CN105730431B - EMU checking cylinder fault monitoring method and fault monitoring system - Google Patents
EMU checking cylinder fault monitoring method and fault monitoring system Download PDFInfo
- Publication number
- CN105730431B CN105730431B CN201610065389.7A CN201610065389A CN105730431B CN 105730431 B CN105730431 B CN 105730431B CN 201610065389 A CN201610065389 A CN 201610065389A CN 105730431 B CN105730431 B CN 105730431B
- Authority
- CN
- China
- Prior art keywords
- variance
- variable
- sample
- test sample
- measurement data
- 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
- B60—VEHICLES IN GENERAL
- B60T—VEHICLE BRAKE CONTROL SYSTEMS OR PARTS THEREOF; BRAKE CONTROL SYSTEMS OR PARTS THEREOF, IN GENERAL; ARRANGEMENT OF BRAKING ELEMENTS ON VEHICLES IN GENERAL; PORTABLE DEVICES FOR PREVENTING UNWANTED MOVEMENT OF VEHICLES; VEHICLE MODIFICATIONS TO FACILITATE COOLING OF BRAKES
- B60T17/00—Component parts, details, or accessories of power brake systems not covered by groups B60T8/00, B60T13/00 or B60T15/00, or presenting other characteristic features
- B60T17/18—Safety devices; Monitoring
- B60T17/22—Devices for monitoring or checking brake systems; Signal devices
- B60T17/228—Devices for monitoring or checking brake systems; Signal devices for railway vehicles
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60T—VEHICLE BRAKE CONTROL SYSTEMS OR PARTS THEREOF; BRAKE CONTROL SYSTEMS OR PARTS THEREOF, IN GENERAL; ARRANGEMENT OF BRAKING ELEMENTS ON VEHICLES IN GENERAL; PORTABLE DEVICES FOR PREVENTING UNWANTED MOVEMENT OF VEHICLES; VEHICLE MODIFICATIONS TO FACILITATE COOLING OF BRAKES
- B60T17/00—Component parts, details, or accessories of power brake systems not covered by groups B60T8/00, B60T13/00 or B60T15/00, or presenting other characteristic features
- B60T17/18—Safety devices; Monitoring
- B60T17/22—Devices for monitoring or checking brake systems; Signal devices
- B60T17/221—Procedure or apparatus for checking or keeping in a correct functioning condition of brake systems
Landscapes
- Engineering & Computer Science (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Valves And Accessory Devices For Braking Systems (AREA)
Abstract
The invention discloses a kind of EMU checking cylinder fault monitoring method and fault monitoring system, which includes, and obtains each brake-cylinder pressure measurement data composing training data set for including braking procedure under nominal situation;Training data is calculated to concentrate variance between the variable of each sample and determine the threshold value of variance between variable;The pressure measurement data at each checking cylinder current time is gathered in real time as test sample, and using the threshold decision of variance test sample between the variable whether comprising faulty;When being judged as including in the event of failure, the reconstruct contribution margin of each brake-cylinder pressure measurement data in the test sample, and the checking cylinder that the checking cylinder with maximum reconstructed contribution margin is determined as breaking down are asked for successively.This method can be detected and separate to failure in time, indicate the checking cylinder or place pipeline being abnormal, and as the prediction of catastrophe failure, certain reference be provided with replacing for the on-call maintenance of component, so as to be effectively prevented from the generation of even more serious failure.
Description
Technical field
The present invention relates to industry monitoring and fault diagnosis field, more particularly to a kind of EMU checking cylinder fault monitoring method
And fault monitoring system.
Background technology
With the continuous improvement of the EMU speed of service, people are also constantly increasing the demand of EMU safety in operation
Add, this just proposes EMU braking technology the requirement of higher.Though the dynamic braking technology of EMU achieves length in recent years
Foot development, but air damping technology is still a part indispensable in EMU braking technology, to ensureing EMU safety
Operation is extremely important.
Brake rigging of the checking cylinder as EMU Pneumatic brake systems, collectively forms with brake clamp, brake disc
The execution links of Pneumatic brake systems.Wherein, critical component of the checking cylinder as EMU Pneumatic brake systems, to EMU
Realize that reliable and secure braking plays a decisive role.
To meet the function (such as braking distance, braking deceleration) of EMU braking and performance (such as anti-impulsion limitation, passenger
Comfort level) demand, during braking/alleviation, usually require that EMU often saves the brake-cylinder pressure on multiple axles of train
Within a certain period of time while rise/fall is to some pressure set points, and allows certain control error.If applied in braking
Add, keep the pipeline during either alleviating where above-mentioned checking cylinder or checking cylinder to break down, then the braking of EMU
Performance even braking function will be affected.Let out for example, if gas occurs for pipeline where some checking cylinder or the checking cylinder
Leak failure, then the brake-cylinder pressure will be less than setting value, so that the corresponding axle brake force of the checking cylinder is less than other axis
Brake force, and then influence train braking performance.Furthermore the reason such as component capabilities degeneration, system caused by long-time service
Dynamic cylinder is it can also happen that the slow failure of inflation/exhaust.Serious checking cylinder inflation/slow failure of exhaust will cause checking cylinder pressure
There is exception in power, so as to trigger existing excessive/too low failure of monitoring system report brake-cylinder pressure, and then causes this section train braking
Power excision even needs the situation for carrying out parking process to occur.Even if slight some checking cylinder inflation/slow failure of exhaust
It can make it that the brake force between each axle of this section train is inconsistent, and then influence braking ability.Therefore, usually pacify in practice
Dress pressure sensor monitors the pressure of checking cylinder in real time, to be diagnosed to be the exception that may occur or serious event in time
Barrier.
