CN111476339A - Rolling bearing fault feature extraction method, intelligent diagnosis method and system - Google Patents
Rolling bearing fault feature extraction method, intelligent diagnosis method and system Download PDFInfo
- Publication number
- CN111476339A CN111476339A CN202010313588.1A CN202010313588A CN111476339A CN 111476339 A CN111476339 A CN 111476339A CN 202010313588 A CN202010313588 A CN 202010313588A CN 111476339 A CN111476339 A CN 111476339A
- Authority
- CN
- China
- Prior art keywords
- test set
- training set
- signal data
- signal
- extracting
- 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
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000000605 extraction Methods 0.000 title claims abstract description 36
- 238000005096 rolling process Methods 0.000 title claims abstract description 29
- 238000003745 diagnosis Methods 0.000 title claims abstract description 16
- 238000012360 testing method Methods 0.000 claims abstract description 90
- 238000012549 training Methods 0.000 claims abstract description 89
- 239000011159 matrix material Substances 0.000 claims abstract description 62
- 230000009467 reduction Effects 0.000 claims abstract description 28
- 238000001228 spectrum Methods 0.000 claims abstract description 21
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000013144 data compression Methods 0.000 claims abstract description 9
- 238000000354 decomposition reaction Methods 0.000 claims description 35
- 239000002245 particle Substances 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 15
- 239000013598 vector Substances 0.000 claims description 9
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 4
- 238000007405 data analysis Methods 0.000 abstract 1
- 238000012706 support-vector machine Methods 0.000 description 20
- 230000006870 function Effects 0.000 description 11
- 238000005457 optimization Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 208000032544 Cicatrix Diseases 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 238000010892 electric spark Methods 0.000 description 1
- 238000013100 final test Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 231100000241 scar Toxicity 0.000 description 1
- 230000037387 scars Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000010183 spectrum analysis Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Acoustics & Sound (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention provides a rolling bearing fault feature extraction method, an intelligent diagnosis method and a system, comprising the following steps: acquiring bearing signal data of the electric driving end, preprocessing the signal data, and dividing the signal data into a test set and a training set; extracting the characteristics of signal data by adopting a wavelet band energy method to respectively obtain a first characteristic matrix of a training set and a first characteristic matrix of a testing set; extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set; extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set; the first, second and third feature matrices are spliced, data compression is carried out through PCA to obtain a training set and a test set after dimensionality reduction, the training set and the test set are input into a classifier to carry out data analysis, and faults are diagnosed.
Description
Technical Field
The invention belongs to the field of mechanical fault diagnosis, and particularly relates to a rolling bearing fault feature extraction method, an intelligent diagnosis method and an intelligent diagnosis system which are integrated with a plurality of feature processing methods and fault identification means.
Background
Mechanical fault diagnosis is a method for automatically supervising and maintaining equipment based on measurable signal characteristics in the shortest period expected by fault occurrence. Rolling bearings are the most important and easily damaged parts of rotating machines, and have been receiving much attention from the industry.
Existing bearing fault diagnosis techniques typically include three steps: collecting digital signals, processing the digital signals, and classifying by a classifier. The collected digital signals are mainly acquired by an acceleration sensor, and a classifier usually adopts a machine supervised learning algorithm to perform fault identification, and although the algorithm is quite perfect at present, the fault identification by adopting the supervised learning algorithm depends on the extracted signal characteristics to a great extent. The more complete and representative the extracted features are, the stronger the fault recognition capability is, and the traditional method for extracting the features by using a single feature often cannot obtain comprehensive and effective features, so that the classifier classification is difficult to express good generalization capability.
Disclosure of Invention
Aiming at the defects in the prior art, the invention adopts various methods to extract the features from different angles and carries out dimension reduction processing by a certain means, thus ensuring the completeness and sparsity of the extracted features, and in the classification and identification stage of a classifier, aiming at the parameter optimization problem of an SVM (support vector machine), the invention provides the PSO (particle swarm optimization) method to accelerate the parameter optimization speed.
In a first aspect, the invention provides a rolling bearing fault feature extraction method, which comprises the following steps: acquiring bearing signal data of the electric driving end, preprocessing the signal data, and dividing the signal data into a test set and a training set;
extracting the characteristics of signal data by adopting a wavelet band energy method to respectively obtain a first characteristic matrix of a training set and a first characteristic matrix of a testing set;
extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set;
extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set;
and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction, namely the bearing fault characteristics.
