CN110160781A - A kind of the test set reconstruct and prediction technique of rotating machinery fault classification - Google Patents

A kind of the test set reconstruct and prediction technique of rotating machinery fault classification Download PDF

Info

Publication number
CN110160781A
CN110160781A CN201910525563.5A CN201910525563A CN110160781A CN 110160781 A CN110160781 A CN 110160781A CN 201910525563 A CN201910525563 A CN 201910525563A CN 110160781 A CN110160781 A CN 110160781A
Authority
CN
China
Prior art keywords
rotating machinery
sample
test set
machinery fault
reconstruct
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
CN201910525563.5A
Other languages
Chinese (zh)
Other versions
CN110160781B (en
Inventor
王世华
张清华
周东华
韩建宇
陈旭
吴淦洲
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.)
Guangdong University of Petrochemical Technology
Original Assignee
Guangdong University of Petrochemical Technology
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 Guangdong University of Petrochemical Technology filed Critical Guangdong University of Petrochemical Technology
Priority to CN201910525563.5A priority Critical patent/CN110160781B/en
Publication of CN110160781A publication Critical patent/CN110160781A/en
Application granted granted Critical
Publication of CN110160781B publication Critical patent/CN110160781B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G01M13/02—Gearings; Transmission mechanisms
    • G01M13/028—Acoustic or vibration analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines

Landscapes

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

Abstract

Test set reconstruct and prediction technique the invention discloses a kind of classification of rotating machinery fault, comprising the following steps: S01 extracts the vibration signal of failure part using acceleration transducer, obtains rotating machinery fault data；S02, the index feature of building rotating machinery fault classification, five waveform, pulse, nargin, peak value and kurtosis dimensionless indexs of rotating machinery are calculated using software；S03 selects two kinds of sorting algorithms of SVM and KNN to carry out failure modes；The test set of rotating machinery fault classification is reconstructed in S04；S05, the prediction label for establishing sample to be tested under new test set determine method.Test set is reconstructed in the test set reconstruct and prediction technique of a kind of rotating machinery fault classification provided by the invention, the main correlation using between fault sample, can be improved rotating machinery fault classification accuracy.

Description

A kind of the test set reconstruct and prediction technique of rotating machinery fault classification
Technical field
Test set reconstruct and prediction technique the present invention relates to a kind of classification of rotating machinery fault, belong to rotating machinery fault With wireless sensor fault diagnosis field.
Background technique
Wireless sensor network (WSN Wireless Sensor Network) is that have perception, meter by numerous distributions Calculate the small-scale wireless network constituted by way of self-organizing with the microsensor node of communication capacity.Sensor Node is shared out the work and helped one another by mutual, can real-time perception, monitoring and monitoring object or surrounding ring in acquisition distributed areas The information in border, and send picker to.Wireless sensor network node is integrated with sensor, data processing unit and communication mould Block constitutes network by way of self-organizing, according to concrete application require configuration multiple sensors, can measure including temperature, Numerous object informations such as humidity, vibration, noise, light intensity, pressure, the size of mobile object, speed and direction.Rotating machinery It is easy to happen failure, if cannot find in time failure and debug, it is most likely that cause the industrial accident of Large-scale machine set, Great economic asset is caused to lose.
Whether the operating status of the rotating machineries such as bearing and gear normally directly influences adding for entire heavy mechanical equipment The performance parameters such as work precision, operational reliability and service life.Carry out condition monitoring for rotating machinery and fault diagnosis, prediction technique is ground Studying carefully is the basis for guaranteeing that mechanical equipment operational safety is stable, and current diagnostic techniques and signal analysis method are to industrial production application In mechanical equipment running state real-time monitoring and fault diagnosis played important function.With industrial enlargement, it is Systemization and informationization require to be increasingly stringenter to the safe operation of critical component rotating machinery, to the failure of gear and bearing into It is about to identification and diagnosis has great importance.
Method for diagnosing faults is generally divided into 3 classes: Knowledge based engineering method for diagnosing faults, the fault diagnosis based on analytic modell analytical model Method and method for diagnosing faults based on signal processing.Knowledge based engineering method for diagnosing faults is adapted to or be not easy to establish machine Manage model, number of sensors deficiency, the system of poor information.Method based on analytic modell analytical model be adapted to number of sensors it is sufficient, The system of information abundance needs to fully understand the mechanism of process, and can establish accurate quantitative math-model.Knowledge based and Method for diagnosing faults based on parsing is usually used in monitoring the less situation of procedure parameter.With the increase of procedure parameter, at present compared with It is the method for diagnosing faults based on data-driven for popular method for diagnosing faults.Method for diagnosing faults pair based on data-driven Process monitoring data is analyzed and processed, the case where the priori knowledge for not depending on system is with system accurate model is required no knowledge about Lower completion fault diagnosis.
In existing Research on fault diagnosis method, extraction and the failure modes algorithm of fault data collection feature are focused primarily upon It improves, and the present invention is then mainly reconstructed test set using the correlation between fault sample, improves failure modes effect Fruit.
Existing Research on fault diagnosis method focuses primarily upon the extraction of fault data collection feature and changing for failure modes algorithm Into failure modes effect is not ideal enough.
Summary of the invention
The technical problem to be solved by the present invention is in view of the deficiencies of the prior art, provide one kind and mainly utilize fault sample Between correlation test set is reconstructed, can be improved the favourable turn tool failure modes of rotating machinery fault classification accuracy Test set reconstruct and prediction technique.
In order to solve the above technical problems, the technical solution adopted by the present invention are as follows:
A kind of the test set reconstruct and prediction technique of rotating machinery fault classification, comprising the following steps:
S01 extracts the vibration signal of failure part using acceleration transducer, obtains rotating machinery fault data；
S02, the index feature of building rotating machinery fault classification, using software calculate the waveform of rotating machinery, pulse, Five nargin, peak value and kurtosis dimensionless indexs；
S03 selects two kinds of sorting algorithms of SVM and KNN to carry out failure modes；
The test set of rotating machinery fault classification is reconstructed in S04；
S05, the prediction label for establishing sample to be tested under new test set determine method.
The acquisition of rotating machinery fault data uses EMT490 data collector.
In S02, waveform, pulse, nargin, the specific calculation of peak value and kurtosis index are as follows:
Waveform index:
Peak index:
Pulse index:
Margin index:
Kurtosis index:
In formula, XrmsFor the root-mean-square value of signal,For the average amplitude of signal, XmaxFor the maximum value of signal, XrFor letter Number root amplitude, β is kurtosis, N indicates to constitute the signal sampling point number of sample, XiIndicate that vibration amplitude, i indicate the subscript index of sampled point.
In S03, failure modes structure meets pii> pij, (j ≠ i, i, j=1,2 ..., n) is set up, wherein pijIndicate practical Type is the ratio that i and sorting algorithm are predicted as type j, piiIndicate the ratio that actual type is i and sorting algorithm is predicted as type i Example, n are the number of types of failure modes.
