CN107192565A - A kind of synchronization detecting method of subway vehicle door system exception operating mode and component degradation - Google Patents
A kind of synchronization detecting method of subway vehicle door system exception operating mode and component degradation Download PDFInfo
- Publication number
- CN107192565A CN107192565A CN201710378336.5A CN201710378336A CN107192565A CN 107192565 A CN107192565 A CN 107192565A CN 201710378336 A CN201710378336 A CN 201710378336A CN 107192565 A CN107192565 A CN 107192565A
- Authority
- CN
- China
- Prior art keywords
- data
- mrow
- msub
- door system
- 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.)
- Granted
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
- G01M17/00—Testing of vehicles
- G01M17/08—Railway vehicles
Abstract
Pretreatment operation is carried out the invention discloses the synchronization detecting method of a kind of subway vehicle door system exception operating mode and component degradation, including to the data of subway train-door system;Based on MeanShift algorithms, to pretreated subway vehicle door system data, the calculating of the density center point of normal data and degraded data is carried out respectively;It is adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect abnormal Outlier Data;Obtaining widget degradation information, and carry out maintenance guidance.The adverse effect that the abnormity point caused present invention, avoiding unusual service condition is caused to Hierarchical Clustering, possibility is provided to build accurately and reliably system unit degradation model, pass through component degradation characteristic matching matching degree, maintenance reference is provided to maintenance personal, reduce system maintenance cost, it is readily appreciated that, it is convenient to implement, have a good application prospect.
Description
Technical field
The present invention relates to the synchronization detecting method of a kind of subway vehicle door system exception operating mode and component degradation, belong to track
Vehicle Detection technical field.
Background technology
With the fast development of urban track traffic, subway is as the important component of urban transportation, and it is to people's work
The influence make, lived is also increasing, and railcar door system is as the key components of subway, and its high fault rate is over the ground
The safe operation of iron constitutes serious threat, so needing to detect the degenerate case of door system part, occurs event in door system
System can be carried out effectively, with targetedly safeguarding before barrier, it is to avoid the generation of the system failure.Simultaneously as subway
During operation by artificially extrude etc. outside environmental elements disturbed, the unusual service condition such as the sporadic failure of system and data acquisition abnormity
Generation, can cause, when carrying out modeling analysis based on data, have the generation of Outlier Data, these abnormal operating modes can be done again
Door system modeling is disturbed, causes model to occur uncertain, and then causes the degeneration that is difficult to detect door system part.So, if
The unusual service condition and component degradation of railcar door system can be synchronously detected, is examined while door system component degradation is detected
The operating mode of system exception is measured, and feeds back to door system, it is to avoid outlier interference system component degradation caused by unusual service condition
Detection, with regard to the accuracy of component degradation model and the validity of component degradation detection can be improved.
At present, the detection technique in terms of the component degradation of railcar door system, existing miscellaneous part are still not based on
Deterioration detecting, such as patent be used for automotive system event-driven fault diagnosis framework (CN102200487A), by than
Failure code when occurring compared with event of failure, to carry out the identification of failure, it needs the diagnostic code and mark of predefined system
Standard, process is complicated and needs to set more parameter.And a kind of lithium battery degeneration discrimination method of patent and degeneration warning system
(CN106199443A), for the performance degradation of lithium battery, the indexs such as electric current, voltage, the temperature of battery is measured, whether it is surpassed
Go out threshold value as degradation criteria, and the unusual service condition for railcar door system and component degradation detection, it is that two differences are general
Read, moreover, this method is not accurate enough, and this kind of large scale system with complex component of subway train-door system can not be fitted
With.
The content of the invention
In the prior art can not be to the synchronous unusual service condition for detecting railcar door system the invention aims to overcome
And the problem of component degradation.The subway vehicle door system exception operating mode of the present invention and the synchronization detecting method of component degradation, in inspection
While measuring system unit and degenerate, complete detection to system exception operating mode (abnormal Outlier Data), it is to avoid unusual service condition
The adverse effect that the abnormity point caused is caused to Hierarchical Clustering, to build accurately and reliably system unit degradation model, provide can
Can, by component degradation characteristic matching matching degree, maintenance reference is provided to maintenance personal, reduces system maintenance cost, easily reason
Solution, it is convenient to implement, have a good application prospect.
In order to achieve the above object, the technical solution adopted in the present invention is:
A kind of synchronization detecting method of subway vehicle door system exception operating mode and component degradation, it is characterised in that:Including with
Lower step,
Step (A), the data to subway train-door system carry out pretreatment operation, including temporal signatures extraction process, frequency
The standardization of characteristic of field extraction process, data;
Step (B), based on MeanShift algorithms, to pretreated subway vehicle door system data, is carried out normal respectively
The calculating of the density center point of data and degraded data;
Step (C), it is adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect
Abnormal Outlier Data;
Step (D), is divided into normal data, degraded data and abnormal Outlier Data, acquisition unit according to step (C) by data set
Part degradation information, and carry out maintenance guidance.