Existing monitoring system is generally taken based on the event that single argument overload alarm mechanism is likely to occur EMU checking cylinder
Barrier is monitored, main to consider beyond larger excessive/too low failure of brake-cylinder pressure of setting value.This kind of failure is usually by very
Serious checking cylinder or place pipeline leak, and can not inflate/be vented caused by checking cylinder dysfunction or seriously
Inflation/exhaust slowly and cause.For the slight leakage of checking cylinder or place pipeline, or due to checking cylinder hydraulic performance decline and
The caused slow failure of inflation/exhaust, existing monitoring logic do not relate to.
To sum up, the multiple checking cylinders for often saving train to EMU there is an urgent need for a kind of new method for diagnosing faults are supervised online
Control, has timely completed the detection and separation of checking cylinder failure.
The content of the invention
The first technical problem to be solved by the present invention be need to provide a kind of new method for diagnosing faults it is every to EMU
The multiple checking cylinders for saving train carry out on-line monitoring, have timely completed the detection and separation of checking cylinder failure.
In order to solve the above-mentioned technical problem, embodiments herein provide firstly a kind of EMU checking cylinder malfunction monitoring
Method, including, obtain each brake-cylinder pressure measurement data composing training data set for including braking procedure under nominal situation;Calculate
Training data concentrates variance between the variable of each sample and determines the threshold value of variance between variable;Gather in real time each checking cylinder it is current when
As test sample, and using the threshold decision of variance between the variable, whether the test sample includes the pressure measurement data at quarter
It is faulty;When being judged as including in the event of failure, the weight of each brake-cylinder pressure measurement data in the test sample is asked for successively
Structure contribution margin, and the checking cylinder that the checking cylinder with maximum reconstructed contribution margin is determined as breaking down.
Preferably, in variance between calculating training data and concentrating the variable of each sample and the threshold value of variance between variable is determined
Step includes:The average of sample is calculated, the poor quadratic sum of each measurement data and the average is used as the sample using in sample
Variable between variance;Variance between the variable of each sample of acquisition training data concentration successively, by the maximum of variance between the variable
Value is determined as the threshold value of variance between variable.
Preferably, according to variance var between the following expression calculating variable:
Wherein, TCF (i) represents measurement data corresponding with i-th of checking cylinder in sample,Represent the average of sample, n
Represent the number of checking cylinder included in sample.
Preferably, whether the threshold decision of the variance test sample is included in faulty step between using the variable
Including, variance between the variable of the test sample is calculated, and between the variable compared with the threshold value of variance, when the test sample
Variable between variance be more than the variable between variance threshold value when, judge the test sample include it is faulty;When the test sample
Variable between variance be less than or equal to the variable between variance threshold value when, judge that the test sample does not include failure.
Preferably, in the test sample is asked for successively each brake-cylinder pressure measurement data reconstruct contribution margin step
Suddenly include:The measurement data set is replaced with the average of other measurement data in test sample in addition to some measurement data
Into reconstructed sample, ask for successively in the test sample corresponding to variance between the variable of the reconstructed sample of each checking cylinder;By institute
The difference of variance between the variable of variance and each reconstructed sample between the variable of test sample is stated as each checking cylinder pressure in test sample
The reconstruct contribution margin of power measurement data.
Embodiments herein additionally provides a kind of EMU checking cylinder fault monitoring system, including, data acquisition module,
It obtains each brake-cylinder pressure measurement data composing training data set for including braking procedure under nominal situation;Modeling module, its
Training data is calculated to concentrate variance between the variable of each sample and determine the threshold value of variance between variable;Detection module, it is adopted in real time
Collect the pressure measurement data at each checking cylinder current time as test sample, and should using the threshold decision of variance between the variable
Whether test sample is comprising faulty;Separation module, it is asked in the test sample successively when being judged as including in the event of failure
The reconstruct contribution margin of each brake-cylinder pressure measurement data, and the checking cylinder with maximum reconstructed contribution margin is determined as generation event
The checking cylinder of barrier.
Preferably, modeling module calculates training data according to following steps and concentrates variance between the variable of each sample and determine
The threshold value of variance between variable:Calculate the average of sample, using in sample the poor quadratic sum of each measurement data and the average as
Variance between the variable of the sample;Variance between the variable of each sample of acquisition training data concentration successively, by variance between the variable
Maximum be determined as the threshold value of variance between variable.
