CN114427972A - Rolling bearing performance degradation feature extraction method and system - Google Patents
Rolling bearing performance degradation feature extraction method and system Download PDFInfo
- Publication number
- CN114427972A CN114427972A CN202210119902.1A CN202210119902A CN114427972A CN 114427972 A CN114427972 A CN 114427972A CN 202210119902 A CN202210119902 A CN 202210119902A CN 114427972 A CN114427972 A CN 114427972A
- Authority
- CN
- China
- Prior art keywords
- factor
- optimal
- rolling bearing
- wavelet transform
- entropy
- 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
- 238000005096 rolling process Methods 0.000 title claims abstract description 84
- 238000006731 degradation reaction Methods 0.000 title claims abstract description 29
- 230000015556 catabolic process Effects 0.000 title claims abstract description 27
- 238000000605 extraction Methods 0.000 title claims abstract description 20
- 230000009466 transformation Effects 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 238000010845 search algorithm Methods 0.000 claims abstract description 11
- 238000000034 method Methods 0.000 claims description 23
- 238000009826 distribution Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000012545 processing Methods 0.000 claims description 14
- 238000000354 decomposition reaction Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000010606 normalization Methods 0.000 claims description 5
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 9
- 238000012544 monitoring process Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 238000001514 detection method Methods 0.000 description 6
- 238000003860 storage Methods 0.000 description 5
- 230000035945 sensitivity Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 230000009467 reduction Effects 0.000 description 3
- 230000001174 ascending effect Effects 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000015654 memory Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000001902 propagating effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
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
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/148—Wavelet transforms
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computational Mathematics (AREA)
- Pure & Applied Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Acoustics & Sound (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a rolling bearing performance degradation characteristic extraction method and system, which are used for decomposing rolling bearing vibration signals respectively based on adjustable wavelet transformation of different Q factors to obtain wavelet transformation coefficients of various frequency bands under different Q factors; calculating the Shannon entropy of each Q factor, and determining the minimum Shannon entropy as the optimal Q factor; determining the optimal wavelet transform coefficient of the optimal Q factor by adopting a split-augmented Lagrange search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor; calculating a sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transformation coefficient of the optimal Q factor; calculating a permutation entropy value corresponding to the sparse representation result according to a permutation entropy algorithm; and determining an arrangement entropy curve capable of reflecting the performance state change of the rolling bearing according to the arrangement entropy value.
Description
Technical Field
The invention relates to the technical field of bearing performance monitoring, in particular to a method and a system for extracting performance degradation characteristics of a rolling bearing.
Background
The rolling bearing is widely applied to various rotary machines, and the statistical result shows that the faults of rotary machine equipment caused by the rolling bearing account for about 30 percent, so that the performance degradation trend of the rolling bearing is effectively monitored by extracting the performance degradation characteristics of the rolling bearing, the initial faults are recognized as early as possible, the optimal maintenance time is determined, and the rolling bearing is crucial to improving the utilization rate and the use safety of various rotary machines on the basis of avoiding the unexpected shutdown of the equipment. Among various monitoring information sources, the vibration signal of the rolling bearing contains information of damage existence and fault type of the rolling bearing, and is a powerful signal source for detecting performance degradation of the rolling bearing.
In the prior art, the entropy value of a vibration signal of a rolling bearing is calculated by using an array entropy algorithm, so that the performance degradation of the rolling bearing is evaluated, the array entropy algorithm is proposed by ChristophBandt and the like and is used for measuring the complexity of a one-dimensional time sequence, but the array entropy algorithm can only ensure that the anti-noise capability is strong to small noise, while the vibration signal of the rolling bearing often contains a large amount of noise components, and the characteristic components of the rolling bearing at the initial stage of the fault are extremely easy to be submerged by the noise signal and are not enough to obviously change the signal complexity, so that the fault detection time lag is caused.
Aiming at the problems in the prior art, the invention provides a rolling bearing performance degradation feature extraction method and system based on self-adaptive Q factor signal sparse decomposition.
Disclosure of Invention
The invention aims to provide a rolling bearing performance degradation characteristic extraction method and system, which solve the problem that the characteristic components of a rolling bearing at the initial fault stage are easily submerged by noise signals and are not enough to obviously change the signal complexity, so that the fault detection time lags.
