CN112347854B - Rolling bearing fault diagnosis method, system, storage medium, equipment and application - Google Patents
Rolling bearing fault diagnosis method, system, storage medium, equipment and application Download PDFInfo
- Publication number
- CN112347854B CN112347854B CN202011083925.9A CN202011083925A CN112347854B CN 112347854 B CN112347854 B CN 112347854B CN 202011083925 A CN202011083925 A CN 202011083925A CN 112347854 B CN112347854 B CN 112347854B
- Authority
- CN
- China
- Prior art keywords
- whale
- carrying
- rolling bearing
- neighborhood
- lssvm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000005096 rolling process Methods 0.000 title claims abstract description 43
- 238000003860 storage Methods 0.000 title claims abstract description 8
- 241000283153 Cetacea Species 0.000 claims abstract description 113
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 46
- 239000013598 vector Substances 0.000 claims abstract description 37
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 34
- 238000012360 testing method Methods 0.000 claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000004891 communication Methods 0.000 claims abstract description 9
- 238000012163 sequencing technique Methods 0.000 claims abstract description 7
- 239000003795 chemical substances by application Substances 0.000 claims description 27
- 238000005457 optimization Methods 0.000 claims description 19
- 230000006870 function Effects 0.000 claims description 18
- 238000011161 development Methods 0.000 claims description 6
- 238000006073 displacement reaction Methods 0.000 claims description 6
- 238000012706 support-vector machine Methods 0.000 claims description 5
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 4
- 230000006872 improvement Effects 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 abstract 1
- 238000004458 analytical method Methods 0.000 description 5
- 230000007547 defect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000012843 least square support vector machine Methods 0.000 description 5
- 238000001228 spectrum Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000013135 deep learning Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 239000002245 particle Substances 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 230000001174 ascending effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 241000282461 Canis lupus Species 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000012634 fragment Substances 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000005469 granulation Methods 0.000 description 1
- 230000003179 granulation Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000011159 matrix material Substances 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
- 230000003287 optical effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000002028 premature Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000011273 social behavior Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- 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
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Acoustics & Sound (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Signal Processing (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention belongs to the technical field of bearing vibration signal identification, and discloses a rolling bearing fault diagnosis method, a system, a storage medium, equipment and application, wherein original signals of four states of a bearing are collected, and signal decomposition is carried out by utilizing a VMD (virtual machine tool) to obtain each IMF (inertial measurement unit) component; extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample; initializing population scale, iteration times and self-adaptive weight values of a whale algorithm; establishing an LSSVM model by using the initialization parameters; calculating a corresponding fitness value of each whale, and sequencing according to the fitness value; carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the neighborhood, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula; and outputting whale positions with the best fitness as parameters of the LSSVM for training, and carrying out fault classification on the test set. The invention has better fault classification performance and higher accuracy.
Description
Technical Field
The invention belongs to the technical field of bearing vibration signal identification, and particularly relates to a rolling bearing fault diagnosis method, a system, a storage medium, equipment and application.
Background
At present: the rolling bearing is used as an important component of rotary mechanical equipment and is one of the parts most prone to faults, due to the fact that working environment is poor and resonance and other problems are prone to occurring, the rolling bearing has important significance in timely and accurate detection and diagnosis, the characteristics of signals of the rolling bearing are difficult to identify, the current common diagnosis method is that vibration signals are firstly collected through sensors, then the vibration signals are processed through signal processing methods such as Fourier transformation and wavelet transformation, the characteristics of the vibration signals are extracted, mode identification and classification diagnosis are finally carried out, an empirical mode decomposition method (Empirical Mode Decomposition, EMD) has self-adaption, the method is very suitable for analyzing non-stable and nonlinear data, however, the EMD has serious mode aliasing phenomenon, in order to solve the problem, wu and Huang propose a novel integrated mode empirical decomposition method (Ensemble Empirical Mode Decomposition, EEMD), the problem of aliasing is ingeniously solved by adding white noise on original signals, the defect that the number of sets of white noise is difficult to select, the problem is 35 Konstantin Dragomiretskiy, the problem of decomposition (VMD) is solved, the problem of the optimal mode aliasing is solved, the problem of the frequency aliasing problem is solved, the optimal decomposition mode is solved, the problem is solved, and the frequency aliasing problem is not solved, and the problem is difficult to be the problem is solved, such that the problem of determining that the decomposition mode has no optimal decomposition effect is solved, and the problem is solved by an optimal.
The first step of fault diagnosis is to extract effective information from vibration signals, entropy is taken as a tool for describing fault complexity, and in recent years, the fault diagnosis method based on entropy has fuzzy entropy, permutation entropy, multi-scale permutation entropy and the like, and the multi-scale permutation entropy can observe a time sequence from different angles, so that effective information is extracted, robustness is improved, and classification capability of a classifier is improved [9] Therefore, the invention extracts fault characteristics by using variational modal decomposition and multi-scale permutation entropy,
the support vector machine has better fault diagnosis result under the condition of small samples, but has long training time, fast LSSVM learning speed and excellent classification performance, is used by students in power transformer fault diagnosis modeling to obtain better effect, but parameters directly influence the power transformer fault diagnosis effect, wang Tiantian optimizes the parameters of the penalty factors and the kernel functions of the wavelet kernel extreme learning machine by using an improved gray wolf algorithm, effectively improves the classification precision, introduces disturbance terms in the traditional particle swarm, introduces chaos ideas in the particle swarm in the Hitachi, enhances the diversity and global searching capability of the particle swarm,
through the above analysis, the problems and defects existing in the prior art are as follows:
(1) At present, fault characteristics in a bearing vibration signal are difficult to extract.