Preferably, modeling module is according to variance var between the following expression calculating variable:
Wherein, TCF (i) represents measurement data corresponding with i-th of checking cylinder in sample,Represent the average of sample, n
Represent the number of checking cylinder included in sample.
Preferably, detection module judges that test sample whether comprising faulty, calculates the test sample according to following steps
Variable between variance, and between the variable compared with the threshold value of variance, when variance is more than institute between the variable of the test sample
When stating the threshold value of variance between variable, it is faulty to judge that the test sample includes;When between the variable of the test sample variance be less than etc.
Between the variable during threshold value of variance, judge that the test sample does not include failure.
Preferably, separation module asks for each brake-cylinder pressure measurement number in the test sample according to following steps successively
According to reconstruct contribution margin:The measurement is replaced with the average of other measurement data in test sample in addition to some measurement data
Data form reconstructed sample, ask for successively in the test sample corresponding to side between the variable of the reconstructed sample of each checking cylinder
Difference;The difference of variance between the variable of variance and each reconstructed sample between the variable of the test sample is respectively made as in test sample
The reconstruct contribution margin of dynamic cylinder pressure measurement data.
Compared with prior art, one or more of such scheme embodiment can have the following advantages that or beneficial to effect
Fruit:
The feature of measurement data is modeled by using this Measure Indexes of variance between the variable of sample, can be timely
Ground is detected and separates to failure, indicates the checking cylinder or place pipeline being abnormal, and is portion as the prediction of catastrophe failure
The on-call maintenance of part provides certain reference with replacing, so as to be effectively prevented from the generation of even more serious failure.
Other advantages, the target of the present invention, and feature will be illustrated in the following description to a certain extent, and
And to a certain extent, based on will be apparent to those skilled in the art to investigating hereafter, Huo Zheke
To be instructed from the practice of the present invention.The target and other advantages of the present invention can be wanted by following specification, right
Specifically noted structure is sought in book, and attached drawing to realize and obtain.
Brief description of the drawings
Attached drawing is used for providing to the technical solution of the application or further understanding for the prior art, and constitution instruction
A part.Wherein, the attached drawing for expressing the embodiment of the present application is used for the technical side for explaining the application together with embodiments herein
Case, but do not form the limitation to technical scheme.
Fig. 1 is the flow diagram according to the EMU checking cylinder fault monitoring method of one embodiment of the invention;
Fig. 2 is that the 1# checking cylinders in an example of the invention inflate slow failure detection result schematic diagram;
Fig. 3 is that the 1# checking cylinders in an example of the invention inflate slow fault reconstruction result schematic diagram;
Fig. 4 is the 3# checking cylinder principal fault testing result schematic diagrames in an example of the invention;
Fig. 5 is the 3# checking cylinder principal fault separating resulting schematic diagrames in an example of the invention;
Fig. 6 is the structure diagram according to the EMU checking cylinder fault monitoring system of another embodiment of the present invention.
Embodiment
Carry out the embodiment that the present invention will be described in detail below with reference to accompanying drawings and embodiments, how the present invention is applied whereby
Technological means solves technical problem, and that reaches relevant art effect realizes that process can fully understand and implement according to this.This Shen
Each feature that please be in embodiment and embodiment, can be combined with each other under the premise of not colliding, the technical solution formed
Within protection scope of the present invention.
In general, the often section train of EMU includes two bogies, each bogie includes two axis (i.e. two pairs take turns to),
Multiple checking cylinders are separately installed with this four axis.Under air supply system air feed, the pressure of checking cylinder is controlled by air damping is
System is regulated and controled according to different brake requests.Then, brake-cylinder pressure is ultimately converted to train braking by brake rigging
Power, even stops so as to fulfill the Reduced Speed Now of EMU.
Checking cylinder it is an object of the invention to often save four axis of train to EMU implements online joint-monitoring, in time
Detect the slight leakage due to checking cylinder or place pipeline, or inflation/exhaust caused by checking cylinder hydraulic performance decline
Slow failure, and position the checking cylinder to break down.For convenience of hereafter stating, equally, we will be often saved on each axis of train
One or more checking cylinder of installation is considered as a checking cylinder.That is, equally think that EMU is often saved on four axis of train
Four checking cylinders are included altogether.A pressure sensor is installed on each axis, to monitor the brake-cylinder pressure of the axis.EMU
Often saving the checking cylinder installed on four axis of train has similar performance, meanwhile, air brake control system is intended to brake force
It is evenly distributed on four axis of every section train.Therefore, do not occur in EMU Pneumatic brake systems checking cylinder or place pipeline
In the case of exception, four brake-cylinder pressures for often saving train are consistent substantially.
Fig. 1 is according to the flow diagram of the EMU checking cylinder fault monitoring method of one embodiment of the invention, this method
Comprise the following steps:
Step S110, obtains each brake-cylinder pressure measurement data composing training data for including braking procedure under nominal situation
Collection.
