CN109598048A - A kind of lubrication degradation prediction technique of track vehicle door system - Google Patents
A kind of lubrication degradation prediction technique of track vehicle door system Download PDFInfo
- Publication number
- CN109598048A CN109598048A CN201811425099.4A CN201811425099A CN109598048A CN 109598048 A CN109598048 A CN 109598048A CN 201811425099 A CN201811425099 A CN 201811425099A CN 109598048 A CN109598048 A CN 109598048A
- Authority
- CN
- China
- Prior art keywords
- door system
- vehicle door
- track vehicle
- data
- lubrication degradation
- 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
Links
- 238000005461 lubrication Methods 0.000 title claims abstract description 68
- 230000015556 catabolic process Effects 0.000 title claims abstract description 67
- 238000006731 degradation reaction Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000007637 random forest analysis Methods 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 15
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 11
- 238000000605 extraction Methods 0.000 claims abstract description 6
- 230000008569 process Effects 0.000 claims description 17
- 238000003066 decision tree Methods 0.000 claims description 12
- 239000000284 extract Substances 0.000 claims description 10
- 230000002123 temporal effect Effects 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000010224 classification analysis Methods 0.000 claims description 6
- 238000012952 Resampling Methods 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000007405 data analysis Methods 0.000 claims description 2
- 230000005611 electricity Effects 0.000 claims description 2
- 230000009467 reduction Effects 0.000 description 5
- 230000009466 transformation Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000003862 health status Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007850 degeneration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 230000001050 lubricating effect Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
- G06F17/148—Wavelet transforms
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
A kind of lubrication degradation prediction technique of track vehicle door system of the invention, the first pretreatment of progress motor acquisition data, then to the characteristics extraction of track vehicle door system；And lubrication degradation model is established with random forests algorithm, it is finally lubricated degradation prediction and compares the training pattern that the characteristic value of extraction obtains, diagnose lubrication degradation state locating for current orbit train-door system.The present invention has arrived the purpose of prediction train-door system lubrication degradation, while according to online prediction result it can be found that unknown lubrication degradation state, helping to take appropriate measures in time avoids the generation of the system failure.
Description
Technical field
The present invention relates to urban rail transit technology field, the lubrication degradation of specifically a kind of track vehicle door system is predicted
Method.
Background technique
Under the overall situation of transport need, urban rail transit vehicles occupy in entire City Rail Transit System equipment
Consequence.At the same time, rail traffic safety and guaranteed reliability are also faced with stern challenge.Wherein rail vehicle
Door is one of component the most used in urban track traffic operation.It is frequent due to switching, door device Frequent Troubles are caused,
It makes troubles to passenger's trip, at the same time safety of urban transit operation is caused to seriously affect.According to statistics it is found that
During entire metro operation, train fault number accounts for 35% or more of sum, wherein crucial subsystem of the door device as train
System, number of faults rank the first in each system of train, account for 50% or so of train fault sum.In door contact interrupter door process
In, due to rubbing bearing it is that oil film bearings are excessively thin, rather than the factors such as fatigue rupture, so studying oil film
It is most important.Oil film thickness has very big influence to working performance, and oil-film force subjects the load that rotor gives bearing.
Subway vehicle door system unit is more, is multi-specialized comprehensive product, is related to machinery, electronics, computer, control
System, material etc. are multi-field.So detecting its lubrication degradation state using traditional based on model or rule-based method
Also become more and more difficult.The current shape of system can be monitored by the data that analysis system acquires based on the method for data
State, and by the difference of analysis real time data and normal data, the lubrication degradation state of forecasting system can be carried out, to avoid possibility
The system failure of appearance；Wide application is thus played in theoretical research and engineer application based on the method for data.At present still
The lubrication degradation for being not based on track vehicle door system and the technology in terms of system health status.The lubrication of existing other systems
The more and different degrees of lubrication degradation of the component of the prediction technique of degeneration, railcar door system causes different abnormal shapes
State can not carry out the detection of subway vehicle door system lubrication degenerate state by this method.Therefore, discovery system is common in time
Early stage lubrication degradation sign carries out real-time lubricating status monitoring and provides lubrication degradation condition discrimination, to raising door device
Safety, reliability, reduce failure rate be of great significance.
Summary of the invention
Goal of the invention: in order to overcome the deficiencies in the prior art, the present invention provides a kind of track vehicle door system
Lubrication degradation prediction technique.
