CN112036458A - Fault diagnosis method for rolling bearing - Google Patents

Fault diagnosis method for rolling bearing Download PDF

Info

Publication number
CN112036458A
CN112036458A CN202010848070.8A CN202010848070A CN112036458A CN 112036458 A CN112036458 A CN 112036458A CN 202010848070 A CN202010848070 A CN 202010848070A CN 112036458 A CN112036458 A CN 112036458A
Authority
CN
China
Prior art keywords
fault
optimal transport
transport distance
bearing
matrix
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010848070.8A
Other languages
Chinese (zh)
Other versions
CN112036458B (en
Inventor
陈云龙雨
李靖超
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.)
Shanghai Dianji University
Original Assignee
Shanghai Dianji 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 Shanghai Dianji University filed Critical Shanghai Dianji University
Priority to CN202010848070.8A priority Critical patent/CN112036458B/en
Publication of CN112036458A publication Critical patent/CN112036458A/en
Application granted granted Critical
Publication of CN112036458B publication Critical patent/CN112036458B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification

Landscapes

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

Abstract

The invention discloses a fault diagnosis method for a rolling bearing, which comprises the following steps: setting respective failure criterion probability distributions Yi(ii) a Respectively solving each signal X to be measurediAnd the fault reference probability distribution YiThe optimal transport distance therebetween; establishing an optimal transport distance matrix, and taking the optimal transport distance matrix as a fault characteristic matrix T; and classifying and judging the fault characteristic matrix T by using a classifier and outputting a diagnosis result. The invention can effectively diagnose the bearing fault under the conditions of less sample quantity and incomplete sample, has high accuracy, short time consumption and good real-time performance, and is not limited by a classifier; the method can accurately predict the fault of unknown working conditions, and solves the problem of bearing fault diagnosis under complex working conditions in the actual operation process.