In order to achieve the purpose, the invention provides the following scheme:
a rolling bearing performance degradation feature extraction method comprises the following steps:
decomposing the vibration signals of the rolling bearing based on adjustable wavelet transformation of different Q factors to obtain wavelet transformation coefficients of various frequency bands under different Q factors; the vibration signal of the rolling bearing is a vibration signal of the rolling bearing in an operation time period;
processing the wavelet transformation coefficient of each frequency band under each Q factor into the distribution probability of characteristic energy in each frequency band, and calculating to obtain the Shannon entropy of each Q factor;
determining an optimal Q factor according to Shannon entropy of all Q factors; the optimal Q factor is the Q factor with the minimum Shannon entropy;
determining the optimal wavelet transform coefficient of the optimal Q factor by adopting a split-augmented Lagrange search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor;
calculating a sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transformation coefficient of the optimal Q factor;
and calculating the arrangement entropy value corresponding to the sparse representation result according to an arrangement entropy algorithm.
Optionally, the processing the wavelet transform coefficient of each frequency band under each Q factor into the distribution probability of the characteristic energy in each frequency band, and calculating to obtain the Shannon entropy of each Q factor specifically includes:
the wavelet transformation coefficient of each frequency band under each Q factor is expressed as the distribution probability of the characteristic energy under the corresponding Q factor in each frequency band according to the following formula;
wherein M is the number of decomposition layers, N is the total number of wavelet transform coefficients corresponding to each layer, and W*(i, j) is the j wavelet transform coefficient of the i wavelet transform of the i layer, PiIs the distribution probability of characteristic energy in the i-th layer frequency band and satisfies
Calculating Shannon entropy of the Q factors according to the distribution probability of the characteristic energy at each frequency band under each Q factor, wherein the Shannon entropy is as follows:
where H (p) is the Shannon entropy corresponding to the Q factor.
Optionally, the determining, according to the wavelet transform coefficient of each frequency band under the optimal Q factor, the optimal wavelet transform coefficient of the optimal Q factor by using a split-augmented lagrange search algorithm specifically includes:
adopting a splitting and amplifying Lagrange search algorithm, and updating the wavelet transform coefficient of each frequency band through iteration to obtain an optimal wavelet transform coefficient which enables an objective function F (W) to be minimum, wherein the objective function F (W) is as follows:
wherein W is wavelet transform coefficient, y is vibration signal of rolling bearing, TQWT-1Representing the inverse of a Q-factor adjustable wavelet transform, lambdajFor regularization parameters, M is the number of decomposition layers, | about | calculation of luminance1Is 11Norm, | × | calry2Is 12And (4) norm.
Optionally, the following formula is adopted to calculate the sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transform coefficient of the optimal Q factor:
wherein, W*And the optimal wavelet transform coefficient is the optimal Q factor.
Optionally, the calculating a permutation entropy value corresponding to the sparse representation result according to a permutation entropy algorithm specifically includes:
dividing the sparse representation result into a plurality of subsequences;
according to the TAKENS theorem, performing phase space reconstruction on each subsequence according to preset embedding time m and delay time tau to obtain a phase space matrix; each row element in the phase space matrix is a reconstruction component;
sequencing elements in each reconstruction component of the phase space matrix respectively to obtain a symbol sequence of each reconstruction component; the symbol sequence is a sequence formed by the original column indexes of all elements after the elements of the corresponding reconstruction components are sequenced;
and counting the probability of occurrence of all the arrangement sequences in each reconstruction component of the phase space matrix, and calculating to obtain an arrangement entropy value.
Optionally, after obtaining the permutation entropy value, the method further comprises: and carrying out normalization processing on the permutation entropy value so that the value of the permutation entropy value is between 0 and 1.
Optionally, the method further comprises:
calculating the mean value mu and the variance sigma of the arrangement entropy values of the rolling bearing in a normal state period;
setting the upper limit and the lower limit of an alarm threshold value as follows according to the mean value mu and the variance sigma: μ ± k σ, k is an arbitrary positive number determined according to the chebyshev inequality.
Corresponding to the rolling bearing performance degradation feature extraction method, in another aspect, the invention also provides a rolling bearing performance degradation feature extraction system, which is executed by a processor to execute the rolling bearing performance degradation feature extraction method.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
according to the rolling bearing performance degradation characteristic extraction method and system provided by the invention, the rolling bearing vibration signals are decomposed based on adjustable wavelet transformation of different Q factors respectively, so that wavelet transformation coefficients of various frequency bands under different Q factors are obtained; calculating the Shannon entropy of each Q factor, and determining the minimum Shannon entropy as the optimal Q factor; determining the optimal wavelet transform coefficient of the optimal Q factor by adopting a split-augmented Lagrange search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor; calculating a sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transformation coefficient of the optimal Q factor; calculating a permutation entropy value corresponding to the sparse representation result according to a permutation entropy algorithm; and determining an arrangement entropy curve capable of reflecting the performance state change of the rolling bearing according to the arrangement entropy value. The invention adopts a signal sparse decomposition method based on Q-factor adjustable wavelet transform to perform noise reduction processing on the vibration signals of the rolling bearing, reduces the interference of noise components on the signal complexity, enhances the sensitivity of the permutation entropy to the early tiny change of the rolling bearing, realizes the effective monitoring of the whole performance degradation process of the rolling bearing, detects the potential fault of the rolling bearing as early as possible, and improves the use safety of the rolling bearing.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required in the embodiments will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a flowchart of a rolling bearing performance degradation characteristic extraction method according to embodiment 1 of the present invention;
FIG. 2 is a flowchart of step A6 in the method provided in embodiment 1 of the present invention;
fig. 3 is a schematic structural diagram of a rolling bearing performance degradation feature extraction system provided in embodiment 2 of the present invention.