Specifically, in the case of EMU Pneumatic brake systems fault-free, braking overall process is run multiple times, including lead
Draw, brake application, keep gentle solution preocess.And the pressure measurement data of corresponding four checking cylinders is gathered for often section train,
Composing training data set.For example, it is assumed that training dataset X, every a line that training data is concentrated represent a sample, i.e.,
Include the row vector of four brake-cylinder pressure measured values of sampling instant.The matrix column number of training dataset X is 4, OK
Number is determined by the quantity of the sample gathered in practice.
The brake-cylinder pressure data that the method for the embodiment of the present invention is produced under normal operating conditions using EMU carry out
Modeling, these pressure datas should include a variety of operating statuses of EMU, and such as traction, braking apply, braking is kept and braking is slow
Solution preocess.These modeling datas are highly susceptible to obtaining, and modeling process is without the accurate mathematical model of checking cylinder, without failure
Pressure measurement data under situation, easy to practical engineering application.
Step S120, calculates training data and concentrates variance between the variable of each sample and determine the threshold value of variance between variable.
Specifically, the average of sample is calculated first, then with each measurement data in sample and the sample that is calculated
The poor quadratic sum of average is as variance between the variable of the sample.Still by taking the sample comprising four checking cylinder measurement data as an example,
Note TCF=[TCF (1), TCF (2), TCF (3), TCF (4)] represents that any one brake-cylinder pressure measures sample, sample TCF
In four brake-cylinder pressures average valueDefined according to expression formula (1):
In formula, TCF (i) represents measurement data corresponding with i-th of checking cylinder in sample, more generally, when each sample
In when including n checking cylinder, averageAs shown in expression formula (2):
Variance var is defined according to expression formula (3) between the variable of the sample:
When including four checking cylinders in each sample, n takes 4.
Variance between the variable of each sample is obtained in training dataset X according to above-mentioned expression formula successively, by variance between variable
Maximum be determined as the threshold value of variance between variable, can be represented with following expression (4):
In formula, δ2The threshold value of variance, var between variablekVariance between the variable of any sample of expression training data concentration, k tables
Show sample number.
The degree of scatter measured using each brake-cylinder pressure of each sampling instant of this Measure Indexes measurement of variance between variable,
Consequently facilitating the failure that some checking cylinder or place pipeline occur in real time is monitored.
Step S130, gathers the pressure measurement data at each checking cylinder current time as test sample in real time, and utilizes institute
The threshold decision of the variance test sample is stated between variable whether comprising faulty.
To monitor the whether faulty generation of EMU, under EMU real time execution operating mode, each system at current time is gathered
The pressure measurement data of dynamic cylinder is used as test sample, and the threshold value of variance carries out test sample former between calculated variable
The step of barrier detection, specifically includes, and variance between the variable of the test sample, and the variance between variable are calculated using preceding method
Threshold value is compared, and when variance is more than the threshold value of variance between variable between the variable of the test sample, judges the test sample bag
Containing faulty.When variance is less than or equal to the threshold value of variance between variable between the variable of the test sample, the test sample is judged not
Include failure.
Specifically, the pressure measurement data for remembering the EMU checking cylinder at gathered current time is TCFt=[TCFt(1),
TCFt(2),TCFt(3),TCFt(4)], TCF is calculated according to preceding methodtVariable between variance vartProcess it is as follows:
In formula,Represent test sample TCFtComprising four brake-cylinder pressures be averaged
Value.If vartMore than the threshold value δ of variance between variable2, then it is assumed that the test sample includes faulty.If vartIt is less than or equal to
δ2, then it is assumed that the test sample does not include failure.
Step S140, when being judged as including in the event of failure, asks for each brake-cylinder pressure in the test sample and surveys successively
Measure the reconstruct contribution margin of data, and the checking cylinder that the checking cylinder with maximum reconstructed contribution margin is determined as breaking down.
The purpose of the step is, if detecting test sample TCFtIn comprising faulty, then in test sample TCFtComprising
Four brake-cylinder pressure measurement data in determine to be most likely to occur the checking cylinder of failure, and then realize the separation of failure.
Test sample is reconstructed first.Specifically, with other in test sample in addition to some measurement data
The average of measurement data replaces measurement data composition reconstructed sample.For example, TCFt(i) it is test sample TCFtIn the i-th (i
=1,2,3,4) measurement data corresponding to a checking cylinder, when the reconstructed sample of first checking cylinder of calculating, first according to expression
Formula (6) calculates the average of other three measurement data in addition to the measurement data corresponding to first checking cylinder:
TCFt(1) '=(TCFt(2)+TCFt(3)+TCFt(4))/3 (6)
Substitute position of the measurement data in former test sample corresponding to first checking cylinder with the average being calculated again
Put, the reconstructed sample on first measurement data of formation is
According to the method described above, each measurement data in test sample is reconstructed respectively.
Next, ask for the reconstructed sample for corresponding to each checking cylinder in test sample successively according to expression formula (3) or (5)
Variable between variance.Remember i-th of reconstructed sampleVariable between variance beIt is recognised thatNo more than originally not
The test sample TCF of reconstructtVariable between variance vart。
The difference of variance between the variable of variance and each reconstructed sample between the variable of former test sample is defined as checking cylinder pressure
The reconstruct contribution margin of power measurement data, as shown in expression formula (7):
In formula, ConiRepresent test sample TCFtThe reconstruct contribution margin of middle ith measurement data.