In a second aspect, the invention further provides an intelligent diagnosis method for rolling bearing faults, which includes the steps of obtaining a training set and a test set after dimensionality reduction by using the rolling bearing fault feature extraction method in the first aspect, inputting the training set and the test set after dimensionality reduction into a classifier for training, and diagnosing the bearing faults;
the specific steps of inputting the training set and the test set after the dimensionality reduction into the classifier for training comprise: initializing a particle swarm, and initializing the speed and the position of particles;
calculating the fitness of the particles, and calculating the SVM recognition accuracy under the current penalty term coefficient and the kernel function width;
searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not; if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
In a third aspect, the present invention further provides a rolling bearing fault feature extraction system, including:
an acquisition module: configured to collect bearing signal data at the electric drive end;
a preprocessing module: configured to pre-process the signal data, divide the signal data into a test set and a training set;
a feature extraction module: the method comprises the steps that the characteristics of signal data are extracted by adopting a wavelet frequency band energy method, and first characteristic matrixes of a training set and a test set are obtained respectively; extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set; extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set; and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction.
In a fourth aspect, the present invention further provides a computer-readable storage medium for storing computer instructions, which when executed by a processor, implement the rolling bearing fault feature extraction method according to the first aspect.
In a fifth aspect, the invention further provides an intelligent diagnosis system for rolling bearing faults, which comprises a classifier, and the acquisition module, the preprocessing module and the feature extraction module in the third aspect; the classifier configured to: inputting a training set and a test set after dimension reduction of a feature extraction module; initializing a particle swarm, and initializing a penalty term coefficient of the particle and the speed and position of the kernel function width; calculating the SVM identification accuracy under the current punishment item coefficient and the kernel function width; searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not; if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
Compared with the prior art, the invention has the following beneficial effects:
1. the method adopts multiple methods to extract the characteristics of the signals, adopts PCA (singular value decomposition) to perform dimensionality reduction, and can effectively ensure the completeness and sparsity of the characteristics of the signals and improve the generalization accuracy rate of the characteristic extraction compared with the characteristic extraction performed by a single method.
2. Aiming at the problem of parameter optimization of the SVM, the optimization speed of two important parameters C and g of the SVM is realized by utilizing a PSO algorithm, and compared with a conventional SVM classifier, the method can effectively avoid the problems of over-fitting and under-fitting and improve the generalization capability of the classifier.
3. The invention can extract accurate and representative characteristics from a large amount of data through preprocessing, adopts a plurality of methods to extract the characteristics from different angles, and carries out dimension reduction processing through a certain means, thereby ensuring the completeness and sparsity of the extracted characteristics, and in the classification and identification stage of a classifier, aiming at the parameter optimization problem of SVM (support vector machine), the invention provides the PSO (particle swarm optimization) to accelerate the parameter optimization speed.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the application and, together with the description, serve to explain the application and are not intended to limit the application.
FIG. 1 is a schematic flow chart of the bearing vibration signal processing of the present invention;
FIG. 2 is a wavelet decomposition tree of the present invention;
FIG. 3 is a wavelet packet decomposition tree of the present invention;
FIG. 4 is a flowchart of the SVM-PSO algorithm of the present invention.
The specific implementation mode is as follows:
the invention is further described with reference to the following figures and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
In the present invention, terms such as "upper", "lower", "left", "right", "front", "rear", "vertical", "horizontal", "side", "bottom", and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only terms of relationships determined for convenience of describing structural relationships of the parts or elements of the present invention, and are not intended to refer to any parts or elements of the present invention, and are not to be construed as limiting the present invention.
In the present invention, terms such as "fixedly connected", "connected", and the like are to be understood in a broad sense, and may be a fixed connection, or may be an integral connection or a detachable connection; may be directly connected or indirectly connected through an intermediate. The specific meanings of the above terms in the present invention can be determined according to specific situations by persons skilled in the relevant scientific or technical field, and are not to be construed as limiting the present invention.
Example 1
In order to extract accurate and representative features from a large amount of data, the invention adopts a plurality of methods to extract the features from different angles and carries out dimension reduction processing by a certain means, thus ensuring the completeness and sparsity of the extracted features, and in the classification and identification stage of a classifier, aiming at the parameter optimization problem of an SVM (support vector machine), the invention proposes to adopt PSO (particle swarm optimization) to accelerate the parameter optimization speed.
In order to achieve the above object, the solution of the present invention is as follows:
the rolling bearing fault feature extraction method comprises the following steps: acquiring bearing signal data of the electric driving end, preprocessing the signal data, and dividing the signal data into a test set and a training set;
extracting the characteristics of signal data by adopting a wavelet band energy method to respectively obtain a first characteristic matrix of a training set and a first characteristic matrix of a testing set;
extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set;
extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set;
and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction, namely the bearing fault characteristics.
