CN114564619A - Fault diagnosis method, recording medium and system for motor bearing - Google Patents

Fault diagnosis method, recording medium and system for motor bearing Download PDF

Info

Publication number
CN114564619A
CN114564619A CN202210055670.8A CN202210055670A CN114564619A CN 114564619 A CN114564619 A CN 114564619A CN 202210055670 A CN202210055670 A CN 202210055670A CN 114564619 A CN114564619 A CN 114564619A
Authority
CN
China
Prior art keywords
sequence
motor bearing
fault diagnosis
fault
diagnosing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210055670.8A
Other languages
Chinese (zh)
Inventor
李忠
欧阳斌
徐兴华
崔小鹏
梁英杰
邱少华
平作为
曾德林
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.)
Naval University of Engineering PLA
Original Assignee
Naval University of Engineering PLA
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 Naval University of Engineering PLA filed Critical Naval University of Engineering PLA
Priority to CN202210055670.8A priority Critical patent/CN114564619A/en
Publication of CN114564619A publication Critical patent/CN114564619A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/901—Indexing; Data structures therefor; Storage structures
    • G06F16/9027—Trees
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/903—Querying
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Acoustics & Sound (AREA)
  • Computational Linguistics (AREA)
  • Software Systems (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention belongs to the technical field of fault diagnosis, and particularly relates to a fault diagnosis method of a motor bearing, which comprises the following steps: dividing data, reducing dimension, modeling, extracting characteristic vectors, generating a database after screening, and performing matching diagnosis. The invention provides a sectional linear representation method for carrying out dimensionality reduction on data, gives consideration to the overall trend and local important characteristics of a time sequence from modeling to database generation, and solves the problem that large-scale time sequences have huge calculation and storage expenses in similarity measurement. The invention also provides a non-transient readable recording medium storing the program of the method and a system comprising the medium, wherein the program can be called by a processing circuit to execute the method.

Description

Fault diagnosis method, recording medium and system for motor bearing
Technical Field
The invention belongs to the technical field of fault diagnosis and discloses a fault diagnosis method of a motor bearing, a recording medium and a system which are stored with a program capable of executing the method.
Background
As an important driving force in modern industrial processes, motors are used in a very wide range. The severe working environment can cause the motor to break down, the industrial production process is influenced slightly, and the life and property safety is endangered seriously, so that the motor fault diagnosis research is enhanced to have great significance. The bearing is one of the core components of the motor system, and mainly achieves the effect of bearing load. Statistically, more than 40% of motor faults occur in bearing parts, and the importance of the motor faults on the safe and stable operation of the whole motor system is self-evident. People can roughly divide the fault diagnosis research of the motor bearing into five stages according to time sequence: in the first stage, whether a fault occurs is analyzed and judged mainly by a simple instrument, the subjectivity of experience judgment is strong, and the diagnosis accuracy rate needs to be improved; in the second stage, the fault diagnosis is realized by using a spectrum analyzer, and the method is difficult to extract signal characteristics and has higher requirements on equipment and technicians; the third stage utilizes the pulse impact technology to realize fault diagnosis, and the instrument is accurate, efficient, convenient to use and still widely used at present; in the fourth stage, fault diagnosis is realized by utilizing a resonance demodulation technology, and the resonance demodulation technology is particularly sensitive to high-frequency resonance amplitude caused by shock waves and is very wide in application; the fifth stage is an intelligent stage, intelligent algorithms such as a neural network and data mining are applied to production practice, and bearing fault diagnosis and the intelligent algorithms are organically combined.
Because the vibration signal of the motor bearing has the characteristics of nonlinearity and instability, the fault diagnosis and the health assessment of the motor bearing are difficult, and a single time domain or frequency domain analysis method is not suitable any more. The method for analyzing and diagnosing by using instruments in the first stage to the fourth stage needs manual participation, is low in efficiency and strong in subjectivity, and the fifth stage applies intelligent algorithms such as a neural network and data mining to production practice, so that the automation level of fault diagnosis can be improved, but the conditions of low diagnosis speed and inaccurate diagnosis are caused because the time sequence acquired by a sensor of a power system is large in scale and occupies more calculation and storage resources.
Disclosure of Invention
Aiming at the problems, the invention provides a fault diagnosis method of a motor bearing, which comprises the following steps:
s1, dividing the obtained original time sequence data of the motor bearing into a training set, a verification set and a test set in a layered sampling mode;
s2, performing dimensionality reduction on the original time sequence data by applying a piecewise linear representation algorithm;
s3, building a feature extraction model through the training set, and optimizing parameters through the verification set until the feature extraction model is converged;
s4, extracting the feature vectors of all sequences in the test set by applying a feature extraction model;
s5, organizing indexes of the feature vectors through an M-Tree search Tree, and adding sequences meeting similarity conditions into a candidate set;
and S6, measuring the similarity between the sequence to be queried and each sequence in the candidate set by applying a dynamic time warping mode, so as to match abnormal states and realize fault diagnosis.
Preferably, the hierarchical sampling mode is as follows: dividing the raw time sequence data of the motor bearing into a training set, a verification set and a test set according to the ratio of 7:1: 2.
Further, the piecewise linear representation algorithm in the step S2 includes:
s2.1, extracting a fluctuation point sequence of the original time series data;
s2.2, on the basis of the fluctuation point sequence, comparing set thresholds, and screening to obtain a key point sequence;
s2.3, reducing the dimension of the original time sequence by adopting a segmentation aggregation approximation method;
and S2.4, fusing the key points and the time sequence after dimensionality reduction according to the index of the key point sequence in the original time sequence to obtain a KP-PAA sequence.
Wherein the step S2.1 is that the original time series X ═ X1,x2,…,xnArranging three adjacent points in the data according to the time increasing order if xiThe following conditions are satisfied:
Figure BDA0003476385370000021
then call xiAnd extracting all fluctuation points as fluctuation points to obtain the fluctuation point sequence.
Marking the fluctuation point sequence as XfFirstly, a proper screening threshold value epsilon is set according to the oscillation amplitude of the original sequence0(ii) a Then the fluctuation point sequence XfThe first value as the comparison point a; if the subsequent fluctuation point xfiThe following conditions are satisfied:
abs(xfi-a)＞ε0 1(2)
will fluctuate point xfiAdding the index and the value of the key point sequence into the key point sequence, updating the value of the comparison point a until all the fluctuation points are screened, and finally obtaining the key point sequence Xk。
The segmented aggregation approximation (PAA) method in S2.3 is:
PAA represents the original time series by equally dividing the time series and using the mean of the segmented series. For example, the original time series X ═ X1,x2,…,xnEqually-spaced and segmented averaging is carried out according to the interval k, and the average is converted into a time sequence Y ═ Y1,y2,…,ymWhere n > m, time-series data reduction and features are achievedAnd (5) symbolizing. Any element y in the new sequenceiSatisfies the following conditions:
Figure BDA0003476385370000031
the method for fusing the key points and the PAA in S2.4 comprises the following steps:
its segmentation in the PAA is found according to the index of the keypoint in the original time series. And replacing the average value of the PAA segment where the key point is positioned with the numerical value of the key point to obtain a final segment linear representation sequence KP-PAA after fusion.
The feature extraction model built in S3 is:
s3.1, referring to an Encoder (Encoder) part of a transform model, and taking an N-layer encoding layer as a main body part of the model;
s3.2 introduces a time embedding layer by taking advantage of design skills of ViT models. The time sequence is convoluted along the time dimension by utilizing the one-dimensional convolution with the same step length and kernel size, so that the length of the sequence can be reduced, and meanwhile, rich local characteristic information can be extracted;
s3.3, by taking the design skill of the Informmer model as a reference, a distillation layer of the Informmer is added between every two coding layers, and the length of the sequence is halved through a one-dimensional convolution layer and a one-dimensional pooling layer, so that the self-attention distillation effect is achieved.
S3.4 output sequence x at time embedding layer by using BERT model design technique as referenceeHead of (2) concatenates a randomly initialized feature token xfForming an input sequence xinAnd then fed to the encoder. And then extracting the feature token from the head of the output sequence of the encoder, and outputting a feature vector F after passing through a full connection layer.
The method for extracting the time-series characteristics through the characteristic extraction model in the step S4 includes:
all the test set sequences after the preliminary dimension reduction are input into the trained feature extraction model according to batches, the Batch Size (Batch Size) is set to be 64, and each sequence outputs 32-dimensional feature vectors.
The method for M-Tree search Tree organization index in S5 comprises the following steps:
the method comprises the steps of taking a feature vector output by a feature extraction model as an actual object, organizing a measurement space by an M-Tree based on two attributes including a coverage radius and a distance between the M-Tree and a father node, adopting an Euclidean distance meeting a distance triangle inequality as a measurement function to quickly construct an index, filtering most dissimilar sequences through a threshold value r, and adding the sequences meeting the similarity condition into a candidate set.
The method for implementing fault diagnosis in S6 includes:
s6.1, measuring the similarity between the query sequence in the state to be determined and all candidate sequences in the candidate set by applying a DTW algorithm, wherein the smaller the DTW distance is, the more similar the query sequence and the candidate sequences are;
s6.2, arranging the DTW distances in an ascending order, and taking the sequence with the minimum distance as the most similar matching result;
and S6.3, according to the most similar sequence obtained by searching and matching, determining the state abnormality of the current query sequence by comparing the tags, and realizing fault diagnosis.
The beneficial technical effects of the invention are as follows:
the invention provides a novel piecewise linear representation method, and simultaneously considers the whole trend and local important characteristics of the time sequence, thereby solving the problem of huge calculation and storage overhead of large-scale time sequences in similarity measurement. The time sequence features with strong separability can be automatically extracted by the feature extraction model built based on the Transformer, the similarity search efficiency is greatly improved by the M-Tree index, and the fault diagnosis efficiency and accuracy can be effectively improved. The method has simple steps and clear logic, provides a new technology for the fault diagnosis of the motor bearing from the angle of time series similarity measurement, and has good practical engineering application value.
Another aspect of the present invention provides a non-transitory readable recording medium storing one or more programs including instructions that, when executed, cause a processing circuit to execute a method for diagnosing a fault of a motor bearing as described above.
In another aspect of the present invention, a system for diagnosing a fault of a motor bearing is provided, which includes a processing circuit and a memory electrically coupled thereto, where the memory is configured to store at least one program, the program includes a plurality of instructions, and the processing circuit executes the program to perform the method for diagnosing a fault of a motor bearing.
Drawings
Fig. 1 is a schematic diagram of a fluctuation point extracted from fault sample data in an embodiment of the present invention.
Fig. 2 is a schematic diagram of key points screened from fault sample data in the embodiment of the present invention.
Fig. 3 is a schematic diagram of fault sample data after being piecewise-linearly represented by KP-PAA in the embodiment of the present invention.
Fig. 4 is a schematic structural diagram of a feature extraction model in the embodiment of the present invention.
FIG. 5 is a flowchart of a method for diagnosing a fault of a motor bearing according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention, and the described embodiments are some embodiments, but not all embodiments, of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any new work, are within the scope of the present invention.
Referring to the attached drawings in the specification, the embodiment provides a specific implementation process of a motor bearing fault diagnosis method based on similarity measurement, and the steps are as follows:
s1, a bearing data set disclosed by American West university of storage is used as an implementation object, vibration signals are collected through an accelerometer, the accelerometer is connected to a shell through a magnetic base and placed at a driving end of the shell of a motor, and bearing fault data are collected at a sampling rate of 48 kHz. The fault types comprise three bearing rolling element faults, three bearing inner ring faults, three bearing outer ring faults and a normal state. Each time sequence has about 24 ten thousand measuring points, 1000 4800-dimensional sequences are extracted from an original data set by adopting a sliding window sampling method to serve as a new experimental data set, and the new experimental data set is divided into a training set, a verification set and a test set according to the proportion of 7:1:2 by adopting a layered sampling method.
S2, carrying out preliminary dimension reduction on the original large-scale time sequence by applying a piecewise linearity representation algorithm KP-PAA;
s2.1 time series X ═ X1,x2,…,xnThere are nine variation patterns among three adjacent points according to the time increasing order. If xiThe following conditions are satisfied:
Figure BDA0003476385370000061
then call xiFor the fluctuation point, a fluctuation point sequence X can be extracted from the original time sequence according to the definition of the fluctuation pointfTaking a bearing inner ring fault sample as an example, the extracted fluctuation points are shown as the five-pointed star points in fig. 1;
s2.2 setting a suitable screening threshold epsilon according to the oscillation amplitude of the original sequence0In this example, let0Is 0.3. Taking a first value in the fluctuation point sequence as a comparison point a; if the subsequent fluctuation point xfiThe following conditions are satisfied:
abs(xfi-a)＞ε0 7(8)
will fluctuate point xfiAdding the index and the value of the key point sequence into the key point sequence, updating the value of the comparison point a until all the fluctuation points are screened, and finally obtaining the key point sequence Xk. The key points obtained by screening are shown as dots in figure 2;
s2.3 converting a sequence of 4800-long fault samples X to { X }1,x2,…,x4800Equally-spaced and segmented mean values are obtained according to the interval 10, and the mean values are converted into a new sequence Y with the length of 480 ═ Y1,y2,…,y480}, wherein:
Figure BDA0003476385370000071
s2.4, according to the index of the key point in the original time sequence, finding the segment where the key point is located in the PAA, and then replacing the average value of the segment where the key point is located with the value of the key point to obtain a final segment linear representation sequence KP-PAA after fusion, as shown in FIG. 3.
S3, referring to an encoder part of a Transformer, and by using design skills of variants of the encoder part, such as an ViT model, an Informer model and a BERT model, introducing a time embedding layer, a distillation layer and a feature token, and constructing an efficient feature extraction model, wherein the model structure is shown in FIG. 4;
s4, inputting the test set sequences subjected to preliminary dimension reduction into a trained feature extraction model according to batches to further extract time sequence features, setting the Batch Size (Batch Size) to be 64, and outputting 32-dimensional feature vectors by each sequence;
and S5, taking the feature vector output by the feature extraction model as an actual object, organizing a measurement space by the M-Tree based on two attributes including a coverage radius and a distance between the M-Tree and a father node, and adopting an Euclidean distance meeting a distance triangle inequality as a measurement function so as to quickly construct an index. The screening threshold r is set to 2, most of the dissimilar sequences can be filtered out, and the sequences meeting the similarity condition are added into the candidate set.
S6, measuring the similarity between the query sequence in the state to be determined and all candidate sequences in the candidate set by applying a DTW algorithm, wherein the smaller the DTW distance is, the more similar the two sequences are, and taking the sequence with the minimum distance as the most similar matching result. And determining the abnormal state of the current query sequence by comparing the labels of the sequence samples, realizing fault diagnosis, and calculating the accuracy of fault classification by using a 1NN classifier.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computers, usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.), and having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The technical scheme of the invention is that the method steps are compiled into a program and then the program is stored in a hard disk or other non-transient storage media to form the non-transient readable recording medium; the storage medium is electrically connected with a computer processor, and fault diagnosis of the motor bearing can be completed through data processing, so that the technical scheme of the fault diagnosis system of the motor bearing is formed.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that changes may be made in the embodiments and/or equivalents thereof without departing from the spirit and scope of the invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. A fault diagnosis method for a motor bearing is characterized by comprising the following steps:
s1, dividing the obtained original time sequence data of the motor bearing into a training set, a verification set and a test set in a layered sampling mode;
s2, performing dimensionality reduction on the original time sequence data by applying a piecewise linear representation algorithm;
s3, building a feature extraction model through the training set, and optimizing parameters through the verification set until the feature extraction model is converged;
s4, extracting the feature vectors of all sequences in the test set by applying a feature extraction model;
s5, organizing indexes of the feature vectors through an M-Tree search Tree, and adding sequences meeting similarity conditions into a candidate set;
and S6, measuring the similarity between the sequence to be queried and each sequence in the candidate set by applying a dynamic time warping mode, so as to match abnormal states and realize fault diagnosis.
2. The method for diagnosing the fault of the motor bearing according to the claim 1, wherein the hierarchical sampling mode is as follows: dividing the raw time sequence data of the motor bearing into a training set, a verification set and a test set according to the ratio of 7:1: 2.
3. The method for diagnosing the fault of the motor bearing according to claim 2, wherein the piecewise linear representation algorithm in the step S2 includes:
s2.1, extracting a fluctuation point sequence of the original time series data;
s2.2, on the basis of the fluctuation point sequence, comparing set thresholds, and screening to obtain a key point sequence;
s2.3, reducing the dimension of the original time sequence by adopting a segmentation aggregation approximation method;
and S2.4, fusing the key points and the time sequence after dimensionality reduction according to the index of the key point sequence in the original time sequence to obtain a KP-PAA sequence.
4. The method of claim 3, wherein the step S2.1 comprises:
the original time series data X ═ { X ═ X1,x2,…,xnArranging three adjacent points in time increasing order if x isiThe following conditions are satisfied:
Figure FDA0003476385360000021
then call xiAnd extracting all fluctuation points as fluctuation points to obtain the fluctuation point sequence.
5. A non-transitory readable recording medium storing one or more programs containing instructions which, when executed, cause a processing circuit to perform a method of diagnosing a fault in a motor bearing according to any one of claims 1 to 4.
6. A system for diagnosing a failure of a motor bearing, comprising a processing circuit and a memory electrically coupled thereto, wherein the memory is configured to store at least one program, the program comprising a plurality of instructions, and the processing circuit executes the program to perform a method for diagnosing a failure of a motor bearing as claimed in claim 4.
CN202210055670.8A 2022-01-18 2022-01-18 Fault diagnosis method, recording medium and system for motor bearing Pending CN114564619A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210055670.8A CN114564619A (en) 2022-01-18 2022-01-18 Fault diagnosis method, recording medium and system for motor bearing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210055670.8A CN114564619A (en) 2022-01-18 2022-01-18 Fault diagnosis method, recording medium and system for motor bearing

Publications (1)

Publication Number Publication Date
CN114564619A true CN114564619A (en) 2022-05-31

Family

ID=81711515

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210055670.8A Pending CN114564619A (en) 2022-01-18 2022-01-18 Fault diagnosis method, recording medium and system for motor bearing

Country Status (1)

Country Link
CN (1) CN114564619A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117708098A (en) * 2024-02-05 2024-03-15 中国第一汽车股份有限公司 Battery fault diagnosis method, device, electronic equipment and storage medium

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117708098A (en) * 2024-02-05 2024-03-15 中国第一汽车股份有限公司 Battery fault diagnosis method, device, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
CN107368809B (en) A kind of bearing fault sorting technique based on rarefaction representation and dictionary learning
Zheng et al. A rolling bearing fault diagnosis method based on multi-scale fuzzy entropy and variable predictive model-based class discrimination
CN105275833B (en) CEEMD (Complementary Empirical Mode Decomposition)-STFT (Short-Time Fourier Transform) time-frequency information entropy and multi-SVM (Support Vector Machine) based fault diagnosis method for centrifugal pump
Xiong et al. Application of convolutional neural network and data preprocessing by mutual dimensionless and similar gram matrix in fault diagnosis
Zheng et al. Refined time-shift multiscale normalised dispersion entropy and its application to fault diagnosis of rolling bearing
CN115510965A (en) Bearing imbalance fault diagnosis method based on generated data fusion
CN108197647A (en) A kind of Fast Speed Clustering of automobile starter durable test data
CN116593157A (en) Complex working condition gear fault diagnosis method based on matching element learning under small sample
CN113008559A (en) Bearing fault diagnosis method and system based on sparse self-encoder and Softmax
CN115452381A (en) Rolling bearing fault diagnosis method under strong noise interference based on multi-modal decomposition and long-short term memory network
CN114564619A (en) Fault diagnosis method, recording medium and system for motor bearing
CN111458145A (en) Cable car rolling bearing fault diagnosis method based on road map characteristics
CN110879802A (en) Log pattern extraction and matching method
CN110929761A (en) Balance method for collecting samples in situation awareness framework of intelligent system security system
CN112541524B (en) BP-Adaboost multisource information motor fault diagnosis method based on attention mechanism improvement
CN112362292B (en) Method for anomaly detection of wind tunnel test data
CN111382792B (en) Rolling bearing fault diagnosis method based on double-sparse dictionary sparse representation
CN111786999B (en) Intrusion behavior detection method, device, equipment and storage medium
CN114912458A (en) Emotion analysis method and device and computer readable medium
CN117078960A (en) Near infrared spectrum analysis method and system based on image feature extraction
Pan et al. Unsupervised two-stage root-cause analysis for integrated systems
CN114781450B (en) Train rolling bearing state identification method based on parameter optimization MOMEDA-MIA-CNN
CN113033683B (en) Industrial system working condition monitoring method and system based on static and dynamic joint analysis
CN114088400B (en) Rolling bearing fault diagnosis method based on envelope permutation entropy
CN111914652B (en) Face feature binary coding and recognition method based on two-layer self-organizing neural network

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