CN109033582A - A kind of bullet train multi parameter intallingent threshold value criterion - Google Patents

A kind of bullet train multi parameter intallingent threshold value criterion Download PDF

Info

Publication number
CN109033582A
CN109033582A CN201810761712.3A CN201810761712A CN109033582A CN 109033582 A CN109033582 A CN 109033582A CN 201810761712 A CN201810761712 A CN 201810761712A CN 109033582 A CN109033582 A CN 109033582A
Authority
CN
China
Prior art keywords
parameter
monitoring data
bullet train
inter
condition
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
CN201810761712.3A
Other languages
Chinese (zh)
Inventor
申中杰
宫云杰
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.)
Xi'an Intemet Information Technology Co Ltd
Original Assignee
Xi'an Intemet Information Technology Co Ltd
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 Xi'an Intemet Information Technology Co Ltd filed Critical Xi'an Intemet Information Technology Co Ltd
Priority to CN201810761712.3A priority Critical patent/CN109033582A/en
Publication of CN109033582A publication Critical patent/CN109033582A/en
Priority to CN201910510051.1A priority patent/CN110781553B/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
    • Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Emergency Alarm Devices (AREA)
  • Testing Or Calibration Of Command Recording Devices (AREA)

Abstract

The invention discloses a kind of bullet train multi parameter intallingent threshold value criterions, acquire the duty parameter and Condition Monitoring Data of bullet train first, divide according to duty parameter to Condition Monitoring Data；Condition Monitoring Data is analyzed followed by fuzzy kernel clustering, obtains all kinds of centre coordinates, inter- object distance；Finally regression relation between calculating centre coordinate, inter- object distance and operating condition multi-parameter, and then the threshold value under the various operating conditions of intelligence computation, carry out intelligent alarm to bullet train.This method is simple and easy, the determination suitable for the bullet train monitoring data intelligence threshold value under variable working condition.