(2) The problem of least square support vector machine model parameter optimization in the prior bearing fault diagnosis.
The difficulty of solving the problems and the defects is as follows:
(1) Complexity of background noise: the environment of the field of the general equipment is quite complex, a large number of interference signals are accompanied around the field when the machine works, and the signals have periodicity and have strong randomness. Sometimes the acoustic signal emitted by the device is even swamped by the ambient noise, so it is still very difficult to extract the fault signal from the complex environment.
(2) The least square support vector machine model has the defects that the computational complexity is approximately three times of the number of samples, and the computational complexity is very large.
The meaning of solving the problems and the defects is as follows:
(1) By analyzing the vibration signals of the bearing, fault feature extraction of the vibration signals and identification of the bearing state are researched by using methods such as time-frequency analysis and the like, and early fault detection and fault feature identification of the rolling bearing are realized.
(2) By optimizing the parameters of the least square support vector machine, the model prediction precision can meet the requirement, the operation amount is reduced, and the calculation time is shortened.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a rolling bearing fault diagnosis method, a system, a storage medium, equipment and application.
The invention is realized in that a rolling bearing fault diagnosis method comprises the following steps:
and collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing the VMD to obtain each IMF component. VMD isDetermining modality u by iteratively searching for a variational model optimal solution k (t) and its corresponding center frequency w k And bandwidth. Each mode is a finite bandwidth with a center frequency (i.e., a certain width in the frequency domain), and the sum of all modes is the source signal. And converting the constraint problem into the non-constraint problem by adopting a quadratic penalty and a Lagrangian multiplier for solving the optimal solution, solving the non-constraint problem by adopting an alternate direction multiplier method, and finally obtaining all modes of signal decomposition through iterative updating. Among all modes of the decomposition are a mode containing a main signal and a mode containing noise. And reconstructing the mode containing the main signals, thereby achieving the denoising effect.
And extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample. The multi-scale permutation entropy is an improvement of permutation entropy and is characterized in that permutation entropy signals are coarsened, coarse-grained fragments are calculated respectively, and therefore description of one-dimensional signals in multiple dimensions is achieved. Firstly, coarsening a time sequence, and performing permutation entropy calculation under each subsequence after the coarsening process is completed. Each sub-sequence is subjected to phase space reconstruction, and the time sequence of the sub-sequence is calculated. And then, arranging the time reconstruction sequences in an ascending order, and calculating the occurrence times of each arrangement type and the corresponding frequency thereof.
Initializing whale algorithm population scale, maximum iteration number and adaptive weight value, and upper position boundary ub= [ ub ] of whale population 1 ,ub 2 ,…,ub n ]And lower bound lb= [ lb 1 ,lb 2 ,…,lb n ]. Whale populations may be represented by the following matrix;
wherein w represents the position of the whale population, w i,j The position size of the ith whale position in the j-th dimension is represented. w (w) i,j The update can be performed with the following formula:
w i,j ＝(ub(i)-lb(i))×rand(i.j)+lb(i)
wherein: w (w) i,j Representing the size of the position of the ith whale in the j-th dimension, lb (i) and ub (i) represent the upper and lower bounds, respectively, of the ith whale position, and rand (i.j) represents a randomly generated number between 0 and 1.
Establishing an LSSVM model by using the initialization parameters: firstly, setting the value ranges of two parameters, wherein the ranges of penalty factors and kernel parameters are [0.01,1000] and [0.01,100], the whale population size is set to 30, and the maximum iteration number is 100.
Calculating a corresponding fitness value of each whale, sorting according to the fitness value, and selecting N whales as a next generation population: and selecting 180 groups of the determined 300 groups of data for model training, reserving the rest 120 groups as a test set, calculating the fitness of each whale through a fitness function, comparing, finding out an individual with the optimal fitness value, and determining the position, the optimal penalty factor and the nuclear parameter at the moment.
Carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the field, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula: each whale with a seat head is provided with a grid form formed by 4 neighbors up, down, left and right, exchanges information with the whales with the seat head up, down, left and right, one whale with the seat head finds out the whales around the influence of the optimal solution command, realizes the full utilization of information of each whale in the population, guides the population to evolve towards multiple directions so as to maintain the diversity of the population, further avoids the phenomenon of premature, and has global property, and the convergence speed and precision are also ensured.
Repeatedly calculating the corresponding fitness value of each whale, carrying out neighborhood search by adopting a von Neumann topological structure until the maximum iteration number is reached, outputting whale positions with the best fitness as parameters of the LSSVM for training, and carrying out fault classification on the test set.
Further, the formula for updating the position of the rolling bearing fault diagnosis method is as follows:
w＝1-exp(1-t)；
the local optimum in the von neumann grid is updated by using the global optimum corresponding to the current iteration times, and the local optimum can be accelerated to gradually coincide with the global optimum by adopting an adaptive weight method along with the progress of iteration.
The position of whale was updated as follows:
wherein: d' = |x * (t) -X (t) represents the distance from the ith whale to the prey, b is a constant defining a logarithmic spiral shape, and l is [ -1,1]Random numbers in (a);
the location update equation is shown below:
D＝|CX rand (t)-X(t)|；
X(t+1)＝X rand (t)-AD；
wherein: x is X rand Representing a random position vector selected from the current population, reducing the parameter a from 2 to 0 for development and exploration, respectively, and selecting A as a random search agent when |A| > 1; when |A| < 1, A is selected as the optimal solution for updating the location of the search agent;
the location update formula is as follows:
w＝1-exp(1-t)；
wherein: w=0.5;the optimal position of the whale search agent i in the Von Neumann topological neighborhood L (i) is obtained, namely the historical optimal position of each whale search agent in the neighborhood corresponding to the individual is obtained by the minimum value of the fitness function of each whale search agent; g represents a global optimum position.
