CN107215357B - A kind of switch breakdown prediction technique - Google Patents
A kind of switch breakdown prediction technique Download PDFInfo
- Publication number
- CN107215357B CN107215357B CN201710375981.1A CN201710375981A CN107215357B CN 107215357 B CN107215357 B CN 107215357B CN 201710375981 A CN201710375981 A CN 201710375981A CN 107215357 B CN107215357 B CN 107215357B
- Authority
- CN
- China
- Prior art keywords
- curve
- prediction
- track switch
- regular event
- switch
- 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.)
- Active
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61L—GUIDING RAILWAY TRAFFIC; ENSURING THE SAFETY OF RAILWAY TRAFFIC
- B61L23/00—Control, warning, or like safety means along the route or between vehicles or vehicle trains
- B61L23/04—Control, warning, or like safety means along the route or between vehicles or vehicle trains for monitoring the mechanical state of the route
- B61L23/042—Track changes detection
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
Abstract
The present invention provides a kind of switch breakdown prediction techniques, wherein this method includes acquiring the continuous n times regular event curve of same track switch;Characteristic features are extracted to each regular event curve;Prediction model is established respectively to the homogenous characteristics of N items regular event curve;Track switch is calculated separately based on prediction model the Y times(Y>N)Value when action per category feature;Fault diagnosis is carried out to the Y times action prediction curve of track switch, judges whether failure.Track switch next stage working condition and remaining life can be predicted by the invention, offer reference for maintenance decision optimization, realize switch breakdown early prevention, reduce track switch accident rate, improve system reliability, ensure traffic safety.
Description
Technical field
The present invention relates to field of track traffic, and in particular to a kind of switch breakdown prediction technique.
Background technology
Track switch is that train is transferred to or crosses essential line facility when another strand of track from one track, is rail
One important component and the highest equipment of failure rate in road.Once track switch is broken down, compulsory exercise cannot be completed,
It is light then temporary parking causes time delays;The derailing of heavy then compartment causes casualties.
China Railway operation is frequent, operating environment is changeable easily leads to switch breakdown, but existing Switch monitor condition with
And repair means fall behind, these situations easily cause driving burst accident or influence driving efficiency.Railway department is to prevent at present
Accident occurs, and arranges professional to inspect periodically and repair track switch, needs a large amount of man power and material, and inefficiency.It is existing
Have in technology, can not also accomplish, to switch breakdown Accurate Prediction, also not proposing quickly and effectively solution.So based on being adopted
The historical datas such as the track switch operation curve of collection and fault message, using scientific method predict track switch next stage working condition and
Remaining life finds and coordinates to implement corresponding maintenance support measure before its failure in time.It predicts to believe according to switch breakdown
Breath is that maintenance decision optimization is offered reference, and rationally adjusts maintenance intervals, reduces " repair is superfluous " and " repair is insufficient ", ensure driving
The important measure of safety and passenger survival safety.
Invention content
The present invention provides a kind of switch breakdown prediction techniques, realize to track switch next stage working condition and remaining life
Prediction, the problem of optimization maintenance decision adjusts maintenance intervals, reduces in existing repair " repair superfluous " and " repair deficiency ".
A kind of switch breakdown prediction technique proposed by the present invention, includes the following steps:
(1):Acquire the continuous n times regular event curve of same track switch;
(2):Characteristic features are extracted to each regular event curve;Characteristic features have 10, specially:T1For road
The current maxima of trouble unlocking phases, T2For the current value corresponding time;T3Electricity during being fallen after rise for track switch unlocking phases electric current
Flow minimum value, T4The position current value corresponding time;T5The electric current mode value in stage, T are converted for track switch6It is corresponded to for the current value
At the time of;T7Enter moment in locking stage corresponding current value, T from the conversion stage for track switch8At the time of correspondence for the current value;T9
End point i.e. this moment current value, which is acted, for track switch is reduced to zero, T10For the current value corresponding time;
(3):Prediction model is established respectively to the homogenous characteristics in the characteristic features of N items regular event curve extraction;
(4):Based on step(3)Prediction model calculates separately track switch the Y times(Y>N)Characteristic value when action per category feature;
(5):The prediction of every category feature of the Y times action of track switch is worth to based on the feature per category feature obtained by prediction model
Curve;
(6):To step(5)The Y times action prediction curve of obtained track switch carries out fault diagnosis, judges whether failure.