In S04, specific steps are test sample correlation length l to be determined, if sample according to prediction effect to be achieved (1), (2) sample ..., sample (k) are former test set sample, carry out the reconstruct of test sample collection as follows:
Wherein, S1,S2,…,SkFor the test set sample set rebuild.
In S05, the prediction label of each test sample in sample set is determined first with sorting algorithm, establishes following throwing Ticket mechanism will be predicted as same type of number and count in sample set, if the statistical magnitude of prediction label h is most in sample set Big and unique, then the label of the test sample collection newly constructed is h；If the maximum more than one of prediction label statistical magnitude, with Machine select the maximum label of statistical magnitude as prediction label.
Beneficial effects of the present invention: the test set that the present invention provides a kind of rotating machinery fault classification reconstructs and prediction side Method is reconstructed test set using the correlation between fault sample, and the prediction of test sample is established using voting mechanism The determination method of label, greatly improves the predictablity rate of test sample.
Detailed description of the invention
Fig. 1 is a kind of flow chart of the test set reconstruct and prediction technique of rotating machinery fault classification of the present invention.
Specific embodiment
The present invention will be further described below with reference to examples, and following embodiment is only used for clearly illustrating this hair Bright technical solution, and not intended to limit the protection scope of the present invention.
As shown in Figure 1, the test set reconstruct and prediction technique of a kind of rotating machinery fault classification provided by the invention, including Following steps:
Step 1 installs different rolling bearings and gearbox fault part on this experiment porch, and vibration signal is by pacifying Acceleration transducer on bearing block extracts, and fault vibration signal is acquired by EMT490 data collector.Experiment Parameter is as follows: revolving speed 1000r/min, sample frequency 1000Hz, and the sampling number for calculating dimensionless index is 1024 points.
Collected vibration acceleration fault data is imported on computer by step 2, and using MATLAB software to it It is read out.The rotating machinery fault of this experiment mainly includes five kinds: gear wheel hypodontia, pinion gear hypodontia, bearing inner race abrasion, Bearing outer ring abrasion and normal condition, sample the vibration signal under this five kinds of states, every kind of state samples number respectively It is 102400.
Step 3 calculates 5 dimensionless indexs: waveform index, peak index, arteries and veins using continuous 1024 sampled points Rush index, margin index, kurtosis index, and with the decision category of rotating machinery operating condition marker samples in embodiment step 2 Property value；
Waveform index:
Peak index:
Pulse index:
Margin index:
Kurtosis index:
In formula, XrmsFor the root-mean-square value of signal,For the average amplitude of signal, XmaxFor the maximum value of signal, XrFor letter Number root amplitude, β is kurtosis, N Indicate the signal sampling point number of composition sample, XiIndicate that vibration amplitude, i indicate the subscript index of sampled point.
Step 4 selects two kinds of sorting algorithms of SVM and KNN to carry out classification prediction to failure, wherein randomly choosing each event Hinder the sample of type 40% as training set, remaining is used 60% as test sample to carry out classification prediction.Failure modes structure Meet pii> pij, (j ≠ i, i, j=1,2 ..., n) is set up, wherein pijIndicate that actual type is i and sorting algorithm is predicted as class The ratio of type j, piiIndicate the ratio that actual type is i and sorting algorithm is predicted as type i, n is the number of types of failure modes.
Step 5 carries out test set reconstruct according to the step of summary of the invention three, and specific steps are according to be achieved pre- Effect is surveyed, determines test sample correlation length l, if sample (1), sample (2) ..., sample (k) are former test set sample This, carries out the reconstruct of test sample collection as follows:
Wherein, S1,S2,…,SkFor the test set sample set rebuild.
It is 60, then as l=10, the test set number of samples of reconstruct to each fault type original test sample number It is 51, then as l=20, the test set number of samples of reconstruct is 41, then as l=30, the test set sample of reconstruct Number is 31.
Step 6, the prediction label determination side of sample to be tested under the new test set provided according to the step of summary of the invention five Method counts the predictablity rate of former test set and reconstruct test set.It determines in sample set first with sorting algorithm and respectively tests The prediction label of sample establishes following voting mechanism, and same type of number will be predicted as in sample set and is counted, if sample The statistical magnitude of this concentration prediction label h is maximum and unique, then the label of the test sample collection newly constructed is h；If prediction label The maximum more than one of statistical magnitude, then randomly choose the prediction label that a maximum label of statistical magnitude is used as, and classification is quasi- True rate comparison the results are shown in Table 1.
Table 1
As it can be seen from table 1 for using two kinds of sorting algorithms of KNN algorithm and support vector machines, the failure of rotating machinery The predictablity rate that reconstruct test is gathered all is higher than the classification accuracy of former test set, and classification accuracy is with association length The increase of degree and increase.
The above is only a preferred embodiment of the present invention, it should be pointed out that: for the ordinary skill people of the art For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered It is considered as protection scope of the present invention.

Claims (6)

1. the test set reconstruct and prediction technique of a kind of rotating machinery fault classification, it is characterised in that: the following steps are included:
S01 extracts the vibration signal of failure part using acceleration transducer, obtains rotating machinery fault data；
S02, the index feature of building rotating machinery fault classification calculate the waveform of rotating machinery, pulse, abundant using software Five degree, peak value and kurtosis dimensionless indexs；
S03 selects two kinds of sorting algorithms of SVM and KNN to carry out failure modes；
The test set of rotating machinery fault classification is reconstructed in S04；
S05, the prediction label for establishing sample to be tested under new test set determine method.
2. a kind of test set reconstruct of rotating machinery fault classification according to claim 1 and prediction technique, feature exist In: the acquisition of rotating machinery fault data uses EMT490 data collector.
3. a kind of test set reconstruct of rotating machinery fault classification according to claim 1 and prediction technique, feature exist In: in S02, waveform, pulse, nargin, the specific calculation of peak value and kurtosis index are as follows:
Waveform index:
Peak index:
Pulse index:
Margin index:
Kurtosis index:
In formula, XrmsFor the root-mean-square value of signal,For the average amplitude of signal, XmaxFor the maximum value of signal, XrFor the side of signal Root range value, β are kurtosis, N table Show the signal sampling point number for constituting sample, XiIndicate that vibration amplitude, i indicate the subscript index of sampled point.
4. a kind of test set reconstruct of rotating machinery fault classification according to claim 1 and prediction technique, feature exist In: in S03, failure modes structure meets pii> pij, (j ≠ i, i, j=1,2 ..., n) is set up, wherein pijIndicate actual type For i, sorting algorithm is predicted as the ratio of type j, piiIndicate the ratio that actual type is i and sorting algorithm is predicted as type i, n For the number of types of failure modes.
5. a kind of test set reconstruct of rotating machinery fault classification according to claim 1 and prediction technique, feature exist In: in S04, specific steps are to determine test sample correlation length l according to prediction effect to be achieved, if sample (1), Sample (2) ..., sample (k) are former test set sample, carry out the reconstruct of test sample collection as follows:
Wherein, S1,S2,…,SkFor the test set sample set rebuild.
6. a kind of test set reconstruct of rotating machinery fault classification according to claim 1 and prediction technique, feature exist In: in S05, the prediction label of each test sample in sample set is determined first with sorting algorithm, establishes following voting machine System, will be predicted as same type of number and counts in sample set, if in sample set the statistical magnitude of prediction label h it is maximum and Uniquely, then the label of the test sample collection newly constructed is h；It is random to select if the maximum more than one of prediction label statistical magnitude Select the prediction label that a maximum label of statistical magnitude is used as.
CN201910525563.5A 2019-06-18 2019-06-18 Test set reconstruction and prediction method for rotary machine fault classification Active CN110160781B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910525563.5A CN110160781B (en) 2019-06-18 2019-06-18 Test set reconstruction and prediction method for rotary machine fault classification

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910525563.5A CN110160781B (en) 2019-06-18 2019-06-18 Test set reconstruction and prediction method for rotary machine fault classification

Publications (2)

Publication Number Publication Date
CN110160781A true CN110160781A (en) 2019-08-23
CN110160781B CN110160781B (en) 2020-12-04

Family

ID=67625801

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910525563.5A Active CN110160781B (en) 2019-06-18 2019-06-18 Test set reconstruction and prediction method for rotary machine fault classification

Country Status (1)

Country Link
CN (1) CN110160781B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183344A (en) * 2020-09-28 2021-01-05 广东石油化工学院 Large unit friction fault analysis method and system based on waveform and dimensionless learning
CN112270227A (en) * 2020-10-16 2021-01-26 广东石油化工学院 Oil film whirl and friction concurrent fault analysis method and analysis system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106371427A (en) * 2016-10-28 2017-02-01 浙江大学 Industrial process fault classification method based on analytic hierarchy process and fuzzy fusion
CN106646096A (en) * 2016-11-15 2017-05-10 国网四川省电力公司广安供电公司 Transformer fault classification and identification method based on vibration analysis method
CN108073158A (en) * 2017-12-05 2018-05-25 上海电机学院 Based on PCA and KNN density algorithm Wind turbines Method for Bearing Fault Diagnosis
CN109323860A (en) * 2018-10-31 2019-02-12 广东石油化工学院 A kind of rotating machinery gearbox fault data set optimization method
CN109489977A (en) * 2018-12-28 2019-03-19 西安工程大学 Method for Bearing Fault Diagnosis based on KNN-AdaBoost

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106371427A (en) * 2016-10-28 2017-02-01 浙江大学 Industrial process fault classification method based on analytic hierarchy process and fuzzy fusion
CN106646096A (en) * 2016-11-15 2017-05-10 国网四川省电力公司广安供电公司 Transformer fault classification and identification method based on vibration analysis method
CN108073158A (en) * 2017-12-05 2018-05-25 上海电机学院 Based on PCA and KNN density algorithm Wind turbines Method for Bearing Fault Diagnosis
CN109323860A (en) * 2018-10-31 2019-02-12 广东石油化工学院 A kind of rotating machinery gearbox fault data set optimization method
CN109489977A (en) * 2018-12-28 2019-03-19 西安工程大学 Method for Bearing Fault Diagnosis based on KNN-AdaBoost

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112183344A (en) * 2020-09-28 2021-01-05 广东石油化工学院 Large unit friction fault analysis method and system based on waveform and dimensionless learning
CN112270227A (en) * 2020-10-16 2021-01-26 广东石油化工学院 Oil film whirl and friction concurrent fault analysis method and analysis system

Also Published As

Publication number Publication date
CN110160781B (en) 2020-12-04

Similar Documents

Publication Publication Date Title
Sinitsin et al. Intelligent bearing fault diagnosis method combining mixed input and hybrid CNN-MLP model
CN106769052B (en) A kind of mechanical system rolling bearing intelligent failure diagnosis method based on clustering
CN110110768B (en) Rolling bearing fault diagnosis method based on parallel feature learning and multiple classifiers
CN103048137B (en) Fault diagnosis method of rolling bearing under variable working conditions
CN104614179B (en) A kind of gearbox of wind turbine state monitoring method
CN112179691A (en) Mechanical equipment running state abnormity detection system and method based on counterstudy strategy
CN110160781A (en) A kind of the test set reconstruct and prediction technique of rotating machinery fault classification
Chen et al. Automated function generation of symptom parameters and application to fault diagnosis of machinery under variable operating conditions
CN103558955B (en) Multi-object state monitor method and system based on multi-object multi-state monitor
CN107473036A (en) Elevator traction machine remote online detects and diagnostic system and its checkout and diagnosis method
CN103471849A (en) Bearing fault diagnosis system of multi-layer relevance vector machine on basis of dual combination
CN108267312A (en) A kind of subway train bearing intelligent diagnostic method based on fast search algorithm
CN106932200A (en) The Gaussian process multiple features intelligent diagnosing method of rolling bearing fault
CN111811819A (en) Bearing fault diagnosis method and device based on machine learning
CN110108474B (en) Online monitoring and evaluating method and system for operation stability of rotary machine
Euldji et al. Decision-making based on decision tree for ball bearing monitoring
CN114417924A (en) Rolling bearing fault diagnosis method based on undirected graph adjacency matrix of mixed features
CN109726893A (en) A kind of determination and evaluation method of rotating machinery fault data set attribute weight
CN111046790A (en) Pump bearing fault diagnosis method
CN109598309A (en) A kind of detection system and its monitoring method of mental package press machine
CN209783910U (en) Intelligent fault diagnosis system for nuclear power pump bearing
CN107704712A (en) Mechanical failure diagnostic method and system based on vector spectrum feature extraction
CN106249146B (en) A kind of analysis of cigar mill's electric motor operation state and method for early warning
CN117213858A (en) Bearing fault diagnosis method based on multi-sensor feature fusion
CN112326246A (en) Bearing safety state online monitoring method based on periodic data and nuclear density estimation

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