Description

A kind of bullet train multi parameter intallingent threshold value criterion
Technical field
The invention belongs to monitoring, diagnosing fields, and in particular to a kind of bullet train multi parameter intallingent threshold value criterion.
Background technique
With the high speed development of China railways, bullet train has become the important symbol that the high-end manufacturing industry of China emerges. However, bullet train belongs to typical complex Mechatronic Systems, in a distributed manner, that networking mode is integrated with mechanical, electrical, air and heat etc. is multiple The component of physical domain leads to failure manifestation mode highly complexization with a variety of physical action complex interactions between component.High speed arranges The maintenance of vehicle, which is generally continued to use, disregards the labor-intensive planned maintenance system that cost ensures safety, it has also become restricts China's high-speed rail hair The bottleneck of exhibition or even outlet, for this purpose, railway maintenance ensures that department's following maintenance mode especially set out should be in accurate perception train Under the premise of state, gradually to State Maintenance system transition, to ensure operational safety, improve maintenance efficiency, meet it is domestic and Overseas maintenance support demand.
In the real-time state monitoring and failure diagnostic process of train, how to formulate threshold value is a great problem in industry. Main reason is that: high speed usually high revolving speed, strong vibration, big stress adverse circumstances under work, single threshold value index is difficult To adapt to the variation of operating condition.If threshold value is excessively high, it can only alarm under the limiting case of maximum (top) speed, biggish fail to report can be generated Rate does not have the effect of real-time early warning, alarm.If threshold value is too low, though slow-speed of revolution operating condition is adapted to, once high rotary speed working, Normal condition is mistakenly considered failure, will lead to very big Maintenance Resource waste, and bring personnel panic, influences to work normally.Therefore, The threshold value determination of bullet train monitoring data need to be from Real-time Monitoring Data, and adapts to the variation of various operating conditions.
Summary of the invention
The object of the present invention is to provide a kind of bullet train multi parameter intallingent threshold value criterions to melt from monitoring data Enter operating condition multi-parameter, intelligence determines the monitoring threshold under various working, realizes the Realtime Alerts of the adaptive operating condition of bullet train.
To achieve the goals above, the technical solution adopted by the present invention is that:
1) duty parameter and Condition Monitoring Data for acquiring bullet train carry out Condition Monitoring Data according to duty parameter It divides；
2) Condition Monitoring Data is analyzed using fuzzy kernel clustering, obtains all kinds of centre coordinates, inter- object distance；
3) regression relation between calculating centre coordinate, inter- object distance and operating condition multi-parameter, and then the various operating conditions of intelligence computation Under threshold value, to bullet train carry out intelligent alarm.
The step 1) specifically includes the following contents:
Firstly, acquiring the Condition Monitoring Data under various operating conditions, corresponding duty parameter acquires together, monitoring state number According to predominantly vibration acceleration signal, duty parameter mainly includes revolving speed, load, temperature etc..
Then, the statistical nature of Condition Monitoring Data is calculated: as root-mean-square value, root amplitude, degree of skewness index, kurtosis refer to Mark, frequency domain amplitude feature, the collecting and distributing feature of frequency etc., and Condition Monitoring Data is simply divided according to duty parameter.
The step 2) specifically includes the following contents:
Firstly, utilizing the sample set X={ x of the monitoring data statistical nature building N × d acquired in claim 21, x2,…,xN, wherein N is the number of monitoring data time series；D is the species number of statistical nature；xi, i=1,2 ..., N is dimension Number is the sample of d.Sample set can be divided into J class according to duty parameter, define degree of membership μijPresentation class is as a result, its meaning is sample This xiA possibility that belonging to jth class, j=1,2 ..., J.
Then, the objective function of fuzzy kernel clustering analysis is provided:
Wherein U=(μij)J×NFor subordinated-degree matrix, V={ v1,v2,…,vJ, vjFor the centre coordinate of jth class, m >=1 is Weighted Index.
To solve the problems, such as that original sample is nonlinear, nonlinear function φ (x) is introduced original sample and maps to higher-dimension sky Between in, construct kernel function K (xj,vi)=φ (xj)T·φ(xj) replace xi·vj, obtain following formula
Above formula is solved, degree of membership μ is obtainedijWith class centre coordinate vj