In the present invention, step(1)Described in the same track switch of acquisition continuous n times regular event curve be from microcomputer monitoring system
The track switch operation curve data or image extracted in system, or be the track switch operation curve data or image in paper document.
In the present invention, step(1)The continuous n times regular event curve of the same track switch of the acquisition, it is specific as follows:
(1a):Same track switch continuous N time operation curve is acquired from microcomputer detecting system(M>N);
(1b):Damage curve is removed, normalized curve is retained;
(1c):Curve is labeled as { curve 1, curve 2, curve 3 ... curve N } sequentially in time.
In the present invention, step(3)Described in the prediction model established be BP neural network, be as follows:
(3a):The be operating normally data of every category feature of curve of track switch n times are normalized respectively, using as follows
Formula:
Wherein, xminFor the minimum number in data sequence;xmaxFor the maximum number in data sequence;
(3b):The characteristic value per category feature is chosen respectively as mode input sample;
(3c):Three layers of BP neural network model are built, determine input layer, hidden layer and output layer neuron number;
(3d):Distinguish assignment to input layer, hidden layer and each connection weight of output layer, determines constructed BP neural network
Target error, learning rate, frequency of training, error function and the hidden layer of model and the activation primitive of output layer;
(3e)According to(3d)Setting, BP neural network model obtain prediction data;
(3f):According to the real data, prediction data, error function of BP neural network model to each neuron of output layer
Partial derivative, the connection weight of hidden layer to output layer, hidden layer output error function pair hidden layer each neuron partial derivative
And the output of each neuron of hidden layer, connection weight is corrected, relative error is calculatedE, using following formula:
Wherein, Ypred(t) it is BP neural network output valve, Yreal(t) it is actual value;
(3g):When relative error reaches default precision or frequency of training more than setting maximum times, this time prediction terminates, obtains
To the predicted value of feature, otherwise repeatedly step(3e)It arrives(3g).
In the present invention, step(5)The prediction that the Y times action of track switch is worth to based on feature obtained by prediction model is bent
Line, it is specific as follows:
(5a):Based on the respective predicted value of 10 category feature of turnout curve, according to(T1,T2),(T3,T4),(T5,T6),(T7,
T8)(T9,T10)Mode forms 5 characteristic points, with origin(0,0)The basic point of constituent curve fitting;
(5b):It based on principle of least square method, is carried out curve fitting using fitting function, obtains the Y times action of track switch
Prediction curve.
In the present invention, step(6)Described in fault diagnosis is carried out to the Y times action prediction curve of track switch, judge whether
Failure, the specific steps are:
(6a):Choose track switch action track switch regular event template curve;
(6b):Set the threshold value of prediction curve and track switch regular event template curve gross accumulation distance;
(6c):Using dynamic time warping algorithm calculate the gross accumulation of prediction curve and normal electric current template curve away from
From;
(6d):Calculate step(6c)Gained gross accumulation distance is more than step(6b)Set threshold value, then it represents that track switch Y
Secondary action is broken down, otherwise, it means that not breaking down.
In the present invention, step(6c)Described in dynamic time warping algorithm calculated curve distance, the specific steps are:
(6c1):Track switch acts the Y times prediction curve and is represented by T={ T(1), T(2)... ..., T(n)... ..., T(N),
N is the sequential label of prediction curve time series, and n=1 is prediction curve time series starting point, and n=N is prediction curve time series
Terminal, T(n)For the value of the prediction curve time series;
(6c2):Track switch regular event template curve is represented by R={ R(1), R(2)... ..., R(m)... ..., R(M), m
For the sequential label for the template curve time series that is operating normally, m=1 is regular event template curve time series starting point, and m=M is
Be operating normally template curve time series terminal, R(m)For the value of the regular event template curve time series;
(6c3):Each sequential label n of prediction curve time series is marked in horizontal axis, it is normal dynamic to mark representative in the longitudinal axis
Make each sequential label m of template curve time series, drawing some co-ordinations by the rounded coordinate of these sequential labels can
A network is formed, all lattice points are followed successively by(1,1)... ...,(N, m)... ...,(N, M), search from(1,1)It arrives(N, M)Most
Shortest path;
(6c4):When path passes through(N, m)Afterwards, it is next by lattice point can only be(n,m+1),(N+1, m),(N+1, m+
1), selection(N, m)Minimum range to next lattice point is optimal path, is calculated(1,1)It arrives(N, M)Accumulation minimum range;
(6c5):The Euclidean distance that prediction curve time series T is calculated between the template curve time series R that is operating normally;
(6c6):Starting point(1,1)To terminal(N, M)Total accumulation distance be starting point(1,1)To terminal(N, M)Product
The sum of Euclidean distance between tired minimum range, prediction curve time series T and regular event template curve time series R;
(6c7):Total accumulation distance is about small, indicates that the similarity of prediction curve and the template curve that is operating normally is higher.
The beneficial effects of the present invention are:The historical datas such as track switch operation curve and fault message, using scientific method
It predicts track switch next stage working condition and remaining life, judges whether to break down, be according to switch breakdown prediction result
Maintenance decision optimization is offered reference, and maintenance intervals are rationally adjusted, and is found in time before its failure and is coordinated the corresponding dimension of implementation
The problems such as repairing safeguard, reducing " repair is superfluous " and " repair is insufficient ", reduces track switch accident rate, improves system reliability, protect
Demonstrate,prove traffic safety.
Description of the drawings
It, below will be to specific in order to illustrate more clearly of the specific embodiment of the invention or technical solution in the prior art
Embodiment or attached drawing needed to be used in the description of the prior art are briefly described, it should be apparent that, in being described below
Attached drawing is some embodiments of the present invention, for those of ordinary skill in the art, before not making the creative labor
It puts, other drawings may also be obtained based on these drawings.
Fig. 1 is the flow chart of switch breakdown prediction technique according to the ... of the embodiment of the present invention;
Fig. 2 is the 100th article of track switch regular event current curve feature extraction figure according to the ... of the embodiment of the present invention;
Fig. 3 is Switch current curvilinear characteristic T according to the ... of the embodiment of the present invention7Prediction model;
Fig. 4 is that track switch according to the ... of the embodiment of the present invention acts the 1001st action prediction current curve;
Fig. 5 is that track switch according to the ... of the embodiment of the present invention acts the 1100th action prediction current curve;
Fig. 6 is track switch prediction curve Troubleshooting Flowchart.
Specific implementation mode
Technical scheme of the present invention is clearly and completely described below in conjunction with attached drawing, it is clear that described implementation
Example is a part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill
The every other embodiment that personnel are obtained without making creative work, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that term " first ", " second ", " third " are used for description purposes only,
It is not understood to indicate or imply relative importance.
As long as in addition, technical characteristic involved in invention described below different embodiments non-structure each other
It can be combined with each other at conflict.
Embodiment 1
A kind of switch breakdown prediction technique is provided in the present embodiment, and Fig. 1 is that failure according to the ... of the embodiment of the present invention is pre-
The flow chart of survey method, as shown in Figure 1, the flow chart includes the following steps:
Step S11 acquires the continuous n times regular event curve of same track switch;
Step S12 extracts characteristic features to each operation curve;
Step S13 establishes prediction model respectively to the homogenous characteristics of N curve;
Step S14 calculates separately track switch the Y times based on prediction model(Y>N)Value when action per category feature;
Step S15 is worth to the prediction curve of the Y times action of track switch based on feature obtained by prediction model;
Step S16 carries out fault diagnosis to the Y times action prediction curve of track switch, judges whether failure.
Through the above steps, it is obtained using scientific method based on the same track switch continuous several times regular event curve acquired
Track switch next stage failure predication result and predicting residual useful life, compared with the prior art in, it is regular by enough staff
The high cost and poor efficiency brought are overhauled, above-mentioned steps reduce the feelings of " repair is superfluous " and " repair is insufficient " in the prior art
Condition, and maintenance intervals are rationally adjusted according to prediction result, it predicts and coordinates to implement corresponding maintenance support in time before the failure
Measure reduces track switch accident rate, improves traffic safety.
It is illustrated with reference to a specific alternative embodiment.
(One)Same track switch continuous several times action current curve is acquired from microcomputer detecting system, removes damage curve, is protected
Normalized curve is stayed, multigroup curve is labeled as { curve 1, curve 2, curve 3 ... curve 1000 } sequentially in time.
(Two)As shown in Fig. 2, acting normal current curve for the 100th article of track switch, 10 representatives are extracted to the operation curve
Property feature and obtain its corresponding value, respectively:T1For the current maxima of track switch unlocking phases, T2It is corresponding for the current value
Time;T3Current minimum during being fallen after rise for track switch unlocking phases electric current, T4The position current value corresponding time;T5For track switch
The electric current mode value in conversion stage, T6At the time of correspondence for the current value;T7Enter the moment in locking stage from the conversion stage for track switch
Corresponding current value, T8At the time of correspondence for the current value, T9End point i.e. this moment current value, which is acted, for track switch is reduced to zero, T10For
The current value corresponding time similarly carries out feature extraction to other 999 track switches action normal current curve successively, finally obtains
The characteristic obtained is as follows, wherein N=1000.
(Three):Prediction model is established respectively to extracting 10 curvilinear characteristics based on BP neural network, chooses turnout curve
T7For establish prediction model, be as follows:
(1) to curvilinear characteristic T71000 corresponding characteristic values are normalized, and as mode input
Sample;
(2)Three layers of BP neural network model are built, set input layer number as 10, hidden layer neuron number is
8, output layer neuron number is 1;
(3)It is 0.5 to each connection weight assignment, target error is set as 0.0001, and learning efficiency is set as 0.01, network
Frequency of training is set as 10000 times;
(4)The activation primitive of hidden layer uses tan-sigmoid transmission functions, the activation primitive of output layer to be passed using linear
Delivery function;Error function uses;
(5)According to BP neural network desired output, reality output, error function to each neuron partial derivative of output layer,
Hidden layer is to the connection weight of output layer, each neuron partial derivative and hidden layer of the output error function pair hidden layer of hidden layer
Connection weight is corrected in the output of each neuron, is calculated relative error and is used following formula:
Wherein, Ypred(t) it is BP neural network output valve, Yreal(t) it is actual value;
(6) it is 0.032664 to calculate relative error magnitudes, and frequency of training is more than 10000 at this time, and model prediction terminates;
(7)Feature T when obtaining the 1001st action of track switch7Predicted value be 1.587;
(8)Repeat step(1)It arrives(7), the predicted value of other 9 features when obtaining the 1001st action of track switch successively, 10
The value of a feature is respectively 4.732,0.4,1.506,1.1,1.323,4.1,1.587,6.9,0,8.8.
(Four)According to the predicted value of 10 features when the action of the track switch obtained the 1001st time, according to(T1,T2),(T3,
T4),(T5,T6),(T7,T8)(T9,T10)Mode forms 5 characteristic points, with origin(0,0)The basic point of constituent curve fitting;
It based on principle of least square method, is carried out curve fitting using polynomial fit function, fitting letter is constantly adjusted according to fitting result
The value of the degree of polynomial in number, best fitting result is chosen, the prediction curve of the 1001st action of track switch is obtained, such as Fig. 4 institutes
Show.
(Five)According to top step 3 to four, predict that the prediction curve for obtaining the 1100th time is as shown in Figure 5 successively.
(Six)It is template curve to select track switch the 1st article of current curve that be operating normally, and sets prediction curve and template curve is total
The threshold value for accumulating distance is 35.
(Seven)Track switch is calculated using dynamic time warping algorithm and acts the 1001st prediction curve and track switch regular event the 1st
The gross accumulation distance of current curve, steps are as follows:
(1)The 1001st prediction curve is acted by way of trouble, is represented by T={ T(1), T(2)... ..., T(116), T(1)=
0, T(2)=0.4325 ... ..., T(116)=0.09189;
(2)Track switch the 1st article of current curve that be operating normally is represented by R={ R(1), R(2)... ..., R(214), R(1)=0,
R(2)=0.22356 ... ..., R(144)=0.01029;
(3)Each sequential label 116 that track switch acts the 1001st prediction curve time series is marked in horizontal axis, in the longitudinal axis
The each sequential label 214 for marking template curve time series, some are drawn in length and breadth by the rounded coordinate of these sequential labels
Line can form a network, and all lattice points are followed successively by(1,1)... ...,(116,214), search(1,1)It arrives(116,214)Most
Shortest path;
(4):Path passes through(1,1)Afterwards, it is next by lattice point can only be(1,2),(2,1),(2,2), can be calculated
(1,1)It arrives(116,214)Accumulation minimum range be 17.7823;
(5):It can be calculated the Euclidean distance between plot against time sequence T and normal representation plot against time sequence R to be identified
It is 1.2834;
(6):Starting point(1,1)To terminal(116,214)Total accumulation distance be 13.0657, the 1001st action of track switch
The gross accumulation of prediction curve and template curve distance is 18.0657.
(Eight)According to step(Seven)Using same method, with this be calculated the 1001st action prediction curve of track switch with
The gross accumulation distance of template curve is 94.6732;
(Nine)The gross accumulation of the 1001st action prediction curve of track switch and template curve distance is less than threshold value, prediction track switch the
1001 actions are not broken down;The gross accumulation of the 1100th prediction curve of track switch and template curve distance is more than threshold value, prediction
The 1100th action of track switch can break down.
It should be understood by those skilled in the art that, the embodiment of the present invention can be provided as method, system or computer program
Product.Therefore, complete hardware embodiment, complete software embodiment or reality combining software and hardware aspects can be used in the present invention
Apply the form of example.Moreover, the present invention can be used in one or more wherein include computer usable program code computer
Usable storage medium(Including but not limited to magnetic disk storage, CD-ROM, optical memory etc.)The computer program of upper implementation produces
The form of product.
The present invention be with reference to according to the method for the embodiment of the present invention, equipment(System)And the flow of computer program product
Figure and/or block diagram describe.It should be understood that can be realized by computer program instructions every first-class in flowchart and/or the block diagram
The combination of flow and/or box in journey and/or box and flowchart and/or the block diagram.These computer programs can be provided
Instruct the processor of all-purpose computer, special purpose computer, Embedded Processor or other programmable data processing devices to produce
A raw machine so that the instruction executed by computer or the processor of other programmable data processing devices is generated for real
The device for the function of being specified in present one flow of flow chart or one box of multiple flows and/or block diagram or multiple boxes.
These computer program instructions, which may also be stored in, can guide computer or other programmable data processing devices with spy
Determine in the computer-readable memory that mode works so that instruction generation stored in the computer readable memory includes referring to
Enable the manufacture of device, the command device realize in one flow of flow chart or multiple flows and/or one box of block diagram or
The function of being specified in multiple boxes.
These computer program instructions also can be loaded onto a computer or other programmable data processing device so that count
Series of operation steps are executed on calculation machine or other programmable devices to generate computer implemented processing, in computer or
The instruction executed on other programmable devices is provided for realizing in one flow of flow chart or multiple flows and/or block diagram one
The step of function of being specified in a box or multiple boxes.
Obviously, the above embodiments are merely examples for clarifying the description, and does not limit the embodiments.It is right
For those of ordinary skill in the art, can also make on the basis of the above description it is other it is various forms of variation or
It changes.There is no necessity and possibility to exhaust all the enbodiments.And it is extended from this it is obvious variation or
It changes still within the protection scope of the invention.
Claims (7)
1. a kind of switch breakdown prediction technique, which is characterized in that include the following steps:
(1):Acquire the continuous n times regular event curve of same track switch;
(2):Characteristic features are extracted to each regular event curve;Characteristic features have 10, specially:T1It is solved for track switch
The current maxima in lock stage, T2For the current value corresponding time;T3Most for electric current during the falling of track switch unlocking phases electric current
Small value, T4For the current value corresponding time;T5The electric current mode value in stage, T are converted for track switch6When corresponding for the current value
It carves;T7Enter moment in locking stage corresponding current value, T from the conversion stage for track switch8At the time of correspondence for the current value;T9For road
Trouble action terminates the corresponding current value of point moment, T10For the current value corresponding time;
(3):Prediction model is established respectively to the homogenous characteristics in the characteristic features of N items regular event curve extraction;
(4):Based on step(3)Characteristic value when prediction model calculates separately the Y times action of track switch per category feature, wherein Y>N;
(5):The prediction curve of the Y times action of track switch is worth to based on the feature per category feature obtained by prediction model;
(6):To step(5)The Y times action prediction curve of obtained track switch carries out fault diagnosis, judges whether failure.
2. switch breakdown prediction technique according to claim 1, which is characterized in that step(1)Described in acquisition it is same
The continuous n times regular event curve of track switch is the track switch operation curve data extracted from microcomputer detecting system or image, or is paper
Track switch operation curve data in matter file or image.
3. switch breakdown prediction technique according to claim 1, which is characterized in that step(1)The acquisition is with along with
Branch off continuous n times regular event curve, it is specific as follows:
(1a):Same track switch continuous N time operation curve, wherein M> are acquired from microcomputer detecting system;N;
(1b):Damage curve is removed, normalized curve is retained;
(1c):Curve is labeled as { curve 1, curve 2, curve 3 ... curve N } sequentially in time.
4. switch breakdown prediction technique according to claim 1, which is characterized in that step(3)Described in the prediction established
Model is BP neural network, is as follows:
(3a):The data of every category feature of track switch n times regular event curve are normalized respectively, utilize following public affairs
Formula:
Wherein, xminFor the minimum number in data sequence;xmaxFor the maximum number in data sequence;
(3b):The characteristic value per category feature is chosen respectively as mode input sample;
(3c):Three layers of BP neural network model are built, determine input layer, hidden layer and output layer neuron number;
(3d):Distinguish assignment to input layer, hidden layer and each connection weight of output layer, determines constructed BP neural network model
Target error, learning rate, frequency of training, error function and hidden layer and output layer activation primitive;
(3e)According to step(3d)Setting, BP neural network model obtain prediction data;
(3f):According to the real data, prediction data, error function of BP neural network model to each neuron local derviation of output layer
Number, each neuron partial derivative of output error function pair hidden layer of the connection weight of hidden layer to output layer, hidden layer and hidden
Connection weight is corrected in the output of each neuron containing layer, calculates relative errorE, using following formula:
Wherein, Ypred(t) it is BP neural network output valve, Yreal(t) it is actual value;
(3g):When relative error reaches default precision or frequency of training more than setting maximum times, this time prediction terminates, and obtains spy
The predicted value of sign, otherwise repeatedly step(3e)It arrives(3g).
5. switch breakdown prediction technique according to claim 1, which is characterized in that step(5)It is described based on prediction mould
Feature obtained by type is worth to the prediction curve of the Y times action of track switch, specific as follows:
(5a):Based on the respective predicted value of 10 category feature of turnout curve, according to(T1,T2),(T3,T4),(T5,T6),(T7,T8)
(T9,T10)Mode forms 5 characteristic points, with origin(0,0)The basic point of constituent curve fitting;
(5b):It based on principle of least square method, is carried out curve fitting using fitting function, obtains the prediction of the Y times action of track switch
Curve.
6. switch breakdown prediction technique according to claim 1, which is characterized in that step(6)Described in track switch Y
Secondary action prediction curve carries out fault diagnosis, judges whether failure, the specific steps are:
(6a):Choose track switch regular event template curve;
(6b):Set the threshold value of prediction curve and track switch regular event template curve gross accumulation distance;
(6c):The gross accumulation distance of prediction curve and track switch regular event template curve is calculated using dynamic time warping algorithm;
(6d):Calculate step(6c)Gained gross accumulation distance is more than step(6b)Set threshold value, then it represents that track switch the Y times is dynamic
It breaks down, otherwise, it means that not breaking down.