Technical solution: in order to solve the above technical problems, a kind of lubrication degradation of track vehicle door system of the invention is predicted
Method, comprising the following steps:
Step 1, motor acquire the pretreatment of data: the data that track vehicle door system acquires being corrected, are remembered again
The record starting point that data acquire during switch gate each time and terminating point；The starting point of train-door system is corrected to simultaneously same
Position, the acquisition data of the track vehicle door system after being corrected；
The characteristics extraction of track vehicle door system: step 2 carries out segment processing, application to switch gate process each time
Descriptive statistics amount carries out data analysis, obtains the characteristic value that can completely embody track vehicle door system state, all characteristic values
Form system features collection；
The foundation of lubrication degradation model: step 3 uses random forests algorithm, to the lubrication degradation of track vehicle door system
State carries out off-line modeling, carries out classification point by a variety of lubrication degradation states and normal condition to track vehicle door system
Analysis, obtains the training pattern of different lubrication degradation states；
Lubrication degradation prediction: step 4 characteristic value and the obtained training pattern of step 3 that step 2 is extracted is carried out pair
Than diagnosing lubrication degradation state locating for current orbit train-door system.
In step 2, the segment processing to switch gate process each time is switch gate process to be divided into 5 sections each time, is determined
Parameter of electric machine value on each section, obtains system features collection；Wherein, described 5 sections include raising speed section, high regime, braking section, jogging
Section and back segment, parameter of electric machine value include position, speed and current value in place.Each section in 5 sections is extracted into 6 time domain spies respectively
Sign, 6 temporal signatures include maximum value, minimum value, mean value, variance, the skewness and kurtosis of the parameter of electric machine value in correspondent section, institute
State 6 temporal signatures composition temporal signatures collection.
Wherein, each frequency is extracted using wavelet-decomposing method in the subspace that system features collection is resolved into multiple independent frequency domains
The energy Frequency Domain Energy of band.
Wherein, using 3 layers of wavelet decomposition structure, to the opening position of track vehicle door system motor acquisition, enabling speed,
Gate current signal, shutdown position, door closing speed and shutdown current signal carry out wavelet decomposition respectively, obtain multiple sub-bands
Energy and the frequency domain character collection for determining train-door system.
In step 3, off-line modeling is carried out with lubrication degradation state of the random forests algorithm to track vehicle door system,
Include:
By bootstrap resampling technique, it is concentrated with from original training and repeats with putting back to randomly select n
Bootstrap sample set repeats k times, and the bootstrap sample set extracted every time is the training set of decision tree growth, k extraction
The sample not being pumped to during sample constitutes the outer data of bag, carries out test assessment to classification performance inside forest；At random
Segmentation candidates feature is chosen, keeps classifier different in structure；It is every decision tree output knot that random forest, which exports result,
The multiple combinations form of fruit, every decision tree provide the prediction category of sample to be estimated, according to the how many decision forecast samples of ballot
Final ownership class, to establish the lubrication degradation model of track vehicle door system.
The step 4 specifically includes the following steps:
Step 4.1: the characteristic value and normal data extract to step 2 carry out classification analysis, analyze the correct of classification results
Rate: if accuracy is lower than threshold value, illustrate that track vehicle door system is in normal condition；Otherwise a Unknown Model is obtained, and
Enter step 4.2；
Step 4.2: Unknown Model being matched with the training pattern that step 3 obtains, is carried out according to random forests algorithm
Prediction obtains the lubrication degradation state that track vehicle door system is presently in.
The step 2 further includes before carrying out segment processing to switch gate process each time to track vehicle door system
Main track data carry out pretreated process, the pretreatment includes the alignment and problem data removal of match line data, described
Problem data includes less than the data of preset normal data threshold value and not in the data of normal data range.
The utility model has the advantages that the invention has the following advantages:
(1) for the present invention using the Mathematical Modeling Methods based on data, the data based on sensor acquisition carry out feature
Extract and Optimal improvements, enable extract state of the data characteristics than more completely embodying system.
(2) according to the system features after extraction process, mathematical modeling is carried out with random forests algorithm, based on normal data and
The data variance that system lubrication is degenerated obtains the profit of system with the state of sorted result judgement current system with classification
Sliding degradation model is more clear the state for intuitively having reacted train-door system；
(3) according to the lubrication degradation model established offline, the mould of degeneration is lubricated by the prediction result obtained online
Type matching has achieved the purpose that prediction train-door system lubrication degradation, while according to online prediction result it can be found that unknown
Lubrication degradation state, helping to take appropriate measures in time avoids the generation of the system failure.