Description

Fault diagnosis method for rolling bearing
Technical Field
The invention relates to the field of mechanical faults, in particular to real-time fault diagnosis of a bearing vibration signal.
Background
The fault diagnosis of the rolling bearing mainly provides support for reliable bearing operation, and the existing methods are mainly divided into four types: the method comprises a time domain analysis method, a frequency domain analysis method, a time frequency analysis method and a method based on an intelligent optimization algorithm, wherein the time domain analysis is to analyze the composition and characteristic quantity of a signal according to the time waveform of the signal, and common time domain indexes comprise a peak value, an effective value, a pulse, a margin, a kurtosis and the like, but the time domain indexes are poor in precision and are easily influenced by noise; the frequency domain analysis method is to transform a fault signal into a frequency domain, and analyzes a frequency spectrum, wherein the commonly used methods comprise Fourier transform, spectral kurtosis, cepstrum and the like, and the problems of precision and parameter selection exist; the time-frequency analysis method maps time-domain signals to a time-frequency plane, reflects the time-frequency joint characteristics of non-stationary signals, is commonly used for wavelet decomposition, empirical mode decomposition, local mean decomposition and the like, but has the problems of mode aliasing phenomenon, time-frequency distribution deviation and the like; methods based on intelligent optimization algorithms mainly include genetic algorithms, convolutional neural networks, deep confidence networks and the like, and the methods require a large number of samples and long training time and are difficult to diagnose in real time.
The time domain analysis method usually needs a plurality of time domain indexes to extract characteristic quantities, the characteristic quantities cannot accurately reflect the difference of faults, and the prediction accuracy is general; the frequency domain analysis method and the time frequency analysis method need to carry out a transform domain, and need to design various complex algorithms for iteration, so that the diagnosis period is longer; the method based on the intelligent optimization algorithm has huge requirements on the basic data quantity of the fault, and the constructed model lacks physical significance and needs a long-time training model. In addition, when the bearing works under various working conditions, and particularly under the working condition that the sample does not exist, the model is difficult to adapt to change, and the identification accuracy rate is greatly reduced. Therefore, how to solve the problems of low accuracy, long training time and poor real-time performance in the bearing fault diagnosis process, especially when the bearing works under an unknown working condition, is the problem to be solved at present.
Disclosure of Invention
The invention aims to solve the technical problems of low accuracy, long training time and poor real-time performance in the bearing fault diagnosis process, particularly when a bearing works under an unknown working condition, and provides a rolling bearing fault diagnosis method based on an optimal transport theory.
The invention solves the technical problems through the following technical scheme:
a fault diagnosis method for a rolling bearing, the fault diagnosis method comprising:
setting respective failure criterion probability distributions Yi；
Respectively solving each signal X to be measurediAnd the fault reference probabilitiesDistribution YiThe optimal transport distance therebetween;
establishing an optimal transport distance matrix, and taking the optimal transport distance matrix as a fault characteristic matrix T;
and classifying and judging the fault characteristic matrix T by using a classifier and outputting a diagnosis result.
Further, acquiring the reference probability distribution Y of each faultiThe method comprises the following steps:
taking the obtained different state data of the bearing as a low-dimensional manifold of high-dimensional data in an optimal transport model;
the probability distribution of the different state data of the bearing respectively corresponds to different subclasses of the low-dimensional manifold;
obtaining the reference probability distribution Y of each faulti。
Further, obtaining the optimal transport distance includes:
defining a non-negative measurable function C (X, Y) in X by Y space;
obtaining marginal probability U and V, projection mapping pix＝U，πy＝V；
Obtaining the optimal transport distance Xi＝{∫X×YC(x，y)dπ(x，y)}。
Preferably, the establishing the optimal distance matrix comprises:
respectively acquiring the optimal transport distance of each fault;
and establishing the optimal transport distance matrix.
Further, the faults may occur under different operating conditions.
On the basis of the common knowledge in the field, the above preferred conditions can be combined randomly to obtain the preferred embodiments of the invention.
The positive progress effects of the invention are as follows: the bearing fault diagnosis method has the advantages that the bearing fault can be effectively diagnosed under the conditions of less sample quantity and incomplete samples, the accuracy is high, the time consumption is short, the real-time performance is good, and the method is not limited by a classifier; the method can accurately predict the fault of unknown working conditions, and solves the problem of bearing fault diagnosis under complex working conditions in the actual operation process.
Drawings
FIG. 1 is a flowchart of a method in one embodiment of a rolling bearing fault diagnosis method of the present invention;
FIG. 2 is a diagram of optimal transport distances in four states in an embodiment of a rolling bearing fault diagnosis method according to the present invention;
fig. 3 is a graph showing the change of the identification rate with the k value in an embodiment of the rolling bearing fault diagnosis method of the present invention.
Detailed Description
To facilitate an understanding of the present application, the present application will now be described more fully with reference to the accompanying drawings. Preferred embodiments of the present application are shown in the drawings. This application may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
It will be understood that when an element is referred to as being "connected" to another element, it can be directly connected to the other element and be integral therewith, or intervening elements may also be present. The terms "mounted," "one end," "the other end," and the like are used herein for illustrative purposes only.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.
Fig. 1 is a flowchart of a method for determining a rolling bearing failure according to an embodiment of the present invention:
s01: setting respective failure criterion probability distributions Yi；
In one example, the probability distributions for different states of the bearing correspond to different subclasses of the same manifold Ω, and the invention is divided intoRespectively selecting a group of random data of four faults as reference probability distribution Yi。
S02: respectively solving each signal X to be measurediAnd the fault reference probability distribution YiThe optimal transport distance therebetween;
in one example, defining C as a non-negative measurable function in X Y space, C (X, Y) representing the quality from the starting point X e X to the end point Y e Y, the invention selects the absolute value of the difference value of the C function; marginal probability of transmission scheme is equal to u and v, projection mapping pix＝u，πyV; optimal transmission x1 ═ min { [ integral ] nX×YC(x,y)dπ(x,y)}。
S03: establishing an optimal transport distance matrix, and taking the optimal transport distance matrix as a fault characteristic matrix T;
in one example, the process II is repeated, four groups of optimal transport distances are respectively obtained, that is, x1, x2, x3, and x4, and the obtained optimal transport distance matrix is used as the feature matrix T of the fault [ x1, x2, x3, and x4 ].
S04: classifying and judging the fault characteristic matrix T by using a classifier and outputting a diagnosis result;
in an optional example, experimental data of bearing testing center of Kaiser Sichu university, USA, is adopted, and the data of measuring points of a motor driving end are used, faults with the size of 0.18mm are artificially set on an inner ring, an outer ring and a rolling body of a bearing, the reference rotating speed is 1796r/min, the sampling frequency is 12kHz, the sampling time is 10s, wherein 4 faults are divided, and each fault corresponds to four working conditions. In the invention, 0.167s is used as a window, 120 groups of data in total of four states under 0hp are used in a training set, 960 groups of samples in total of a data set containing three unknown working conditions are used in a test set, and as shown in table 1:
TABLE 1 bearing data of Kaiser university of West reservoir
Bearing condition 0hp 1hp 2hp 3hp
Normal state 60 60 60 60
Inner ring failure 60 60 60 60
Ball failure 60 60 60 60
Outer ring failure 60 60 60 60
Number of samples 240 240 240 240
As shown in FIG. 2, the optimal transport distances of four different states have obvious difference and are quite stable in the characteristics of each fault, the lowest rhombus is a characteristic value in the same state, the optimal transport distance is basically 0, the self-similarity is reflected, and the theoretical derivation is met.
In an alternative example, the feature matrix is tested by using KNN (K-nearest Neighbors) algorithm, the diagnosis time of 480 samples is 0.178s, the accuracy is 100%, and the K value is 1 to 19, as shown in fig. 3, the accuracy is 100%.
In an alternative example, the feature matrix is tested using a random forest algorithm with a diagnosis time of 0.658s and a 100% accuracy.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that this is by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.