Then, the reconstruct contribution margin of all four measurement data is asked for based on expression formula (7), and there will be maximum reconstructed tribute
The checking cylinder for offering value is determined as the checking cylinder to break down.
Test sample, which is reconstructed, actually makes sample after reconstructVariable between varianceReach minimum value.
This reconstruction strategy is more advantageous to fault reconstruction.Assume that in fact that it is that i-th of checking cylinder in test sample is faulty, then test
I-th of component in sample will be more with the deviation of its excess-three component, so that test sample is between larger variable
Variance, i.e. vartIt is larger.I-th of reconstructed sampleI-th of component by its excess-three normal component average replace,
SoVariable between varianceJust become very little, understand ith measurement data by with very big according to expression formula (7)
Reconstruct contribution margin.If not i-th of checking cylinder is reconstructed, variance still has larger take between the variable of reconstructed sample at this time
Value, so that corresponding contribution margin is not very big.Therefore, this strategy based on reconstruct usually has in the embodiment of the present invention
Preferable fault reconstruction effect.
Further, detection and separation algorithm are fairly simple used by the method for the embodiment of the present invention, are simply related to
The computing that several variables are averaged with variance, easy to be realized with computer.Meanwhile this method often saves EMU the four of train
A checking cylinder real time and on line monitoring, inflation/exhaust caused by some checking cylinder due to reasons such as ageing of performance (correspond to braking
Application/release of brake process) slow failure, and the principal fault that checking cylinder or place pipeline occur is more sensitive, can and
When detect failure and position the checking cylinder that breaks down.And four brakings that on-line monitoring strategy collects each moment
Cylinder pressure measurement data is handled and analyzed, and real-time performance is good, and failure can be detected and be separated in time, instruction
The checking cylinder or place pipeline being abnormal, as the prediction of catastrophe failure, provide necessarily for the on-call maintenance of component with replacing
With reference to so as to be effectively prevented from the generation of even more serious failure.
Each step of the embodiment of the present invention is illustrated below by an example.This example is based on car Qingdao in China
The Braking System for Multiple Units united test platform of Co., Ltd of cubic vehicle research institute, for the training dataset modeled and is used for
The test data of displaying fault detect and separating effect is all from the actual operating data of testing stand.In EMU normal operation feelings
Under condition, the brake-cylinder pressure data composing training data set of testing stand monitoring software record.Artificially apply some on testing stand
Checking cylinder inflation/slow failure of exhaust and checking cylinder principal fault, corresponding checking cylinder is recorded using testing stand monitoring software
Pressure data, forms the test data for including failure.Detailed process is as follows:
(1) produce and collect training data
Braking System for Multiple Units united test platform based on Chinese Zhong Che Qingdao SiFang Vehicle Research Institute Co., Ltd, in sky
Under air braking system fault-free situation, operation braking overall process (draws, brakes application, braking is kept, release of brake), and profit
Corresponding brake-cylinder pressure measurement data is recorded with testing stand monitoring of software.Without loss of generality, during retardation test, with normal
Exemplified by 7 grades of brakings under braking mode, the checking cylinder measurement data of No. 1 car is gathered, runs 30 braking overall processes altogether, and
Record corresponding data, composing training data set.
(2) the threshold value δ of variance between variable is determined2
For above-mentioned training dataset, variance between the variable of each sample is calculated using expression formula (3).Obtain all samples
Variable between after covariance information, the threshold value δ of variance between variable is determined based on expression formula (4)2.In this example, threshold value δ2For
8.25。
(3) produce and collect the test data comprising failure, carry out fault detect
In this example, without loss of generality, two kinds of checking cylinder failures of arteface, are respectively 1# checking cylinders inflation/exhaust
The principal fault of pipeline where slow failure and 3# checking cylinders.
First, the measurement data for including 1# checking cylinders inflation/slow failure of exhaust, the number are produced and collected using testing stand
150 samples are shared according to collection, contain traction, braking applies and braking keeps three phases.The test data set is included
Each test sample, variance between its variable is calculated using expression formula (5), and with the threshold value δ in (2) step2=8.25 carry out
Compare, judge whether faulty generation.
Fig. 2 illustrates covariance information and threshold value between the variables of this 150 test samples, and wherein threshold value 8.25 is shown by dotted line
Go out.In fig. 2, in order to more clearly show the less sample information of variance value between variable, we by ordinate (between variable side
Difference) value shows and is restricted to 100.As seen in Figure 2, in the 36th sample between the 57th sample, variance between variable
Value exceeds threshold value, there is abnormal generation.Variance value is zero between the variable of preceding 34 samples, and EMU is in traction shape at this time
State, the pressure of four checking cylinders is zero.Since the 58th sample, variance value is returned within threshold value between variable, four systems
Dynamic cylinder pressure recovery is consistent.Knowable to being analyzed with reference to certain priori Fig. 2, (the 36th sample is to the 57th for failure
Between sample) occur, in the braking application stage, to indicate that some checking cylinder and other checking cylinder pressurizations are inconsistent, i.e. inflation is slow
Slow failure.When the pressure of four checking cylinders reaches near setting value (the 58th sample starts), system recovers normal, measurement
Index (variance between variable) is returned within threshold value.