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(A), temporal signatures extraction process, comprises the following steps,
(A11) the temporal signatures value of door system data, including reflection central tendency information, are extracted with statistical analysis technique
First order statistic, be expressed as average;Reflect the second-order statistic of dispersion tendency information, be expressed as standard deviation;Reflect statistical
The high-order statistic of cloth shape, is expressed as the degree of bias, kurtosis;
(A21) stage extraction, is carried out to the temporal signatures value for extracting door system data, a switch gate process is divided into
Raising speed, high speed, reduction of speed, jogging section and back segment in place, points of five stage statistical analysis electric currents, rotating speed, position and acceleration
Maximum, minimum value, average, variance, the degree of bias, kurtosis.
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(A), frequency domain character extraction process, comprises the following steps,
(A12) frequency domain character of door system electric current, is extracted, using the method for wavelet transformation, the energy of multiple frequency domains is extracted
Feature;
(A22), by WAVELET PACKET DECOMPOSITION, the energy of each frequency band is isolated, as one of frequency domain character of extraction, is come
Reflect the performance degradation situation of door system.
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(A), the standardization of data, carries out data normalization using the similar Z-score methods standardized, makes different variable variances
Normalizing, realizes nondimensionalization.
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(B), based on MeanShift algorithms, to pretreated subway vehicle door system data, normal data and degeneration number are carried out respectively
According to density center point calculating, comprise the following steps,
(B1), using MeanShift algorithms, ask for the density center point of normal data and degraded data, MeanShift to
Shown in the formal definition of amount, such as formula (1),
Wherein, the statistical nature vector x that ith switch gated data is obtainediRepresent, x represents the statistical nature of datum mark
Vector, ShIt is the higher-dimension ball region that a radius is h;(xi- x) it is sample point xiRelative to datum mark x offset;MSh(x)
It is to fall into region ShIn offset vector of the t sample point relative to point x average;
(B2) gaussian kernel function G (x), is introduced, monotonously distance and the power of sample point between reflected sample point and datum mark
The relation of weight, now, vector M Sh(x) turn to, such as shown in formula (2),
Wherein, x represents the statistical nature vector of datum mark, and h is the bandwidth of kernel function, is met | | MSh(x) | | less than certain appearance
Perhaps error condition is can to obtain the stable state cluster centre point most converged to;
(B3) algorithm, described by (B1) and (B2), respectively obtains the density center point of normal data and degraded data
C1、C2。
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(C) it is, adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect abnormal Outlier Data,
Comprise the following steps,
(C1), respectively from the density center point of normal data and degraded data, search radius R is expanded, to data set
In point classified, define data sample minimum range dminFor data sample point other sample points into data set
Shown in degree of the peeling off OF of minimum Eustachian distance, then sample, such as formula (3),
(C2) corresponding classification, is assigned to non-classified data sample, expands search radius R, until the degree of peeling off of sample
OF is more than threshold value 0.3, and the radius when search radius of each class is the kind judging of the last unfiled point of completion is most counted at last
It is divided into two classes according to sample, the data that tag along sort is not assigned are abnormal Outlier Data.
Foregoing subway vehicle door system exception operating mode and the synchronization detecting method of component degradation, it is characterised in that:Step
(D) data set, is divided into by normal data, degraded data and abnormal Outlier Data according to step (C), obtaining widget degradation information,
And maintenance guidance is carried out, comprise the following steps,
(D1) data set, is divided into by normal data, degraded data and abnormal Outlier Data according to step (C), by normal
The density center point C of data and degraded data1、C2Carry out the analysis of the component degradation of door system;
(D2), by comparing the density center point C of normal data and degraded data1、C2, obtain component degradation model
FEATURE, by the corresponding C of each characteristic value1、C2Difference sorted from big to small, take preceding 20 variation characteristic amounts, obtain known to
Component degradation causes corresponding feature to change, and it is increase or reduction to obtain variation tendency;
(D3), when carrying out component degradation experiment offline, according to the degradation model FEATURE for obtaining different component degradations,
It is comprising 20 feature sequence numbers and positive and negative model vector, when handling the real time data of door system, by real time data
The algorithm process based on Meanshift is carried out with normal data, the unknown vector f eature of real time data generation is obtained, from
Search corresponding component degradation type in the model library set up offline, component degradation characteristic matching degree, according to formula (4),
Wherein, len (FEATURE) is the number of the length, i.e. variation characteristic value of degradation model, len (feature ∩
FEATURE the model vector) obtained for real time data and the model vector variation characteristic value identical number in off-line model storehouse;
(D4) the degeneration part of current door system, is provided according to aspect of model matching degree, maintenance personal is showed, reduces dimension
Repair the time, realize that maintenance is instructed.