Based on above-mentioned formula, the step of fuzzy kernel clustering algorithm are as follows:
1. setting cluster numbers J, fuzzy clustering exponent m and outage threshold ε；
2. according to the above parameter initialization subordinated-degree matrix U；
3. calculating target function judges whether to reach suspension condition, if it is satisfied, end of clustering, otherwise continues.
4. updating fuzzy clustering center and fuzzy membership matrix according to formula (3) and formula (4), and turn to step 3..
Finally, the inter- object distance of jth class cluster is
dj=max | | xl-vj| |, l=1,2 ..., L (11)
The step 3) specifically includes the following contents:
Firstly, by class heart coordinate V={ v1,v2,…,vJ, inter- object distance D={ d1,d2,…,dJAnd operating condition multi-parameter E= {e1,e2,…,esBuilding multivariate nonlinear regression analysis model:
Wherein ekIt include various operating conditions for J dimensional vector.Function f is solved using data automatic Fitting1、f2, obtain class heart seat The functional relation between V, inter- object distance D and operating condition multi-parameter E is marked, various operating conditions then can be obtained and correspond between inter- object distance D Relationship.
Then declared working condition J is provided0Under the conditions of, inter- object distance50% it is considered as alarm threshold beyond normal its, then counts Calculate the corresponding alarm threshold d' of various operating conditionsj, f1、f2Inverse function acquire, and then realize to the intelligent alarm of bullet train.
Due to a kind of bullet train multi parameter intallingent threshold value criterion of the present invention, have following differences in the significant of conventional method Advantage:
1) the information fusion of monitoring data is realized based on fuzzy kernel clustering analysis method, and to the monitoring number under each operating condition According to self-adaption cluster has been carried out, the monitoring data under each operating condition are unfolded under the same dimension, convenient for formulating monitoring threshold.
2) operating condition multi-parameter is incorporated during the determination of threshold value, realizes threshold value automatically updating with operating condition, intelligent recognition Bullet train failure.
3) it calculates simply, arithmetic speed is fast, convenient for promoting in practice in engineering.
Detailed description of the invention
Fig. 1 show High-speed Train Bearing signal primitive character；
Fig. 2 show High-speed Train Bearing signal ambiguity kernel clustering result；
Fig. 3 show High-speed Train Bearing signal to the distance of the class heart；
Specific embodiment
High-speed Train Bearing signal primitive character shown in referring to Fig.1, in the vibration acceleration letter that simulator stand obtains It number calculates and to obtain, abscissa is chronomere min, and since each statistical nature meaning is different, ordinate is without unit.Each statistics Characteristic difference is very big, is difficult to show under another scale, and very big with the influence of operating condition.
With reference to High-speed Train Bearing signal ambiguity kernel clustering shown in Fig. 2 as a result, adaptively being carried out according to operating condition to feature Classification lays the foundation to formulate threshold value under various operating conditions, and abscissa, ordinate are synthetic parameters, without apparent physics Meaning.
With reference to the distance of High-speed Train Bearing signal shown in Fig. 3 to the class heart, the influence of the operating condition got rid of, in the same dimension Degree is lower to be unfolded signal characteristic.Abscissa is chronomere min, and ordinate is distance.
Attached drawing is specific embodiments of the present invention；
The contents of the present invention are described in further detail with reference to the accompanying drawing:
1) duty parameter and Condition Monitoring Data for acquiring bullet train carry out Condition Monitoring Data according to duty parameter It divides；
2) Condition Monitoring Data is analyzed using fuzzy kernel clustering, obtains all kinds of centre coordinates, inter- object distance；
3) regression relation between calculating centre coordinate, inter- object distance and operating condition multi-parameter, and then the various operating conditions of intelligence computation Under threshold value, to bullet train carry out intelligent alarm.
The step 1) specifically includes the following contents:
Firstly, acquiring the Condition Monitoring Data under various operating conditions, corresponding duty parameter acquires together, monitoring state number According to predominantly vibration acceleration signal, duty parameter mainly includes revolving speed, load, temperature etc..
Then, the statistical nature of Condition Monitoring Data is calculated: as root-mean-square value, root amplitude, degree of skewness index, kurtosis refer to Mark, frequency domain amplitude feature, the collecting and distributing feature of frequency etc., and Condition Monitoring Data is simply divided according to duty parameter.
1. statistical nature of table
The step 2) specifically includes the following contents:
Firstly, utilizing the sample set X={ x of the monitoring data statistical nature building N × d acquired in claim 21, x2,…,xN, wherein N is the number of monitoring data time series；D is the species number of statistical nature；xi=(T1,T2,…,T6),i =1,2 ..., N is the sample that dimension is d=6.Sample set can be divided into J class according to duty parameter, define degree of membership μijIt indicates to divide Class is as a result, its meaning is sample xiA possibility that belonging to jth class, j=1,2 ..., J.
Then, the objective function of fuzzy kernel clustering analysis is provided:
Wherein U=(μij)J×NFor subordinated-degree matrix, V={ v1,v2,…,vJ, vjFor the centre coordinate of jth class, m >=1 is Weighted Index.
To solve the problems, such as that original sample is nonlinear, nonlinear function φ (x) is introduced original sample and maps to higher-dimension sky Between in, construct kernel function K (xj,vi)=φ (xj)T·φ(xj) replace xi·vj, obtain following formula
Above formula is solved, degree of membership μ is obtainedijWith class centre coordinate vj
Based on above-mentioned formula, the step of fuzzy kernel clustering algorithm are as follows:
1. setting cluster numbers J, fuzzy clustering exponent m and outage threshold ε；
2. according to the above parameter initialization subordinated-degree matrix U；
3. calculating target function judges whether to reach suspension condition, if it is satisfied, end of clustering, otherwise continues.
4. updating fuzzy clustering center and fuzzy membership matrix according to formula (3) and formula (4), and turn to step 3..
Jth class cluster inter- object distance be
dj=max | | xl-vj| |, l=1,2 ..., L (17)
The step 3) specifically includes the following contents:
Firstly, by class heart coordinate V={ v1,v2,…,vJ, inter- object distance D={ d1,d2,…,dJAnd operating condition multi-parameter E= {e1,e2,...,esBuilding multivariate nonlinear regression analysis model:
Wherein ekIt include various operating conditions for J dimensional vector.Function f is solved using data automatic Fitting1、f2, obtain class heart seat The functional relation between V, inter- object distance D and operating condition multi-parameter E is marked, various operating conditions then can be obtained and correspond between inter- object distance D Relationship.
Then declared working condition J is provided0Under the conditions of, inter- object distance50% it is considered as alarm threshold beyond normal its, then counts Calculate the corresponding alarm threshold d' of various operating conditionsj, f1、f2Inverse function acquire, and then realize to the intelligent alarm of bullet train.
Embodiment:
This embodiment gives specific implementation process of the present invention in High-speed Train Bearing test, while demonstrating the hair Bright validity.Test is divided into operating condition in following 8 according to four kinds of revolving speeds, two kinds of load, and as shown in table 2, wherein operating condition 6 is specified Operating condition.
2 bullet train operating condition of table
High-speed Train Bearing vibration acceleration counting statistics characteristic value is acquired, as shown in Figure 1, two parts are broadly divided into, frequency Domain amplitude achievement data is larger, shows on the top of figure, remaining index is smaller, shows in figure lower part.Further to show There is a partial enlarged view in lower left quarter in the trend for showing lower part index, can be it is better seen that root-mean-square value, root amplitude And kurtosis index is with the situation of change of operating condition.
Information fusion is carried out to statistical nature using fuzzy kernel clustering, can be clustered according to operating condition, in high dimensional feature It is unfolded in space, as a result as shown in Figure 2.Calculate each sample to the class heart distance, as shown in Figure 3.Therefore in various operating condition lower classes Maximum value under distance various operating conditions as shown in Fig. 3, as shown in table 3.
Each operating condition inter- object distance of table 3
Operating condition 1 Operating condition 2 Operating condition 3 Operating condition 4 Operating condition 5 Operating condition 6 Operating condition 7 Operating condition 8
5.08 7.96 11.99 14.05 9.22 13.01 8.75 12.56
Wherein, operating condition 6 is declared working condition, and under declared working condition, inter- object distance is considered as alarm threshold beyond 50%, is passed through The multivariate nonlinear regression analysis model for solving class heart coordinate V, inter- object distance D and operating condition multi-parameter E, finally calculates under various operating conditions Alarm in class distance as shown in table 4, realize to the intelligent alarm of bullet train.
Each condition alarm inter- object distance of table 4
Operating condition 1 Operating condition 2 Operating condition 3 Operating condition 4 Operating condition 5 Operating condition 6 Operating condition 7 Operating condition 8
7.36 11.8 16.75 21.22 12.18 19.65 12.21 18.83

Claims (4)

1. a kind of bullet train multi parameter intallingent threshold value criterion, it is characterised in that:
1) duty parameter and Condition Monitoring Data for acquiring bullet train, draw Condition Monitoring Data according to duty parameter Point；
2) Condition Monitoring Data is analyzed using fuzzy kernel clustering, obtains all kinds of centre coordinates, inter- object distance；
3) regression relation between calculating centre coordinate, inter- object distance and operating condition multi-parameter, and then under the various operating conditions of intelligence computation Threshold value carries out intelligent alarm to bullet train.
2. a kind of bullet train multi parameter intallingent threshold value criterion according to claim 1, which is characterized in that described acquisition The duty parameter and Condition Monitoring Data of bullet train divide Condition Monitoring Data according to duty parameter, including following Step:
Firstly, acquiring the Condition Monitoring Data under various operating conditions, corresponding duty parameter acquires together, monitoring state data master It to be vibration acceleration signal, duty parameter mainly includes revolving speed, load, temperature etc..
Then, calculate the statistical nature of Condition Monitoring Data: as root-mean-square value, root amplitude, degree of skewness index, kurtosis index, Collecting and distributing feature of frequency domain amplitude feature, frequency etc., and Condition Monitoring Data is simply divided according to duty parameter.
3. a kind of bullet train multi parameter intallingent threshold value criterion according to claim 1, which is characterized in that described utilization Fuzzy kernel clustering analyzes Condition Monitoring Data, obtains all kinds of centre coordinates, inter- object distance, comprising the following steps:
Firstly, utilizing the sample set X={ x of the monitoring data statistical nature building N × d acquired in claim 21,x2,…, xN, wherein N is the number of monitoring data time series；D is the species number of statistical nature；xi, i=1,2 ..., N is that dimension is d Sample.Sample set can be divided into J class according to duty parameter, define degree of membership μijPresentation class is as a result, its meaning is sample xi A possibility that belonging to jth class, j=1,2 ..., J.
Then, the objective function of fuzzy kernel clustering analysis is provided:
Wherein U=(μij)J×NFor subordinated-degree matrix, V={ v1,v2,…,vJ, vjFor the centre coordinate of jth class, m >=1 is weighting Index.
To solve the problems, such as that original sample is nonlinear, it is introduced into nonlinear function φ (x) and original sample is mapped in higher dimensional space, Construct kernel function K (xj,vi)=φ (xj)T·φ(xj) replace xi·vj, obtain following formula
Above formula is solved, degree of membership μ is obtainedijWith class centre coordinate vj
Based on above-mentioned formula, the step of fuzzy kernel clustering algorithm are as follows:
1. setting cluster numbers J, fuzzy clustering exponent m and outage threshold ε；
2. according to the above parameter initialization subordinated-degree matrix U；
3. calculating target function judges whether to reach suspension condition, if it is satisfied, end of clustering, otherwise continues.
4. updating fuzzy clustering center and fuzzy membership matrix according to formula (3) and formula (4), and turn to step 3..
Finally, the inter- object distance of jth class cluster is
dj=max | | xl-vj| |, l=1,2 ..., L (5).
4. a kind of bullet train multi parameter intallingent threshold value criterion according to claim 1, which is characterized in that described calculating Regression relation between centre coordinate, inter- object distance and operating condition multi-parameter, and then the threshold value under the various operating conditions of intelligence computation, to high speed Train carries out intelligent alarm, comprising the following steps:
Firstly, by class heart coordinate V={ v1,v2,…,vJ, inter- object distance D={ d1,d2,…,dJAnd operating condition multi-parameter E={ e1, e2,…,esBuilding multivariate nonlinear regression analysis model:
Wherein ekIt include various operating conditions for J dimensional vector.Function f is solved using data automatic Fitting1、f2, obtain class heart coordinate V, Functional relation between inter- object distance D and operating condition multi-parameter E then can be obtained various operating conditions and correspond to pass between inter- object distance D System.
Then declared working condition J is provided0Under the conditions of, inter- object distance50% it is considered as alarm threshold beyond normal its, then calculates The corresponding alarm threshold d' of various operating conditionsj, f1、f2Inverse function acquire, and then realize to the intelligent alarm of bullet train.
CN201810761712.3A 2018-07-12 2018-07-12 A kind of bullet train multi parameter intallingent threshold value criterion Pending CN109033582A (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201810761712.3A CN109033582A (en) 2018-07-12 2018-07-12 A kind of bullet train multi parameter intallingent threshold value criterion
CN201910510051.1A CN110781553B (en) 2018-07-12 2019-06-12 Multi-parameter intelligent threshold monitoring method for high-speed train

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810761712.3A CN109033582A (en) 2018-07-12 2018-07-12 A kind of bullet train multi parameter intallingent threshold value criterion

Publications (1)

Publication Number Publication Date
CN109033582A true CN109033582A (en) 2018-12-18

Family

ID=64642350

Family Applications (2)

Application Number Title Priority Date Filing Date
CN201810761712.3A Pending CN109033582A (en) 2018-07-12 2018-07-12 A kind of bullet train multi parameter intallingent threshold value criterion
CN201910510051.1A Active CN110781553B (en) 2018-07-12 2019-06-12 Multi-parameter intelligent threshold monitoring method for high-speed train

Family Applications After (1)

Application Number Title Priority Date Filing Date
CN201910510051.1A Active CN110781553B (en) 2018-07-12 2019-06-12 Multi-parameter intelligent threshold monitoring method for high-speed train

Country Status (1)

Country Link
CN (2) CN109033582A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111539374A (en) * 2020-05-07 2020-08-14 上海工程技术大学 Rail train bearing fault diagnosis system and method based on multidimensional data space

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113297291A (en) * 2021-05-08 2021-08-24 上海电气风电集团股份有限公司 Monitoring method, monitoring system, readable storage medium and wind driven generator
CN114323707B (en) * 2022-01-04 2023-07-07 中车株洲电力机车有限公司 Magnetic levitation train and vibration signal calculation method, simulation generation method and device thereof

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB0318339D0 (en) * 2003-08-05 2003-09-10 Oxford Biosignals Ltd Installation condition monitoring system
CN102636991A (en) * 2012-04-18 2012-08-15 国电科学技术研究院 Method for optimizing running parameters of thermal power unit and based on fuzzy set association rule
CN103455635A (en) * 2013-09-24 2013-12-18 华北电力大学 Thermal process soft sensor modeling method based on least squares and support vector machine ensemble
CN104392071B (en) * 2014-12-12 2017-09-29 北京交通大学 A kind of bullet train system security assessment method based on complex network
CN106203856A (en) * 2016-07-18 2016-12-07 交通运输部公路科学研究所 A kind of Combined Principal Components analysis and the vehicle driving-cycle formulating method of Fuzzy c-means Clustering

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111539374A (en) * 2020-05-07 2020-08-14 上海工程技术大学 Rail train bearing fault diagnosis system and method based on multidimensional data space

Also Published As

Publication number Publication date
CN110781553B (en) 2023-08-25
CN110781553A (en) 2020-02-11

Similar Documents

Publication Publication Date Title
CN106124175B (en) A kind of compressor valve method for diagnosing faults based on Bayesian network
CN109033582A (en) A kind of bullet train multi parameter intallingent threshold value criterion
CN110377465A (en) A kind of method for detecting abnormality of vehicle-mounted CAN bus
CN104832418B (en) A kind of based on local mean value conversion and the Fault Diagnosis of Hydraulic Pump method of Softmax
CN112819059B (en) Rolling bearing fault diagnosis method based on popular retention transfer learning
CN109708907B (en) Equipment fault feature extraction method based on envelope information
Yan et al. Fault diagnosis of rotating machinery equipped with multiple sensors using space-time fragments
CN113542241B (en) Intrusion detection method and device based on CNN-BiGRU hybrid model
CN105894024A (en) Possibility fuzzy c mean clustering algorithm based on multiple kernels
CN104596780A (en) Diagnosis method for sensor faults of motor train unit braking system
CN101178703A (en) Failure diagnosis chart clustering method based on network dividing
CN105137324B (en) A kind of more detection point failure component localization methods based on emulation disaggregated model
CN105225523A (en) A kind of parking space state detection method and device
CN108267312A (en) A kind of subway train bearing intelligent diagnostic method based on fast search algorithm
CN109933040A (en) Fault monitoring method based on level density peaks cluster and most like mode
CN107392979B (en) The two dimensional visible state composition and quantitative analysis index method of time series
CN104622446A (en) Heart rate variability signal optimization method based on KHM clustering algorithm
CN105956318B (en) Based on the wind power plant group of planes division methods for improving division H-K clustering method
CN105590167A (en) Method and device for analyzing electric field multivariate operating data
Khalid et al. Brain abnormalities segmentation performances contrasting: adaptive network-based fuzzy inference system (ANFIS) vs K-nearest neighbors (k-NN) vs fuzzy c-means (FCM)
CN104239411B (en) A kind of detection method of the lattice-shaped radar based on color, position cluster and Corner Detection
CN108414228B (en) Based on averagely more granularity decision rough sets and NNBC Method for Bearing Fault Diagnosis
CN114861749A (en) Low-sample bearing fault diagnosis method based on depth prototype network
Zhang et al. An outlier detection algorithm based on clustering analysis
CN110110795A (en) Image classification method and device

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
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20181218