Then, the measurement data set for including 3# checking cylinder gas pressure leak failures, the data are produced and collected using testing stand
Collection shares 200 samples, contains traction, braking applies and braking keeps three phases.Wherein, failure is added keeps in braking
Stage, i.e. four brake-cylinder pressures reach after stable state sometime.Each test sample included to the test data set,
Calculate variance between its variable using expression formula (5), and with the threshold value δ in (2) step2=8.25 are compared, and determine whether
Failure occurs.
Fig. 3 illustrates covariance information and threshold value between the variables of this 200 test samples, and wherein threshold value 8.25 is drawn by dotted line
Go out.In figure 3, preceding 29 samples are in traction state, and four brake-cylinder pressures are zero, thus variance value between its variable
It is zero.In the previous stage (the 30th sample to the 104th sample) that braking applies and braking is kept, covariance information between variable
Without departing from threshold value, system worked well.The latter half (the 105th sample to the 200th sample) of holding process is being braked,
Variance value exceeds threshold value between the variable of test sample, there is abnormal generation.At this time, some brake-cylinder pressure and other checking cylinder pressures
Power is inconsistent, may indicate that there occurs gas pressure leak failure for a certain checking cylinder or place pipeline.
(4) fault reconstruction after fault detect
Four checking cylinders that the method for the present invention saves EMU one on train carry out real-time combined monitoring, are detecting failure
Afterwards, it is also necessary to determine it is which checking cylinder variable may have occurred exception, that is, complete fault reconstruction.Event is included for each
The test sample of barrier, the reconstruct contribution margin of each measurement data in the test sample is calculated using expression formula (7).And then according to this
Test sample determines the specific checking cylinder to break down, realizes fault reconstruction.
For the measurement data set that slow failure is inflated comprising 1# checking cylinders, failure is determined using above-mentioned fault reconstruction strategy
Failure variable in sample (the 36th sample to the 57th sample), the results are shown in Figure 4 for fault reconstruction.As can be seen that the event
Barrier separation strategy is correctly by fault location to 1# checking cylinders.For the measurement data for including 3# checking cylinder gas pressure leak failures
Collection, the failure variable in fault sample (the 105th sample to the 200th sample) is determined using above-mentioned fault reconstruction strategy, therefore
It is as shown in Figure 5 to hinder separating resulting.As can be seen that the fault reconstruction strategy is correctly by fault location to 3# checking cylinders.Further
Prove, the fault detect of the embodiment of the present invention and separation method are simple and effective.
Fig. 6 is the structure diagram of EMU checking cylinder fault monitoring system according to another embodiment of the present invention, below
Illustrate that each several part of the fault monitoring system forms with reference to figure 6.
As shown in fig. 6, the system includes data acquisition module 61, modeling module 62, detection module 63 and separation module 64.
Wherein, training module 61 performs the operation of step S110, and modeling module 62 performs the operation of step S120, and detection module 63 performs
The operation of step S130, separation module 64 perform the operation of step S140.No longer it is developed in details herein.
Although disclosed herein embodiment as above, the content only to facilitate understand the present invention and adopt
Embodiment, is not limited to the present invention.Any those skilled in the art to which this invention pertains, are not departing from this
On the premise of the disclosed spirit and scope of invention, any modification and change can be made in the implementing form and in details,
But the scope of patent protection of the present invention, still should be subject to the scope of the claims as defined in the appended claims.
Claims (2)
1. a kind of EMU checking cylinder fault monitoring method, including:
Obtain each brake-cylinder pressure measurement data composing training data set for including braking procedure under nominal situation;
Training data is calculated to concentrate variance between the variable of each sample and determine the threshold value of variance between variable;
The pressure measurement data at each checking cylinder current time is gathered in real time as test sample, and utilizes variance between the variable
Whether the threshold decision test sample is comprising faulty;
When being judged as including in the event of failure, the reconstruct tribute of each brake-cylinder pressure measurement data in the test sample is asked for successively
Offer value, and the checking cylinder that the checking cylinder with maximum reconstructed contribution margin is determined as breaking down;
Include in the step of threshold value of variance between variance and definite variable between calculating the variable of each sample of training data concentration:
The average of sample is calculated, the poor quadratic sum of each measurement data and the average is as between the variable of the sample using in sample
Variance;With specific reference to variance var between the following expression calculating variable:
<mrow>
<mi>var</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<mi>T</mi>
<mi>C</mi>
<mi>F</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<mover>
<mrow>
<mi>T</mi>
<mi>C</mi>
<mi>F</mi>
</mrow>
<mo>&OverBar;</mo>
</mover>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
Wherein, TCF (i) represents measurement data corresponding with i-th of checking cylinder in sample,Represent the average of sample, n is represented
The number of checking cylinder included in sample;
Variance between the variable of each sample of acquisition training data concentration successively, the maximum of variance between the variable is determined as becoming
The threshold value of variance between amount;
Whether include in the threshold decision using variance between the variable test sample comprising faulty step, calculating should
Variance between the variable of test sample, and between the variable compared with the threshold value of variance,
When variance is more than the threshold value of variance between the variable between the variable of the test sample, judge that the test sample includes event
Barrier;
When variance is less than or equal to the threshold value of variance between the variable between the variable of the test sample, judge that the test sample is not wrapped
Containing failure;
Include in the step of reconstruct contribution margin of each brake-cylinder pressure measurement data in asking for the test sample successively:
The measurement data is replaced with the average of other measurement data in test sample in addition to some measurement data and forms weight
Structure sample, is asked in the test sample corresponding to variance between the variable of the reconstructed sample of each checking cylinder successively;
Using the difference of variance between the variable of variance and each reconstructed sample between the variable of the test sample as each in test sample
The reconstruct contribution margin of brake-cylinder pressure measurement data.