Symbol interpretation: 1: an adjustable wavelet transform unit; 2: a Shannon entropy calculation unit; 3: an optimal Q factor determination unit; 4: an optimal wavelet transform coefficient determination unit; 5: a sparse representation unit; 6: a permutation entropy calculation unit; 7: an alarm threshold determination unit.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention aims to provide a rolling bearing performance degradation characteristic extraction method and system, which solve the problem that the characteristic components of a rolling bearing at the initial fault stage are easily submerged by noise signals and are not enough to obviously change the signal complexity, so that the fault detection time lags.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Example 1:
as shown in the flowchart of fig. 1, the present embodiment provides a rolling bearing performance degradation feature extraction method, including the following steps:
a1, decomposing the vibration signals of the rolling bearing based on the adjustable wavelet transform of different Q factors to obtain the wavelet transform coefficients of each frequency band under different Q factors; the vibration signal of the rolling bearing is the vibration signal of the rolling bearing in the operation time period.
A2, processing the wavelet transform coefficients of each frequency band under each Q factor into the distribution probability of the characteristic energy in each frequency band, and calculating to obtain the Shannon entropy of each Q factor, wherein the step A2 specifically comprises the following steps:
the wavelet transform coefficient of each band at each Q factor is expressed as the distribution probability of the characteristic energy at each band at the corresponding Q factor according to the following formula.
Wherein M is the number of decomposition layers, N is the total number of wavelet transform coefficients corresponding to each layer, and W*(i, j) is the j wavelet transform coefficient of the i wavelet transform of the i layer, PiFor characteristic energy at i-th layer frequencyDistribution probability of the band, and satisfy
Calculating Shannon entropy of the Q factors according to the distribution probability of the characteristic energy at each frequency band under each Q factor, wherein the Shannon entropy is as follows:
where H (p) is the Shannon entropy corresponding to the Q factor.
A3, determining an optimal Q factor according to Shannon entropy of all Q factors; the optimal Q factor is the Q factor with the minimum Shannon entropy.
A4, determining the optimal wavelet transform coefficient of the optimal Q factor by adopting a split-augmented Lagrange search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor, wherein the step A4 specifically comprises the following steps:
adopting a splitting and amplifying Lagrange search algorithm, and updating the wavelet transform coefficient of each frequency band through iteration to obtain an optimal wavelet transform coefficient which enables an objective function F (W) to be minimum, wherein the objective function F (W) is as follows:
wherein W is wavelet transform coefficient, y is vibration signal of rolling bearing, TQWT-1Representing the inverse of a Q-factor adjustable wavelet transform, lambdajFor regularization parameters, M is the number of decomposition layers, | about | calculation of luminance1Is 11Norm, | × | luminance2Is 12And (4) norm.
A5, calculating a sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transform coefficient of the optimal Q factor by adopting the following formula:
wherein, the first and the second end of the pipe are connected with each other,W*and the optimal wavelet transform coefficient is the optimal Q factor.
A6, calculating a permutation entropy value corresponding to the sparse representation result according to a permutation entropy algorithm, as shown in fig. 2, step a6 specifically includes:
a61, dividing the sparse representation result into a plurality of subsequences.
A62, according to the TAKENS theorem, performing phase space reconstruction on each subsequence according to preset embedding time m and delay time tau to obtain a phase space matrix; each row element in the phase space matrix is a reconstruction component.
A63, respectively sequencing elements in each reconstruction component of the phase space matrix to obtain a symbol sequence of each reconstruction component; the symbol sequence is a sequence formed by the original column indexes of all elements after the elements of the corresponding reconstruction components are sequenced.
And A64, counting the probability of occurrence of all the arrangement sequences in each reconstruction component of the phase space matrix, and calculating to obtain an arrangement entropy value.
A65, in order to facilitate comparison of different time series complexity in practical application, after obtaining the permutation entropy, the method further includes: and carrying out normalization processing on the permutation entropy value so that the value of the permutation entropy value is between 0 and 1.
A7, calculating the mean value mu and the variance sigma of the arrangement entropy values of the rolling bearing in the normal state period; setting the upper limit and the lower limit of an alarm threshold value as follows according to the mean value mu and the variance sigma: and mu +/-k sigma, which is taken as a basis for judging whether the rolling bearing state is normal, wherein k is any positive number determined according to the Chebyshev inequality.
The invention adopts a signal sparse decomposition method based on Q-factor adjustable wavelet transform to perform noise reduction processing on the vibration signals of the rolling bearing, reduces the interference of noise components on the signal complexity, enhances the sensitivity of the permutation entropy to the early tiny change of the rolling bearing, realizes the effective monitoring of the whole performance degradation process of the rolling bearing, detects the potential fault of the rolling bearing as early as possible, and improves the use safety of the rolling bearing.
Example 2:
in addition, the method of embodiment 1 of the present invention can also be implemented by means of the architecture of a rolling bearing performance degradation feature extraction system shown in fig. 3. As shown in fig. 3, the rolling bearing performance degradation feature extraction system may include an adjustable wavelet transform unit 1, a Shannon entropy calculation unit 2, an optimal Q factor determination unit 3, an optimal wavelet transform coefficient determination unit 4, a sparse representation unit 5, an arrangement entropy calculation unit 6, and an alarm threshold determination unit 7; some modules may also have sub-units for implementing their functions. Of course, the architecture shown in fig. 3 is merely exemplary, and one or at least two components of the system shown in fig. 3 may be omitted as needed in order to implement different functions.
Portions of the technology may be considered "articles of manufacture" or "articles of manufacture" in the form of executable code and/or associated data embodied in or carried out by a computer readable medium. Tangible, non-transitory storage media may include memory or storage for use by any computer, processor, or similar device or associated module. For example, various semiconductor memories, tape drives, disk drives, or any similar device capable of providing a storage function for software.
All or a portion of the software may sometimes communicate over a network, such as the internet or other communication network. Such communication may load software from one computer device or processor to another. For example: from a server or host computer of the video object detection device to a hardware platform of a computer environment, or other computer environment implementing a system, or similar functionality related to providing information needed for object detection. Thus, another medium capable of transferring software elements may also be used as a physical connection between local devices, such as optical, electrical, electromagnetic waves, etc., propagating through cables, optical cables, air, etc. The physical medium used for the carrier wave, such as an electric, wireless or optical cable or the like, may also be considered as the medium carrying the software. As used herein, unless limited to a tangible "storage" medium, other terms referring to a computer or machine "readable medium" refer to media that participate in the execution of any instructions by a processor.
Example 3:
the rolling bearing performance degradation characteristic extraction method provided by the present invention is described in detail below by way of specific examples.
When the performance degradation trend of the rolling bearing is reflected by using the permutation entropy, the higher sensitivity of the permutation entropy to the state change of the rolling bearing can be ensured only by reducing or eliminating the interference caused by noise components to the maximum extent. Taking the rolling bearing fault simulation signal as an example, after white noise is superimposed according to different signal-to-noise ratios, the corresponding arrangement entropy values are shown in table 1 below,
TABLE 1
The obvious difference of the arrangement entropy values of the rolling bearing signals under the condition of no noise interference can be seen, and the arrangement entropy values are increased along with the enhancement of noise components, so that the signal arrangement entropy values are easily increased due to the existence of noise, the identification of the rolling bearing state is interfered, and the necessity of calculating the arrangement entropy after the de-noising processing of the signals is highlighted; the method comprises the following steps:
step 1: and (5) carrying out sparse decomposition on the signals. And for the collected vibration signals, obtaining the optimal wavelet transform coefficient of each frequency band corresponding to each Q factor by adopting a signal sparse decomposition method based on Q factor adjustable wavelet transform.
Step 2: an optimal Q factor is determined. Processing the optimal wavelet transform coefficient of each frequency band into the distribution probability of characteristic energy in each frequency band, calculating the Shannon entropy corresponding to each Q factor, selecting the Q factor corresponding to the minimum Shannon entropy from the Shannon entropy, and calculating the Shannon entropy as follows:
in the formula, PiIs an uncertain probability distribution, satisfies that the sum of all uncertain probabilities is 1, i.e.
For the Q-factor adjustable wavelet transform, if the jth coefficient of the ith layer wavelet transform is marked as W*(i, j), then PiThe calculation expression of (a) is:
in the formula, M is the total number of the filter banks; and N is the total number of the wavelet transform coefficients corresponding to each layer.