7. switch breakdown prediction technique according to claim 6, which is characterized in that step(6c)The specific steps are:
(6c1):Track switch acts the Y times prediction curve and is expressed as T={ T(1), T(2)... ..., T(n)... ..., T(N), n is pre-
The sequential label of plot against time sequence is surveyed, n=1 is prediction curve time series starting point, and n=N is prediction curve time series terminal,
T(n)For the value of the prediction curve time series;
(6c2):Track switch regular event template curve is expressed as R={ R(1), R(2)... ..., R(m)... ..., R(M), m is normal
The sequential label of template curve time series is acted, m=1 is regular event template curve time series starting point, and m=M is normal dynamic
Make template curve time series terminal, R(m)For the value of the regular event template curve time series;
(6c3):Each sequential label n of prediction curve time series is marked in horizontal axis, representative regular event mould is marked in the longitudinal axis
Each sequential label m of plate plot against time sequence draws some co-ordinations by the rounded coordinate of these sequential labels and forms one
A network, all lattice points are followed successively by(1,1)... ...,(N, m)... ...,(N, M), search from(1,1)It arrives(N, M)Optimal road
Diameter;
(6c4):When path passes through(N, m)Afterwards, it is next by lattice point can only be(n,m+1),(N+1, m),(N+1, m+1),
Selection(N, m)Minimum range to next lattice point is optimal path, is calculated(1,1)It arrives(N, M)Accumulation minimum range;
(6c5):The Euclidean distance that prediction curve time series T is calculated between the template curve time series R that is operating normally;
(6c6):Starting point(1,1)To terminal(N, M)Total accumulation distance be starting point(1,1)To terminal(N, M)Accumulation most
The sum of Euclidean distance between small distance, prediction curve time series T and regular event template curve time series R;
(6c7):Total accumulation distance is about small, indicates that the similarity of prediction curve and the template curve that is operating normally is higher.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710375981.1A CN107215357B (en) | 2017-05-25 | 2017-05-25 | A kind of switch breakdown prediction technique |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710375981.1A CN107215357B (en) | 2017-05-25 | 2017-05-25 | A kind of switch breakdown prediction technique |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107215357A CN107215357A (en) | 2017-09-29 |
CN107215357B true CN107215357B (en) | 2018-10-26 |
Family
ID=59944753
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710375981.1A Active CN107215357B (en) | 2017-05-25 | 2017-05-25 | A kind of switch breakdown prediction technique |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107215357B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108182386B (en) * | 2017-12-13 | 2021-12-17 | 河南辉煌科技股份有限公司 | Automatic generation method of turnout standard curve |
CN108052628B (en) * | 2017-12-19 | 2020-07-14 | 河北省科学院应用数学研究所 | Turnout starting current detection method, system and terminal equipment |
CN108238066B (en) * | 2017-12-22 | 2019-11-05 | 同济大学 | Track switch operation curve template choosing method and its application |
CN110954734B (en) * | 2018-09-26 | 2022-03-25 | 北京国铁盛阳技术有限公司 | Fault diagnosis method, device, equipment and storage medium |
CN109501834B (en) * | 2018-11-28 | 2020-09-25 | 交控科技股份有限公司 | Method and device for predicting fault of turnout switch machine |
CN109633237B (en) * | 2018-12-14 | 2022-01-28 | 卡斯柯信号有限公司 | Method and equipment for correcting total current meter reading of turnout based on least square method |
CN111105147B (en) * | 2019-12-02 | 2021-12-21 | 北京交通大学 | Turnout health state assessment method based on dynamic time warping |
US11577763B2 (en) * | 2020-03-06 | 2023-02-14 | Alstom Transport Technologies | Method and controller for determining the relationship between a track-circuit transmitted current signal and a railway vehicle location on a railway track |
CN111915061B (en) * | 2020-07-06 | 2022-06-07 | 同济大学 | Switch action current curve prediction method and fault discrimination method thereof |
CN111881950A (en) * | 2020-07-10 | 2020-11-03 | 交控科技股份有限公司 | Method and device for representing characteristics of current time sequence of turnout switch machine |
CN112036505A (en) * | 2020-09-17 | 2020-12-04 | 广西交控智维科技发展有限公司 | Method and device for determining equipment state of turnout switch machine and electronic equipment |
CN112084673B (en) * | 2020-09-17 | 2023-01-31 | 广西交控智维科技发展有限公司 | Automatic setting method and device for switch friction current |
CN112381308A (en) * | 2020-11-20 | 2021-02-19 | 广州新科佳都科技有限公司 | Training of current prediction model, current prediction method and related device |
CN113627496A (en) * | 2021-07-27 | 2021-11-09 | 交控科技股份有限公司 | Method, device, electronic equipment and readable storage medium for predicting fault of turnout switch machine |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104805742A (en) * | 2014-01-24 | 2015-07-29 | 中国铁道科学研究院 | Method and device for dynamic detection of railway turnout states |
CN105260595A (en) * | 2015-04-02 | 2016-01-20 | 北京交通大学 | Feature extraction method for switch action current curve and switch fault diagnosis method |
CN105787511A (en) * | 2016-02-26 | 2016-07-20 | 清华大学 | Track switch fault diagnosis method and system based on support vector machine |
CN105882687A (en) * | 2016-05-13 | 2016-08-24 | 南京雅信科技集团有限公司 | Method for analyzing categories of faults of point machines |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103129584B (en) * | 2011-11-30 | 2016-02-03 | 国际商业机器公司 | Switch monitor method and system, railway line management method and system |
-
2017
- 2017-05-25 CN CN201710375981.1A patent/CN107215357B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104805742A (en) * | 2014-01-24 | 2015-07-29 | 中国铁道科学研究院 | Method and device for dynamic detection of railway turnout states |
CN105260595A (en) * | 2015-04-02 | 2016-01-20 | 北京交通大学 | Feature extraction method for switch action current curve and switch fault diagnosis method |
CN105787511A (en) * | 2016-02-26 | 2016-07-20 | 清华大学 | Track switch fault diagnosis method and system based on support vector machine |
CN105882687A (en) * | 2016-05-13 | 2016-08-24 | 南京雅信科技集团有限公司 | Method for analyzing categories of faults of point machines |
Also Published As
Publication number | Publication date |
---|---|
CN107215357A (en) | 2017-09-29 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107215357B (en) | A kind of switch breakdown prediction technique | |
CN110441629B (en) | Fault diagnosis method and device based on turnout switch machine operating current curve | |
CN107203746B (en) | Turnout fault identification method | |
CN206265072U (en) | The control device of station train travel operation | |
CN105083331B (en) | Operation management system for locomotive running preparation | |
CN104535581B (en) | Steel rail damage detection cart | |
CN111915061B (en) | Switch action current curve prediction method and fault discrimination method thereof | |
CA2395395A1 (en) | A train corridor scheduling process including various cost functions associated with railway operations | |
CN105973457A (en) | China railway high-speed train on-board stability monitoring device and method | |
JP2022543921A (en) | Predictive maintenance method for equipment through distribution map | |
EP3623202A1 (en) | Method for detecting anomalies in an electric traffic system and monitoring system for detecting anomalies in an electric traffic system | |
US11299185B2 (en) | Monitoring infrastructure facilities by means of geo-clustering | |
CN111703476A (en) | Method and system for tracking shunting route and automatically reporting point in non-centralized area | |
Tam et al. | Intelligent optical fibre sensing networks facilitate shift to predictive maintenance in railway systems | |
CN102963400B (en) | Method and device for realizing graphical display of inlet route instructions | |
CN110598750B (en) | Working condition identification method based on switch machine action curve similarity characteristics | |
CN110414739B (en) | Method and device for optimizing overhaul period of urban rail train | |
JP2011173529A (en) | Computer and method for operation arrangement according to train congestion status | |
JP2017081216A (en) | Operation management device | |
CN204389405U (en) | Rail failure detection go-cart | |
JP6082311B2 (en) | Wide area disaster recovery support system and monitoring center in this system | |
CN104574028B (en) | Judge different GPS whether with the movement of same car method | |
CN112257547A (en) | Transformer substation safety measure identification method based on deep learning | |
JP6927835B2 (en) | Railway equipment management system and railway equipment management method | |
CN113137925A (en) | Intelligent dynamic displacement measurement system based on image recognition |
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 |