Further, the step of preprocessing the signal comprises: the signal sequence to be used in the bearing data signal is F1～FiWill FiAre appropriately divided to form l signal sequences, i.e. Fi＝{fi1，fi2，fi3…filIn which fi1Dividing a first subset sequence into ith fault sequences; random slave FiSelecting a part of signal subsequence as an original training set; the other part is used as the original test set.
Further, the specific step of extracting the features of the signal data by using the wavelet band energy method includes:
carrying out wavelet decomposition on the bearing signal data;
decomposing different components of the bearing signal data into corresponding frequency bands, and calculating the energy characteristics of each frequency band;
and normalizing the energy feature vector to respectively obtain a first feature matrix of the training set and a first feature matrix of the test set.
Further, the specific step of extracting the features of the signal data by using the wavelet packet-AR spectrum estimation method includes: carrying out wavelet packet decomposition on the bearing signal;
decomposing different components of the signal into corresponding frequency bands, and performing wavelet packet reconstruction on each component to obtain a reconstructed signal;
performing AR spectrum estimation on the reconstructed signal, and calculating AR power spectral density of each frequency band;
and calculating the AR spectrum energy characteristics of each frequency band to respectively obtain a second characteristic matrix of the training set and the test set.
Further, the specific step of extracting the features of the signal data by using the EMD-SVD method includes:
EMD decomposition is carried out on the signals to obtain a plurality of IMF components;
calculating the energy ratio of the IMF components;
setting a threshold value, and extracting a plurality of IMF components with energy ratio exceeding the threshold value;
and compressing the extracted IMF components by adopting SVD, and taking the compressed IMF components as signal characteristics to respectively obtain a third characteristic matrix of the training set and the test set.
Further, the specific step of splicing the first, second and third feature matrices includes obtaining a training set α and a test set β after the splicing process, wherein α ═ α1，α2，α3]A1×M，β＝[β1，β2，β3]A2×MA1 and A2 respectively represent the number of signals in the training set and the test set, and M represents the feature dimension α1、α2And α3First, second and third feature matrices of the training set, β, respectively1、β2And β3First, second and third feature matrices of the test set, respectively.
Further, the specific steps of data compression through PCA include calculation of a covariance matrix of a training data set α, singular value decomposition of the covariance matrix, output of a training set after dimension reduction, and the same processing of a test set β to obtain a test set after dimension reduction.
Example 2
The invention is based on a bearing signal data set at a motor driving end of the university of western medicine storage, and needs to point out that the bearing fault data signals are divided into three types, namely an inner ring, an outer ring and a rolling body, wherein each type comprises artificially-manufactured electric spark scars with the diameters of 0.1778, 0.3556 and 0.5334mm, and the normal bearing data signals are divided into ten cases in total, so that the signal sequence to be used is determined to be F1～FiWhere i is 1, 2 … 10.
Fig. 1 is a flow chart of feature extraction and intelligent diagnosis of a bearing vibration signal according to the present invention, and as shown in fig. 1, the method of feature extraction and intelligent diagnosis of a bearing vibration signal according to the present invention specifically includes the following steps:
(1) data pre-processing
F is to beiAre appropriately divided to form 1 signal sequence, i.e. Fi＝{fi1，fi2，fi3…filIn which fi1And the first subset sequence is formed by dividing the ith fault sequence. Then randomly from FiSelecting a part of signal subsequence as an original training set; the other part is used as the original test set. For representing matrices with training set A1B 1 and test set A2B 2, wherein A1 and A2 are signal numbers, B1 and B2 represent characteristic dimensions of signals
(2) Multi-resolution analysis is carried out by utilizing wavelet transformation to extract energy characteristics of signals in different frequency bands
And (3) carrying out wavelet decomposition on the bearing signal, so as to decompose different components of the signal into corresponding frequency bands, and finally calculating the energy characteristics of each frequency band. The method comprises the following specific steps:
(2-1) wavelet decomposition: selecting discrete wavelet functions
j ∈ Z (1) is a wavelet basis function, and the signal is decomposed, and the specific mathematical expression is as follows:
where j is the decomposition scale, x is the decomposition position parameter, W2jf (x) is the wavelet decomposition coefficient of signal f at decomposition scale 2j, position x.
And (3) performing M1 layer decomposition on the signal by taking the formula (2) wavelet decomposition mathematical expression as a theoretical basis, and extracting the signal characteristics contained in each leaf node. Assuming that the original signal is f, the nth node of the mth layer of the wavelet decomposition tree (as shown in fig. 2) is represented by (M, n), where M is 0, 1, 2, 3 … M1, and n is 0, 1; each node represents a certain frequency band characteristic, (0, 0) represents an original signal, (1, 0) and (1, 1) respectively represent a low-frequency coefficient S of wavelet decomposition10And high frequency systemNumber S11Analogy (m, n) represents the coefficient S of the nth node of the mth layermn。
(2-2) calculating band energy: calculating the energy E of each leaf node of the signal fmnThe calculation formula is as follows:
in the formula, the value of n is restricted by M, and M is more than 0 and less than or equal to M1; when M < M1, n is 1; when M is M1, n is 0 or n is 1 (in practical terms, the M1 level signal decomposition results in a low frequency leaf node and a high frequency leaf node). When k represents SmnCharacteristic dimension of
(2-3) feature vector normalization: feature vector
Normalized by the energy ratio of each frequency band
Is determined by the characteristic of the one or more characteristic components,
as follows:
wherein the value ranges of m and n are the same as those of the formula (3).
(2-4) obtaining a feature matrix, extracting features of each signal according to the steps to finally obtain α of the first A1M 11β of training set a2 × M11And (5) testing the set.
(3) Method for extracting AR spectrum energy characteristics of signal in each frequency band by utilizing wavelet packet-AR spectrum estimation method
And (3) carrying out wavelet packet decomposition on the bearing signal, decomposing different components of the signal into corresponding frequency bands, reconstructing each component, and finally calculating the AR spectrum energy characteristics of each frequency band through AR spectrum estimation. The method comprises the following specific steps:
(3-1) wavelet packet decomposition: selecting a proper wavelet basis function, and determining the number of decomposition layers of the signal, wherein the specific mathematical expression is as follows:
wherein k is a translation time factor; h represents a low pass filter coefficient; g represents the high pass filter coefficients;
the m-th decomposition sequence is obtained by decomposing j layers of wavelet packets of the signal f (x);
is the original digital signal f (x).
And (3) carrying out M2 layer decomposition on the signal by taking the formula (5) wavelet packet decomposition mathematical expression as a theoretical basis, and extracting the signal characteristics contained in the leaf node of the last layer. Assuming that the original signal is f, the nth node of the mth layer of the wavelet packet decomposition tree (as shown in fig. 3) is represented by (M, n), where M is 0, 1, 2, 3 … M2, and n is 0, 1M2-1; each node represents a certain frequency band characteristic, (0, 0) represents an original signal, (1, 0) and (1, 1) represent a low-frequency coefficient S10 and a high-frequency coefficient S11 of wavelet decomposition respectively, and the analogy of (m, n) represents an nth node coefficient S of an mth layermn。
(3-2) wavelet packet reconstruction: and performing signal reconstruction according to the coefficient sequence obtained by the M2 th layer. RSmnRepresents the coefficient SmnThe reconstructed signal of (2). Wavelet reconstruction is mathematically expressed as follows:
in the formula:
and
the dual operators of h and g respectively,
the mth reconstructed sequence formed for the j-1 layer reconstruction.
(3-3) AR Spectrum analysis: for RSmnThe reconstructed signal is subjected to AR spectrum estimation to obtain AR spectrum power density P of each frequency bandmnThe method comprises the following steps:
assuming that the reconstructed signal sequence is x (n), the autoregressive model of the sequence can be expressed as:
wherein w (x) is a mean value of zero and a variance of
Normal distribution white noise of (1); q is the order of the model. The AR model parameters aj (j is 1, 2 … N) and are obtained from the knowledge of the relevant signal processing
Then, from the self-transfer function, the power spectral density of the signal x (n) is calculated as:
(3-4) solving the AR spectrum energy of each frequency band signal: let PmnCorresponding energy EpmnThen, there are:
wherein k is a feature dimension; M-M2; n is 0, 1, 2 … 2M2-1。
(3-5) obtaining feature vectors
Energy of each frequency band as
Is determined by the characteristic of the one or more characteristic components,
as follows:
(3-6) obtaining a feature matrix: extracting the characteristics of each signal according to the steps to obtain the size A1 (2)M2α of-1)2Training set and a2 (2)M2β of-1)2And (5) testing the set.
(4) EMD-SVD-based signal feature extraction method
EMD decomposition is carried out on the signals to obtain a plurality of IMF components; calculating the energy ratio of the IMF components; extracting a plurality of IMF components with larger energy proportion; and compressing the extracted IMF component by adopting SVD, and taking the compressed IMF component as a signal feature.
The method comprises the following specific steps:
(4-1) extracting the IMF components with energy sum more than 98% arranged in front by adopting EMD decomposition, which is as follows:
finding all local extreme points of the original signal f (t)
Fitting all the found local extreme points by using an interpolation algorithm to obtain an upper envelope fmax(t) and lower envelope fmin(t)
The mean value a is obtained from the upper and lower envelope curves1
Further passage h1＝f(t)-a1Judgment of h1Whether two conditions of IMF are satisfied, if so, h1Is the first order IMF component; if not, the ratio is h1Repeating the steps a) to d) as new f (t) to obtain h11＝h1-a11Wherein a is11The mean of the upper envelope and the lower envelope of h 1; if h11If not, the steps are repeated until h1kSatisfy the condition, record as C1＝h1kAs a first order IMF component
After separation of the first IMF component, r is obtained1＝f(t)-C1R is to1As a new f (t), then repeating the steps a) to d), calculating for a plurality of times to obtain an IMF component of n order and a residual rnAnd finally realizing EMD decomposition, wherein the formula is as follows:
calculating the energy Eci of different IMF components Ci, wherein i is 1, 2, 3. n, and then finding out the IMF component S which has a larger energy proportion and more than 98% of the energy summ。
(4-2) obtaining an IMF feature matrix Um×τ: each IMF component being Um×τA characteristic component, Um×τAs follows:
Um×τm in the matrix represents the number of vectors, and τ represents the feature dimension.
(4-3) further decomposing the characteristic value through SVD to realize Um×τThe matrix is compressed, and the decomposition formula is as follows:
in the above formula Qm×mIn the form of a left-hand singular matrix,
a right singular matrix; b ism×τIs a singular value matrix, and is a diagonal matrix, which is specifically shown as follows:
the first singular values with larger values are taken as the characteristics of the signals according to the situation, and the extracted characteristic vector is set as
And dimension M3, so we can get:
(4-4) obtaining a feature matrix, extracting features of each signal according to the steps to finally obtain α with the size of A1M 33β of training set a2 × M33And (5) testing the set.
(5) Dimensionality reduction of the resulting dataset using PCA
(5-1) obtaining a training set α and a test set β, wherein α ═ α1，α2，α3]A1×M，β＝[β1，β2，β3]A2×MA1 and a2 respectively indicate the number of signals in the training set and the test set, M indicates the feature dimension, and M is M1+2M2The further expansion of α and β by 1+ M3 is shown in detail as follows:
wherein Tri and Tei represent vectors of a1 x 1 and a2 x 1, respectively.
(5-2) compressing α and β with PCA
The method comprises the following specific steps:
for the training data set α, its covariance matrix Q is calculated, as follows:
singular value decomposition is carried out on the covariance matrix: [ U, S, V ] ═ svd (q), where U represents the left singular matrix, V represents the right singular matrix, and S represents the singular value matrix.
And searching a characteristic ratio and a corresponding k value when the characteristic ratio meets a threshold value, wherein the judgment conditions are as follows:
wherein eviRepresenting the ith singular value of Q.
The first k columns of U are selected as projection matrices Ptr ═ U1, U1, … uk, where the k-th column is the eigenvector corresponding to the k-th singular value.
Outputting the training data set Trian after dimension reductionA1×k＝Ptr×α，k＜M
The same processing is performed on the Test set β to obtain the final Test set TestA2×r＝Pte×β，r＜M
(6) Fault signal classification by SVM method
(6-1) classifier training
As shown in the flowchart of the SVM-PSO algorithm of fig. 4, the following is specifically explained:
(a) initializing a particle swarm, wherein the particle swarm size is N; the number of iterations It; the position of the penalty term coefficient ci is Pci, and the speed is Vci; the position of the kernel function width gi is Pgi, and the speed is Vgi; defining SVM classification error
(b) Calculating the fitness Fit [ i ] of the particles, namely calculating the accuracy of SVM classification under the current ci and gi;
(c) comparing the Fit [ i ] with the individual optimal Pbest [ i ], if Fit [ i ] is greater than Pbest [ i ], then Pbest [ i ] ═ Fit [ i ]; comparing Fit [ i ] with the individual optimal Gbest [ i ], if Fit [ i ] is more than Gbest [ i ], then Gbest [ i ] ═ Fit [ i ]
(d) Updating (Pci, Vci) and (Pgi, Vgi), namely, updating ci and gi; then calculating the fitness Fit [ i ] of the particle, and updating the formula as shown in (13) and (14):
vim＝w·vim(k)+c1r1·(Pim(k)-xid(k))+c2r2·(Pgm(k)-xim(k)) (13)
xim(k+1)＝xim(k)+vim(k+1) (14)
in the above formula, w is called inertia weight, c1 and c2 represent acceleration constants, r1 and r2 are random numbers between (0, 1), and k is the number of iterations.
(e) Judging whether the current error is small enough or the current iteration iter is larger than It, exiting, and outputting parameters c and g and target parameter values of w and b in the target function (15); otherwise, returning to the second step.
Subject to：yi(WTxi+b)≥1-ξi，ξi≥0
(6-2) placing the test set into a classifier for classification
(6-3) analyzing the classification results of the training set Train and the Test set Test
The above is a specific process of the present invention.
In other embodiments, there are provided:
an intelligent diagnosis method for rolling bearing faults comprises the steps of obtaining a training set and a test set after dimensionality reduction by adopting the rolling bearing fault feature extraction method in embodiment 1, inputting the training set and the test set after dimensionality reduction into a classifier for training, and diagnosing the bearing faults.
Further, the specific step of inputting the training set and the test set after the dimension reduction into the classifier for training includes: initializing a particle swarm, and initializing the speed and the position of a penalty term coefficient ci and a kernel function width gi of the particle; calculating the particle fitness (SVM recognition accuracy) under the current penalty term coefficient ci and the kernel function width gi; searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not; if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
A rolling bearing fault feature extraction system comprising:
an acquisition module: configured to collect bearing signal data at the electric drive end;
a preprocessing module: configured to pre-process the signal data, divide the signal data into a test set and a training set;
a feature extraction module: the method comprises the steps that the characteristics of signal data are extracted by adopting a wavelet frequency band energy method, and first characteristic matrixes of a training set and a test set are obtained respectively; extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set; extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set; and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction.
A computer-readable storage medium storing computer instructions which, when executed by a processor, perform the rolling bearing fault feature extraction method according to embodiment 1.
An intelligent diagnosis system for rolling bearing faults comprises a classifier, and the acquisition module, the preprocessing module and the feature extraction module in the embodiment; the classifier configured to: inputting a training set and a test set after dimension reduction of a feature extraction module; initializing a particle swarm, and initializing the speed and the position of the particles c and g; under the current punishment coefficient c and the kernel function width g, calculating the particle fitness (SVM identification accuracy); searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not; if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (10)
1. A rolling bearing fault feature extraction method is characterized by comprising the following steps: acquiring bearing signal data of the electric driving end, preprocessing the signal data, and dividing the signal data into a test set and a training set;
extracting the characteristics of signal data by adopting a wavelet band energy method to respectively obtain a first characteristic matrix of a training set and a first characteristic matrix of a testing set;
extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set;
extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set;
and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction, namely the bearing fault characteristics.
2. The rolling bearing fault signature extraction method of claim 1, wherein the signal preprocessing step comprises: the signal sequence to be used in the bearing data signal is F1～FiWill FiAre appropriately divided to form I signal sequences, i.e. Fi＝{fi1，fi2，fi3…filIn which fi1Dividing a first subset sequence into ith fault sequences; random slave FiSelecting a part of signal subsequence as an original training set; the other part is used as the original test set.
3. The rolling bearing fault feature extraction method according to claim 1, wherein the specific step of extracting the features of the signal data by using the wavelet band energy method comprises: carrying out wavelet decomposition on the bearing signal data;
decomposing different components of the bearing signal data into corresponding frequency bands, and calculating the energy characteristics of each frequency band;
and normalizing the energy feature vector to respectively obtain a first feature matrix of the training set and a first feature matrix of the test set.
4. The rolling bearing fault feature extraction method according to claim 1, wherein the specific step of extracting the features of the signal data by using a wavelet packet-AR spectrum estimation method comprises: carrying out wavelet packet decomposition on the bearing signal;
decomposing different components of the signal into corresponding frequency bands, and performing wavelet packet reconstruction on each component to obtain a reconstructed signal;
performing AR spectrum estimation on the reconstructed signal, and calculating AR spectrum energy characteristics of each frequency band;
and obtaining a feature vector of the AR spectrum energy feature, and respectively obtaining a second feature matrix of the training set and the test set.
5. The rolling bearing fault feature extraction method according to claim 1, wherein the specific step of extracting the features of the signal data by using the EMD-SVD method comprises:
EMD decomposition is carried out on the signals to obtain a plurality of IMF components;
calculating the energy ratio of the IMF components;
setting a threshold value, and extracting a plurality of IMF components with energy ratio exceeding the threshold value;
and compressing the extracted IMF components by adopting SVD, and taking the compressed IMF components as signal characteristics to respectively obtain a third characteristic matrix of the training set and the test set.
6. The rolling bearing fault feature extraction method according to claim 1, wherein the concrete step of splicing the first, second and third feature matrixes comprises obtaining a training set α and a test set β after splicing, wherein α ═ α1，α2，α3]A1×M，β＝[β1，β2，β3]A2×MTables A1 and A2Representing the number of signals in the training set and test set, M representing the feature dimension α1、α2And α3First, second and third feature matrices of the training set, β, respectively1、β2And β3First, second and third feature matrices of the test set, respectively;
the specific steps of data compression through PCA include that a covariance matrix of a training data set α is calculated, singular value decomposition is conducted on the covariance matrix, a training set after dimension reduction is output, and the same processing is conducted on a test set β in the same mode to obtain a test set after dimension reduction.
7. An intelligent diagnosis method for rolling bearing faults is characterized by comprising the steps of obtaining a training set and a test set after dimensionality reduction by adopting the rolling bearing fault feature extraction method according to claims 1-6, inputting the training set and the test set after dimensionality reduction into a classifier for training, and diagnosing the bearing faults;
the specific steps of inputting the training set and the test set after the dimensionality reduction into the classifier for training comprise: initializing a particle swarm, and initializing a penalty term coefficient of the particle and the speed and position of the kernel function width; calculating the SVM identification accuracy under the current punishment item coefficient and the kernel function width;
searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not;
if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
8. A rolling bearing fault feature extraction system comprising:
an acquisition module configured to: collecting bearing signal data of the electric driving end;
a pre-processing module configured to: preprocessing signal data, and dividing the signal data into a test set and a training set;
a feature extraction module configured to: extracting the characteristics of signal data by adopting a wavelet band energy method to respectively obtain a first characteristic matrix of a training set and a first characteristic matrix of a testing set; extracting the characteristics of the signal data by adopting a wavelet packet-AR spectrum estimation method to respectively obtain second characteristic matrixes of the training set and the test set; extracting the characteristics of the signal data by adopting an EMD-SVD method to respectively obtain a third characteristic matrix of the training set and the test set; and splicing the first characteristic matrix, the second characteristic matrix and the third characteristic matrix, and performing data compression through PCA to obtain a training set and a test set after dimensionality reduction.
9. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the rolling bearing fault feature extraction method of claims 1-6.
10. An intelligent diagnosis system for rolling bearing faults is characterized by comprising: an acquisition module, a preprocessing module, a feature extraction module in a classifier and the feature extraction system of claim 8; the classifier configured to: inputting a training set and a test set after dimension reduction of a feature extraction module; initializing a particle swarm, and initializing a penalty term coefficient of the particle and the speed and position of the kernel function width; calculating the SVM identification accuracy under the current punishment item coefficient and the kernel function width; searching an extreme value; updating the particle speed and position; judging whether the SVM classification error meets a termination condition or not; if the termination condition is met, putting the test set into a classifier for classification, and analyzing the classification results of the training set and the test set; if not, returning to the position where the extreme value is found for continuous processing.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010313588.1A CN111476339B (en) | 2020-04-20 | 2020-04-20 | Rolling bearing fault feature extraction method, intelligent diagnosis method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010313588.1A CN111476339B (en) | 2020-04-20 | 2020-04-20 | Rolling bearing fault feature extraction method, intelligent diagnosis method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111476339A true CN111476339A (en) | 2020-07-31 |
| CN111476339B CN111476339B (en) | 2023-05-05 |
Family
ID=71754323
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010313588.1A Active CN111476339B (en) | 2020-04-20 | 2020-04-20 | Rolling bearing fault feature extraction method, intelligent diagnosis method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111476339B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112633333A (en) * | 2020-12-11 | 2021-04-09 | 广州致新电力科技有限公司 | Method for identifying partial discharge defects |
| CN113076844A (en) * | 2021-03-26 | 2021-07-06 | 华中科技大学 | Method for constructing fault diagnosis model of rotating part and application |
| CN113705100A (en) * | 2021-08-31 | 2021-11-26 | 哈尔滨理工大学 | Gas detection method based on combination of temperature modulation detection and PSO-SVM algorithm |
| CN116893314A (en) * | 2023-09-04 | 2023-10-17 | 国网浙江省电力有限公司余姚市供电公司 | Non-invasive power load monitoring method, device, equipment and storage medium |
| CN117686226A (en) * | 2024-02-04 | 2024-03-12 | 南京凯奥思数据技术有限公司 | Automatic bearing fault diagnosis method and system based on energy ratio and energy sum |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180273066A1 (en) * | 2017-03-24 | 2018-09-27 | Canadian Pacific Railway Company | Condition based maintenance of railcar roller bearings using predictive wayside alerts based on acoustic bearing detector measurements |
| CN110220711A (en) * | 2019-05-22 | 2019-09-10 | 北京化工大学 | A kind of piston-mode motor shock characteristic extracting method based on EMD |
-
2020
- 2020-04-20 CN CN202010313588.1A patent/CN111476339B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180273066A1 (en) * | 2017-03-24 | 2018-09-27 | Canadian Pacific Railway Company | Condition based maintenance of railcar roller bearings using predictive wayside alerts based on acoustic bearing detector measurements |
| CN110220711A (en) * | 2019-05-22 | 2019-09-10 | 北京化工大学 | A kind of piston-mode motor shock characteristic extracting method based on EMD |
Non-Patent Citations (2)
| Title |
|---|
| JIAYU CHEN: "An Active Learning Method Based on Uncertainty and Complexity for Gearbox Fault Diagnosis" * |
| 张玲玲: "基于小波包 -AR 谱的变速器轴承故障特征提取" * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112633333A (en) * | 2020-12-11 | 2021-04-09 | 广州致新电力科技有限公司 | Method for identifying partial discharge defects |
| CN113076844A (en) * | 2021-03-26 | 2021-07-06 | 华中科技大学 | Method for constructing fault diagnosis model of rotating part and application |
| CN113076844B (en) * | 2021-03-26 | 2022-05-20 | 华中科技大学 | Method for constructing fault diagnosis model of rotating part and application |
| CN113705100A (en) * | 2021-08-31 | 2021-11-26 | 哈尔滨理工大学 | Gas detection method based on combination of temperature modulation detection and PSO-SVM algorithm |
| CN116893314A (en) * | 2023-09-04 | 2023-10-17 | 国网浙江省电力有限公司余姚市供电公司 | Non-invasive power load monitoring method, device, equipment and storage medium |
| CN117686226A (en) * | 2024-02-04 | 2024-03-12 | 南京凯奥思数据技术有限公司 | Automatic bearing fault diagnosis method and system based on energy ratio and energy sum |
| CN117686226B (en) * | 2024-02-04 | 2024-04-16 | 南京凯奥思数据技术有限公司 | Automatic bearing fault diagnosis method and system based on energy ratio and energy sum |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111476339B (en) | 2023-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111476339B (en) | Rolling bearing fault feature extraction method, intelligent diagnosis method and system | |
| CN108319962B (en) | Tool wear monitoring method based on convolutional neural network | |
| CN108827605B (en) | Mechanical fault feature automatic extraction method based on improved sparse filtering | |
| CN111523509B (en) | Equipment fault diagnosis and health monitoring method integrating physical and depth expression characteristics | |
| CN111460956B (en) | Unbalanced electrocardiogram sample classification method based on data enhancement and loss weighting | |
| CN112254964A (en) | Rolling bearing fault diagnosis method based on rapid multi-scale convolution neural network | |
| CN105841961A (en) | Bearing fault diagnosis method based on Morlet wavelet transformation and convolutional neural network | |
| CN114237046B (en) | Partial discharge pattern recognition method based on SIFT data feature extraction algorithm and BP neural network model | |
| CN113158722A (en) | Rotary machine fault diagnosis method based on multi-scale deep neural network | |
| CN112577743A (en) | Rolling bearing fault diagnosis method based on maximum local boundary criterion | |
| Masood et al. | Differential evolution based advised SVM for histopathalogical image analysis for skin cancer detection | |
| CN114169377A (en) | G-MSCNN-based fault diagnosis method for rolling bearing in noisy environment | |
| CN115290326A (en) | Rolling bearing fault intelligent diagnosis method | |
| CN110598628B (en) | Electromyographic signal hand motion recognition method based on integrated deep learning | |
| CN115859142A (en) | Small sample rolling bearing fault diagnosis method based on convolution transformer generation countermeasure network | |
| CN113295413B (en) | Traction motor bearing fault diagnosis method based on indirect signals | |
| CN116337449A (en) | Sparse self-coding fault diagnosis method and system based on information fusion | |
| CN115238796A (en) | Motor imagery electroencephalogram signal classification method based on parallel DAMSCN-LSTM | |
| CN112945556B (en) | Bearing fault diagnosis method based on wavelet packet decomposition and optimal rejection classification strategy | |
| CN113758709A (en) | Rolling bearing fault diagnosis method and system combining edge calculation and deep learning | |
| CN112151067B (en) | Digital audio tampering passive detection method based on convolutional neural network | |
| CN111677674A (en) | Fault detection method and device | |
| CN116340812A (en) | Transformer partial discharge fault mode identification method and system | |
| CN115273908B (en) | Live pig cough voice recognition method based on classifier fusion | |
| CN116429426A (en) | Bearing fault diagnosis method, device and medium for multi-domain feature fusion |
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 | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20231122 Address after: No. 2151, 21st Floor, Building 2, No. 35 Jianguo Road, Chaoyang District, Beijing, 100123 Patentee after: Beijing Xiniu Technology Co.,Ltd. Address before: 250014 No. 88, Wenhua East Road, Lixia District, Shandong, Ji'nan Patentee before: SHANDONG NORMAL University |