Claims (5)

1. A fault diagnosis method for a rolling bearing, characterized by comprising:
setting respective failure criterion probability distributions Yi；
Respectively solving each signal X to be measurediAnd the fault reference probability distribution YiThe optimal transport distance therebetween;
establishing an optimal transport distance matrix, and taking the optimal transport distance matrix as a fault characteristic matrix T;
and classifying and judging the fault characteristic matrix T by using a classifier and outputting a diagnosis result.
2. A rolling bearing failure diagnosis method according to claim 1, whereinIn that the probability distribution Y of each fault criterion is obtainediThe method comprises the following steps:
taking the obtained different state data of the bearing as a low-dimensional manifold of high-dimensional data in an optimal transport model;
the probability distribution of the different state data of the bearing respectively corresponds to different subclasses of the low-dimensional manifold;
obtaining the reference probability distribution Y of each faulti。
3. The rolling bearing fault diagnosis method according to claim 1, wherein obtaining the optimal transport distance comprises:
defining a non-negative measurable function C (X, Y) in X by Y space;
obtaining marginal probability U and V, projection mapping pix＝U，πy＝V；
Obtaining the optimal transport distance Xi＝{∫X×YC(x，y)dπ(x，y)}。
4. A rolling bearing fault diagnosis method according to claim 3, characterized in that constructing the optimal distance matrix comprises:
respectively acquiring the optimal transport distance of each fault;
and establishing the optimal transport distance matrix.
5. A method for diagnosing faults of rolling bearings according to any one of claims 1 to 4, wherein the faults are likely to occur under different operating conditions.
CN202010848070.8A 2020-08-21 2020-08-21 Rolling bearing fault diagnosis method Active CN112036458B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010848070.8A CN112036458B (en) 2020-08-21 2020-08-21 Rolling bearing fault diagnosis method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010848070.8A CN112036458B (en) 2020-08-21 2020-08-21 Rolling bearing fault diagnosis method

Publications (2)

Publication Number Publication Date
CN112036458A true CN112036458A (en) 2020-12-04
CN112036458B CN112036458B (en) 2023-05-23

Family

ID=73581687

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010848070.8A Active CN112036458B (en) 2020-08-21 2020-08-21 Rolling bearing fault diagnosis method

Country Status (1)

Country Link
CN (1) CN112036458B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113822228A (en) * 2021-10-27 2021-12-21 南京大学 User expression recognition method and system based on continuous learning

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104133732A (en) * 2014-06-23 2014-11-05 合肥工业大学 Fault-tolerant method aiming at TSV fault grading in 3D NoC
CN110749443A (en) * 2019-11-27 2020-02-04 济南大学 Rolling bearing fault diagnosis method and system based on high-order origin moment

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104133732A (en) * 2014-06-23 2014-11-05 合肥工业大学 Fault-tolerant method aiming at TSV fault grading in 3D NoC
CN110749443A (en) * 2019-11-27 2020-02-04 济南大学 Rolling bearing fault diagnosis method and system based on high-order origin moment

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CHAO CHEN 等: "Topic Correlation Analysis for Bearing Fault Diagnosis Under Variable Operating Conditions", 《12TH INTERNATIONAL CONFERENCE ON DAMAGE ASSESSMENT OF STRUCTURES》 *
杨超: "多工况过程迁移建模方法与应用研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》 *
马丽涛 等: "最优传输理论及其在图像处理中的应用", 《运筹学学报》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113822228A (en) * 2021-10-27 2021-12-21 南京大学 User expression recognition method and system based on continuous learning
CN113822228B (en) * 2021-10-27 2024-03-22 南京大学 User expression recognition method and system based on continuous learning

Also Published As

Publication number Publication date
CN112036458B (en) 2023-05-23

Similar Documents

Publication Publication Date Title
CN112418013B (en) Complex working condition bearing fault diagnosis method based on meta-learning under small sample
CN109829402B (en) GS-SVM-based bearing damage degree diagnosis method under different working conditions
CN112084974B (en) Multi-label rolling bearing fault diagnosis method based on meta-learning
WO2021135630A1 (en) Rolling bearing fault diagnosis method based on grcmse and manifold learning
CN108388860B (en) Aero-engine rolling bearing fault diagnosis method based on power entropy spectrum-random forest
CN104198184A (en) Bearing fault diagnostic method based on second generation wavelet transform and BP neural network
CN109827777B (en) Rolling bearing fault prediction method based on partial least square method extreme learning machine
CN108760266B (en) The virtual degeneration index building method of mechanical key component based on learning distance metric
CN111562108A (en) Rolling bearing intelligent fault diagnosis method based on CNN and FCMC
WO2023137807A1 (en) Rolling bearing class imbalance fault diagnosis method and system
CN111412977A (en) Preprocessing method for vibration sensing data of mechanical equipment
CN110243603B (en) Rolling bearing fault diagnosis method based on Welch conversion-radial basis function neural network
CN110595778B (en) Wind turbine generator bearing fault diagnosis method based on MMF and IGRA
CN109932179A (en) A kind of rolling bearing fault testing method based on the reconstruct of DS Adaptive spectra
CN110849625A (en) Bearing fault diagnosis method under variable working condition based on mixed entropy and joint distribution adaptation
CN112364706A (en) Small sample bearing fault diagnosis method based on class imbalance
CN113268833A (en) Migration fault diagnosis method based on deep joint distribution alignment
CN106198020A (en) Wind turbines bearing failure diagnosis method based on subspace and fuzzy C-means clustering
CN105241665A (en) Rolling bearing fault diagnosis method based on IRBFNN-AdaBoost classifier
CN112182912B (en) Manufacturing equipment spindle bearing health assessment method based on probability description and spectrum analysis
CN115127806A (en) Gear box fault diagnosis method and device based on multi-sensor vibration signals
CN112036458A (en) Fault diagnosis method for rolling bearing
CN112541510A (en) Intelligent fault diagnosis method based on multi-channel time series data
CN114118219A (en) Data-driven real-time abnormal detection method for health state of long-term power-on equipment
CN103728135A (en) Bearing fault feature extraction and diagnosis method of non-negative matrix factorization

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