And step 3: the signal is optimally sparse represented. Calculating sparse representation result of signal y (t) according to wavelet transform coefficient of each frequency band corresponding to optimal Q factor
The calculation expression is as follows:
in the formula TQWT-1Representing an inverse of the Q-factor adjustable wavelet transform; w*An optimal transform coefficient for minimizing an objective function is obtained by iteratively updating a wavelet transform coefficient W.
And (3) carrying out noise reduction on the signal y (t) based on a Q-factor adjustable wavelet transform signal sparse decomposition method, and estimating the sparse distribution of the real signal x (t). In order to minimize signal reconstruction errors and obtain optimal sparse representation, the method introduces a common basis tracking denoising algorithm, defines the sparse representation problem of signals as a constrained extreme value problem, and uses a regularization term (namely, a minimized objective function) which is l of wavelet transform coefficients of each frequency band1The norm sum combined with the residual component energy, the mathematical expression is:
in the formula (I), the compound is shown in the specification,w is a wavelet transformation coefficient, and y is a vibration signal of the rolling bearing; TQWT-1Representing an inverse of the Q-factor adjustable wavelet transform; lambda [ alpha ]jIs a regularization parameter; m is the number of decomposition layers (i.e., the total number of filter banks); l |. electrically ventilated margin1Is 11Norm, | × | luminance2Is 12And (4) norm. The solution of the formula adopts a split augmented Lagrange search algorithm, and obtains an optimal transformation coefficient W which minimizes the objective function by iteratively updating the wavelet transformation coefficient W*Thereby ensuring an estimate of the true signal
No distortion is generated while obtaining an optimum noise suppression effect. Calculating out
The expression of (a) is:
and 4, step 4: and calculating permutation entropy. Calculation according to permutation entropy algorithm
Corresponding permutation entropy value Hp；
The permutation entropy is proposed by ChristophBandt et al, is used for measuring the complexity of a one-dimensional time sequence, and is widely applied to the aspects of weather forecast, earthquake monitoring, medical signal detection and the like by virtue of the advantage of sensitivity to dynamic mutation of a complex system.
The steps of calculating the time sequence complexity of the vibration signal of the rolling bearing by using the permutation entropy are summarized as follows:
step 4.1: a vibration signal time sequence is initialized. Dividing a time sequence { x (i) }, i ═ 1,2, …, N } into a plurality of subsequences which may or may not overlap with each other;
step 4.2: the phase space is reconstructed. According to the TAKENS theorem, phase space reconstruction is carried out on each subsequence according to preset embedding time m and delay time tau, and the vector space of the ith subsequence is obtained as follows: { x (i), x (i + τ), x (i +2 τ), …, x (i + (m-1) τ) };
performing phase space reconstruction on a given time sequence { x (i) }, i ═ 1,2, …, N } according to the theorem of TAKENS to obtain a matrix Y:
wherein m is the embedding dimension; τ is the delay time; k represents the total number of reconstructed components; x (j) is the jth row component of the reconstruction matrix.
Step 4.3: the elements in the reconstructed components are sorted. Rearranging m elements in { x (i), x (i + τ), x (i +2 τ), …, x (i + (m-1) τ) } in ascending order, if equal elements exist in the reconstructed components, determining the arranging order according to the initial positions, and the arranging result is: x (i + (j)1-1)τ),x(i+(j2-1)τ),…,x(i+(jm-1) τ) to obtain a sequence of symbols for each of all vectors. The implementation process is as follows:
each row in the matrix Y is taken as a reconstruction component, and K ═ N- (m-1) τ components are reconstructed in total. Rearranging the elements in the reconstructed components { x (i), x (i + τ), x (i +2 τ), …, x (i + (m-1) τ) }, in ascending order, with the result:
x(i+(j1-1)τ)≤x(i+(j2-1)τ)≤x(i+(jm-1)τ)
in the formula, j1,j2,…,jmIs the index of the column in which the corresponding element in the reconstructed component is located.
If there are equal elements in the reconstructed components, that is: x (i + (j)p-1)τ)＝x(i+(jq-1) τ) according to jpAnd jqMagnitude ordering of values, i.e. when jp≤jqThen, x (i + (j) is determined from the abovep-1)τ)≤x(i+(jq-1) τ). Finally, for each row of the matrix Y, a set of symbol sequences can be obtained, namely:
S(l)＝{j1,j2,…,jm},l＝1,2,…,K
in the process of m-dimensional phase space mapping, at most m! Seed of a species of riceDifferent symbol sequences j1,j2,…,jmThe symbol sequence S (l) is one of the permutations.
Step 4.4: and calculating permutation entropy. Counting the probability of all the sequence occurrences in the time sequence reconstruction component, and calculating the sequence entropy value Hp(m) the expression is:
in the formula, P1,P2,…,PKWhen K is less than or equal to m! Under the condition (2), a probability value appears in each symbol sequence.
