. 301

Patentee before: JIANGSU University

TR01 Transfer of patent right