CN111428418A - Bearing fault diagnosis method and device, computer equipment and storage medium - Google Patents
Bearing fault diagnosis method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111428418A CN111428418A CN202010127805.8A CN202010127805A CN111428418A CN 111428418 A CN111428418 A CN 111428418A CN 202010127805 A CN202010127805 A CN 202010127805A CN 111428418 A CN111428418 A CN 111428418A
- Authority
- CN
- China
- Prior art keywords
- svm
- bearing
- kernel function
- parameter
- vibration signal
- 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
- 238000000034 method Methods 0.000 title claims abstract description 67
- 238000003745 diagnosis Methods 0.000 title claims abstract description 33
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000012706 support-vector machine Methods 0.000 claims abstract description 67
- 238000005457 optimization Methods 0.000 claims abstract description 51
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 42
- 238000012360 testing method Methods 0.000 claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 27
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 24
- 239000013598 vector Substances 0.000 claims abstract description 16
- 230000010354 integration Effects 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000010606 normalization Methods 0.000 claims abstract description 9
- 241000282461 Canis lupus Species 0.000 claims description 62
- 230000006870 function Effects 0.000 claims description 49
- 238000004590 computer program Methods 0.000 claims description 13
- 230000003044 adaptive effect Effects 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 6
- 239000002245 particle Substances 0.000 claims description 6
- 238000002790 cross-validation Methods 0.000 claims description 5
- 238000001514 detection method Methods 0.000 claims description 2
- 238000005096 rolling process Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000010845 search algorithm Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000002068 genetic effect Effects 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010892 electric spark Methods 0.000 description 1
- 238000004870 electrical engineering Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000003754 machining Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000003045 statistical classification method Methods 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
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Abstract
The invention discloses a bearing fault diagnosis method, a bearing fault diagnosis device, computer equipment and a storage medium, wherein a vibration signal of a bearing is obtained, and an IMF modal component of the vibration signal is extracted through a complete integration empirical mode decomposition (CEEMDAN) method of self-adaptive white noise; determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM); updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the group intelligent optimization algorithm; the SVM determines the fault result of the bearing according to the test set, the punishment parameter and the kernel function parameter, so that the problems of low efficiency and low precision of bearing fault diagnosis are solved, and the efficiency and the precision of bearing fault diagnosis are improved.
Description
Technical Field
The invention relates to the field of computer software, in particular to a bearing fault diagnosis method and device, computer equipment and a storage medium.
Background
The rolling bearing is a key component in the operation process of mechanical equipment, and if the rolling bearing is in failure, the normal operation of other components is influenced, and even the whole system is broken down. Therefore, the rapid, accurate and easy detection of the existence and severity of the bearing fault is of great significance.
An Ensemble Empirical Mode Decomposition (EEMD) is a time-frequency domain signal analysis method widely used in fault diagnosis, which can adaptively decompose a bearing vibration signal, but residual noise exists in a signal reconstructed by the EEMD, and although the reconstruction error can be reduced by increasing the integration times, the workload is very large.
In the aspect of pattern recognition by using fault features, a Support Vector Machine (SVM for short) can obtain good classification and popularization capability under the condition of less samples, and has applicability to the problems of small samples, non-linearity, high-dimensional pattern recognition and the like. Therefore, SVM is considered to be a very potential classification technique, which is widely applied to the field of mechanical fault diagnosis. In fault diagnosis by using an SVM, the diagnosis accuracy has a large relationship with a kernel function parameter and a penalty coefficient, so that the Optimization problem of the parameter is particularly important, and in the related art, methods for optimizing the SVM parameter mainly include a grid search method, a cross validation method, a Genetic Algorithm (GA for short), and a Particle Swarm Algorithm (PSO for short), but the methods have low Optimization efficiency and low accuracy.
Aiming at the problems of low efficiency and low precision of bearing fault diagnosis in the related technology, no effective solution is provided at present.
Disclosure of Invention
The embodiment of the invention at least solves the problems of low efficiency and low precision of bearing fault diagnosis in the related art.
According to an aspect of the present invention, there is provided a method of diagnosing a bearing failure, the method comprising:
obtaining a vibration signal of a bearing, and extracting an IMF modal component of the vibration signal by a complete integration empirical mode decomposition (CEEMDAN) method of self-adaptive white noise;
determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM);
updating punishment parameters and kernel function parameters of a classifier of the SVM according to the training set and the group intelligent optimization algorithm;
and the SVM determines a fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter.
In one embodiment, updating the penalty parameters and the kernel function parameters of the classifier of the SVM according to the training set and group intelligence optimization algorithm comprises:
updating the grey wolf position in the grey wolf optimization algorithm GWO through the inertia weight and the flight speed in the particle swarm optimization PSO;
and performing multiple iterations according to the updated wolf location to determine a penalty parameter and a kernel function parameter of the classifier of the SVM.
In one embodiment, the extracting the modal component of the vibration signal by the fully integrated empirical mode decomposition method CEEMDAN of adaptive white noise, and the determining the energy entropy of the modal component includes:
extracting modal components of a preset number of the vibration signals by a complete integration empirical mode decomposition method CEEMDAN of self-adaptive white noise;
and determining Shannon energy entropy according to the modal components of the preset number.
In one embodiment, the SVM determining the fault result of the bearing according to the test set, the penalty parameter, and the kernel function parameter includes:
and the SVM counts the average accuracy, the average optimization time and the longest and shortest optimization time difference of the test set to serve as judgment standards, and the fault result of the bearing is determined after ten-fold cross validation.
In one embodiment, the kernel function is a radial basis kernel function RBF.
According to another aspect of the present invention, there is also provided a bearing failure diagnosis apparatus, including:
the acquisition module is used for acquiring a vibration signal of the bearing and extracting an IMF modal component of the vibration signal through an integrated empirical mode decomposition algorithm;
the construction module is used for determining the energy entropy of the modal component, carrying out normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM);
the parameter optimization module is used for updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the cluster intelligent optimization algorithm;
and the diagnosis module is used for determining the fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter by the SVM.
In one embodiment, the parameter optimization module is further configured to update the grey wolf position in the grey wolf optimization algorithm GWO through the inertia weight and the "flight speed" in the particle swarm algorithm PSO;
and performing multiple iterations according to the updated wolf location to determine a penalty parameter and a kernel function parameter of the classifier of the SVM.
In one embodiment, the construction module is further configured to extract a preset number of modal components of the vibration signal by a fully integrated empirical mode decomposition method adaptive to white noise CEEMDAN; and determining Shannon energy entropy according to the modal components of the preset number.
According to another aspect of the invention, there is also provided a computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of any of the methods when executing the computer program.
According to another aspect of the invention, there is also provided a computer-readable storage medium, on which a computer program is stored, characterized in that the computer program realizes the steps of any of the methods when executed by a processor.
According to the method, a vibration signal of the bearing is obtained, and IMF modal components of the vibration signal are extracted through a CEEMDAN method of complete integration empirical mode decomposition of self-adaptive white noise; determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM); updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the group intelligent optimization algorithm; the SVM determines the fault result of the bearing according to the test set, the punishment parameter and the kernel function parameter, so that the problems of low efficiency and low precision of bearing fault diagnosis are solved, and the efficiency and the precision of bearing fault diagnosis are improved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:
FIG. 1 is a schematic diagram of a method for diagnosing a bearing fault according to an embodiment of the present invention;
FIG. 2 is a flow chart of a method of diagnosing bearing faults according to an embodiment of the present invention;
FIG. 3 is a schematic flow diagram of a fault diagnosis model according to an embodiment of the invention;
FIG. 4 is a flow chart of a fault diagnosis method according to an embodiment of the present invention;
fig. 5 is a block diagram of a bearing fault diagnosis apparatus according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of a waveform of a rolling element fault CEEMDAN after decomposition according to an embodiment of the present invention;
fig. 7 is a structural diagram of the inside of a computer apparatus according to an embodiment of the present invention.
Detailed Description
The invention will be described in detail hereinafter with reference to the accompanying drawings in conjunction with embodiments. It should be noted that the embodiments and features of the embodiments in the present application may be combined with each other without conflict.
In the present embodiment, an application scenario of a method for diagnosing a bearing fault is provided, and fig. 1 is a schematic diagram of a method for diagnosing a bearing fault according to an embodiment of the present invention, as shown in fig. 1, in the application scenario, a terminal 12 communicates with a server 14 through a network. The terminal 12 can acquire a vibration signal of the bearing and send the vibration signal to the server 14, and the server 14 extracts an IMF modal component of the vibration signal through a complete integration empirical mode decomposition (CEEMDAN) method of adaptive white noise; determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM); updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the group intelligent optimization algorithm; and the SVM determines a fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter. In another embodiment, the terminal 12 performs the diagnosis method of the bearing fault to determine the fault result, wherein the terminal 12 may be, but is not limited to, various personal computers, laptops, smartphones, tablets and portable wearable devices, and the server 14 may be implemented by an independent server or a server cluster composed of a plurality of servers.
In the present embodiment, a method for diagnosing a bearing fault is provided, and fig. 2 is a flowchart of a method for diagnosing a bearing fault according to an embodiment of the present invention, as shown in fig. 2, the method includes the following steps:
step S202, obtaining a vibration signal of the bearing, and extracting an IMF modal component of the vibration signal by a Complete integrated Empirical Mode Decomposition (CEEMDAN) method of Adaptive white Noise, wherein the vibration signal can be a vibration signal of each part of the bearing, for example, the bearing part comprises a bearing inner ring, a bearing outer ring and a rolling body of the bearing;
step S204, determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM), wherein when a bearing fails, the energy of a vibration signal is more concentrated on certain natural frequency, the uncertainty and complexity of energy distribution are small, and the energy entropy can be used as the feature vector to measure, and the SVM is used as a statistical classification method and needs samples of the training set and the test set;
step S206, updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the swarm intelligent optimization algorithm, obtaining the optimized punishment parameters and kernel function parameters by the training set through the swarm intelligent optimization algorithm, and finding better punishment parameters and kernel function parameters through multiple iterations of the swarm intelligent optimization algorithm;
and S208, the SVM determines the fault result of the bearing according to the test set, the punishment parameter and the kernel function parameter, the judgment standard in the test set is selected, and the SVM evaluates the optimization efficiency and the precision of the fault through the optimized punishment parameter and kernel function parameter.
Through the steps S202 to S208, the punishment parameters and the kernel function of the SVM classifier are optimized through the group intelligent optimization algorithm, so that the SVM is converged more quickly and classified optimally, the problems of low efficiency and low precision of bearing fault diagnosis are solved, and the efficiency and the precision of bearing fault diagnosis are improved.
In one embodiment, the mode component of the vibration signal is extracted by a complete integration empirical mode decomposition method CEEMDAN of adaptive white noise, so that the reconstruction error of EEMD is reduced, the problem of low EEMD decomposition efficiency is solved, CEEMDAN is an improvement on EEMD, the decomposition process can effectively overcome the mode mixing problem by using CEEMDAN, the reconstruction error is almost zero, the calculation cost is greatly reduced, and the process of CEEMDAN algorithm is described as follows:
in step S1, each S is decomposed by Empirical Mode Decomposition (EMD)i(t)＝s(t)+0ni(t), (I ═ 1,2.., I), until its first modal component is obtained, and the first CEEMDAN modal component is defined by equation 1 as:
in step S2, when the first phase j is equal to 1, equation 2 calculates a first margin:
at step S3, equation 3 decomposes the EMD for each r1(t)+1(t)E1(ni(t)), (I ═ 1,2.., I), until its first modal component is obtained, while defining a second CEEMDAN modal component:
in step S4, when j is 2, 3.
At step S5, equation 5 decomposes the EMD for each rj(t)+j(t)Ej(ni(t)), (I ═ 1,2.., I), until its first modal component is obtained, while defining the (j +1) th CEEMDAN modal component:
step S6, proceeding to step S4, repeats steps S4 through S6 until the obtained margin is no longer further decomposed by the EMD because it meets the IMF criteria or it has been less than three local extrema. The final margin satisfies equation 6:
where N is the total number of modal components, the target signal can be represented by equation 7:
in summary, it can be seen that the decomposition stage of each modal function can be relied uponjAnd selecting a proper signal-to-noise ratio, wherein the original vibration signal can be accurately reconstructed through CEEMDAN.
In addition, the uniformity of the probability distribution is reflected by the size of the Shannon energy entropy value. For the fault signals of the rotating machinery with the bearing as the main component, the fault or not of the running state of the rotating machinery shows different intrinsic mode complexity: when the rolling bearing fails, the energy in the signal is more concentrated in certain inherent frequency bands, the uncertainty and complexity of energy distribution are smaller, and the Shannon energy entropy value is smaller. Therefore, Shannon energy entropy with strong fault tolerance is selected as a feature vector, and a training set and a testing set of the SVM are constructed.
The advantages of the CEEMDAN and Shannon energy entropy are comprehensively considered, the first 3 IMF components are extracted by adopting the CEEMDAN method, and the Shannon energy entropy is respectively calculated to be used as the characteristic vector, and the algorithm is shown as formula 8 and formula 9:
h in equation 8j(x) For the final CEEMDAN-Shannon energy entropy value, IMFj(i) Is a value after the mode decomposition of the CEEMDAN method.
In the present embodiment, SVM is a classification method of statistical learning theory proposed by Vapnik et al. Given a training sample of (x)i,yi),xi∈Rn,yi∈ { -1, +1} are class labels, an objective function is constructed by using SVM, and the optimal segmentation hyperplane is sought by equation 10:
y＝wTphi (x) + b equation 10
In equation 10, ω represents the hyperplane normal vector and b represents the offset. If the linearity is not separable, the following optimal equation needs to be solved through equation 11:
ξ in equation 11iThe linear irreversibility is introduced as a relaxation variable, and c is a penalty factor and is used for representing a penalty index for error classification. Therefore, c is a key factor for determining the SVM learning ability and the experience risk co-scheduling, if c is too large, excessive learning is caused, so that the generalization ability of the classifier is reduced, otherwise, the classification accuracy of the classifier is too low, and even the whole classifier model fails. By introducing lagrangian function solution formula 12, the optimally classified structural hyperplane is converted into a convex quadratic programming problem:
α in equation 12iIs a Lagrange multiplier, c is a penalty factor, phi (x)i)T·φ(xj) I.e. kernel function, with K (x)i,xj) Expressed, the decision function is expressed by equation 13 as:
in equation 13
And b*Are parameters that determine the best classification hyperplane.
In SVM, the choice of kernel functions has a large impact on SVM performance. Common kernel functions are mainly: a polynomial kernel, a Radial Basis Function (RBF) or a sigmoid kernel, etc. The RBF kernel function only comprises a parameter sigma, is relatively simple to optimize, is beneficial to parameter optimization, and is the most frequently used kernel function. Therefore, the present invention selects a radial basis kernel function (RBF), which is defined by equation 14:
σ in equation 14 is a kernel function parameter, which affects the distribution of the complexity of the sample data in the feature space.
From the analysis, it can be seen that the accuracy of the SVM classifier can be effectively improved by reasonably selecting the penalty parameter c and the RBF kernel function parameter sigma.
The gray wolf optimization algorithm (the grey wolf optizer, abbreviated as GWO) is a group intelligent optimization algorithm proposed in 2014 by simulating the natural gray wolf population leader level and the predation behavior mechanism, which is widely applied to the problems of functional optimization and the like because of low complexity and few control parameters, and widely accepted by simulating the top of the food chain in nature, the gray wolf algorithm GWO references the strict social level system inside the gray wolf, the population is generally divided into four levels α, β and omega, thereby forming a hierarchy of duty tower structure, the gray wolf at the first level of the King wolf tower is the top wolf α, the main decision-making work of the gray wolf is good, all major problems of the whole wolf group are weighted, the gray wolf at the second level is 2, the gray wolf at the first level is the top wolf lead α, the main decision making work of the whole wolf group is good, the important problems of the whole wolf group are solved by the three-level gray wolf search algorithm models, namely the gray wolf search target gray wolf trap location is represented by the following formula 5, the initial gray wolf search algorithm which is the highest when the gray wolf search target of the gray wolf, the gray wolf is found by the three-level of the gray wolf search algorithm, the gray wolf cluster, the gray wolf search algorithm, the gray wolf is represented by the initial search algorithm, the final target of the gray wolf cluster, the gray wolf search algorithm, the gray wolf search target of the gray wolf cluster, the gray wolf search algorithm, the gray wolf, the third stroke of the gray wolf cluster, the third stroke, the gray wolf cluster, the target stroke of the third stroke, the third stroke of the third stroke, the target stroke, the third stroke, the stroke of the stroke, the stroke of the stroke, the stroke of the stroke, the stroke of the stroke, the stroke of the stroke:
D＝|C·Xp(t) -X (t) | equation 15
X(t+1)＝Xp(t) -A.D equation 16 where t represents the current iteration number, A and C are coefficient vectors, XpIs the location vector of the prey, X is the location vector of the wolf, a and C are expressed by equation 17 and equation 18 as follows:
A＝2a·r1a formula 17
C＝2·r2Equation 18
R in equations 17 and 181And r2Are all [0,1]A is a convergence factor that decreases linearly from 2 to 0 with the number of iterations.
The position X of the remaining wolf ω in the population is determined by the positions of α and is expressed by equation 19 and equation 20 as follows:
X(t+1)＝(X1+X2+X3) Equation/3 20
However, GWO algorithm has disadvantages in the optimization process, such as slow convergence speed in the later period, weak local search capability, etc., the present embodiment proposes an IGWO-SVM model, which introduces inertial weight w and "flying" speed v in PSO algorithm to update the gray wolf position, and when the wolf α leads β and the wolf surrounds the prey, the inertial weight w is used to update its position, and the specific content is shown in formula 21:
inertial weight in equation 21
rand () represents a random number between 0 and 1;
the position X of the remaining gray wolf ω in the population is determined by α and the position X is shown in equations 22 and 23:
X(t+1)＝posi(t+1)＝posi(t)+Vi(t +1) formula 23
In this embodiment, fig. 3 is a schematic flow chart of a fault diagnosis model according to an embodiment of the present invention, and fig. 4 is a flow chart of a fault diagnosis method according to an embodiment of the present invention, as shown in fig. 3 and fig. 4, the method specifically includes the following steps:
step S402, inputting a vibration signal and extracting the decomposed modal component by using a CEEMADN algorithm;
s404, selecting IMFs 1-3, solving Shannon energy entropy for the IMFs, normalizing the data extracted by the complete features, and equally dividing the data into a test set and a training set;
and S406, optimizing the parameters of the model input into the IGWO-SVM by the training set, performing ten-fold cross validation for 50 times, and finally selecting the statistical average as the final result.
The implementation method of step S406 is as follows:
step S501, initializing a wolf population, initializing IGWO algorithm parameters a, A, C, w and V, and randomly generating a wolf population position;
step S502, calculating the initial fitness of each gray wolf individual, and selecting α gray wolfs;
step S503, updating the wolf group position, and updating the current wolf position according to the formula 22 and the formula 23;
step S504, updating IGWO algorithm parameters a, A, C, w and V;
step S505, calculating the updated fitness value of each gray wolf and re-determining a new value α;
step S506, comparing the current iteration times with the maximum allowable iteration times, and if the current iteration times are not reached, jumping to step S503 to continue the iteration; otherwise, ending to obtain the global optimal position.
In this embodiment, a bearing fault diagnosis device is further provided, which is used to implement the above embodiments and preferred embodiments, and the description of the device is omitted. As used below, the term "module" may be a combination of software and/or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementations in hardware or a combination of software and hardware are also possible and contemplated.
In an embodiment, there is also provided a bearing fault diagnosis apparatus, and fig. 5 is a block diagram of a bearing fault diagnosis apparatus according to an embodiment of the present invention, as shown in fig. 5, the apparatus includes:
the obtaining module 52 is configured to obtain a vibration signal of the bearing, and extract an IMF modal component of the vibration signal through a complete integrated empirical mode decomposition method CEEMDAN of adaptive white noise;
a constructing module 54, configured to determine an energy entropy of the modal component, perform normalization processing on the energy entropy as a feature vector, and construct a training set and a test set of a support vector machine SVM;
a parameter optimization module 56, configured to update the penalty parameter and the kernel function parameter of the classifier of the SVM according to the training set and the swarm intelligence optimization algorithm;
and a diagnostic module 58, configured to determine a fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter.
Through the device, the construction module 54 and the parameter optimization module 56 optimize punishment parameters and kernel functions of the SVM classifier through a group intelligent optimization algorithm, so that the SVM is converged faster and classified optimally, the problems of low efficiency and low precision of bearing fault diagnosis are solved, and the efficiency and the precision of bearing fault diagnosis are improved.
The parameter optimization module 56 is further configured to update the grey wolf position in the grey wolf optimization algorithm GWO through the inertia weight and the "flight speed" in the particle swarm optimization PSO; and performing multiple iterations according to the updated wolf location to determine the penalty parameter and the kernel function parameter of the classifier of the SVM.
The construction module 54 is further configured to extract a preset number of modal components of the vibration signal by a complete integrated empirical mode decomposition method CEEMDAN of adaptive white noise; and determining Shannon energy entropy according to the preset number of modal components.
In order to verify the feasibility and the effectiveness of the method provided by the embodiment of the invention, the rolling bearing data of the electrical engineering laboratory of Kaiser University of western province (CWRU) is adopted to verify the method, and the superiority of the method is emphasized by comparing different parameter optimization methods.
In the embodiment of the invention, a bearing vibration signal is collected at a rotating speed of 1797r/min as an analysis data sample, vibration data under the whole working condition is obtained at a sampling frequency of 12kH, a single-point fault is arranged on the bearing by using an electric spark machining technology in an experiment, the fault width diameter is 0.1778mm, the fault depth is 0.2794mm, and vibration signals of 4 states, namely normal, inner ring fault, outer ring fault and rolling element fault, are collected.
Taking a rolling element fault signal as an example, fig. 6 is a waveform schematic diagram of a rolling element fault CEEMDAN according to an embodiment of the present invention after decomposition, as shown in fig. 6, the first 3 decomposed IMF components include IMF1, IMF2, and IMF3, Shannon energy entropy is respectively obtained for the first three IMFs of the four vibration signals, each 200 groups of each type of samples are taken, 1/2 is taken as a test set, and 1/2 is taken as a training set. In the embodiment of the invention, preprocessed bearing data samples of four types including normal bearing data, inner bearing data, outer bearing data and rolling element data are respectively numbered as 0,1 bearing data, 2 bearing data and 3 bearing data, Genetic Algorithm (GA for short), PSO Algorithm, GWO Algorithm and IGWO Algorithm are selected to respectively optimize parameters c and sigma of an SVM classifier, each Algorithm is tested for 50 times, and in addition, the initial population N of the GA/PSO/GWO/IGWO is the same as the iteration times t and is 20/100 respectively.
In order to evaluate the performance of the SVM classification model, 3 statistical data of the average accuracy, the average optimization time and the longest and shortest optimization time difference of the test set are counted in the test as the judgment standard, wherein Δ T is the longest optimization time-the shortest optimization time, and the classification results of the four algorithms after ten-fold cross validation are shown in table 1.
TABLE 1
Through the analysis of the table 1, although the PSO-SVM model is most stable and the optimization time is the minimum, the accuracy of a test set is 93.50 percent, the average optimization time is 130.76s, and obviously the recognition accuracy and efficiency are the worst; although the accuracy of the GA-SVM is slightly improved to 94.27% compared with that of the PSO-SVM, the time range is longest and the GA-SVM is most unstable; GWO-SVM has better performance compared with the former two models, no matter the average accuracy, the optimizing average time or the optimizing time is extremely poor, the feasibility and the superiority of the GWO algorithm on SVM parameter optimizing in bearing fault diagnosis are verified, but the GWO algorithm is easy to fall into local optimum, and the defect of large optimizing time extreme difference is also caused. Compared with the GWO-SVM model, the IGWO-SVM model has the advantages that various indexes are improved, the average accuracy is improved by 0.03%, the average time is reduced by 9.37s, and the optimization time range is reduced by 9.76 s. Compared with the comprehensive performance, the IGWO-SVM has the advantages of higher convergence rate, easier achievement of optimal classification, stronger judgment capability on bearing faults and higher identification accuracy.
In one embodiment, a computer device is provided, which may be a server, and fig. 7 is a block diagram of the inside of a computer device according to an embodiment of the present invention, as shown in fig. 7. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing channel corresponding relation table data. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement the above-described bearing fault diagnosis method.
According to another aspect of the present invention, there is also provided a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the above-described method for diagnosing a bearing fault.
It will be understood by those of ordinary skill in the art that all or a portion of the processes of the methods of the embodiments described above may be implemented by a computer program that may be stored on a non-volatile computer-readable storage medium, which when executed, may include the processes of the embodiments of the methods described above, wherein any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of diagnosing a bearing fault, the method comprising:
obtaining a vibration signal of a bearing, and extracting an IMF modal component of the vibration signal by a complete integration empirical mode decomposition (CEEMDAN) method of self-adaptive white noise;
determining the energy entropy of the modal component, performing normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM);
updating punishment parameters and kernel function parameters of a classifier of the SVM according to the training set and the group intelligent optimization algorithm;
and the SVM determines a fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter.
2. The method of claim 1, wherein updating penalty parameters and kernel function parameters of a classifier of the SVM according to the training set and group intelligence optimization algorithm comprises:
updating the grey wolf position in the grey wolf optimization algorithm GWO through the inertia weight and the flight speed in the particle swarm optimization PSO;
and performing multiple iterations according to the updated wolf location to determine a penalty parameter and a kernel function parameter of the classifier of the SVM.
3. The method of claim 1, wherein the extracting modal components of the vibration signal by a fully integrated empirical mode decomposition (CEEMDAN) method of adaptive white noise, and wherein determining the energy entropy of the modal components comprises:
extracting modal components of a preset number of the vibration signals by a complete integration empirical mode decomposition method CEEMDAN of self-adaptive white noise;
and determining Shannon energy entropy according to the modal components of the preset number.
4. The method of claim 1, wherein the SVM determining a fault result for the bearing based on the test set, the penalty parameter, and the kernel function parameter comprises:
and the SVM counts the average accuracy, the average optimization time and the longest and shortest optimization time difference of the test set to serve as judgment standards, and the fault result of the bearing is determined after ten-fold cross validation.
5. The method according to any of claims 1 to 4, wherein the kernel function is a radial basis kernel function (RBF).
6. A diagnostic device for bearing failure, said device comprising:
the system comprises an acquisition module, a detection module and a processing module, wherein the acquisition module is used for acquiring a vibration signal of a bearing and extracting an IMF modal component of the vibration signal by a CEEMDAN method of complete integration of self-adaptive white noise;
the construction module is used for determining the energy entropy of the modal component, carrying out normalization processing by taking the energy entropy as a feature vector, and constructing a training set and a test set of a Support Vector Machine (SVM);
the parameter optimization module is used for updating punishment parameters and kernel function parameters of the classifier of the SVM according to the training set and the cluster intelligent optimization algorithm;
and the diagnosis module is used for determining the fault result of the bearing according to the test set, the penalty parameter and the kernel function parameter by the SVM.
7. The apparatus of claim 6, wherein the parameter optimization module is further configured to update the grayish wolf position in the grayish wolf optimization algorithm GWO through the inertia weight and the "flight speed" in the particle swarm optimization PSO;
and performing multiple iterations according to the updated wolf location to determine a penalty parameter and a kernel function parameter of the classifier of the SVM.
8. The apparatus of claim 6, wherein the construction module is further configured to extract a pre-set number of modal components of the vibration signal by a fully integrated empirical mode decomposition (CEEMDAN) method of adaptive white noise; and determining Shannon energy entropy according to the modal components of the preset number.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 5 are implemented when the computer program is executed by the processor.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010127805.8A CN111428418A (en) | 2020-02-28 | 2020-02-28 | Bearing fault diagnosis method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010127805.8A CN111428418A (en) | 2020-02-28 | 2020-02-28 | Bearing fault diagnosis method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111428418A true CN111428418A (en) | 2020-07-17 |
Family
ID=71547242
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010127805.8A Pending CN111428418A (en) | 2020-02-28 | 2020-02-28 | Bearing fault diagnosis method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111428418A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132191A (en) * | 2020-09-01 | 2020-12-25 | 兰州理工大学 | Intelligent evaluation and identification method for early damage state of rolling bearing |
| CN112183439A (en) * | 2020-10-13 | 2021-01-05 | 上海明略人工智能(集团)有限公司 | Signal feature extraction method, system, computer-readable storage medium and electronic device |
| CN112258587A (en) * | 2020-10-27 | 2021-01-22 | 上海电力大学 | Camera calibration method based on wolf-wolf particle swarm hybrid algorithm |
| CN112836604A (en) * | 2021-01-22 | 2021-05-25 | 合肥工业大学 | Rolling bearing fault diagnosis and classification method, system and equipment based on VMD-SSAE and storage medium thereof |
| CN112945557A (en) * | 2021-01-29 | 2021-06-11 | 福建省特种设备检验研究院 | Slewing bearing fault diagnosis method and device and storage medium |
| CN113063314A (en) * | 2021-03-23 | 2021-07-02 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113191232A (en) * | 2021-04-21 | 2021-07-30 | 西安交通大学 | Electro-hydrostatic actuator fault identification method based on multi-mode homologous features and XGboost model |
| CN113191477A (en) * | 2021-01-29 | 2021-07-30 | 河北建投新能源有限公司 | Fault diagnosis method and device for temperature sensor of basic electrolytic cell |
| CN113203565A (en) * | 2021-03-25 | 2021-08-03 | 长江大学 | Bearing fault identification method and system based on EEMD sparse decomposition |
| CN113571136A (en) * | 2021-06-08 | 2021-10-29 | 广州地铁设计研究院股份有限公司 | Polycyclic aromatic hydrocarbon prediction method, device, medium and equipment |
| CN113642508A (en) * | 2021-08-27 | 2021-11-12 | 中国航空工业集团公司上海航空测控技术研究所 | Bearing fault diagnosis method based on parameter self-adaptive VMD and optimized SVM |
| CN113836802A (en) * | 2021-09-13 | 2021-12-24 | 上海工业自动化仪表研究院有限公司 | Gas turbine sensor fault diagnosis method based on MFO-SVM |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107193212A (en) * | 2017-06-26 | 2017-09-22 | 南京航空航天大学 | Aero-engine non-linear predication control method based on new grey wolf optimized algorithm |
| CN108694390A (en) * | 2018-05-15 | 2018-10-23 | 南京邮电大学 | A kind of cuckoo search improves the modulated signal sorting technique of grey wolf Support Vector Machines Optimized |
| CN109242139A (en) * | 2018-07-23 | 2019-01-18 | 华北电力大学 | A kind of electric power day peak load prediction technique |
| CN110132596A (en) * | 2019-04-24 | 2019-08-16 | 昆明理工大学 | A method of the rolling bearing fault diagnosis based on wavelet packet and GWO-SVM |
| CN110333462A (en) * | 2019-08-08 | 2019-10-15 | 首都师范大学 | A kind of lithium ion battery life-span prediction method under random discharge environment based on DGWO-ELM |
-
2020
- 2020-02-28 CN CN202010127805.8A patent/CN111428418A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107193212A (en) * | 2017-06-26 | 2017-09-22 | 南京航空航天大学 | Aero-engine non-linear predication control method based on new grey wolf optimized algorithm |
| CN108694390A (en) * | 2018-05-15 | 2018-10-23 | 南京邮电大学 | A kind of cuckoo search improves the modulated signal sorting technique of grey wolf Support Vector Machines Optimized |
| CN109242139A (en) * | 2018-07-23 | 2019-01-18 | 华北电力大学 | A kind of electric power day peak load prediction technique |
| CN110132596A (en) * | 2019-04-24 | 2019-08-16 | 昆明理工大学 | A method of the rolling bearing fault diagnosis based on wavelet packet and GWO-SVM |
| CN110333462A (en) * | 2019-08-08 | 2019-10-15 | 首都师范大学 | A kind of lithium ion battery life-span prediction method under random discharge environment based on DGWO-ELM |
Non-Patent Citations (2)
| Title |
|---|
| QASEM AL-TASHI, ET AL: "Binary Optimization Using Hybrid Grey Wolf Optimization for Feature Selection", 《IEEE ACCESS》 * |
| 卓仁雄: "基于CEEMDAN和GWO-SVM的电机滚动轴承故障诊断", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112132191A (en) * | 2020-09-01 | 2020-12-25 | 兰州理工大学 | Intelligent evaluation and identification method for early damage state of rolling bearing |
| CN112183439A (en) * | 2020-10-13 | 2021-01-05 | 上海明略人工智能(集团)有限公司 | Signal feature extraction method, system, computer-readable storage medium and electronic device |
| CN112258587A (en) * | 2020-10-27 | 2021-01-22 | 上海电力大学 | Camera calibration method based on wolf-wolf particle swarm hybrid algorithm |
| CN112258587B (en) * | 2020-10-27 | 2023-07-07 | 上海电力大学 | Camera calibration method based on gray wolf particle swarm mixing algorithm |
| CN112836604A (en) * | 2021-01-22 | 2021-05-25 | 合肥工业大学 | Rolling bearing fault diagnosis and classification method, system and equipment based on VMD-SSAE and storage medium thereof |
| CN112945557A (en) * | 2021-01-29 | 2021-06-11 | 福建省特种设备检验研究院 | Slewing bearing fault diagnosis method and device and storage medium |
| CN112945557B (en) * | 2021-01-29 | 2024-01-12 | 福建省特种设备检验研究院 | Slewing bearing fault diagnosis method, slewing bearing fault diagnosis device and storage medium |
| CN113191477A (en) * | 2021-01-29 | 2021-07-30 | 河北建投新能源有限公司 | Fault diagnosis method and device for temperature sensor of basic electrolytic cell |
| CN113063314B (en) * | 2021-03-23 | 2022-03-22 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113063314A (en) * | 2021-03-23 | 2021-07-02 | 哈尔滨工程大学 | Fault diagnosis method for gun launching system based on SVM (support vector machine) and GA-SVM (genetic algorithm-support vector machine) |
| CN113203565A (en) * | 2021-03-25 | 2021-08-03 | 长江大学 | Bearing fault identification method and system based on EEMD sparse decomposition |
| CN113191232A (en) * | 2021-04-21 | 2021-07-30 | 西安交通大学 | Electro-hydrostatic actuator fault identification method based on multi-mode homologous features and XGboost model |
| CN113571136A (en) * | 2021-06-08 | 2021-10-29 | 广州地铁设计研究院股份有限公司 | Polycyclic aromatic hydrocarbon prediction method, device, medium and equipment |
| CN113642508A (en) * | 2021-08-27 | 2021-11-12 | 中国航空工业集团公司上海航空测控技术研究所 | Bearing fault diagnosis method based on parameter self-adaptive VMD and optimized SVM |
| CN113642508B (en) * | 2021-08-27 | 2024-04-09 | 中国航空工业集团公司上海航空测控技术研究所 | Bearing fault diagnosis method based on parameter self-adaptive VMD and optimized SVM |
| CN113836802A (en) * | 2021-09-13 | 2021-12-24 | 上海工业自动化仪表研究院有限公司 | Gas turbine sensor fault diagnosis method based on MFO-SVM |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111428418A (en) | Bearing fault diagnosis method and device, computer equipment and storage medium | |
| Van et al. | Bearing‐fault diagnosis using non‐local means algorithm and empirical mode decomposition‐based feature extraction and two‐stage feature selection | |
| Pratama et al. | pClass: an effective classifier for streaming examples | |
| Karaboğa et al. | Training ANFIS by using the artificial bee colony algorithm | |
| US11790237B2 (en) | Methods and apparatus to defend against adversarial machine learning | |
| Chen et al. | Convolutional neural network for classifying space target of the same shape by using RCS time series | |
| Zhou et al. | Stacked blockwise combination of interpretable TSK fuzzy classifiers by negative correlation learning | |
| CN110232416A (en) | Equipment failure prediction technique based on HSMM-SVM | |
| CN114254668A (en) | Fault detection method and device based on discharge signal of insulated switchgear | |
| Wang et al. | An improved bearing fault detection strategy based on artificial bee colony algorithm | |
| JP6172317B2 (en) | Method and apparatus for mixed model selection | |
| Rawat et al. | A soft computing genetic-neuro fuzzy approach for data mining and its application to medical diagnosis | |
| Sreekumar et al. | Performance enhancement of the machine‐fault diagnosis system using feature mapping, normalisation and decision fusion | |
| CN116756594A (en) | Method, system, equipment and medium for detecting abnormal points of power grid data | |
| Schrunner et al. | A comparison of supervised approaches for process pattern recognition in analog semiconductor wafer test data | |
| Zhou et al. | ptse: A multi-model ensemble method for probabilistic time series forecasting | |
| Li et al. | Development of a global batch clustering with gradient descent and initial parameters in colour image classification | |
| Hasanbelliu et al. | Online learning using a Bayesian surprise metric | |
| CN114254686A (en) | Method and device for identifying confrontation sample | |
| Zhang et al. | Fault diagnosis model based on fuzzy support vector machine combined with weighted fuzzy clustering | |
| Senthilnath et al. | A novel harmony search-based approach for clustering problems | |
| Yang et al. | A hybrid gene selection and classification approach for microarray data based on clustering and PSO | |
| Płoński et al. | Self-organising maps for classification with metropolis-hastings algorithm for supervision | |
| CN112115268B (en) | Training method and device based on feature encoder, and classifying method and device | |
| CN111949530B (en) | Test result prediction method and device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200717 |