Step 4.5: and (6) normalization processing. For facilitating the comparison of different time sequence complexity in practical application, the pair Hp(m) carrying out normalization processing to make the permutation entropy take a value between 0 and 1. The normalized computational expression is as follows:
and 5: and obtaining an arrangement entropy curve reflecting the performance state change of the rolling bearing. According to the mean value mu and the variance sigma of the array entropy in the initial time period (the rolling bearing is in a normal state), the upper limit and the lower limit of the alarm threshold are set as follows: mu + k sigma, k is any positive number (can be determined according to the Chebyshev inequality), and the positive number is used as a basis for judging whether the rolling bearing state is normal or not.
For the determination of the state monitoring threshold, a dynamic threshold setting method is provided, so that when the rolling bearing state monitoring is implemented, the monitoring threshold can be adaptively adjusted by using newly acquired performance degradation characteristic parameters, the interference of factors such as noise and the like on the monitoring parameters is reduced, and the accuracy of rolling bearing state evaluation is improved. By permutation entropy Hp(m) as a monitoring parameter, and given the current N monitoring values, the dynamic threshold value corresponding to the T period (comprising a plurality of time points T) is
The calculation expression is as follows:
the principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; those skilled in the art will appreciate that the modules or steps of the invention described above can be implemented using general purpose computing apparatus, or alternatively, they can be implemented using program code executable by computing apparatus, such that it is executed by computing apparatus when stored in a storage device, or separately fabricated into integrated circuit modules, or multiple modules or steps thereof can be fabricated into a single integrated circuit module. The present invention is not limited to any specific combination of hardware and software.
Meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A rolling bearing performance degradation feature extraction method is characterized by comprising the following steps:
decomposing the vibration signals of the rolling bearing based on adjustable wavelet transformation of different Q factors to obtain wavelet transformation coefficients of various frequency bands under different Q factors; the vibration signal of the rolling bearing is a vibration signal of the rolling bearing in an operation time period;
processing the wavelet transformation coefficient of each frequency band under each Q factor into the distribution probability of characteristic energy in each frequency band, and calculating to obtain the Shannon entropy of each Q factor;
determining an optimal Q factor according to Shannon entropy of all Q factors; the optimal Q factor is the Q factor with the minimum Shannon entropy;
determining the optimal wavelet transform coefficient of the optimal Q factor by adopting a split-augmented Lagrange search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor;
calculating a sparse representation result of the vibration signal of the rolling bearing according to the optimal wavelet transformation coefficient of the optimal Q factor;
and calculating the arrangement entropy value corresponding to the sparse representation result according to an arrangement entropy algorithm.
2. The method according to claim 1, wherein the processing of the wavelet transform coefficients of each frequency band under each Q factor into the distribution probability of the characteristic energy in each frequency band, and the calculating of the Shannon entropy of each Q factor specifically includes:
the wavelet transformation coefficient of each frequency band under each Q factor is expressed as the distribution probability of the characteristic energy under the corresponding Q factor in each frequency band according to the following formula;
wherein M is the number of decomposition layers, N is the total number of wavelet transform coefficients corresponding to each layer, and W*(i, j) is the j wavelet transform coefficient of the i wavelet transform of the i layer, PiIs the distribution probability of characteristic energy in the i-th layer frequency band and satisfies
Calculating Shannon entropy of the Q factors according to the distribution probability of the characteristic energy at each frequency band under each Q factor, wherein the Shannon entropy is as follows:
where H (p) is the Shannon entropy corresponding to the Q factor.
3. The method according to claim 1, wherein the determining the optimal wavelet transform coefficient of the optimal Q factor by using a split-augmented lagrangian search algorithm according to the wavelet transform coefficient of each frequency band under the optimal Q factor specifically comprises:
adopting a splitting and amplifying Lagrange search algorithm, and iteratively updating the wavelet transform coefficients of each frequency band to obtain an optimal wavelet transform coefficient which minimizes an objective function F (W), wherein the objective function F (W) is as follows:
wherein W is wavelet transform coefficient, y is vibration signal of rolling bearing, TQWT-1Representing the inverse of a Q-factor adjustable wavelet transform, lambdajFor regularization parameters, M is the number of decomposition layers, | about | calculation of luminance1Is 11Norm, | × | luminance2Is 12And (4) norm.
4. The method according to claim 3, characterized in that the sparse representation of the vibration signal of the rolling bearing is calculated from the optimal wavelet transform coefficients of the optimal Q-factor using the following formula:
wherein, W*And the optimal wavelet transform coefficient is the optimal Q factor.
5. The method according to claim 1, wherein the calculating a permutation entropy value corresponding to the sparse representation result according to a permutation entropy algorithm specifically includes:
dividing the sparse representation result into a plurality of subsequences;
according to the TAKENS theorem, performing phase space reconstruction on each subsequence according to preset embedding time m and delay time tau to obtain a phase space matrix; each row element in the phase space matrix is a reconstruction component;
sequencing elements in each reconstruction component of the phase space matrix respectively to obtain a symbol sequence of each reconstruction component; the symbol sequence is a sequence formed by the original column indexes of all elements after all elements of the corresponding reconstruction components are sequenced;
and counting the probability of occurrence of all the arrangement sequences in each reconstruction component of the phase space matrix, and calculating to obtain an arrangement entropy value.
6. The method of claim 1, wherein after obtaining the rank entropy value, the method further comprises: and carrying out normalization processing on the permutation entropy value so that the value of the permutation entropy value is between 0 and 1.
7. The method of claim 1, further comprising:
calculating the mean value mu and the variance sigma of the arrangement entropy values of the rolling bearing in a normal state period;
setting the upper limit and the lower limit of an alarm threshold value as follows according to the mean value mu and the variance sigma: μ ± k σ, k is an arbitrary positive number determined according to the chebyshev inequality.
8. A rolling bearing performance degradation feature extraction system, wherein the system is executed by a processor to perform the steps of the rolling bearing performance degradation feature extraction method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210119902.1A CN114427972A (en) | 2022-02-09 | 2022-02-09 | Rolling bearing performance degradation feature extraction method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210119902.1A CN114427972A (en) | 2022-02-09 | 2022-02-09 | Rolling bearing performance degradation feature extraction method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114427972A true CN114427972A (en) | 2022-05-03 |
Family
ID=81313880
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210119902.1A Pending CN114427972A (en) | 2022-02-09 | 2022-02-09 | Rolling bearing performance degradation feature extraction method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114427972A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103499437A (en) * | 2013-09-11 | 2014-01-08 | 西安交通大学 | Rotary machine fault detection method of dual-tree complex wavelet transformation with adjustable quality factors |
| CN107992112A (en) * | 2017-12-06 | 2018-05-04 | 浙江理工大学 | A kind of control loop performance estimating method and system based on arrangement entropy |
| CN110333078A (en) * | 2019-08-21 | 2019-10-15 | 佛山科学技术学院 | A kind of rolling bearing degenerate state stage determines method |
| CN110836775A (en) * | 2019-12-25 | 2020-02-25 | 济南大学 | Rolling bearing fault identification method and system based on multi-dimensional entropy space distance |
| CN111638055A (en) * | 2020-05-18 | 2020-09-08 | 东南大学 | Gearbox fault diagnosis method based on resonance sparse decomposition improved algorithm |
| CN111914735A (en) * | 2020-07-29 | 2020-11-10 | 天津大学 | Epilepsia electroencephalogram signal feature extraction method based on TQWT and entropy features |
| CN112285703A (en) * | 2020-10-16 | 2021-01-29 | 电子科技大学 | Sea clutter suppression and target detection method |
| AU2020103681A4 (en) * | 2020-11-26 | 2021-02-04 | Anhui University Of Technology | Rolling Bearing Fault Diagnosis Method Based on Fourier Decomposition and Multi-scale Arrangement Entropy Partial Mean Value |
| CN113780056A (en) * | 2021-07-22 | 2021-12-10 | 山东科技大学 | Rolling bearing fault diagnosis method based on vibration signal |
-
2022
- 2022-02-09 CN CN202210119902.1A patent/CN114427972A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103499437A (en) * | 2013-09-11 | 2014-01-08 | 西安交通大学 | Rotary machine fault detection method of dual-tree complex wavelet transformation with adjustable quality factors |
| CN107992112A (en) * | 2017-12-06 | 2018-05-04 | 浙江理工大学 | A kind of control loop performance estimating method and system based on arrangement entropy |
| CN110333078A (en) * | 2019-08-21 | 2019-10-15 | 佛山科学技术学院 | A kind of rolling bearing degenerate state stage determines method |
| CN110836775A (en) * | 2019-12-25 | 2020-02-25 | 济南大学 | Rolling bearing fault identification method and system based on multi-dimensional entropy space distance |
| CN111638055A (en) * | 2020-05-18 | 2020-09-08 | 东南大学 | Gearbox fault diagnosis method based on resonance sparse decomposition improved algorithm |
| CN111914735A (en) * | 2020-07-29 | 2020-11-10 | 天津大学 | Epilepsia electroencephalogram signal feature extraction method based on TQWT and entropy features |
| CN112285703A (en) * | 2020-10-16 | 2021-01-29 | 电子科技大学 | Sea clutter suppression and target detection method |
| AU2020103681A4 (en) * | 2020-11-26 | 2021-02-04 | Anhui University Of Technology | Rolling Bearing Fault Diagnosis Method Based on Fourier Decomposition and Multi-scale Arrangement Entropy Partial Mean Value |
| CN113780056A (en) * | 2021-07-22 | 2021-12-10 | 山东科技大学 | Rolling bearing fault diagnosis method based on vibration signal |
Non-Patent Citations (8)
| Title |
|---|
| KHANDAKER NOMAN等: "Oscillation based permutation entropy calculation as a dynamic nonlinear feature for health monitoring of rolling element bearing", MEASUREMENT, vol. 172, 28 February 2021 (2021-02-28), pages 1 - 14 * |
| 刘朦月: "基于振动信号的电机轴承故障诊断方法研究", 中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑, no. 02, 15 February 2017 (2017-02-15), pages 7 - 18 * |
| 屈梁生著: "机械监测诊断中的理论与方法 屈梁生论文集", 31 March 2009, 西安交通大学出版社, pages: 97 * |
| 李星;于德介;张顶成;: "基于最优品质因子信号共振稀疏分解的滚动轴承故障诊断", 振动工程学报, no. 06, 15 December 2015 (2015-12-15), pages 998 - 1004 * |
| 田福庆等著: "机械故障非平稳特征提取方法及其应用", 30 June 2014, 国防工业出版社, pages: 231 - 232 * |
| 胡易平;安钢;王凯;王传菲;: "基于小波自适应包络解调技术的滚动轴承故障识别研究", 煤矿机械, no. 02, 15 February 2010 (2010-02-15), pages 214 - 217 * |
| 马本栋等: "基于多小波包-能量算子解调的滚动轴承复合故障诊断", 噪声与振动控制, vol. 38, no. 04, 31 August 2018 (2018-08-31), pages 154 - 158 * |
| 高倩;陈晓英;孙丽颖;: "基于稀疏表示的TQWT在低频振荡信号去噪中应用", 电力系统保护与控制, no. 13, 1 July 2016 (2016-07-01), pages 57 - 58 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Sheik Fareed et al. | Fast adaptive and selective mean filter for the removal of high‐density salt and pepper noise | |
| CN111444807B (en) | Target detection method, device, electronic equipment and computer readable medium | |
| CN111932646A (en) | Image processing method for resisting attack | |
| CN116502042A (en) | Power quality disturbance denoising method based on variational modal decomposition and improved wavelet threshold | |
| Samantaray et al. | Neighbourhood decision based impulse noise filter | |
| JP6495292B2 (en) | Method for compressed sensing of streaming data and apparatus for carrying it out | |
| CN116578843A (en) | Centrifugal pump diagnostic model training method, diagnostic method, system, device and medium | |
| CN117290788B (en) | Power distribution network fault identification method and system based on improved wavelet transformation algorithm | |
| CN114970607A (en) | Transformer partial discharge detection method based on deep neural network acoustic emission signal separation | |
| CN113610735A (en) | Hybrid noise removing method for infrared image of power equipment | |
| CN114427972A (en) | Rolling bearing performance degradation feature extraction method and system | |
| CN116299219A (en) | Interference depth characteristic distance measurement combined detection and suppression method | |
| CN112904197A (en) | Parameter identification-based mechanical vibration signal feature extraction method for high-voltage circuit breaker | |
| Yang et al. | Mixed noise removal by residual learning of deep cnn | |
| Aetesam et al. | Proximal approach to denoising hyperspectral images under mixed‐noise model | |
| CN114978586B (en) | Power grid attack detection method and system based on attack genes and electronic equipment | |
| CN116226770A (en) | Time sequence data anomaly detection method and device | |
| Li et al. | Partial discharge signal denoising method based on frequency spectrum clustering and local mean decomposition | |
| CN115758121A (en) | Bearing fault classification method and device, electronic equipment and readable storage medium | |
| Li et al. | A robust fault diagnosis method for rolling bearings based on deep convolutional neural network | |
| CN112559956A (en) | Wavelet threshold adaptive shrinkage method, system, electronic device and storage medium | |
| CN113869358A (en) | Bearing fault diagnosis method based on cyclic correlation entropy and one-dimensional shallow convolution neural network | |
| CN109614937B (en) | Wavelet domain signal enhancement method based on group sparsity | |
| Lu et al. | Image enhancement using convolutional neural network to identify similar patterns | |
| CN117692074B (en) | Low-frequency aliasing noise suppression method suitable for unsteady-state underwater sound target signal |
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 |