The beneficial effects of the invention are as follows:The synchronous detection of the subway vehicle door system exception operating mode and component degradation of the present invention
Method, while detecting that system unit is degenerated, completes detection to system exception operating mode (abnormal Outlier Data), with
Lower feature,
(1) foundation of the present invention independent of physical model, the data gathered using the method for data-driven based on sensor
To carry out the detection of system unit degeneration, it is readily appreciated that, it is convenient to implement;
(2) present invention can be while detecting system component degradation, and (peel off synchronous progress system exception operating mode number extremely
According to) detection, with synchronous real-time, it is to avoid the adverse effect that the abnormity point that unusual service condition is caused is caused to Hierarchical Clustering,
Possibility is provided to build accurately and reliably system unit degradation model;
(3) present invention is modeled by system degradation part on the door, can carry out the detection of component degradation, obtain be
System component degradation information, by component degradation characteristic matching matching degree, gives maintenance personal to provide maintenance reference, reduces system maintenance
Cost, realizes that maintenance is instructed.
Brief description of the drawings
Fig. 1 be the present invention subway vehicle door system exception operating mode and component degradation synchronization detecting method flow chart;
Fig. 2 is the structure chart of the offline carry out component degradation experiment of the present invention.
Fig. 3 is the normal data of the specific embodiment of the invention and the result figure of V-type 8mm off-line modelings;
Fig. 4 is that the normal data and screw mandrel of the specific embodiment of the invention bend the result figure of off-line modeling.
Embodiment
Below in conjunction with Figure of description, the present invention will be further described.Following examples are only used for clearly
Illustrate technical scheme, and can not be limited the scope of the invention with this.
As shown in figure 1, the subway vehicle door system exception operating mode and the synchronization detecting method of component degradation of the present invention, including
Following steps,
Step (A), the data to subway train-door system carry out pretreatment operation, including temporal signatures extraction process, frequency
The standardization of characteristic of field extraction process, data,
Temporal signatures extraction process, comprises the following steps,
(A11) the temporal signatures value of door system data, including reflection central tendency information, are extracted with statistical analysis technique
First order statistic, be expressed as average;Reflect the second-order statistic of dispersion tendency information, be expressed as standard deviation;Reflect statistical
The high-order statistic of cloth shape, is expressed as the degree of bias, kurtosis;
(A21) stage extraction, is carried out to the temporal signatures value for extracting door system data, a switch gate process is divided into
Raising speed, high speed, reduction of speed, jogging section and back segment in place, points of five stage statistical analysis electric currents, rotating speed, position and acceleration
Maximum, minimum value, average, variance, the degree of bias, kurtosis.
Frequency domain character extraction process, comprises the following steps,
(A12) frequency domain character of door system electric current, is extracted, using the method for wavelet transformation, the energy of multiple frequency domains is extracted
Feature;
(A22), by WAVELET PACKET DECOMPOSITION, the energy of each frequency band is isolated, as one of frequency domain character of extraction, is come
Reflect the performance degradation situation of door system;
For example, carrying out 4 wavelet decompositions to the signal of electric current, retain low frequency signal at every layer, high-frequency signal is entered
Next layer of wavelet decomposition of row, 5 sub-band signal Di are obtained, and (i=1,2 ..., 5), sub-band signal Di energy is under
Formula is calculated:
Ei=∑ | Di|2
The standardization of data, carries out data normalization using the similar Z-score methods standardized, makes different variables
Variance normalizing, realizes nondimensionalization.
Step (B), based on MeanShift algorithms, to pretreated subway vehicle door system data, is carried out normal respectively
The calculating of the density center point of data and degraded data, comprises the following steps,
(B1), using MeanShift algorithms, ask for the density center point of normal data and degraded data, MeanShift to
Shown in the formal definition of amount, such as formula (1),
Wherein, the statistical nature vector x that ith switch gated data is obtainediRepresent, x represents the statistical nature of datum mark
Vector, ShIt is the higher-dimension ball region that a radius is h;(xi- x) it is sample point xiRelative to datum mark x offset;MSh(x)
It is to fall into region ShIn offset vector of the t sample point relative to point x average;
(B2) gaussian kernel function G (x), is introduced, monotonously distance and the power of sample point between reflected sample point and datum mark
The relation of weight, now, vector M Sh(x) turn to, such as shown in formula (2),
Wherein, x represents the statistical nature vector of datum mark, and h is the bandwidth of kernel function, is met | | MSh(x) | | less than certain appearance
Perhaps error condition is can to obtain the stable state cluster centre point most converged to;
(B3) algorithm, described by (B1) and (B2), respectively obtains the density center point of normal data and degraded data
C1、C2;
Step (C), it is adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect
Abnormal Outlier Data, comprises the following steps,
(C1), respectively from the density center point of normal data and degraded data, search radius R is expanded, to data set
In point classified, define data sample minimum range dminFor data sample point other sample points into data set
Shown in degree of the peeling off OF of minimum Eustachian distance, then sample, such as formula (3),
(C2) corresponding classification, is assigned to non-classified data sample, expands search radius R, until the degree of peeling off of sample
OF is more than threshold value 0.3, and the radius when search radius of each class is the kind judging of the last unfiled point of completion is most counted at last
It is divided into two classes according to sample, the data that tag along sort is not assigned are abnormal Outlier Data;
Step (D), is divided into normal data, degraded data and abnormal Outlier Data, acquisition unit according to step (C) by data set
Part degradation information, and maintenance guidance is carried out, comprise the following steps,
(D1) data set, is divided into by normal data, degraded data and abnormal Outlier Data according to step (C), by normal
The density center point C of data and degraded data1、C2Carry out the analysis of the component degradation of door system;
(D2), by comparing the density center point C of normal data and degraded data1、C2, obtain component degradation model
FEATURE, by the corresponding C of each characteristic value1、C2Difference sorted from big to small, take preceding 20 variation characteristic amounts, obtain known to
Component degradation causes corresponding feature to change, and it is increase or reduction to obtain variation tendency;
(D3), when carrying out component degradation experiment offline, as shown in Fig. 2 according to the degeneration mould for obtaining different component degradations
Type FEATURE, its be comprising 20 feature sequence numbers and positive and negative model vector, when handling the real time data of door system,
Real time data and normal data are subjected to the algorithm process based on Meanshift, the unknown vector of real time data generation is obtained
Feature, searches corresponding component degradation type, component degradation characteristic matching degree, according to public affairs from the model library set up offline
Shown in formula (4),
Wherein, len (FEATURE) is the number of the length, i.e. variation characteristic value of degradation model, len (feature ∩
FEATURE the model vector) obtained for real time data and the model vector variation characteristic value identical number in off-line model storehouse;
(D4) the degeneration part of current door system, is provided according to aspect of model matching degree, maintenance personal is showed, reduces dimension
Repair the time, realize that maintenance is instructed.
But, if component degradation characteristic matching degree is less than 0.5, it is determined as novel features degradation model, now will be
The model vector feature that line is obtained is added in the algorithm model knowledge base of system, while maintenance personal carries out door system
Maintenance improves knowledge base there is provided component degradation information, and the maintenance for the degradation model next time provides foundation.
The synchronization detecting method of subway vehicle door system exception operating mode and component degradation below according to the present invention, introduces one
Individual specific embodiment, carries out experimental verification on a certain railcar door system, and one group of collection subway train-door system is normal
Data, the door leaf of car door is adjusted afterwards makes the V-type size of car door reach 8mm, gathers the data of V-type mono- group of data of 8mm, synchronous inspection
Result is surveyed, as shown in figure 3, wherein, preceding 16 data are normal data, the 17th data are unusual service condition, and remainder data is to be
Door leaf V-type of uniting 8mm data, degree of membership represents that sample belongs to the degree of normal data, and degree of membership is represented in sample and density for 1
Heart point is overlapped, it is observed that carrying out when setting up of off-line system model, can accurately obtain the degeneration mould of door system
Type, and synchronously detect the outlier of system, it is to avoid interference of the outlier to the degradation model degree of accuracy.Pass through analysis system
Model FEATURE, wherein, door closing procedure total kilometres reduce, door opening process total kilometres increase meets system physical principle, side
The accuracy of system model is demonstrated, such as Fig. 4 provides the cluster result that one group of normal data bends data with screw mandrel,
The component degradation matching degree result of table 1
Set up 4 component degradation models of system, respectively V-type 8mm, screw mandrel bending, intermediate supports offline with algorithm
Loosen and nut be damaged, using corresponding 4 groups of test datas are tested respectively when, test result as shown in table 1, correspondence survey
The characteristic matching degree of data is tried more than 80%, and the characteristic matching degree of erroneous matching is not arranged all less than 10% in table 1
Go out.Thus, it is possible to draw the synchronization detecting method energy of subway vehicle door system exception operating mode proposed by the present invention and component degradation
Enough to provide component degradation information exactly, the personnel that can maintain easily timely and accurately carry out system maintenance, shorten maintenance time,
Maintenance cost is reduced, while avoiding system from further deteriorating causes the generation of failure.
In summary, the synchronization detecting method of subway vehicle door system exception operating mode of the invention and component degradation, in inspection
While measuring system unit degeneration, the detection to system exception operating mode (abnormal Outlier Data) is completed, is had the characteristics that,
(1) foundation of the present invention independent of physical model, the data gathered using the method for data-driven based on sensor
To carry out the detection of system unit degeneration, it is readily appreciated that, it is convenient to implement;
(2) present invention can be while detecting system component degradation, and (peel off synchronous progress system exception operating mode number extremely
According to) detection, with synchronous real-time, it is to avoid the adverse effect that the abnormity point that unusual service condition is caused is caused to Hierarchical Clustering,
Possibility is provided to build accurately and reliably system unit degradation model;
(3) present invention is modeled by system degradation part on the door, can carry out the detection of component degradation, obtain be
System component degradation information, by component degradation characteristic matching matching degree, gives maintenance personal to provide maintenance reference, reduces system maintenance
Cost, realizes that maintenance is instructed.
General principle, principal character and the advantage of the present invention has been shown and described above.The technical staff of the industry should
Understand, the present invention is not limited to the above embodiments, the original for simply illustrating the present invention described in above-described embodiment and specification
Reason, without departing from the spirit and scope of the present invention, various changes and modifications of the present invention are possible, these changes and improvements
It all fall within the protetion scope of the claimed invention.The claimed scope of the invention is by appended claims and its equivalent circle
It is fixed.
Claims (7)
1. the synchronization detecting method of a kind of subway vehicle door system exception operating mode and component degradation, it is characterised in that:Including following
Step,
Step (A), the data to subway train-door system carry out pretreatment operation, including temporal signatures extraction process, frequency domain spy
Levy the standardization of extraction process, data;
Step (B), based on MeanShift algorithms, to pretreated subway vehicle door system data, carries out normal data respectively
With the calculating of the density center point of degraded data;
Step (C), it is adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect exception
Outlier Data;
Step (D), is divided into normal data, degraded data and abnormal Outlier Data, obtaining widget is moved back according to step (C) by data set
Change information, and carry out maintenance guidance.
2. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 1 and component degradation, it is special
Levy and be:Step (A), temporal signatures extraction process comprises the following steps,
(A11) the temporal signatures value of door system data, is extracted with statistical analysis technique, includes the one of reflection central tendency information
Rank statistic, is expressed as average;Reflect the second-order statistic of dispersion tendency information, be expressed as standard deviation;Reflect statistical distribution shape
The high-order statistic of shape, is expressed as the degree of bias, kurtosis;
(A21) stage extraction, is carried out to the temporal signatures value for extracting door system data, a switch gate process is divided into liter
Speed, high speed, reduction of speed, jogging section and back segment in place, points of five stage statistical analysis electric currents, rotating speed, position and acceleration are most
Big value, minimum value, average, variance, the degree of bias, kurtosis.
3. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 1 and component degradation, it is special
Levy and be:Step (A), frequency domain character extraction process comprises the following steps,
(A12) frequency domain character of door system electric current, is extracted, using the method for wavelet transformation, the energy feature of multiple frequency domains is extracted;
(A22), by WAVELET PACKET DECOMPOSITION, the energy of each frequency band is isolated, as one of frequency domain character of extraction, to reflect
The performance degradation situation of door system.
4. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 1 and component degradation, it is special
Levy and be:Step (A), the standardization of data carries out data normalization using the similar Z-score methods standardized, made
Different variable variance normalizings, realize nondimensionalization.
5. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 1 and component degradation, it is special
Levy and be:Step (B), based on MeanShift algorithms, to pretreated subway vehicle door system data, is carried out normal respectively
The calculating of the density center point of data and degraded data, comprises the following steps,
(B1), using MeanShift algorithms, the density center point of normal data and degraded data is asked for, MeanShift vectors
Shown in formal definition, such as formula (1),
<mrow>
<msub>
<mi>MS</mi>
<mi>h</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mn>1</mn>
<mi>t</mi>
</mfrac>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>S</mi>
<mi>h</mi>
</msub>
</mrow>
</munder>
<mrow>
<mo>(</mo>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, the statistical nature vector x that ith switch gated data is obtainediRepresent, x represents the statistical nature vector of datum mark,
ShIt is the higher-dimension ball region that a radius is h;(xi- x) it is sample point xiRelative to datum mark x offset;MSh(x) it is to fall into
Region ShIn offset vector of the t sample point relative to point x average;
(B2) gaussian kernel function G (x), is introduced, monotonously distance and the weight of sample point between reflected sample point and datum mark
Relation, now, vector M Sh(x) turn to, such as shown in formula (2),
<mrow>
<msub>
<mi>MS</mi>
<mi>h</mi>
</msub>
<mrow>
<mo>(</mo>
<mi>x</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>S</mi>
<mi>h</mi>
</msub>
</mrow>
</munder>
<mi>G</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mi>x</mi>
</mrow>
<mi>h</mi>
</mfrac>
<mo>)</mo>
</mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
</mrow>
<mrow>
<munder>
<mo>&Sigma;</mo>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>&Element;</mo>
<msub>
<mi>S</mi>
<mi>h</mi>
</msub>
</mrow>
</munder>
<mi>G</mi>
<mrow>
<mo>(</mo>
<mfrac>
<mrow>
<msub>
<mi>x</mi>
<mi>i</mi>
</msub>
<mo>-</mo>
<mi>x</mi>
</mrow>
<mi>h</mi>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mi>x</mi>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, x represents the statistical nature vector of datum mark, and h is the bandwidth of kernel function, is met | | MSh(x) | | allow to miss less than certain
Poor condition is can to obtain the stable state cluster centre point most converged to;
(B3) algorithm, described by (B1) and (B2), respectively obtains the density center point C of normal data and degraded data1、C2。
6. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 1 and component degradation, it is special
Levy and be:Step (C), it is adaptive to choose density center neighborhood of a point radius, the cluster modeling of complete paired data, and detect different
Normal Outlier Data, comprises the following steps,
(C1), respectively from the density center point of normal data and degraded data, expand search radius R, data are concentrated
Point is classified, and defines the minimum range d of data sampleminFor the minimum of data sample point other sample points into data set
Shown in degree of the peeling off OF of Euclidean distance, then sample, such as formula (3),
<mrow>
<mi>O</mi>
<mi>F</mi>
<mo>=</mo>
<mfrac>
<msub>
<mi>d</mi>
<mrow>
<mi>m</mi>
<mi>i</mi>
<mi>n</mi>
</mrow>
</msub>
<mi>R</mi>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
<mo>;</mo>
</mrow>
(C2) corresponding classification, is assigned to non-classified data sample, expands search radius R, until degree of the peeling off OF of sample is big
In threshold value 0.3, the radius during kind judging that the search radius of each class completes unfiled point to be last, most data sample at last
Originally it is divided into two classes, the data that tag along sort is not assigned are abnormal Outlier Data.
7. the synchronization detecting method of subway vehicle door system exception operating mode according to claim 5 and component degradation, it is special
Levy and be:Step (D), is divided into normal data, degraded data and abnormal Outlier Data, acquisition unit according to step (C) by data set
Part degradation information, and maintenance guidance is carried out, comprise the following steps,
(D1) data set, is divided into by normal data, degraded data and abnormal Outlier Data according to step (C), passes through normal data
With the density center point C of degraded data1、C2Carry out the analysis of the component degradation of door system;
(D2), by comparing the density center point C of normal data and degraded data1、C2, component degradation model FEATURE is obtained,
By the corresponding C of each characteristic value1、C2Difference sorted from big to small, take preceding 20 variation characteristic amounts, obtain known elements degenerate
Corresponding feature is caused to change, and it is increase or reduction to obtain variation tendency;
(D3), when carrying out component degradation experiment offline, according to the degradation model FEATURE for obtaining different component degradations, it is
Comprising 20 feature sequence numbers and positive and negative model vector, when handling the real time data of door system, by real time data with just
Regular data carries out the algorithm process based on Meanshift, the unknown vector f eature of real time data generation is obtained, from offline
Search corresponding component degradation type in the model library of foundation, component degradation characteristic matching degree, according to formula (4),
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>t</mi>
<mi>c</mi>
<mi>h</mi>
<mi>F</mi>
<mi>E</mi>
<mi>A</mi>
<mi>T</mi>
<mi>U</mi>
<mi>R</mi>
<mi>E</mi>
<mo>=</mo>
<mfrac>
<mrow>
<mi>l</mi>
<mi>e</mi>
<mi>n</mi>
<mrow>
<mo>(</mo>
<mrow>
<mi>f</mi>
<mi>e</mi>
<mi>a</mi>
<mi>t</mi>
<mi>u</mi>
<mi>r</mi>
<mi>e</mi>
<mo>&cap;</mo>
<mi>F</mi>
<mi>E</mi>
<mi>A</mi>
<mi>T</mi>
<mi>U</mi>
<mi>R</mi>
<mi>E</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<mi>l</mi>
<mi>e</mi>
<mi>n</mi>
<mrow>
<mo>(</mo>
<mrow>
<mi>F</mi>
<mi>E</mi>
<mi>A</mi>
<mi>T</mi>
<mi>U</mi>
<mi>R</mi>
<mi>E</mi>
</mrow>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
Wherein, len (FEATURE) is the number of the length, i.e. variation characteristic value of degradation model, len (feature ∩
FEATURE the model vector) obtained for real time data and the model vector variation characteristic value identical number in off-line model storehouse;
(D4) the degeneration part of current door system, is provided according to aspect of model matching degree, maintenance personal is showed, when reducing maintenance
Between, realize that maintenance is instructed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710378336.5A CN107192565B (en) | 2017-05-25 | 2017-05-25 | A kind of synchronization detecting method of subway vehicle door system exception operating condition and component degradation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710378336.5A CN107192565B (en) | 2017-05-25 | 2017-05-25 | A kind of synchronization detecting method of subway vehicle door system exception operating condition and component degradation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107192565A true CN107192565A (en) | 2017-09-22 |
CN107192565B CN107192565B (en) | 2019-05-28 |
Family
ID=59875532
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710378336.5A Active CN107192565B (en) | 2017-05-25 | 2017-05-25 | A kind of synchronization detecting method of subway vehicle door system exception operating condition and component degradation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107192565B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108197647A (en) * | 2017-12-28 | 2018-06-22 | 中南大学 | A kind of Fast Speed Clustering of automobile starter durable test data |
CN108629007A (en) * | 2018-05-04 | 2018-10-09 | 上饶市普适科技有限公司 | A method of excluding history enrollment data exceptional value |
CN108827632A (en) * | 2018-03-30 | 2018-11-16 | 刘文飞 | A kind of determination method of communications equipment room equipment health status |
CN109374318A (en) * | 2018-09-27 | 2019-02-22 | 南京康尼机电股份有限公司 | The more door system method for detecting abnormality of rail vehicle and system based on DPC |
CN109580260A (en) * | 2018-11-27 | 2019-04-05 | 南京康尼机电股份有限公司 | A kind of inferior health diagnostic method of track vehicle door system |
CN110398382A (en) * | 2019-06-28 | 2019-11-01 | 南京康尼机电股份有限公司 | A kind of underground automobile door system performance decline detection method based on DPC |
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 |
CN111999580A (en) * | 2020-08-14 | 2020-11-27 | 佳都新太科技股份有限公司 | Detection method and device for subway platform gate, computer equipment and storage medium |
CN112198472A (en) * | 2020-09-29 | 2021-01-08 | 国网上海市电力公司 | Online remote calibration method and system for partial discharge intelligent sensor |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101510230A (en) * | 2009-03-11 | 2009-08-19 | 同济大学 | Vehicle road load emulation method |
CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
CN102187539A (en) * | 2008-09-08 | 2011-09-14 | Abb研究有限公司 | Apparatus and method for adaptive fault detection in mv distribution circuits |
KR20140099986A (en) * | 2013-02-04 | 2014-08-14 | 동서대학교산학협력단 | Colorization-based color image coding method by meanshift segmentation algorithm |
CN104697804A (en) * | 2015-03-24 | 2015-06-10 | 清华大学 | Method and system for detecting and separating intermittent faults of train active suspension system |
CN105046270A (en) * | 2015-06-19 | 2015-11-11 | 上海卓悠网络科技有限公司 | Application classification model constructing method and system and application classification method and system |
CN105241680A (en) * | 2015-08-26 | 2016-01-13 | 电子科技大学 | Health state assessment method for rotary machine based on probability density function |
CN106406295A (en) * | 2016-12-02 | 2017-02-15 | 南京康尼机电股份有限公司 | Rail transit vehicle door system fault diagnosis and early warning method based on multiple conditions |
-
2017
- 2017-05-25 CN CN201710378336.5A patent/CN107192565B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102187539A (en) * | 2008-09-08 | 2011-09-14 | Abb研究有限公司 | Apparatus and method for adaptive fault detection in mv distribution circuits |
CN101510230A (en) * | 2009-03-11 | 2009-08-19 | 同济大学 | Vehicle road load emulation method |
CN102005034A (en) * | 2010-12-01 | 2011-04-06 | 南京大学 | Remote sensing image segmentation method based on region clustering |
KR20140099986A (en) * | 2013-02-04 | 2014-08-14 | 동서대학교산학협력단 | Colorization-based color image coding method by meanshift segmentation algorithm |
CN104697804A (en) * | 2015-03-24 | 2015-06-10 | 清华大学 | Method and system for detecting and separating intermittent faults of train active suspension system |
CN105046270A (en) * | 2015-06-19 | 2015-11-11 | 上海卓悠网络科技有限公司 | Application classification model constructing method and system and application classification method and system |
CN105241680A (en) * | 2015-08-26 | 2016-01-13 | 电子科技大学 | Health state assessment method for rotary machine based on probability density function |
CN106406295A (en) * | 2016-12-02 | 2017-02-15 | 南京康尼机电股份有限公司 | Rail transit vehicle door system fault diagnosis and early warning method based on multiple conditions |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108197647B (en) * | 2017-12-28 | 2021-08-31 | 中南大学 | Rapid clustering method for automobile starter endurance test data |
CN108197647A (en) * | 2017-12-28 | 2018-06-22 | 中南大学 | A kind of Fast Speed Clustering of automobile starter durable test data |
CN108827632A (en) * | 2018-03-30 | 2018-11-16 | 刘文飞 | A kind of determination method of communications equipment room equipment health status |
CN108629007A (en) * | 2018-05-04 | 2018-10-09 | 上饶市普适科技有限公司 | A method of excluding history enrollment data exceptional value |
CN108629007B (en) * | 2018-05-04 | 2022-08-09 | 上饶市普适科技有限公司 | Method for eliminating abnormal values of historical enrollment data |
CN109374318A (en) * | 2018-09-27 | 2019-02-22 | 南京康尼机电股份有限公司 | The more door system method for detecting abnormality of rail vehicle and system based on DPC |
CN109580260A (en) * | 2018-11-27 | 2019-04-05 | 南京康尼机电股份有限公司 | A kind of inferior health diagnostic method of track vehicle door system |
CN110398382A (en) * | 2019-06-28 | 2019-11-01 | 南京康尼机电股份有限公司 | A kind of underground automobile door system performance decline detection method based on DPC |
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 |
CN111999580A (en) * | 2020-08-14 | 2020-11-27 | 佳都新太科技股份有限公司 | Detection method and device for subway platform gate, computer equipment and storage medium |
CN111999580B (en) * | 2020-08-14 | 2023-12-01 | 佳都科技集团股份有限公司 | Method and device for detecting subway platform gate, computer equipment and storage medium |
CN112198472A (en) * | 2020-09-29 | 2021-01-08 | 国网上海市电力公司 | Online remote calibration method and system for partial discharge intelligent sensor |
CN112198472B (en) * | 2020-09-29 | 2023-11-07 | 国网上海市电力公司 | Online remote verification method and system for partial discharge intelligent sensor |
Also Published As
Publication number | Publication date |
---|---|
CN107192565B (en) | 2019-05-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107192565A (en) | A kind of synchronization detecting method of subway vehicle door system exception operating mode and component degradation | |
Li et al. | Data alignments in machinery remaining useful life prediction using deep adversarial neural networks | |
Chen et al. | A just-in-time-learning-aided canonical correlation analysis method for multimode process monitoring and fault detection | |
CN111882446B (en) | Abnormal account detection method based on graph convolution network | |
CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
CN111832225A (en) | Method for constructing driving condition of automobile | |
CN113702895B (en) | Online quantitative evaluation method for error state of voltage transformer | |
CN109684673B (en) | Feature extraction and cluster analysis method for transient stability result of power system | |
CN107153841B (en) | Sub-health prediction method for urban rail transit vehicle door system | |
Mechefske et al. | Fault detection and diagnosis in low speed rolling element bearings Part II: The use of nearest neighbour classification | |
CN113848417A (en) | Rail transit power supply equipment fault prediction method and device | |
CN106778908A (en) | A kind of novelty detection method and apparatus | |
CN108257365A (en) | A kind of industrial alarm designs method based on global nonspecific evidence dynamic fusion | |
CN105241665A (en) | Rolling bearing fault diagnosis method based on IRBFNN-AdaBoost classifier | |
Yang et al. | Enhanced hierarchical symbolic dynamic entropy and maximum mean and covariance discrepancy-based transfer joint matching with Welsh loss for intelligent cross-domain bearing health monitoring | |
CN115221930A (en) | Fault diagnosis method for rolling bearing | |
Zhang et al. | A bearing fault diagnosis method based on multiscale dispersion entropy and GG clustering | |
CN115496108A (en) | Fault monitoring method and system based on manifold learning and big data analysis | |
CN114882069A (en) | Taxi track abnormity detection method based on LSTM network and attention mechanism | |
Wen et al. | A new method for identifying the ball screw degradation level based on the multiple classifier system | |
Yan et al. | Bearing performance degradation assessment based on the continuous-scale mathematical morphological particle and feature fusion | |
KR102362872B1 (en) | Method for refining clean labeled data for artificial intelligence training | |
Li et al. | Source-free domain adaptation framework for fault diagnosis of rotation machinery under data privacy | |
Xie et al. | Detection of wheel diameter difference of railway wagon by ACMD-FBD and optimized MKELM | |
CN114547796B (en) | Ball mill feature fusion fault diagnosis method based on optimized BN network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |