CN110307981B - Bearing fault diagnosis method based on PNN-IFA - Google Patents

Bearing fault diagnosis method based on PNN-IFA Download PDF

Info

Publication number
CN110307981B
CN110307981B CN201910521937.6A CN201910521937A CN110307981B CN 110307981 B CN110307981 B CN 110307981B CN 201910521937 A CN201910521937 A CN 201910521937A CN 110307981 B CN110307981 B CN 110307981B
Authority
CN
China
Prior art keywords
bearing
firefly
smoothing factor
vector
pnn
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.)
Expired - Fee Related
Application number
CN201910521937.6A
Other languages
Chinese (zh)
Other versions
CN110307981A (en
Inventor
宋玉琴
师少达
邓思成
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Polytechnic University
Original Assignee
Xian Polytechnic University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Polytechnic University filed Critical Xian Polytechnic University
Priority to CN201910521937.6A priority Critical patent/CN110307981B/en
Publication of CN110307981A publication Critical patent/CN110307981A/en
Application granted granted Critical
Publication of CN110307981B publication Critical patent/CN110307981B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/047—Probabilistic or stochastic networks

Abstract

The invention discloses a bearing fault diagnosis method based on PNN-IFA, which comprises the steps of specifically constructing a PNN model; obtaining an optimal smoothing factor to be selected through a firefly algorithm; and taking the optimal smoothing factor to be selected as a smoothing factor, and inputting the fault feature vector of each bearing vibration data in the test set into the PNN model for fault diagnosis. The invention provides a bearing fault diagnosis method based on PNN-IFA, which is characterized in that a firefly foraging algorithm is improved, the firefly foraging algorithm is applied to parameter optimization of a probabilistic neural network, and the optimized probabilistic neural network is applied to bearing diagnosis, so that the bearing fault diagnosis accuracy is improved.

Description

Bearing fault diagnosis method based on PNN-IFA
Technical Field
The invention belongs to the technical field of fault detection, and relates to a bearing fault diagnosis method based on PNN-IFA.
Background
The coal mine fully mechanized mining equipment is widely applied to actual engineering production as a large-scale complex engineering machine integrating machinery, electricity and hydraulic pressure, and the most important mechanical part in the large-scale rotating machinery is a rolling bearing which plays a role in fixing and reducing a friction load coefficient in mechanical transmission. However, due to the constraints of working environment and conditions, the bearing is more prone to failure during operation, and accidents and shutdown maintenance are more frequent. Therefore, the method can timely and reliably diagnose the potential faults of the bearing and has profound significance for guaranteeing the long-term operation of the fully mechanized mining equipment. But at present, no method capable of accurately diagnosing the bearing fault can be effectively applied to the engineering field.
Disclosure of Invention
The invention aims to provide a bearing fault diagnosis method based on PNN-IFA, which can improve the accuracy of bearing fault diagnosis.
The technical scheme adopted by the invention is that the bearing fault diagnosis method based on PNN-IFA specifically comprises the following steps:
step 1, constructing a PNN model, and setting an initial value of the PNN; selecting a plurality of smoothing factors to be selected in the range of the smoothing factor parameters, and establishing a parameter set of the smoothing factors to be selected;
collecting bearing vibration data of the multi-section fully-mechanized mining equipment in different fault states, and dividing the bearing vibration data into a training set and a test set; extracting a fault characteristic vector of each section of bearing vibration data;
step 2, taking a first smoothing factor to be selected in the smoothing factor parameter set as a smoothing factor, and calculating the accuracy of the bearing vibration vector diagnosis result corresponding to the smoothing factor through a PNN model;
step 3, calculating the brightness of the accuracy corresponding to the smoothing factor according to the accuracy of the bearing vibration vector diagnosis result;
step 4, taking the next smoothing factor to be selected as a smoothing factor, and repeating the step 2-3 until the brightness of the accuracy rate corresponding to all the smoothing factors to be selected is calculated;
step 5, taking the smoothing factors to be selected as individual fireflies, taking the brightness of the accuracy corresponding to each smoothing factor to be selected as the position information of the fireflies, and obtaining the optimal smoothing factor to be selected through a fireflies algorithm;
and 6, taking the optimal smoothing factor to be selected as a smoothing factor, and inputting the fault characteristic vector of each bearing vibration data in the test set into the PNN model for fault diagnosis.
The invention is also characterized in that:
the multi-section bearing vibration data comprises vibration data when the bearing normally operates, vibration data when the bearing inner ring fails, vibration data when the bearing rolling body fails and vibration data when the bearing outer ring fails.
Extracting the fault characteristic vector of each section of bearing vibration data in the step 1 specifically according to the following method:
step 1.1, using all vibration signals in each section of bearing vibration data as a group of vibration signals to obtain a plurality of groups of vibration signals;
step 1.2, EMD processing is carried out on each vibration signal in each group of vibration signals to form modal components for determining center frequency and bandwidth, and the modal components with the frequency values closest to fault characteristic frequency are selected as optimal modal components from a plurality of modes of each group through EMD envelope spectrums;
step 1.3, respectively calculating the fault feature vector of each optimal modal component, specifically according to the following method:
extracting a root mean square value, a kurtosis value and a skewness value of each modal component under time domain analysis, extracting a mean frequency, a frequency divergence, a sample entropy and an arrangement entropy of each modal component under frequency domain analysis, and then forming a bearing fault feature vector by the root mean square value, the kurtosis value, the skewness value, the mean frequency, the frequency divergence, the sample entropy and the arrangement entropy;
wherein the root mean square value is obtained according to the following formula:
Figure BDA0002096990050000031
wherein, x is an original signal, and N is the number of samples;
the kurtosis is obtained as follows:
Figure BDA0002096990050000032
in the formula, σxIs the standard deviation of the signal;
the skewness value is obtained according to the following formula:
Figure BDA0002096990050000033
the mean frequency is obtained as follows:
Figure BDA0002096990050000034
in the formula, s (k) is a frequency domain signal obtained by Fourier transform of an original signal;
the frequency dispersion is obtained as follows:
Figure BDA0002096990050000041
in the formula, FmIs the mean value of the signal frequency, fkIs the signal frequency;
the sample entropy is obtained as follows:
Figure BDA0002096990050000042
where m is the number of dimensions, typically 1 or 2, r is a given threshold, r is usually chosen to be 0.1 std to 0.25 std, where std is the standard deviation of the original time series, and is empirically set to 3, am(r) is the number of subsequences less than or equal to a threshold r in an m + 1-dimensional vector formed by data according to a time sequence, Bm(r) is the number of subsequences in the m-dimensional vector which are less than or equal to the threshold r;
the permutation entropy is obtained as follows:
Figure BDA0002096990050000043
in the step 2, calculating the accuracy of the bearing vibration vector diagnosis result corresponding to the smoothing factor through the PNN model according to the following steps:
step 2.1, fitting each bearing vibration vector in the training set with each mode layer neuron to obtain a fitting degree;
step 2.2, carrying out weighted average on the fitting degree of each bearing vibration vector and neurons of the same type of mode layer, and respectively obtaining the fitting average value of each bearing vibration vector and neurons of four types of mode layers:
step 2.3, comparing the fitting average values of the four types of mode layer neurons, and taking the largest fitting average value as a diagnosis result of the bearing vibration vector;
and 2.4, judging the accuracy of the diagnosis result of each bearing vibration vector, and simultaneously calculating the accuracy of the diagnosis result of all bearing vibration vectors.
And 3, calculating the brightness of the accuracy corresponding to the smoothing factor according to the following formula:
Figure BDA0002096990050000051
in the formula, N is the number of the smoothing factors to be selected, i is the iteration number, and YiTo a desired accuracy, QiThe accuracy of the bearing vibration vector diagnosis result is obtained.
The step 5 is specifically carried out according to the following steps:
step 5.1, setting the light absorption intensity coefficient gamma, the step factor alpha and the maximum attraction beta of the firefly0Iteration number MaxGeneration, search Range PminAnd Pmax(ii) a Initializing the position information of the firefly, and randomly distributing the smooth factor group to be selected to the range of a search area by calling a related function;
step 5.2, calculating the relative brightness I and the attraction degree beta of the firefly, comparing the fluorescence brightness of the firefly in the neighborhood, and determining the movement direction of the firefly according to the relative brightness;
step 5.3, recalculating the brightness of the firefly according to the updated position of the firefly;
and 5.4, repeating the step 5.2 to the step 5.3 until the brightness of the firefly is not less than a set value or the maximum iteration number, stopping iteration, and taking the smoothing factor to be selected corresponding to the current brightness of the firefly as the optimal smoothing factor.
The invention has the beneficial effects that:
the invention relates to a bearing fault diagnosis method based on PNN-IFA, which improves a firefly foraging algorithm, applies the firefly foraging algorithm to parameter optimization of a probabilistic neural network, and applies the optimized probabilistic neural network to bearing diagnosis, thereby improving the accuracy of bearing fault diagnosis.
Drawings
FIG. 1 is a flow chart of a bearing fault diagnosis method based on PNN-IFA according to the invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention discloses a bearing fault diagnosis method based on PNN-IFA, the flow of which is shown in figure 1 and is specifically carried out according to the following steps:
step 1, constructing a PNN model, and setting an initial value of the PNN; the PNN model network has four layers including an input layer, a mode layer, a category layer and an output layer. The number of the neurons of the input layer is 7, and the neurons respectively correspond to a root mean square value, a kurtosis value, a deviation value, a mean frequency, a frequency divergence, a sample entropy and an arrangement entropy in a bearing fault feature vector; the neurons in the mode layer are divided into four types which respectively correspond to four types of bearing fault types, namely normal operation of a bearing, fault of a bearing inner ring, fault of a bearing rolling body and fault of a bearing outer ring; the number of the neurons of the category layer and the output layer is set to be 4, and the four types of the neurons correspond to the four types of bearing fault types;
selecting a plurality of smoothing factors to be selected in the range of the smoothing factor parameters, and establishing a parameter set of the smoothing factors to be selected;
collecting multi-section bearing vibration data of multi-section fully-mechanized mining equipment in different fault states, and dividing the bearing vibration data into a training set and a test set; extracting a fault characteristic vector of each section of bearing vibration data, and specifically performing the following steps:
step 1.1, using all vibration signals in each section of bearing vibration data as a group of vibration signals to obtain a plurality of groups of vibration signals;
step 1.2, EMD processing is carried out on each vibration signal in each group of vibration signals to form modal components for determining center frequency and bandwidth, and the modal components with the frequency values closest to fault characteristic frequency are selected as optimal modal components from a plurality of modes of each group through EMD envelope spectrums;
step 1.3, respectively calculating the fault feature vector of each optimal modal component, specifically according to the following method:
extracting a root mean square value, a kurtosis value and a skewness value of each modal component under time domain analysis, extracting a mean frequency, a frequency divergence, a sample entropy and an arrangement entropy of each modal component under frequency domain analysis, and then forming a bearing fault feature vector by the root mean square value, the kurtosis value, the skewness value, the mean frequency, the frequency divergence, the sample entropy and the arrangement entropy;
wherein the root mean square value is obtained according to the following formula:
Figure BDA0002096990050000071
wherein, x is an original signal, and N is the number of samples;
the kurtosis is obtained as follows:
Figure BDA0002096990050000072
in the formula, σxIs the standard deviation of the signal;
the skewness value is obtained according to the following formula:
Figure BDA0002096990050000073
the mean frequency is obtained as follows:
Figure BDA0002096990050000074
in the formula, s (k) is a frequency domain signal obtained by Fourier transform of an original signal;
the frequency dispersion is obtained as follows:
Figure BDA0002096990050000075
in the formula, FmIs the mean value of the signal frequency, fkIs the signal frequency;
the sample entropy is obtained as follows:
Figure BDA0002096990050000081
where m is the number of dimensions, typically 1 or 2, r is a given threshold, r is usually chosen to be 0.1 std to 0.25 std, where std is the standard deviation of the original time series, and is empirically set to 3, am(r) is the number of subsequences less than or equal to a threshold r in an m + 1-dimensional vector formed by data according to a time sequence, Bm(r) is the number of subsequences in the m-dimensional vector which are less than or equal to the threshold r;
the permutation entropy is obtained as follows:
Figure BDA0002096990050000082
step 2, using the first smoothing factor to be selected in the smoothing factor parameter set as a smoothing factor, inputting all fault feature vectors in the training set to the PNN model through an input layer, and calculating the accuracy of the bearing vibration vector diagnosis result corresponding to the smoothing factor, specifically according to the following steps:
step 2.1, fitting each bearing vibration vector in the training set with each mode layer neuron to obtain a fitting degree:
Figure BDA0002096990050000083
where σ is the smoothing factor, X is the input training sample, WiThe weight from the input layer to the mode layer;
step 2.2, carrying out weighted average on the fitting degree of each bearing vibration vector and neurons of the same type of mode layer, and respectively obtaining the fitting average value of each bearing vibration vector and neurons of four types of mode layers:
step 2.3, comparing the fitting average values of the four types of mode layer neurons, and taking the largest fitting average value as a diagnosis result of the bearing vibration vector;
and 2.4, judging the accuracy of the diagnosis result of each bearing vibration vector, and simultaneously calculating the accuracy of the diagnosis result of all bearing vibration vectors.
And 3, calculating the brightness of the accuracy corresponding to the smoothing factor according to the accuracy of the bearing vibration vector diagnosis result:
Figure BDA0002096990050000091
in the formula, N is the number of the smoothing factors to be selected, i is the iteration number, and YiTo a desired accuracy, QiThe accuracy of the bearing vibration vector diagnosis result is obtained.
Step 4, taking the next smoothing factor to be selected as a smoothing factor, and repeating the step 2-3 until the brightness of the accuracy rate corresponding to all the smoothing factors to be selected is calculated;
step 5, using the smoothing factors to be selected as individual fireflies, using the brightness of the accuracy corresponding to each smoothing factor to be selected as the position information of the fireflies, obtaining the optimal smoothing factor to be selected through a fireflies algorithm, and specifically performing the following steps:
step 5.1, setting the light absorption intensity coefficient gamma, the step factor alpha and the maximum attraction beta of the firefly0Iteration number MaxGeneration, search Range PminAnd Pmax(ii) a Initializing the position information of the firefly, and randomly distributing the smooth factor group to be selected to the range of a search area by calling a related function;
step 5.2, calculating the relative brightness I, the attraction degree beta and the self-adaptive step length of the firefly, comparing the fluorescence brightness of the firefly in the neighborhood, and determining the movement direction of the firefly according to the relative brightness:
wherein the relative brightness of the firefly is:
Figure BDA0002096990050000092
in the formula IiThe absolute brightness of the firefly i, namely the brightness of the light intensity of the firefly i at the initial position before flying; gamma is a parameter of air absorption, called light absorption coefficient, which has great influence on the convergence rate and optimization effect of fireflies, and the value of gamma belongs to [0.01,100 ] in most cases]；rijThe distance from firefly i to firefly j;
the attraction degree β' is:
Figure BDA0002096990050000101
in the formula, beta0To a maximum attractive force, i.e.Attraction of fireflies, generally beta, at the source location0The value is 1; beta is aminIs the minimum attraction; r isijThe distance from firefly i to firefly j;
the adaptive step size is:
Figure BDA0002096990050000102
in the formula, alpha0Is an initial step size factor and takes the value of [0, 1%]，xi-xbestRepresents the spatial distance, d, of firefly i from the current optimal fireflymaxExpressing the maximum distance between the optimal firefly and the neighborhood firefly;
step 5.3, calculating the updated position of the firefly; brightness of firefly was obtained:
Figure BDA0002096990050000103
wherein:
Figure BDA0002096990050000104
is the location of the firefly i at a point in space,
Figure BDA0002096990050000105
is the location of firefly j at some point in space,
Figure BDA0002096990050000106
is the position of the firefly i after movement, betaijThe attraction degree of the firefly j to the i attraction; alpha is a step size factor and is the minimum distance of each movement of the firefly, and generally alpha belongs to [0,1 ]](ii) a rand is [0,1 ]]Obeying a uniformly distributed random factor. Alpha (rand-1/2) is added random disturbance term
And 5.4, repeating the step 5.2 to the step 5.3 until the brightness of the firefly is not less than a set value or the maximum iteration number, stopping iteration, and taking the smoothing factor to be selected corresponding to the current brightness of the firefly as the optimal smoothing factor.
And 6, taking the optimal smoothing factor to be selected as a smoothing factor, and inputting the fault characteristic vector of each bearing vibration data in the test set into the PNN model for fault diagnosis.
According to the method, the firefly foraging algorithm is improved, so that the search accuracy and the convergence speed of a firefly population are effectively improved, the firefly foraging algorithm is further reliably applied to the fault diagnosis of the rolling bearing, and the PNN parameters are optimized through the firefly algorithm, so that the optimized probabilistic neural network can effectively diagnose the vibration type of the bearing fault.

Claims (4)

1. The bearing fault diagnosis method based on PNN-IFA is characterized by comprising the following steps:
step 1, constructing a PNN model, and setting an initial value of the PNN; selecting a plurality of smoothing factors to be selected in the range of the smoothing factor parameters, and establishing a parameter set of the smoothing factors to be selected;
collecting bearing vibration data of a multi-section fully mechanized mining device in different fault states, and dividing the bearing vibration data into a training set and a test set; extracting a fault characteristic vector of each section of bearing vibration data;
the step 1 of extracting the fault feature vector of each section of bearing vibration data is specifically carried out according to the following method:
step 1.1, using all vibration signals in each section of bearing vibration data as a group of vibration signals to obtain a plurality of groups of vibration signals;
step 1.2, EMD processing is carried out on each vibration signal in each group of vibration signals to form modal components for determining center frequency and bandwidth, and the modal components with the frequency values closest to fault characteristic frequency are selected as optimal modal components from a plurality of modes of each group through EMD envelope spectrums;
step 1.3, respectively calculating the fault feature vector of each optimal modal component, specifically according to the following method:
extracting a root mean square value, a kurtosis value and a skewness value of each modal component under time domain analysis, extracting a mean frequency, a frequency divergence, a sample entropy and an arrangement entropy of each modal component under frequency domain analysis, and then forming a bearing fault feature vector by the root mean square value, the kurtosis value, the skewness value, the mean frequency, the frequency divergence, the sample entropy and the arrangement entropy;
wherein the root mean square value is obtained according to the following formula:
Figure FDA0002814789500000011
wherein, x is an original signal, and N is the number of samples;
the kurtosis is obtained as follows:
Figure FDA0002814789500000021
in the formula, σxIs the standard deviation of the signal;
the skewness value is obtained according to the following formula:
Figure FDA0002814789500000022
the mean frequency is obtained as follows:
Figure FDA0002814789500000023
in the formula, s (k) is a frequency domain signal obtained by Fourier transform of an original signal;
the frequency dispersion is obtained as follows:
Figure FDA0002814789500000024
in the formula, FmIs the mean value of the signal frequency, fkIs the signal frequency;
the sample entropy is obtained as follows:
Figure FDA0002814789500000025
where m is the number of dimensions, typically 1 or 2, r is a given threshold, r is usually chosen to be 0.1 std to 0.25 std, where std is the standard deviation of the original time series, and is empirically set to 3, am(r) is the number of subsequences less than or equal to a threshold r in an m + 1-dimensional vector formed by data according to a time sequence, Bm(r) is the number of subsequences in the m-dimensional vector which are less than or equal to the threshold r;
the permutation entropy is obtained as follows:
Figure FDA0002814789500000031
step 2, taking a first smoothing factor to be selected in the smoothing factor parameter set as a smoothing factor, and calculating the accuracy of the bearing vibration vector diagnosis result corresponding to the smoothing factor through a PNN model;
in the step 2, calculating the accuracy of the bearing vibration vector diagnosis result corresponding to the smoothing factor through the PNN model according to the following steps:
step 2.1, fitting each bearing vibration vector in the training set with each mode layer neuron to obtain a fitting degree;
step 2.2, carrying out weighted average on the fitting degree of each bearing vibration vector and neurons of the same type of mode layer, and respectively obtaining the fitting average value of each bearing vibration vector and neurons of four types of mode layers:
step 2.3, comparing the fitting average values of the four types of mode layer neurons, and taking the largest fitting average value as a diagnosis result of the bearing vibration vector;
step 2.4, judging the accuracy of the diagnosis result of each bearing vibration vector, and simultaneously calculating the accuracy of the diagnosis result of all bearing vibration vectors;
step 3, calculating the brightness of the accuracy corresponding to the smoothing factor according to the accuracy of the bearing vibration vector diagnosis result;
step 4, taking the next smoothing factor to be selected as a smoothing factor, and repeating the step 2-3 until the brightness of the accuracy rate corresponding to all the smoothing factors to be selected is calculated;
step 5, taking the smoothing factors to be selected as individual fireflies, taking the brightness of the accuracy corresponding to each smoothing factor to be selected as the position information of the fireflies, and obtaining the optimal smoothing factor to be selected through a fireflies algorithm;
and 6, taking the optimal smoothing factor to be selected as a smoothing factor, and inputting the fault characteristic vector of each bearing vibration data in the test set into the PNN model for fault diagnosis.
2. The PNN-IFA-based bearing fault diagnosis method as recited in claim 1, wherein the plurality of pieces of bearing vibration data comprise vibration data when a bearing normally operates, vibration data when a bearing inner ring fails, vibration data when a bearing rolling element fails, and vibration data when a bearing outer ring fails.
3. The PNN-IFA-based bearing fault diagnosis method according to claim 1, wherein the brightness of the accuracy corresponding to the smoothing factor is calculated in step 3 according to the following formula:
Figure FDA0002814789500000041
in the formula, N is the number of the smoothing factors to be selected, i is the iteration number, and YiTo a desired accuracy, QiThe accuracy of the bearing vibration vector diagnosis result is obtained.
4. The PNN-IFA-based bearing fault diagnosis method according to claim 1, wherein the step 5 is specifically performed according to the following steps:
step 5.1, setting the light absorption intensity coefficient gamma, the step factor alpha and the maximum attraction beta of the firefly0Iteration number MaxGeneration, search Range PminAnd Pmax(ii) a Initializing the position information of firefly, and selecting the firefly to be selected by calling related functionsThe smooth factor groups are randomly distributed in the range of the search area;
step 5.2, calculating the relative brightness I and the attraction degree beta of the firefly, comparing the fluorescence brightness of the firefly in the neighborhood, and determining the movement direction of the firefly according to the relative brightness;
step 5.3, recalculating the brightness of the firefly according to the updated position of the firefly;
and 5.4, repeating the step 5.2 to the step 5.3 until the brightness of the firefly is not less than a set value or the maximum iteration number, stopping iteration, and taking the smoothing factor to be selected corresponding to the current brightness of the firefly as the optimal smoothing factor.
CN201910521937.6A 2019-06-17 2019-06-17 Bearing fault diagnosis method based on PNN-IFA Expired - Fee Related CN110307981B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910521937.6A CN110307981B (en) 2019-06-17 2019-06-17 Bearing fault diagnosis method based on PNN-IFA

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910521937.6A CN110307981B (en) 2019-06-17 2019-06-17 Bearing fault diagnosis method based on PNN-IFA

Publications (2)

Publication Number Publication Date
CN110307981A CN110307981A (en) 2019-10-08
CN110307981B true CN110307981B (en) 2021-01-26

Family

ID=68077550

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910521937.6A Expired - Fee Related CN110307981B (en) 2019-06-17 2019-06-17 Bearing fault diagnosis method based on PNN-IFA

Country Status (1)

Country Link
CN (1) CN110307981B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110766100B (en) * 2019-12-02 2022-05-20 珠海格力电器股份有限公司 Bearing fault diagnosis model construction method, bearing fault diagnosis method and electronic equipment
CN115060497B (en) * 2022-06-10 2023-07-11 南昌工程学院 Bearing fault diagnosis method based on CEEMD energy entropy and optimized PNN

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DK2053375T3 (en) * 2007-10-24 2010-04-12 Abb Research Ltd Method for detecting and automatically identifying damage to roller bearings
CN105973596A (en) * 2016-05-09 2016-09-28 华锐风电科技(集团)股份有限公司 Fan fault diagnosis method based on LLTSA and PNN
CN106650071A (en) * 2016-12-12 2017-05-10 中国航空工业集团公司上海航空测控技术研究所 Intelligent fault diagnosis method for rolling bearing
CN208207843U (en) * 2018-04-10 2018-12-07 西南石油大学 A kind of slush pump trouble-shooter
CN109489977B (en) * 2018-12-28 2021-03-05 西安工程大学 KNN-AdaBoost-based bearing fault diagnosis method

Also Published As

Publication number Publication date
CN110307981A (en) 2019-10-08

Similar Documents

Publication Publication Date Title
CN112149316B (en) Aero-engine residual life prediction method based on improved CNN model
CN110361176B (en) Intelligent fault diagnosis method based on multitask feature sharing neural network
CN109918752B (en) Mechanical fault diagnosis method, equipment and medium based on migration convolution neural network
CN111797567B (en) Bearing fault classification method and system based on deep learning network
CN109141847B (en) Aircraft system fault diagnosis method based on MSCNN deep learning
CN109308522B (en) GIS fault prediction method based on recurrent neural network
CN111596604A (en) Intelligent fault diagnosis and self-healing control system and method for engineering equipment based on digital twinning
CN113935406B (en) Mechanical equipment unsupervised fault diagnosis method based on countermeasure flow model
Yin et al. Wasserstein generative adversarial network and convolutional neural network (WG-CNN) for bearing fault diagnosis
CN111046916A (en) Motor fault diagnosis method and system based on void convolution capsule network
CN110082136B (en) Rotary machine fault diagnosis method based on cloud genetic algorithm optimization support vector machine
CN106168539A (en) Fault Diagnosis of Roller Bearings based on convolutional neural networks and system
CN113392881B (en) Rotary machine fault diagnosis method based on transfer learning
CN106779063B (en) A kind of hoist braking system method for diagnosing faults based on RBF network
CN110307981B (en) Bearing fault diagnosis method based on PNN-IFA
CN113203566B (en) Motor bearing fault diagnosis method based on one-dimensional data enhancement and CNN
CN110909802A (en) Improved PSO (particle swarm optimization) based fault classification method for optimizing PNN (portable network) smoothing factor
CN112763214A (en) Rolling bearing fault diagnosis method based on multi-label zero-sample learning
CN112132102B (en) Intelligent fault diagnosis method combining deep neural network with artificial bee colony optimization
CN113792758B (en) Rolling bearing fault diagnosis method based on self-supervision learning and clustering
CN112508242A (en) Method for constructing bearing fault location and classification model of wind power generator
CN113496061A (en) SOM network clustering electromechanical device bearing fault analysis method based on transfer learning and manifold distance
Chen et al. Adaptation regularization based on transfer learning for fault diagnosis of rotating machinery under multiple operating conditions
CN107016440B (en) The multiresolution deep neural network intelligent diagnosing method of machine driving failure
CN113092044B (en) Rotary machine fault diagnosis method based on weighted level visible graph

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
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20210126

Termination date: 20210617

CF01 Termination of patent right due to non-payment of annual fee