Detailed description of the invention
Fig. 1 is door device lubrication degradation diagnostic method flow chart of the present invention；
Fig. 2 is random forest method structure chart.
Specific embodiment
The present invention will be further explained with reference to the accompanying drawing.
As shown in Figure 1, the present invention uses random forests algorithm, off-line modeling is carried out to the lubrication degradation state of system, is led to
It crosses and classification analysis is carried out to the 4 kinds of lubrication degradation states and normal condition of system, obtain the training of different lubrication degradation states
Model.And compare the result of online classification with the model library established offline, predict health status locating for current system.
Whole process includes the following steps:
Step 1: the pretreatment of motor acquisition data；
Data prediction includes the alignment and problem data removal of match line data, and described problem data include being less than in advance
Data of the normal data threshold value of setting and not in the data of normal data range.
The data acquired in the present embodiment are the motor corner of rack pitch lubrication degradation experiment, revolving speed and three, electric current
Process variable, therefore data prediction is to be corrected the data that track vehicle door system acquires, the alignment of match line data
It is removed with problem data, described problem data include that the number of sampling points of revolving speed, corner and current data is significantly less than and normally adopts
The data of sample number, or the initial data of corner not in the normal range；
Step 2: it includes extracting every section of motor ginseng respectively that temporal signatures, which extract and extract each section respectively 6 temporal signatures,
The maximum value of numerical value, minimum value, mean value, variance, skewness and kurtosis form system temporal signatures collection, for example, calculate separately enabling and
The raising speed of door closing procedure, high speed, reduction of speed, jogging section and (what is found out is 2 × 3 × 5=30 feature change to the mean value of back segment in place
Amount, has similarly asked maximum value, minimum value, mean value, variance, skewness and kurtosis to come to 180, with two-dimensional matrix X ∈ Rn×pCarry out table
Show, i.e. X=X1,X2,…Xi,Xj,…Xn, wherein n row indicates the number of the data point of switch gate process acquisition, in this implementation
It is 3 in example；P column indicate that the feature extracted is 5 in the present embodiment.
The mean value that data are extracted is described, expression formula is as follows:
Calculate separately the raising speed of enabling and door closing procedure, high speed, reduction of speed, jogging section and the maximum value and minimum of back segment in place
Value: reflection data variation range.
Xmax=max | xi|
Xmin=min | xi|
Calculate separately the raising speed of enabling and door closing procedure, high speed, reduction of speed, jogging section and the variance of back segment in place: description number
According to the departure degree with mean value.
Calculate separately the raising speed of enabling and door closing procedure, high speed, reduction of speed, jogging section and the degree of bias of back segment in place: reflection number
According to the measurement of distribution deflection, Skewness > 0 is known as right avertence state, and the distribution trend of data is that data are located at mostly at this time
Value right side；Skewness < 0 is known as left avertence state, and the distribution deflection situation of data is in contrast；It can recognize when Skewness is close to 0
It is symmetrical for data distribution.
Calculate separately the raising speed of enabling and door closing procedure, high speed, reduction of speed, jogging section and the kurtosis of back segment in place: reflection number
It is uneven according to peak value of the probability density distribution within the scope of mean value.The kurtosis that normal distribution data are presented is 3, if sample number
According to the degree of bias be much larger than 3, indicate the spike steeper of sample data, illustrate in sample data far from mean value data it is more,
So can generally measure the degree that sample data deviates normal distribution with kurtosis.
X in above formulaiFor system state variables value.
System features collection is resolved into the subspace of multiple independent frequency domains by frequency domain character when extracting, using wavelet-decomposing method
Extract the energy Frequency Domain Energy of each frequency band.
Wavelet transformation is the partial transformation method of a kind of time and frequency, carries out partial transformation in short-term in Fourier transformation
On the basis of improve, the size of window can be changed for the variation of frequency, be successfully applied to many fields, such as signal
Processing, image procossing and pattern-recognition etc..Wavelet transformation, which is it with an important characteristic, has part well in frequency domain
Change feature.In wavelet transformation, first have to define two functions.
Wherein,For scaling function, ψ (t) is wavelet function.
Signal x (t) ∈ L2(R) decomposed signal in a wherein small echo subvector space is
In formula,For node, (j, n) corresponding wavelet packet coefficient.
Then the wavelet decomposition of signal x (t) can be written as follow form:
ω in formulan,j,kIt (t) is Orthogonal Wavelets.
Sub-band signalEnergy calculated by following formula:
When the data to middle orbit train-door system of the present invention are analyzed, the high fdrequency component of data is no longer divided
Solution, and the low frequency part of data is continued to decompose.3 layers of wavelet decomposition structure are selected herein.To track vehicle door system motor
Position, speed and the current signal of the switch gate of acquisition carry out wavelet decomposition respectively, obtain ENERGY E 1, E2, E3 of 4 sub-bands
Frequency Domain Energy with E4 as lubrication degradation.
Step 3, the off-line modeling of the lubrication degradation state of system, includes the following steps,
Step 3.1, by bootstrap (bootstrap) resampling technique, be concentrated with from original training put back to repeat with
Machine extracts n sample, repeats k times (i.e. the number of spanning tree is k), the bootstrap sample set extracted every time is raw for decision tree
Long training set, the sample not being pumped to during k sample drawn constitute the outer data (OBB) of bag inside forest to classification
Performance carries out test assessment；
Step 3.2, segmentation candidates feature is randomly selected, keeps classifier different in structure.It is special equipped with M input
Sign, randomly selects m feature, the empirical equation provided according to Liaw usually takes from M input feature vectorI.e.
M takes the subduplicate downward integer of M, then selects 1 spy using certain tactful (such as information gain) from the m feature selected
Levy the disruptive features as the node.The branch for repeating the above process every decision tree classifier of building, until single decision tree
Sample classification or all characteristic attributes of traversal can be described, the value of m remains unchanged in above process.
For random forest using system features value as input, output result can be a variety of groups of every decision tree output result
Conjunction form, most plain mode are exactly linear combination.Every decision tree can provide the prediction category of sample to be estimated, and how much is foundation ballot
Determine the final ownership class of forecast sample, so that the lubrication degradation model of track vehicle door system is established, rule of voting are as follows:
Wherein H (x) is assembled classification model, hkIt (x) is single Decision-Tree Classifier Model, Y is classification ownership, and I () is
Indicator function, Fig. 2 are the method structure chart of random forest.
Step 4 when the lubrication degradation state of in-circuit diagnostic system, carries out classification analysis to online data and normal data,
The classification accuracy rate of analysis classification results R illustrates that system is in normal condition if accuracy is too low；Otherwise it can be concluded that one
A Unknown Model f completes two classification to system health or lubrication degradation, and accuracy rate as a result is as shown in table 1, and enters
In next step；
Car door lubrication degradation state two classification accuracy of the table 1 based on random forest
| Vehicle door status | Accuracy rate |
| Normally | 100% |
| Lubrication degradation | 87.5% |
Unknown Model f is matched with the model in the model library established offline, is carried out according to random forests algorithm pre-
It surveys, the lubrication degradation state that acquisition system is presently in completes more points of the different lubrication degradation states based on random forest
Class, accuracy rate as a result are as shown in table 2:
Car door lubrication degradation state more classification accuracies of the table 2 based on random forest
| Lubrication degradation type | Accuracy rate |
| 1st class | 95.00% |
| 2nd class | 90.00% |
| 3rd class | 90.00% |
| 4th class | 80.00% |
Finally the accuracy rate of lubrication degradation diagnosis is verified, the results showed that accuracy rate is good.The present invention passes through electricity
The collected data characteristics of machine sensor is augmented, extract the correlated characteristic index of time-frequency domain as monitoring vehicle door status feature to
Amount assigns feature vector as the input of random forests algorithm, to the lubrication degradation state of system progress off-line modeling, by being
Several lubrication degradation states and normal condition of system carry out classification analysis, obtain the training pattern of different lubrication degradation states.
The result of final online classification is compared with the model library established offline, diagnoses health status locating for current system.
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 (8)
1. a kind of lubrication degradation prediction technique of track vehicle door system, which comprises the following steps:
Step 1, motor acquire the pretreatment of data: the data that track vehicle door system acquires being corrected, record is every again
The starting point and terminating point that data acquire during switch gate；The starting point of train-door system is corrected to same position simultaneously
It sets, the acquisition data of the track vehicle door system after being corrected；
The characteristics extraction of track vehicle door system: step 2 carries out segment processing to switch gate process each time, using description
Statistic carries out data analysis, obtains the characteristic value that can completely embody track vehicle door system state, all eigenvalue clusters at
System features collection；
The foundation of lubrication degradation model: step 3 uses random forests algorithm, to the lubrication degradation state of track vehicle door system
Off-line modeling is carried out, classification analysis is carried out by a variety of lubrication degradation states and normal condition to track vehicle door system, obtains
To the training pattern of different lubrication degradation states；
Lubrication degradation prediction: step 4 the training pattern that the characteristic value that step 2 is extracted is obtained with step 3 is compared, is examined
Lubrication degradation state locating for disconnected current orbit train-door system.
2. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 1, it is characterised in that: step
In two, the segment processing to switch gate process each time is switch gate process to be divided into 5 sections each time, determines the electricity on each section
Machine parameter value obtains system features collection；Wherein, described 5 sections include raising speed section, high regime, braking section, jogging section and in place after
Section, parameter of electric machine value includes position, speed and current value.
3. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 2, it is characterised in that: by 5
Section in each section respectively extract 6 temporal signatures, 6 temporal signatures include the parameter of electric machine value in correspondent section maximum value,
Minimum value, mean value, variance, skewness and kurtosis, 6 temporal signatures form temporal signatures collection.
4. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 2, it is characterised in that: will be
System feature set resolves into the subspace of multiple independent frequency domains, and the energy frequency domain energy for extracting each frequency band using wavelet-decomposing method is special
Sign.
5. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 4, it is characterised in that: use
3 layers of wavelet decomposition structure to the opening position of track vehicle door system motor acquisition, enabling speed, gate current signal, are closed the door
Position, door closing speed and shutdown current signal carry out wavelet decomposition respectively, obtain the energy of multiple sub-bands and determine vehicle doorn
The frequency domain character collection of system.
6. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 1, it is characterised in that: step
In three, off-line modeling is carried out with lubrication degradation state of the random forests algorithm to track vehicle door system, comprising:
By bootstrap resampling technique, it is concentrated with from original training and repeats with putting back to randomly select n bootstrap sample
This collection repeats k times, and the bootstrap sample set extracted every time is the training set of decision tree growth, the process of k sample drawn
In the sample that is not pumped to constitute the outer data of bag, test assessment is carried out to classification performance inside forest；Randomly select candidate point
Feature is cut, keeps classifier different in structure；It is a variety of groups that every decision tree exports result that random forest, which exports result,
Conjunction form, every decision tree provide the prediction category of sample to be estimated, and the final ownership classes of forecast samples how much are determined according to ballot,
To establish the lubrication degradation model of track vehicle door system.
7. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 6, it is characterised in that: described
Step 4 specifically includes the following steps:
Step 4.1: the characteristic value and normal data extract to step 2 carry out classification analysis, analyze the accuracy of classification results:
If accuracy is lower than threshold value, illustrate that track vehicle door system is in normal condition；Otherwise it obtains a Unknown Model, and enters
Step 4.2；
Step 4.2: Unknown Model being matched with the training pattern that step 3 obtains, is carried out according to random forests algorithm pre-
It surveys, obtains the lubrication degradation state that track vehicle door system is presently in.
8. a kind of lubrication degradation prediction technique of track vehicle door system according to claim 1, it is characterised in that: described
Step 2 further include before carrying out segment processing to switch gate process each time to the main track data of track vehicle door system into
The pretreated process of row, the pretreatment includes the alignment of match line data and problem data removal, described problem data include
Less than the data of preset normal data threshold value and not in the data of normal data range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811425099.4A CN109598048A (en) | 2018-11-27 | 2018-11-27 | A kind of lubrication degradation prediction technique of track vehicle door system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811425099.4A CN109598048A (en) | 2018-11-27 | 2018-11-27 | A kind of lubrication degradation prediction technique of track vehicle door system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109598048A true CN109598048A (en) | 2019-04-09 |
Family
ID=65959042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811425099.4A Pending CN109598048A (en) | 2018-11-27 | 2018-11-27 | A kind of lubrication degradation prediction technique of track vehicle door system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109598048A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717379A (en) * | 2019-08-28 | 2020-01-21 | 南京康尼机电股份有限公司 | Health assessment method for subway car door key components based on feature fusion |
| CN111947954A (en) * | 2020-07-17 | 2020-11-17 | 南京康尼机电股份有限公司 | Method and system for diagnosing urban rail door system fault or sub-health |
| CN113821866A (en) * | 2020-06-19 | 2021-12-21 | 南京康尼机电股份有限公司 | Method for predicting residual life of finger protection adhesive tape of urban rail door system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372748A (en) * | 2016-08-29 | 2017-02-01 | 上海交通大学 | Hard-rock tunnel boring machine boring efficiency prediction method |
| CN107153841A (en) * | 2017-04-24 | 2017-09-12 | 南京康尼机电股份有限公司 | A kind of inferior health Forecasting Methodology of urban rail transit vehicles door system |
| CN107563425A (en) * | 2017-08-24 | 2018-01-09 | 长安大学 | A kind of method for building up of the tunnel operation state sensor model based on random forest |
| CN108691678A (en) * | 2017-04-05 | 2018-10-23 | 通用汽车环球科技运作有限责任公司 | Detection and the method and system for alleviating sensor degradation |
-
2018
- 2018-11-27 CN CN201811425099.4A patent/CN109598048A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106372748A (en) * | 2016-08-29 | 2017-02-01 | 上海交通大学 | Hard-rock tunnel boring machine boring efficiency prediction method |
| CN108691678A (en) * | 2017-04-05 | 2018-10-23 | 通用汽车环球科技运作有限责任公司 | Detection and the method and system for alleviating sensor degradation |
| CN107153841A (en) * | 2017-04-24 | 2017-09-12 | 南京康尼机电股份有限公司 | A kind of inferior health Forecasting Methodology of urban rail transit vehicles door system |
| CN107563425A (en) * | 2017-08-24 | 2018-01-09 | 长安大学 | A kind of method for building up of the tunnel operation state sensor model based on random forest |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110717379A (en) * | 2019-08-28 | 2020-01-21 | 南京康尼机电股份有限公司 | Health assessment method for subway car door key components based on feature fusion |
| CN113821866A (en) * | 2020-06-19 | 2021-12-21 | 南京康尼机电股份有限公司 | Method for predicting residual life of finger protection adhesive tape of urban rail door system |
| CN113821866B (en) * | 2020-06-19 | 2024-01-19 | 南京康尼机电股份有限公司 | Method for predicting residual life of finger protection adhesive tape of urban rail door system |
| CN111947954A (en) * | 2020-07-17 | 2020-11-17 | 南京康尼机电股份有限公司 | Method and system for diagnosing urban rail door system fault or sub-health |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Data alignments in machinery remaining useful life prediction using deep adversarial neural networks | |
| Atamuradov et al. | Railway point machine prognostics based on feature fusion and health state assessment | |
| Li et al. | A deep learning driven method for fault classification and degradation assessment in mechanical equipment | |
| Soualhi et al. | Prognosis of bearing failures using hidden Markov models and the adaptive neuro-fuzzy inference system | |
| CN105718876B (en) | A kind of appraisal procedure of ball-screw health status | |
| Soualhi et al. | Pattern recognition method of fault diagnostics based on a new health indicator for smart manufacturing | |
| CN109460618A (en) | A kind of rolling bearing remaining life on-line prediction method and system | |
| CN109598048A (en) | A kind of lubrication degradation prediction technique of track vehicle door system | |
| CN104832418B (en) | A kind of based on local mean value conversion and the Fault Diagnosis of Hydraulic Pump method of Softmax | |
| CN108709744B (en) | Motor bearings method for diagnosing faults under a kind of varying load operating condition | |
| CN109597396B (en) | A kind of distribution transforming on-line fault diagnosis method based on high amount of traffic and transfer learning | |
| CN109374318A (en) | The more door system method for detecting abnormality of rail vehicle and system based on DPC | |
| Li et al. | Parallel multi-fusion convolutional neural networks based fault diagnosis of rotating machinery under noisy environments | |
| CN109580260A (en) | A kind of inferior health diagnostic method of track vehicle door system | |
| CN105022912A (en) | Rolling bearing fault prediction method based on wavelet principal component analysis | |
| Bie et al. | An integrated approach based on improved CEEMDAN and LSTM deep learning neural network for fault diagnosis of reciprocating pump | |
| CN107121285A (en) | A kind of bearing vibration signal fault feature extracting method | |
| CN108267312A (en) | A kind of subway train bearing intelligent diagnostic method based on fast search algorithm | |
| Hwang et al. | Application of cepstrum and neural network to bearing fault detection | |
| Khlaief et al. | Feature engineering for ball bearing combined-fault detection and diagnostic | |
| CN115496108A (en) | Fault monitoring method and system based on manifold learning and big data analysis | |
| CN111678699A (en) | Early fault monitoring and diagnosing method and system for rolling bearing | |
| Sadoughi et al. | A deep learning approach for failure prognostics of rolling element bearings | |
| Hasegawa et al. | Adaptive training of vibration-based anomaly detector for wind turbine condition monitoring | |
| Liang et al. | Generalized composite multiscale diversity entropy and its application for fault diagnosis of rolling bearing in automotive production line |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190409 |