Furthermore, the rolling bearing fault diagnosis method is characterized in that the whale optimization algorithm is improved by using a von Neumann topological structure and self-adaptive weights, the self-adaptive weights are introduced into a displacement formula, along with iterative updating, the local optimum gradually coincides with the global optimum, and the position updating formula is as follows:
w＝1-exp(1-t)；
wherein: w=0.5;the optimal position of the whale search agent i in the Von Neumann topological neighborhood L (i) is obtained, namely the historical optimal position of each whale search agent in the neighborhood corresponding to the individual is obtained by the minimum value of the fitness function of each whale search agent; g represents a global optimum position.
It is a further object of the present invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of:
collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing a VMD to obtain each IMF component;
extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample;
initializing population scale, iteration times and self-adaptive weight values of a whale algorithm;
establishing an LSSVM model by using the initialization parameters;
calculating a corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as a next generation population;
carrying out neighborhood search by adopting a von neumann topological structure, carrying out information exchange in the field, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula;
repeatedly calculating the corresponding fitness value of each whale, carrying out neighborhood search by adopting a von Neumann topological structure until the maximum iteration number is reached, outputting whale positions with the best fitness as parameters of the LSSVM for training, and carrying out fault classification on the test set.
Another object of the present invention is to provide a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of:
collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing a VMD to obtain each IMF component;
extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample;
initializing population scale, iteration times and self-adaptive weight values of a whale algorithm;
establishing an LSSVM model by using the initialization parameters;
calculating a corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as a next generation population;
carrying out neighborhood search by adopting a von neumann topological structure, carrying out information exchange in the field, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula;
repeatedly calculating the corresponding fitness value of each whale, carrying out neighborhood search by adopting a von Neumann topological structure until the maximum iteration number is reached, outputting whale positions with the best fitness as parameters of the LSSVM for training, and carrying out fault classification on the test set.
Another object of the present invention is to provide a rolling bearing failure diagnosis system implementing the rolling bearing failure diagnosis method, the rolling bearing failure diagnosis system comprising:
the IMF component acquisition module is used for acquiring original signals of the bearing in four states, and performing signal decomposition by utilizing the VMD to obtain each IMF component;
the feature vector set construction module is used for extracting signal features by using multi-scale permutation entropy, constructing a feature vector set and dividing the feature vector set into a training sample and a test sample;
the initialization module is used for initializing population scale, iteration times and self-adaptive weight values of the whale algorithm;
the model building module is used for building an LSSVM model by using the initialization parameters;
the fitness value calculation module is used for calculating the corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as the population of the next generation;
the position updating module is used for carrying out neighborhood searching by adopting a von neumann topological structure, carrying out information communication in the field, finding out the optimal whale in the neighborhood, and carrying out position updating according to a formula;
and the fault classification module is used for outputting whale positions with the best fitness as parameters of the LSSVM for training and carrying out fault classification on the test set.
Another object of the present invention is to provide an automotive rotating platform, which is equipped with the rolling bearing fault diagnosis system.
Another object of the present invention is to provide a rotary restaurant platform mounted with the rolling bearing fault diagnosis system.
Another object of the present invention is to provide a rotary advertising platform mounted with the rolling bearing fault diagnosis system.
Another object of the present invention is to provide a rotary electric windmill equipped with the rolling bearing failure diagnosis system.
By combining all the technical schemes, the invention has the advantages and positive effects that: the invention adopts a fault classification method based on an improved whale algorithm optimized least square support vector machine model, adopts variation modal decomposition and multi-scale permutation entropy to extract signal fault characteristics, aims at the problems of slow convergence speed and low precision of whale algorithm (Whale Optimization Algorithm, WOA), introduces von Neumann topological structure and self-adaptive weight to improve, and can properly adjust the balance between global searching capability and local searching capability; the improved whale algorithm is adopted to optimize the parameters and penalty factors of the LSSVM kernel function, a rolling bearing fault diagnosis model is established, and the result shows that the fault classification performance of the method is better and the accuracy is higher. Aiming at the problem of optimizing the parameters of a least square support vector machine model in the prior bearing fault diagnosis, the invention provides a fault diagnosis method for optimizing an LSSVM model based on an improved whale algorithm, which comprises the steps of firstly decomposing a vibration signal into a plurality of IMF components by utilizing variation modal decomposition, and calculating the multiscale permutation entropy of each IMF component; then, inputting the feature vector into an improved whale algorithm optimized LSSVM model, and establishing a bearing fault diagnosis model; finally, the method is compared and analyzed with other five methods, the feasibility and the effectiveness of the method are verified,
drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following description will briefly explain the drawings needed in the embodiments of the present application, and it is obvious that the drawings described below are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a rolling bearing fault diagnosis method provided by an embodiment of the present invention.
FIG. 2 is a schematic diagram of a fault diagnosis system for a rolling bearing according to an embodiment of the present invention;
in fig. 2: 1. an IMF component acquisition module; 2. the feature vector set construction module; 3. initializing a module; 4. a model building module; 5. the fitness value calculating module; 6. a location update module; 7. and a fault classification module.
Fig. 3 is a schematic diagram of a topology provided by an embodiment of the present invention.
Fig. 4 is a normal state spectrogram provided by an embodiment of the present invention.
Fig. 5 is a graph of an inner ring failure spectrum provided by an embodiment of the present invention.
Fig. 6 is a graph of an outer ring failure spectrum provided by an embodiment of the present invention.
Fig. 7 is a graph of a rolling element failure spectrum provided by an embodiment of the present invention.
Fig. 8 is a diagram of a VMD decomposition diagnosis result provided by an embodiment of the present invention.
Fig. 9 is a diagram of an EMD decomposition diagnosis result provided by an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In view of the problems existing in the prior art, the present invention provides a method, a system, a storage medium, a device and an application for diagnosing faults of a rolling bearing, and the present invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the method for diagnosing the faults of the rolling bearing provided by the invention comprises the following steps:
s101: collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing a VMD to obtain each IMF component;
s102: extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample;
s103: initializing population scale, iteration times and self-adaptive weight values of a whale algorithm;
s104: establishing an LSSVM model by using the initialization parameters;
s105: calculating a corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as a next generation population;
s106: carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the neighborhood, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula;
s107: and repeating the step S105 and the step S106 until the maximum iteration times are reached, outputting the whale position with the best fitness as the parameter of the LSSVM for training, and carrying out fault classification on the test set.
Those skilled in the art may also implement other steps in the method for diagnosing a rolling bearing fault provided by the present invention, and the method for diagnosing a rolling bearing fault provided by the present invention in fig. 1 is merely a specific embodiment.
As shown in fig. 2, the rolling bearing fault diagnosis system provided by the present invention includes:
the IMF component acquisition module 1 is used for acquiring original signals of the bearing in four states, and performing signal decomposition by utilizing the VMD to obtain each IMF component.
The feature vector set constructing module 2 is used for extracting signal features by using multi-scale permutation entropy, constructing a feature vector set and dividing the feature vector set into a training sample and a test sample.
And the initialization module 3 is used for initializing the population scale, the iteration times and the self-adaptive weight value of the whale algorithm.
The model building module 4 is configured to build an LSSVM model using the initialization parameters.
The fitness value calculating module 5 is used for calculating the corresponding fitness value of each whale, sorting according to the fitness value, and selecting N whales as the population of the next generation.
And the position updating module 6 is used for carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the neighborhood, finding out the optimal whale in the neighborhood and carrying out position updating according to a formula.
The fault classification module 7 is used for outputting whale positions with the best fitness as parameters of the LSSVM for training, and performing fault classification on the test set.
The technical scheme of the invention is further described below with reference to the accompanying drawings.
1 theory basis
1.1 principle of decomposition of variation modes
After VMD decomposition, the original signal becomes several modes, and the natural mode function is defined as amplitude modulation and frequency modulation signal, as shown in formula (1):
u k (t)＝A k (t)cos(φ k (t)) (1)
wherein A is k (t) is the instantaneous amplitude, φ k (t) is the phase and φ' k (t)≥0，w k (t)＝φ' k (t) is the center frequency, the instantaneous amplitude and the instantaneous frequency change more slowly than the phase change, each mode has sparsity after VMD decomposition, the sparsity is determined by the bandwidth of the spectral domain in which the mode is positioned, and in order to estimate the bandwidth of each mode, hilbert transform is firstly carried out on each mode to obtain a single-side frequency spectrum; then adding an exponential function at the respective center frequency, and transferring each modal spectrum to a baseband; finally, the bandwidth is estimated by calculating the square norm of the gradient of the demodulated signal, the variation constraint problem being that, under the constraint that the sum of each modal component is equal to the signal:
in the formula, delta is Dirac distribution, which represents convolution operation, in order to obtain the optimal solution of the problem, a quadratic penalty term alpha and a Lagrangian multiplier lambda are introduced to convert the variation constraint problem into an unconstrained variation problem, and the VMD converts the optimal solution of the original problem into saddle points for solving an augmented Lagrangian equation through a multiplier alternating direction method to obtain K modal components.
1.2 Multi-scale permutation entropy principle
The permutation entropy algorithm is a method for measuring the complexity of time sequence, and the algorithm is described as follows: setting a one-dimensional time sequence X= { X (1), X (2), … and X (n) }, then carrying out phase space reconstruction on X (i) by adopting a phase space reconstruction delay coordinate method, and taking m continuous sampling points of each sampling point to obtain a reconstruction vector of an m-dimensional space of the point X (i), wherein m is an embedding dimension, and l is delay time; to x% i ) The elements of the reconstruction vector X (i) are arranged in ascending order, each X (i) being mapped to D (l) = { j 1 ,j 2 ,…,j m I=1, 2, …, k (k.ltoreq.m|), and the frequency of occurrence of each case is calculated as the probability P 1 ,P 2 ,…,P k According to Shannon's theorem, the permutation entropy of the normalized time series is calculated:
in actual bearing fault diagnosis, the conventional permutation entropy does not well consider different time scales possibly existing in a time sequence, and information contained in a single scale is often incomplete, so that multi-scale analysis is needed, more hidden information is found from analysis problems of multiple time scales, and non-overlapping coarse granulation is performed on a one-dimensional time sequence to obtain the time sequence:
where τ represents a time scale factor and then phase space reconstruction delay coordinate method is used for the pairAnd carrying out phase space reconstruction.
1.3 least squares support vector machine classification principle
The LSSVM is an extension of the support vector machine that uses a linear least squares criterion of the loss function instead of the inequality constraint, the basic principle being: given a set of samplesWherein x is i ∈R n Is the sum of the input vectors y i E R is the output value of the corresponding sample i, via a non-linear mapping +.>Mapping the data of the original feature space into a high-dimensional space is as follows:
wherein: w represents the weight vector, b represents the error, and in the original space, the least squares support vector machine with equality constraints can be expressed as:
wherein the method comprises the steps of γ Is a penalty factor, e i As a relaxation variable, the lagrangian function L can be expressed as:
wherein alpha is i Is a Lagrangian multiplier, solved according to The Karush-Kuhn-Tucker (KKT) condition:
by elimination of variables w and e i An optimization problem can be converted into a linear solution problem:
wherein: q= [1, …,1] T ，A＝[α 1 ,…,α N ] T ，Y＝[y 1 ,…,y N ] T According to the Mercer condition, the kernel function can be expressed as:
the LSSVM model may be expressed as:
there are several different types of Mercer kernel functions, such as sigmoid, polynomials and Radial Basis Functions (RBFs), which are common choices of kernel functions, requiring few parameters to be set and having good overall performance, so the present invention selects RBFs as kernel functions:
therefore, the LSSVM model requires optimization of two parameters, namely, the parameter σ of the Gaussian radial basis function 2 And a penalty factor gamma.
2 parameter optimization based on whale optimization algorithm
2.1 optimization algorithm for whales
Whale optimization algorithm Whale Optimization Algorithm (WOA) is a novel natural heuristic optimization algorithm proposed by Seyedali, mirjalili and Andrew Lewis in 2016 that simulates the social behavior of whales in the seat, and for an unknown location of the optimization design in the search space, the current best candidate solution is the best solution in the target prey or near WOA algorithm, defining the best search agent, and the other search agents will attempt to update their locations to find the best search agent, the positional equation is as follows:
D＝|CX * (t)-X(t)| (13)
X(t+1)＝X * (t)-AD (14)
A＝2ar-a (15)
C＝2r (16)
wherein: t represents the current iteration number, a and C are coefficient vectors, X is the position vector of the best solution obtained, X is the position vector, r is a random number between [0,1], a decreases linearly from 2 to 0 during the iteration, and whale is killed during the development phase, assuming a 50% likelihood of choosing between the reduced surround mechanism and spiral mode to update the whale's position, the equation is as follows:
wherein: d' = |x * (t) -X (t) | represents the distance from the ith whale to the prey, b is a constant defining a logarithmic spiral shape, and l is [ -1,1]Is used for the random number in the random number code,
unlike the development phase, where the seeker whales randomly search according to each other's location during the exploration phase, the present invention uses a random value greater than 1 or less than-1 to force the seeker agent away from the reference whale, the present invention updates the seeker agent's location according to a randomly selected seeker agent instead of the best seeker agent currently found during the exploration phase, allowing the WOA algorithm to perform a global search, the location update equation is as follows:
D＝|CX rand (t)-X(t)| (18)
X(t+1)＝X rand (t)-AD (19)
wherein: x is X rand Representing a random position vector selected from the current population, reducing the parameter a from 2 to 0 for development and exploration, respectively, and selecting A as a random search agent when |A| > 1; when |A| < 1, A is selected as the optimal solution for updating the location of the search agent according to p The WOA can be switched between spiral or circular movements and finally the WOA algorithm is terminated by meeting a termination criterion.
2.2 improvement of optimization algorithm for whales
Individuals in whale shoals can communicate with other individuals to share information owned by the whales, namely, a neighborhood topology is formed among the individuals, various neighborhood topologies are adopted in the past, and different neighborhood topologies have different communication capacities, so that the performance of a final solution is further improved.
The four typical neighborhood topologies are global association, ring topology, star topology and von neumann topology respectively, wherein the convergence speed of the first three topologies is very high, but the former three topologies are prone to falling into local optimum, von neumann forms a rectangular lattice topology, as shown in fig. 3, each individual communicates with four neighbors, the optimal solution found by one individual only affects the surrounding four neighbors, thereby keeping the diversity of the individual, avoiding falling into precocity, compared with the other three topologies, the convergence speed is guaranteed, the WOA algorithm can easily solve the problem of unimodal optimization, but when the problem of peak multidimensional optimization is processed, the obtained solution is not very good, so that in order to solve the problems of slow convergence speed, low precision and the like of the WOA algorithm, the invention uses von neumann topology and self-adaptive weights to improve the whale algorithm, the position update of whales can be found by analyzing the position update formula of the whales, the position update of whales is influenced by the global optimum solution, the position update can be carried out along with the global optimum solution, the global optimum solution can be prevented from falling into precocity, the best update can be guided by the following the global optimum formula, and the optimal displacement can not be gradually overlapped with the optimal position update formula, and the optimal displacement can be led to the optimal displacement is gradually and the best update formula along with the local update formula, and the best displacement is also can be gradually led along with the local update:
w＝1-exp(1-t) (20)
wherein: w=0.5;the optimal position of the whale search agent i in the Von Neumann topological neighborhood L (i) is obtained, namely the historical optimal position of each whale search agent in the neighborhood corresponding to the individual is obtained by the minimum value of the fitness function of each whale search agent; g represents a global optimum position.
2.3 whale optimization algorithm steps
The bearing fault diagnosis method based on VMD decomposition and IWOA-LSSVM comprises the following specific steps:
step1: and collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing the VMD to obtain each IMF component.
step2: and extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample.
step3: initializing population scale, iteration times and self-adaptive weight values of a whale algorithm.
step4: and establishing an LSSVM model by using the initialization parameters.
step5: calculating the corresponding fitness value of each whale, sorting according to the fitness value, and selecting N whales as the next generation population.
step6: carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the field, finding out the optimal whale in the neighborhood, and carrying out position update according to formulas (17), (18), (19), (20) and (21).
step7: repeating the steps 5 and 6 until the maximum iteration times are reached, outputting whale positions with the best fitness as parameters of the LSSVM for training, and then carrying out fault classification on the test set.
The technical effects of the present invention will be described in detail with reference to experiments.
1. Experimental results and analysis
In order to test the performance of the IWOA-LSSVM, the invention adopts bearing data sets provided by Kassi university of America to carry out experiments and compares the bearing data sets with other five algorithms, wherein the failure diameters are respectively 0.007 inch, 0.014 inch and 0.021 inch, in the data sets, drive end bearing data (only the drive end bearing data are considered) and fan end bearing data are arranged, all the drive end bearing types are SKF 6205, vibration data are collected by an accelerometer, the sampling frequency is 12kHz, in actual fault diagnosis, the fault types of the bearings are sometimes known, the diameters of the faults are also known, in order to test the performance of the IWOA-LSSVM model for diagnosing different fault diameters, classification labels are set to be 10, namely the healthy bearing corresponds to one label, other three fault types correspond to three labels respectively, and specific class labels are shown in a table 1,
table 1 bearing failure data
/>
2. Feature extraction
When the VMD is adopted for signal decomposition, the K value of the modal decomposition number needs to be determined, when the K value is smaller, information can be lost, and when the K value is larger, the problem of modal aliasing can occur.
Table 2 modal center frequencies for different K values
The spectrograms of the normal state, the inner ring fault, the outer ring fault and the rolling body fault are shown in fig. 4 to 7, according to the intelligent fault diagnosis method provided by the invention, fault characteristic vectors are firstly extracted from original vibration signals by utilizing MPE, and the time scale and the embedding dimension are set to be 6 through the study of the time scale and the embedding dimension, so that 4 modal center frequencies are obtained by decomposing each state, and the phenomenon of decomposition does not occur.
3. Fault identification
By inputting training samples into the IWOA-LSSVM, the optimal parameter sigma is obtained 2 For 97.54 and 632.36 penalty factor gamma, the best parameters of the LSSVM model are obtained, then the feature vectors extracted from the multi-scale permutation entropy can be identified by using the IWOA-LSSVM, in order to illustrate the advantage of signal decomposition by the VMD, fig. 8 and 9 respectively show the fault diagnosis results of the VMD and the EMD in four states, wherein each label takes 30 sets of data, the first 12 sets are training samples, the second 18 sets are test samples, the specific average accuracy is shown in table 3,
TABLE 3 fault diagnosis results based on VMD and EMD decomposition
In order to test the performance of the optimization method provided by the invention, IWOA-LSSVM, GA-LSSVM, PSO-LSSVM, WOA-LSSVM, PSO-SVM and GA-SVM are subjected to classification recognition on faults, wherein in the GA-LSSVM and GA-SVM algorithms, the cross probability is 0.7, and the genetic probability is 0.05; in the PSO-LSSVM and the PSO-SVM, the minimum weight is 0.5, the maximum weight is 0.9, the local search parameter is 0.5, the global search parameter is 0.9, after the parameters are set, the model can be trained and diagnosed, the test results are shown in the table 4,
TABLE 4 bearing fault diagnosis comparative experiments
As can be seen from Table 4, the IWA-LSSVM compares the SVM model optimized by the GA and PSO algorithms, the former is superior to the latter two in diagnosis of three fault types, the improved whale algorithm optimizes LSSVM parameters, the performance of the algorithm can be improved to a certain extent, the IWA-LSSVM is compared with other three algorithm optimizes LSSVM, the IWA-LSSVM diagnosis accuracy is superior to other algorithms, and the IWA-LSSVM algorithm has certain advantages in the current fault diagnosis.
The invention mainly tests bearing fault diagnosis of a rotary machine neighborhood, firstly, fault feature extraction is carried out by utilizing multi-scale arrangement entropy, then, an improved whale algorithm is introduced into an LSSVM model for fault diagnosis, and by using a U.S. Kassi Chu Da bearing fault data set for testing, the IWOA-LSSVM fault diagnosis accuracy is found to be high, the bearing health state diagnosis accuracy reaches 100%, the inner ring fault diagnosis accuracy reaches 98.1%, the rolling body fault diagnosis accuracy reaches 96.3%, the outer ring fault diagnosis accuracy reaches 100%, the diagnosis capability meets the actual requirements of engineering, finally, the accuracy of the provided optimization algorithm and the superiority of the provided method are verified through comparison with other five algorithms, in addition, with the development of artificial intelligence, big data and deep learning, the theory and system of the three aspects are more and more perfect, and the deep learning is likely to replace machine learning in the near future, so the bearing fault diagnosis based on the deep learning becomes a research hotspot.
It should be noted that the embodiments of the present invention can be realized in hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those of ordinary skill in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The device of the present invention and its modules may be implemented by hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., as well as software executed by various types of processors, or by a combination of the above hardware circuitry and software, such as firmware.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (9)
1. A rolling bearing failure diagnosis method, characterized by comprising:
collecting original signals of the bearing in four states, and carrying out signal decomposition by utilizing a VMD to obtain each IMF component;
extracting signal characteristics by using multi-scale permutation entropy, constructing a characteristic vector set, and dividing the characteristic vector set into a training sample and a test sample;
initializing population scale, iteration times and self-adaptive weight values of a whale algorithm;
establishing an LSSVM model by using the initialization parameters;
calculating a corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as a next generation population;
carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the neighborhood, finding out the optimal whale in the neighborhood, and carrying out position update according to a formula;
repeatedly calculating the corresponding fitness value of each whale, carrying out neighborhood search by adopting a von Neumann topological structure until the maximum iteration number is reached, outputting whale positions with the best fitness as parameters of an LSSVM (least squares support vector machine) for training, and carrying out fault classification on a test set;
the formula for updating the position of the rolling bearing fault diagnosis method is as follows:
the position of whale was updated as follows:
wherein: d' = |x * (t) -X (t) | represents the distance from the ith whale to the prey, b is a constant defining a logarithmic spiral shape, and l is [ -1,1]Random numbers in (a);
the location update equation is shown below:
D＝|CX rand (t)-X(t)|；
X(t+1)＝X rand (t)-AD；
wherein: x is X rand Representing a random position vector selected from the current population, reducing the parameter a from 2 to 0 for development and exploration, respectively, and selecting A as a random search agent when |A| > 1; when |A| < 1, A is selected as the optimal solution for updating the location of the search agent;
the location update formula is as follows:
w＝1-exp(1-t)；
wherein: w=0.5;is a seat headThe optimal position of whale search agent i in von neumann topological neighborhood L (i), i.e., the historical optimal position of each whale search agent in the neighborhood corresponding to the individual with the minimum of fitness function; g represents a global optimum position.
2. The rolling bearing fault diagnosis method according to claim 1, characterized in that the rolling bearing fault diagnosis method is an improvement of whale optimization algorithm, the whale algorithm is improved by using von neumann topology structure and adaptive weight, the adaptive weight is introduced into a displacement formula, with iterative updating, local optimum gradually coincides with global optimum, and the position updating formula is as follows:
w＝1-exp(1-t)；
wherein: w=0.5;the optimal position of the whale search agent i in the Von Neumann topological neighborhood L (i) is obtained, namely the historical optimal position of each whale search agent in the neighborhood corresponding to the individual is obtained by the minimum value of the fitness function of each whale search agent; g represents a global optimum position.
3. A computer device, characterized in that it comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the rolling bearing fault diagnosis method according to any one of claims 1-2.
4. A computer-readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the rolling bearing failure diagnosis method according to any one of claims 1 to 2.
5. A rolling bearing failure diagnosis system that implements the rolling bearing failure diagnosis method according to any one of claims 1 to 2, characterized in that the rolling bearing failure diagnosis system includes:
the IMF component acquisition module is used for acquiring original signals of the bearing in four states, and performing signal decomposition by utilizing the VMD to obtain each IMF component;
the feature vector set construction module is used for extracting signal features by using multi-scale permutation entropy, constructing a feature vector set and dividing the feature vector set into a training sample and a test sample;
the initialization module is used for initializing population scale, iteration times and self-adaptive weight values of the whale algorithm;
the model building module is used for building an LSSVM model by using the initialization parameters;
the fitness value calculation module is used for calculating the corresponding fitness value of each whale, sequencing according to the fitness value, and selecting N whales as the population of the next generation;
the position updating module is used for carrying out neighborhood search by adopting a von neumann topological structure, carrying out information communication in the neighborhood, finding out the optimal whale in the neighborhood, and carrying out position updating according to a formula;
and the fault classification module is used for outputting whale positions with the best fitness as parameters of the LSSVM for training and carrying out fault classification on the test set.
6. An automotive rotating platform, characterized in that it is equipped with the rolling bearing failure diagnosis system according to claim 5.
7. A rotary restaurant platform, characterized in that it is equipped with the rolling bearing failure diagnosis system of claim 5.
8. A rotary advertising platform, wherein the rotary advertising platform is equipped with the rolling bearing failure diagnosis system of claim 5.
9. A rotary electric windmill, characterized in that it is equipped with the rolling bearing failure diagnosis system according to claim 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011083925.9A CN112347854B (en) | 2020-10-12 | 2020-10-12 | Rolling bearing fault diagnosis method, system, storage medium, equipment and application |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011083925.9A CN112347854B (en) | 2020-10-12 | 2020-10-12 | Rolling bearing fault diagnosis method, system, storage medium, equipment and application |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112347854A CN112347854A (en) | 2021-02-09 |
| CN112347854B true CN112347854B (en) | 2024-03-29 |
Family
ID=74361708
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011083925.9A Active CN112347854B (en) | 2020-10-12 | 2020-10-12 | Rolling bearing fault diagnosis method, system, storage medium, equipment and application |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112347854B (en) |
Families Citing this family (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113514783B (en) * | 2021-03-16 | 2023-12-15 | 国网江苏省电力有限公司南京供电分公司 | Method for detecting and judging mechanical state of transformer winding under no-load closing |
| CN113241087B (en) * | 2021-05-10 | 2022-05-27 | 青岛科技大学 | Underwater sound signal denoising method based on combination of improved VMD and improved wavelet threshold method |
| CN113344406A (en) * | 2021-06-21 | 2021-09-03 | 国网山东省电力公司电力科学研究院 | Power quality monitoring reliability assessment method for intelligent fusion terminal in distribution network area |
| CN113358357B (en) * | 2021-06-25 | 2023-08-04 | 上海应用技术大学 | Fault diagnosis method for rolling bearing of numerical control machine tool |
| CN113435314B (en) * | 2021-06-25 | 2023-02-21 | 陈珅 | Rolling bearing acoustic signal early fault sensitivity characteristic screening method and system |
| CN113591652A (en) * | 2021-07-22 | 2021-11-02 | 中南大学 | Electromagnetic signal processing method and device based on optimized variational modal decomposition |
| CN114139293A (en) * | 2021-08-18 | 2022-03-04 | 福州大学 | Rolling bearing fault diagnosis method and system |
| CN113591401B (en) * | 2021-08-24 | 2023-10-20 | 华北电力大学（保定） | Power transformer data cleaning method based on time sequence decomposition |
| CN113496091B (en) * | 2021-09-07 | 2021-12-24 | 西安航天动力研究所 | Method for simulating contact dynamic characteristics of high-speed heavy-load ball bearing of liquid rocket engine |
| CN114077847B (en) * | 2021-10-27 | 2022-08-19 | 河北建投能源投资股份有限公司 | Variable working condition rolling bearing fault diagnosis method, system, equipment and medium |
| CN114441172B (en) * | 2021-12-28 | 2023-12-15 | 东南大学 | Rolling bearing fault vibration signal analysis method |
| CN114169110B (en) * | 2022-02-10 | 2022-06-07 | 中国人民解放军海军工程大学 | Motor bearing fault diagnosis method based on feature optimization and GWAA-XGboost |
| CN115096590B (en) * | 2022-05-23 | 2023-08-15 | 燕山大学 | Rolling bearing fault diagnosis method based on IWOA-ELM |
| CN115113038B (en) * | 2022-08-19 | 2022-12-30 | 国网江西省电力有限公司电力科学研究院 | Circuit breaker fault diagnosis method based on current signal phase space reconstruction |
| CN115876476B (en) * | 2023-02-16 | 2023-05-16 | 山东科技大学 | Rolling bearing fault diagnosis method, system, computer device and storage medium |
| CN116384627A (en) * | 2023-03-24 | 2023-07-04 | 中国自然资源航空物探遥感中心 | Geological disaster evaluation method based on machine learning |
| CN117235506B (en) * | 2023-11-10 | 2024-02-13 | 四川大学 | Signal extraction method and device based on phase space reconstruction |
| CN117574057B (en) * | 2023-11-24 | 2024-05-28 | 昆明理工大学 | Intelligent fault diagnosis method for vertical water pump unit |
| CN117609768A (en) * | 2024-01-23 | 2024-02-27 | 昆明理工大学 | Vertical water pump unit fault diagnosis method based on improved whale algorithm |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019153388A1 (en) * | 2018-02-12 | 2019-08-15 | 大连理工大学 | Power spectral entropy random forest-based aeroengine rolling bearing fault diagnosis method |
| CN111062533A (en) * | 2019-12-16 | 2020-04-24 | 国家能源集团谏壁发电厂 | Fan fault prediction method based on whale optimization algorithm optimization weighted least square support vector machine |
| CN111272429A (en) * | 2020-03-04 | 2020-06-12 | 贵州大学 | Bearing fault diagnosis method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109933881A (en) * | 2019-03-06 | 2019-06-25 | 武汉大学 | A kind of Fault Diagnosis of Power Electronic Circuits method based on optimization deepness belief network |
-
2020
- 2020-10-12 CN CN202011083925.9A patent/CN112347854B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2019153388A1 (en) * | 2018-02-12 | 2019-08-15 | 大连理工大学 | Power spectral entropy random forest-based aeroengine rolling bearing fault diagnosis method |
| CN111062533A (en) * | 2019-12-16 | 2020-04-24 | 国家能源集团谏壁发电厂 | Fan fault prediction method based on whale optimization algorithm optimization weighted least square support vector machine |
| CN111272429A (en) * | 2020-03-04 | 2020-06-12 | 贵州大学 | Bearing fault diagnosis method |
Non-Patent Citations (2)
| Title |
|---|
| 周志朋 ; 谢冬梅 ; .风电齿轮箱复合故障诊断方法研究.河南科技.2020,(第17期),全文. * |
| 徐继亚 ; 王艳 ; 纪志成 ; .基于鲸鱼算法优化WKELM的滚动轴承故障诊断.系统仿真学报.2017,(第09期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112347854A (en) | 2021-02-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112347854B (en) | Rolling bearing fault diagnosis method, system, storage medium, equipment and application | |
| Sahu et al. | A lightweight multi-section CNN for lung nodule classification and malignancy estimation | |
| CN110006649B (en) | Bearing fault diagnosis method based on improved ant lion algorithm and support vector machine | |
| JP4556120B2 (en) | Information processing apparatus and method, and program | |
| EP4235514A2 (en) | Methods, systems, articles of manufacture and apparatus to map workloads | |
| CN111062533A (en) | Fan fault prediction method based on whale optimization algorithm optimization weighted least square support vector machine | |
| Xu et al. | An imbalanced fault diagnosis method for rolling bearing based on semi-supervised conditional generative adversarial network with spectral normalization | |
| Moldovan | Transfer learning based method for two-step skin cancer images classification | |
| CN112149705A (en) | Method and system for training classification model, computer equipment and storage medium | |
| EP3493154A1 (en) | Segmentation system for segmenting an object in an image | |
| Wang et al. | Intelligent fault diagnosis for planetary gearbox using transferable deep q network under variable conditions with small training data | |
| Aydin et al. | Artificial immune classifier with swarm learning | |
| CN114169110B (en) | Motor bearing fault diagnosis method based on feature optimization and GWAA-XGboost | |
| Fei | The Hybrid Method of VMD‐PSR‐SVD and Improved Binary PSO‐KNN for Fault Diagnosis of Bearing | |
| Zhu et al. | Mix contrast for COVID-19 mild-to-critical prediction | |
| CN114298132A (en) | Wind power prediction method and device and electronic equipment | |
| CN112633385A (en) | Model training method, data generation method and device | |
| Lemikhova et al. | Transfer learning for few-shot plants recognition: Antarctic station greenhouse use-case | |
| JP2022140857A (en) | Methods and apparatus for processing image data for machine vision | |
| CN114462446A (en) | Rolling bearing fault diagnosis method based on vibration signal and electronic equipment | |
| JP4556121B2 (en) | Information processing apparatus and method, and program | |
| He et al. | An under-sampling method of unbalanced data classification based on support vector | |
| Yuan et al. | A novel fault diagnosis method for second-order bandpass filter circuit based on TQWT-CNN | |
| Venkatesan et al. | Automatic Classification of Diseases From X-Ray Images Using Xception Deep Convolution Neural Networks | |
| Hou et al. | Bearing fault diagnosis based on variational mode decomposition and modified CNN |
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 |