2. a kind of EMU checking cylinder fault monitoring system, including:
Data acquisition module, it obtains each brake-cylinder pressure measurement data composing training number for including braking procedure under nominal situation
According to collection;
Modeling module, it calculates training data and concentrates variance between the variable of each sample and determine the threshold value of variance between variable;Tool
Body includes:
The average of sample is calculated, the poor quadratic sum of each measurement data and the average is as between the variable of the sample using in sample
Variance;With specific reference to variance var between the following expression calculating variable:
<mrow>
<mi>var</mi>
<mo>=</mo>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>i</mi>
<mo>=</mo>
<mn>1</mn>
</mrow>
<mi>n</mi>
</munderover>
<msup>
<mrow>
<mo>(</mo>
<mi>T</mi>
<mi>C</mi>
<mi>F</mi>
<mo>(</mo>
<mi>i</mi>
<mo>)</mo>
<mo>-</mo>
<mover>
<mrow>
<mi>T</mi>
<mi>C</mi>
<mi>F</mi>
</mrow>
<mo>&OverBar;</mo>
</mover>
<mo>)</mo>
</mrow>
<mn>2</mn>
</msup>
</mrow>
Wherein, TCF (i) represents measurement data corresponding with i-th of checking cylinder in sample,Represent the average of sample, n is represented
The number of checking cylinder included in sample;
Variance between the variable of each sample of acquisition training data concentration successively, the maximum of variance between the variable is determined as becoming
The threshold value of variance between amount;
Detection module, it gathers the pressure measurement data at each checking cylinder current time as test sample in real time, and described in utilization
Whether the threshold decision of the variance test sample is comprising faulty between variable;Specifically include:
Variance between the variable of the test sample is calculated, and between the variable compared with the threshold value of variance,
When variance is more than the threshold value of variance between the variable between the variable of the test sample, judge that the test sample includes event
Barrier;
When variance is less than or equal to the threshold value of variance between the variable between the variable of the test sample, judge that the test sample is not wrapped
Containing failure;
Separation module, it asks for each brake-cylinder pressure measurement in the test sample successively when being judged as including in the event of failure
The reconstruct contribution margin of data, and the checking cylinder that the checking cylinder with maximum reconstructed contribution margin is determined as breaking down;Its basis
Following steps ask for the reconstruct contribution margin of each brake-cylinder pressure measurement data in the test sample successively:
The measurement data is replaced with the average of other measurement data in test sample in addition to some measurement data and forms weight
Structure sample, is asked in the test sample corresponding to variance between the variable of the reconstructed sample of each checking cylinder successively;
Using the difference of variance between the variable of variance and each reconstructed sample between the variable of the test sample as each in test sample
The reconstruct contribution margin of brake-cylinder pressure measurement data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610065389.7A CN105730431B (en) | 2016-01-29 | 2016-01-29 | EMU checking cylinder fault monitoring method and fault monitoring system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610065389.7A CN105730431B (en) | 2016-01-29 | 2016-01-29 | EMU checking cylinder fault monitoring method and fault monitoring system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105730431A CN105730431A (en) | 2016-07-06 |
CN105730431B true CN105730431B (en) | 2018-04-20 |
Family
ID=56247122
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610065389.7A Active CN105730431B (en) | 2016-01-29 | 2016-01-29 | EMU checking cylinder fault monitoring method and fault monitoring system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105730431B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106525466B (en) * | 2016-10-14 | 2019-02-22 | 清华大学 | A kind of Braking System for Multiple Units critical component robust filtering method and system |
CN110293949B (en) * | 2019-06-06 | 2021-09-24 | 山东科技大学 | Method for detecting tiny faults of air brake system of high-speed train |
CN110843755B (en) * | 2019-11-19 | 2021-09-28 | 奇瑞汽车股份有限公司 | Method and equipment for estimating braking pressure of electric automobile |
CN111114519B (en) * | 2020-01-03 | 2022-04-08 | 中国铁路郑州局集团有限公司科学技术研究所 | Railway vehicle brake fault prediction method and health management system |
CN112061093B (en) * | 2020-09-21 | 2021-07-20 | 中车青岛四方车辆研究所有限公司 | Locomotive braking system for realizing automatic driving and realization method thereof |
CN112208506B (en) * | 2020-09-23 | 2021-07-06 | 西南交通大学 | Intelligent fault detection method for air brake device of railway wagon |
CN112345918B (en) * | 2020-10-30 | 2023-02-24 | 中车青岛四方机车车辆股份有限公司 | Multi-contact circuit state detection device, system, method, loop and rail vehicle |
CN113511183B (en) * | 2021-07-15 | 2022-05-17 | 山东科技大学 | Optimization criterion-based early fault separation method for air brake system of high-speed train |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102662321A (en) * | 2012-03-23 | 2012-09-12 | 清华大学 | Online updating method of principal component analysis monitoring model |
CN103776480A (en) * | 2014-01-29 | 2014-05-07 | 清华大学 | Small-fault detection method and device based on multiple moving average |
CN104182642A (en) * | 2014-08-28 | 2014-12-03 | 清华大学 | Sparse representation based fault detection method |
CN104316729A (en) * | 2014-11-13 | 2015-01-28 | 成都运达科技股份有限公司 | Self-diagnosis method of acceleration sensors for locomotive bogie detection |
CN104864985A (en) * | 2015-05-21 | 2015-08-26 | 南车株洲电力机车研究所有限公司 | Train axle temperature sensor fault detection method and apparatus |
CN105204496A (en) * | 2015-09-25 | 2015-12-30 | 清华大学 | Sensor fault diagnosing method and system for air braking control system of motor train unit |
-
2016
- 2016-01-29 CN CN201610065389.7A patent/CN105730431B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102662321A (en) * | 2012-03-23 | 2012-09-12 | 清华大学 | Online updating method of principal component analysis monitoring model |
CN103776480A (en) * | 2014-01-29 | 2014-05-07 | 清华大学 | Small-fault detection method and device based on multiple moving average |
CN104182642A (en) * | 2014-08-28 | 2014-12-03 | 清华大学 | Sparse representation based fault detection method |
CN104316729A (en) * | 2014-11-13 | 2015-01-28 | 成都运达科技股份有限公司 | Self-diagnosis method of acceleration sensors for locomotive bogie detection |
CN104864985A (en) * | 2015-05-21 | 2015-08-26 | 南车株洲电力机车研究所有限公司 | Train axle temperature sensor fault detection method and apparatus |
CN105204496A (en) * | 2015-09-25 | 2015-12-30 | 清华大学 | Sensor fault diagnosing method and system for air braking control system of motor train unit |
Also Published As
Publication number | Publication date |
---|---|
CN105730431A (en) | 2016-07-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105730431B (en) | EMU checking cylinder fault monitoring method and fault monitoring system | |
CN105204496B (en) | The method and system of EMUs air brake control system sensor fault diagnosis | |
CN110293949B (en) | Method for detecting tiny faults of air brake system of high-speed train | |
DK2288530T3 (en) | Process for monitoring at least one system parameter affecting the operating behavior of vehicles or vehicles | |
KR101829645B1 (en) | Device and method for monitoring railway vehicle and method for monitoring using the same | |
CA2620188C (en) | Landing load monitor for aircraft landing gear | |
EP0391047B1 (en) | Brake-testing and -monitoring procedure in railway cars, and arrangement for executing the procedure | |
US8260486B2 (en) | Method and apparatus for diagnosis of sensors faults using adaptive fuzzy logic | |
US10005566B2 (en) | Systems and methods for detection of dragging brake | |
CA2847219C (en) | Brake control device for a brake system, brake system, rail vehicle and method for operating a brake system | |
CN109466534B (en) | Train braking fault judgment method | |
CN111071291B (en) | Train wheel set monitoring system and train wheel set monitoring method | |
CN111238725A (en) | Fault diagnosis method, device and system for pressure sensor of electro-pneumatic brake system | |
CN109435936B (en) | Dual-channel data transmission train braking monitoring system | |
Green | A comparison of stopping distance performance for motorcycles equipped with ABS, CBS and conventional hydraulic brake systems | |
CN113390654A (en) | Train braking fault diagnosis method based on multi-source information fusion | |
CN107139915B (en) | System for testing air pressure anti-lock performance of vehicle | |
Senabre et al. | Comparative analysis of vehicle brake data in the ministry of transport test on the roller brake tester and on flat ground | |
Ji et al. | Fault detection of EMU brake cylinder | |
Du et al. | Fault Detection and Isolation for Brake Rotor Thickness Variation | |
CN113511183B (en) | Optimization criterion-based early fault separation method for air brake system of high-speed train | |
CN207280740U (en) | A kind of car traffic track brake beam forms product product pull force tester | |
Kim et al. | Measurement of the disc-pad friction coefficient for mechanical brakes using direct and indirect methods | |
CN113557198A (en) | Aircraft brake control system | |
Duan et al. | Study on fault-tolerant filter algorithm for integrated navigation system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |