CN104316729B - Self-diagnosis method of acceleration sensors for locomotive bogie detection - Google Patents
Self-diagnosis method of acceleration sensors for locomotive bogie detection Download PDFInfo
- Publication number
- CN104316729B CN104316729B CN201410635543.0A CN201410635543A CN104316729B CN 104316729 B CN104316729 B CN 104316729B CN 201410635543 A CN201410635543 A CN 201410635543A CN 104316729 B CN104316729 B CN 104316729B
- Authority
- CN
- China
- Prior art keywords
- measuring point
- acceleration transducer
- vibration
- average
- detection
- 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
Landscapes
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a self-diagnosis method of acceleration sensors for locomotive bogie detection. The method comprises the following sequential steps: firstly, dividing bogie detection points into a plurality of monitoring arrays, wherein each monitoring array detects one type of bearings correspondingly, and each detection point in each monitoring array corresponds to one bearing needing to be detected; secondly, installing one acceleration sensor on each bearing needing to be detected; thirdly, establishing RBF neural network predictors of the acceleration sensors of K detection points in each monitoring array respectively; fourthly, setting the mean value threshold value and the variance threshold value of shock signals of the acceleration sensors during the open-circuit fault, the mean value threshold value and the variance threshold value of the shock signals of the acceleration sensors during the short-circuit fault, and the distance error threshold value of the acceleration sensors; fifthly, completing routing inspection of the multiple monitoring arrays and carrying out fault diagnosis when a locomotive bogie works. By means of the method, the short-circuit fault, the open-circuit fault and other faults of the acceleration sensors can be quickly detected, and the method is of great significance on the safety of locomotive bogie detection.
Description
Technical field
The present invention relates to fault diagnosis technology, the specifically self diagnosis of rolling stock bogie detection acceleration transducer
Method.
Background technology
Bogie, as the important component part of rolling stock, plays the important function such as carrying, traction, traveling and braking,
It is to determine that rolling stock runs safety and the most key part of dynamic performance.The bearing of rolling stock bogie is general
For rolling bearing, it plays a part to bear load and transmission load.Under the operation condition of high-speed overload, rolling stock turns to
The bearing of frame once breaks down, and is easily caused hot axle, the generation fired axle, cut the accidents such as axle, therefore, in order to ensure rolling stock peace
Full reliability service, carries out detecting particularly important to the bearing of bogie.
The fault detect commonly used vibratory impulse detection method of rolling stock bogie bearing, the tool of vibratory impulse detection method
Body operating process is: installs multiple detection arrays on bogie, each detection array comprises to be arranged on bogie symmetric position
Multiple acceleration transducers of bearing, carry out the fault of bogie bearing by the vibration and shock signal that acceleration transducer gathers
Detection.The source that acceleration transducer obtains as impact information, is that rolling stock bogie bearing fault detection is indispensable
Part, the accuracy of its measurement data has vital impact to testing result.Due to rolling stock bogie bearing
Working environment is complicated, severe, is also at complexity, in rugged environment when acceleration transducer is mounted thereto, therefore, realize plus
The fault detect of velocity sensor oneself state is particularly significant.
On current rolling stock bogie bearing, the online fault detection technique of acceleration frame sensor is concentrated mainly on and adds
The detection of velocity sensor failure fault, detection technique mainly taps acceleration transducer using artificial or particular tool, and root
According to the comparing difference and to realize of surge waveform value and ideal value of acceleration transducer output, real-time and accuracy are poor.Comprehensive
Upper described, the limitation of fault detection type of acceleration transducer and detection technique on existing rolling stock bogie bearing
Defect, makes the working condition of acceleration transducer cannot get in time detection, has had a strong impact on the reality of bogie bearing fault detection
When property and accuracy are it is impossible to guarantee the traffic safety of locomotive.
Content of the invention
It is an object of the invention to overcoming the deficiencies in the prior art, there is provided a kind of rolling stock bogie detection accelerates
The self-diagnosing method of degree sensor, it fault to acceleration transducer promptly and accurately can be identified the row it is ensured that locomotive
Car safety.
The present invention solves the above problems and is achieved through the following technical solutions: rolling stock bogie detection acceleration
The self-diagnosing method of sensor, comprises the following steps:
Step one, according to the feature of same type bearing symmetry distribution detected on rolling stock bogie, bogie is examined
Survey measuring point and be divided into multiple monitoring arrays, the quantity phase of quantity required detection bearing type with same bogie of monitoring array
With, each monitoring array corresponding detection one class bearing, each monitoring array includes k and assumes symmetrical measuring point, each survey
Bearing is detected, wherein, k is the even number more than or equal to 4 needed for corresponding one of point;
On step 2, respectively detection bearing needed for each, an acceleration transducer is installed;
Step 3, set up the respective rbf neutral net of k measuring point acceleration transducer in each monitoring array respectively offline
Fallout predictor;
Vibration when vibration and shock signal average threshold value, short trouble when step 4, setting acceleration transducer short trouble
Vibration and shock signal variance threshold when vibration and shock signal average threshold value, open fault when impact signal variance threshold values, open fault
Value and the range error threshold value of acceleration transducer;
Step 5, when rolling stock bogie works, realize multiple monitoring the patrolling and examining and carry out acceleration sensing of array
Device fault diagnosis.
Further, set up each in described step 3 offline and monitor the respective rbf of k measuring point acceleration transducer of array
The concrete operation step of neural network prediction device is as follows: obtains the training sample of n each measuring point acceleration transducer of moment, root
According to training sample, adjust weight matrix according to gradient descent method, after reaching setting accuracy, represent the rbf nerve of each measuring point
The pre- then device of network is set up offline and is completed;Wherein, the training sample in each measuring point a certain moment includes this measuring point corresponding moment
Input data and output data, the acquisition of training sample is l by taking the length of this measuring point remaining measuring point synchronization outer sampling
Vibration and shock signal average as the input data of the rbf neural network prediction device of this measuring point, and take this measuring point with for the moment
Carve the output data as the rbf neural network prediction device of this measuring point for the average of the vibration and shock signal that the length gathering is l.
The present invention is directed to the data in a time window, and its average can effectively reflect the amplitude of signal, and variance can be anti-
Reflect the stability of signal, based on its output signal during sensor generation open fault close to acceleration transducer output valve
The form of expression of big value, has formulated the standard of open sensor fault diagnosis, based on sensor be short-circuited fault when its output
Signal, close to 0, has formulated the standard of short circuit sensor fault diagnosis.Particularly as follows: acceleration transducer leaves in described step 4
The threshold value of road fault and short trouble sets in the following ways: vibration and shock signal average during acceleration transducer open fault
Threshold value is, during acceleration transducer short trouble, vibration and shock signal average threshold value is, acceleration transducer
When open circuit or short trouble, vibration and shock signal variance threshold values are, wherein,For the output maximum of acceleration transducer,,.
Further, described rolling stock bogie detection measuring point is divided into three monitoring arrays, and three monitoring arrays are respectively
Monitor array and seize position monitoring array for axle box position monitoring array, motor position, each monitoring array comprises symmetrical on bogie
4 similar detection bearings of distribution, corresponding 4 measuring points.
Further, the bearing of described rolling stock bogie is provided with speed probe, described acceleration transducer
After speed probe output pulse signal, synchronous acquisition monitors the vibration and shock signal of k measuring point in array.
Further, following steps are specifically included during the detection realizing a certain monitoring array in described step 5:
Step 5.1, the synchronous vibration and shock signal obtaining the corresponding k acceleration transducer generation of this monitoring array, and
Calculate average and the variance of this k measuring point vibration and shock signal of monitoring array;
Step 5.2, the average of k measuring point and variance are contrasted with the average threshold value setting and variance threshold values, and root
Judge that acceleration transducer whether there is open fault or short trouble according to comparing result, if there is open fault or short circuit event
Barrier, sends acceleration transducer open circuit or short trouble is reported to the police, and enters next and monitors array detection, otherwise enters next step and enters
This monitoring array acceleration transducer other fault diagnosis of row;
The distance of step 5.3, k measuring point vibration and shock signal mean-max of calculating and minimum of a value, if this distance is more than
The range error threshold value of the acceleration transducer setting, then enter next step, and the acceleration otherwise exiting this monitoring array passes
Sensor fault diagnosis and enter next monitoring array detection;
Step 5.4, average m of k measuring point vibration and shock signal average of calculating, and obtain and maximum the shaking of average m deviation
The measuring point a of dynamic impact signal average, and using except measuring point a remaining vibration and shock signal average as measuring point a rbf neutral net
The estimate of the input value On-line Estimation measuring point a vibration and shock signal average of fallout predictor, if measuring point a vibration and shock signal average
The range error threshold value that error between estimate and actual average is more than the acceleration transducer setting then enters next step,
Otherwise exit the acceleration transducer fault diagnosis of this monitoring array and enter the detection of next monitoring array;
Step 5.5, the multi-group data of k measuring point of synchronous acquisition, and calculate Estimation of Mean value and the reality of measuring point a multi-group data
The mean error of border average, if mean error is more than the range error threshold value of the acceleration transducer setting, the acceleration of measuring point a
There is fault in degree sensor, otherwise, the acceleration transducer detection of measuring point a is normal.
The present invention is symmetrical on same bogie based on k acceleration transducer, the vibration punching of k measuring point parallel acquisition
The impact Magnitude Difference hitting signal is maintained in certain scope, if it is too high or too low certain measuring point ballistic throw degree, can be just
Step judges that the shock transducer of this measuring point deviation or skew to the impact of fault, and then realizes the isolation of trouble point.Then
Based on locomotive with the redundancy between k sensor collection information of measuring point symmetrical on bogie, you can be sensed by k-1
The information of device provides the redundancy estimate of a remaining sensor, impacts average using k-1 measuring point of non-faulting and trains
Rbf neural network prediction device realize the estimation that average is impacted in trouble point, if the error of estimate and actual measured value exceedes by mistake
Difference limen value is then it represents that the sensor of fault measuring point exists extremely.Certainly die to exclude the non-sensor such as ambient noise further
The abnormal erroneous judgement leading to sensor fault of impact that barrier causes, takes inspection policies immediately: detect the setting time cycle immediately
The multigroup mean error estimating average and actual average of internal fault measuring point, if mean error is still greater than error threshold, sends event
The detection information of barrier measuring point sensor fault;Otherwise, there is fault in the sensor of measuring point of fixing a breakdown.
Further, judge in described step 5.2 that open fault and short trouble are realized in the following ways: first determine whether
Vibration and shock signal average threshold value when whether the average of vibration and shock signal is more than open fault, and when variance is less than open fault
Vibration and shock signal variance threshold values, if then judge that corresponding measuring point acceleration transducer has open fault, if otherwise judging certain
Whether the average of individual measuring point vibration and shock signal is less than vibration and shock signal average threshold value during short trouble, and variance is less than short circuit
Vibration and shock signal variance threshold values during fault, if then judge that corresponding measuring point acceleration transducer has short trouble.
In sum, the method have the advantages that (1) present invention can quick detection to go out acceleration transducer short
Road, open fault, can quickly realize the detection of the faults such as precise decreasing, deviation and the skew of acceleration transducer again, for
Find that the fault of rolling stock bogie acceleration transducer provides reliable and effective means, further ensures locomotive in time
Traffic safety.
(2) there is no dismounting, the superiority of frequently detection, for guaranteeing acceleration transducer detection number during present invention application
According to correctness, particularly to rolling stock bogie monitoring security there is important function.
Brief description
Fig. 1 is the diagnostic flow chart of the specific embodiment of one monitoring array of the present invention;
Fig. 2 is acceleration transducer short circuit, open fault diagnostic flow chart in Fig. 1;
Fig. 3 is non-shorting, the non-open fault diagnostic flow chart of acceleration transducer in Fig. 1.
Specific embodiment
With reference to embodiment and accompanying drawing, the present invention is done with detailed description further, but embodiments of the present invention
Not limited to this.
Embodiment:
The self-diagnosing method of rolling stock bogie detection acceleration transducer, the following steps including carrying out successively:
Step one, according to the feature of same type bearing symmetry distribution detected on rolling stock bogie, bogie detection measuring point is divided
Become multiple monitoring arrays, the quantity of quantity detection required with the same bogie bearing type of monitoring array is identical, each prison
Survey array corresponding detection one class bearing, each monitoring array includes k and assumes symmetrical measuring point, corresponding one of each measuring point
Required detection bearing, wherein, k is the even number more than or equal to 4;One is installed on step 2, respectively detection bearing needed for each
Individual acceleration transducer;Step 3, set up the respective rbf of k measuring point acceleration transducer in each monitoring array respectively offline
Neural network prediction device;Vibration and shock signal average threshold value, short trouble when step 4, setting acceleration transducer short trouble
When vibration and shock signal variance threshold values, open fault when vibration and shock signal average threshold value, open fault when vibration and shock signal
The range error threshold value of variance threshold values and acceleration transducer;Step 5, when rolling stock bogie works, realize many
Individual monitoring the patrolling and examining and carry out acceleration transducer fault diagnosis of array.The quantity monitoring array in the present embodiment is three, three
Individual monitoring array is respectively axle box position monitoring array, motor position monitoring array and seizes position monitoring array, and each monitors array bag
Containing symmetrical 4 similar detection bearings on bogie, that is, each monitoring array includes 4 and assumes symmetrical measuring point.
One speed probe is provided with the bearing of rolling stock bogie in the present embodiment, the acceleration sensing of each monitoring array
Device synchronous acquisition after speed probe output pulse signal monitors the vibration and shock signal of 4 measuring points in array.This enforcement
During example application, three monitoring arrays are numbered in order, circulation during detection is carried out.
The present embodiment also needs to set up four surveys in each monitoring array offline before the vibration and shock signal gathering each measuring point
Point acceleration transducer respective rbf neural network prediction device, wherein, each measuring point rbf neutral net in each monitoring array
The process of setting up of fallout predictor is: obtains the training sample of n each measuring point acceleration transducer of moment, according to training sample, presses
Adjust weight matrix according to gradient descent method, after reaching setting accuracy, represent the pre- then device of rbf neutral net of each measuring point from
Line is set up and is completed.Wherein, the training sample in each measuring point a certain moment includes input data and the output in this measuring point corresponding moment
Data, the acquisition of training sample is by the vibration and shock signal taking the length that this measuring point remaining measuring point synchronization outer is sampled to be l
Average as the input data of the rbf neural network prediction device of this measuring point, and take this measuring point synchronization to play the length of collection
The average of the vibration and shock signal for l is as the output data of the rbf neural network prediction device of this measuring point.Wherein, the present embodiment
Middle l value is 200.The concrete operation step of four rbf neural network prediction device foundation is as follows: four acceleration transducers are divided
Not numbering is 1,2,3 and 4, and the corresponding measuring point of four acceleration transducers is respectively measuring point 1,2,3 and 4, and synchronized sampling is from the moment
t1When rising, length during four acceleration transducer normal works is the vibration and shock signal of l, takes the vibratory impulse of measuring point 1,2,3
The average of signal is as the input of the rbf neural network prediction device of measuring point 4, the average conduct of the vibration and shock signal of measuring point 4
The output of the rbf neural network prediction device of measuring point 4, one group of training sample of rbf neural network prediction device of composition measuring point 4;Take survey
The average of the vibration and shock signal of point 2,3,4 is as the input of the rbf neural network prediction device of measuring point 1, the vibration punching of measuring point 1
Hit the output as the rbf neural network prediction device of measuring point 1 for the average of signal, the rbf neural network prediction device one of composition measuring point 1
Group training sample;Take measuring point 1,3,4 vibration and shock signal average as the rbf neural network prediction device of measuring point 2 input,
The average of the vibration and shock signal of measuring point 2, as the output of the rbf neural network prediction device of measuring point 2, forms the rbf god of measuring point 2
Through one group of training sample of neural network forecast device;Take measuring point 1,2,4 vibration and shock signal average as measuring point 3 rbf nerve net
The input of network fallout predictor, the average of the vibration and shock signal of measuring point 3 as the output of the rbf neural network prediction device of measuring point 3,
One group of training sample of rbf neural network prediction device of composition measuring point 3;The like until obtained n group training sample.Last root
According to n group training sample off-line training rbf neural network prediction device.
The present embodiment also needs to preset acceleration transducer short trouble before the vibration and shock signal gathering each measuring point
When vibration and shock signal average threshold value, short trouble when vibration and shock signal variance threshold values, open fault when vibration and shock signal
The range error threshold value of vibration and shock signal variance threshold values and acceleration transducer when average threshold value, open fault.Acceleration
The threshold value of open sensor fault and short trouble sets in the following ways: vibratory impulse during acceleration transducer open fault
Signal average threshold value is, during acceleration transducer short trouble, vibration and shock signal average threshold value is, accelerate
When degree open sensor or short trouble, vibration and shock signal variance threshold values are, wherein,Output for acceleration transducer
Maximum,,.
As shown in figure 1, the present embodiment specifically includes following steps when realizing the detection of a certain monitoring array: step one,
Gather the vibration and shock signal of each measuring point this monitoring array Nei and carry out average and variance calculating;Step 2, diagnose each and add
Velocity sensor whether there is open fault or short trouble, if there is open fault or short trouble, sends open sensor
Or short trouble warning, enter next and monitor array detection, if no open fault or short trouble, enter next step;Step
Rapid three, diagnose each acceleration transducer and whether there is other faults.
As shown in Fig. 2 the present embodiment diagnose certain monitoring array in each acceleration transducer whether there is open fault or
Short trouble comprises the following steps: step a.1, judge that whether the average of each measuring point is more than open fault average threshold value and variance
Less than open fault variance threshold values, if so, then show the acceleration transducer open fault of corresponding measuring point and exit fault diagnosis,
If otherwise entering next step;Step a.2, judge average whether less than short trouble average threshold value and variance is less than short trouble
Variance threshold values, if then showing the acceleration transducer short trouble of corresponding measuring point and exiting fault diagnosis, if otherwise to acceleration
Degree sensor other fault diagnosis.Wherein, judge that the concrete operation step of open fault or short trouble is as follows: by four measuring points
The average of vibration and shock signal and variance are all contrasted with the average threshold value setting and variance threshold values, first determine whether vibratory impulse
Whether the average of signal is more than vibration and shock signal average threshold value during open fault, and vibratory impulse when variance is less than open fault
Signal variance threshold value, if then judging that corresponding measuring point acceleration transducer has open fault, if otherwise judging, certain measuring point shakes
Whether the average of dynamic impact signal is less than vibration and shock signal average threshold value during short trouble, and shakes when variance is less than short trouble
Dynamic impact signal variance threshold values, if then judge that corresponding measuring point acceleration transducer has short trouble.
As shown in figure 3, the present embodiment diagnoses each acceleration transducer in certain monitoring array whether there is other fault bags
Include following steps: step b.1, calculate ultimate range and average m of four measuring point averages;Step b.2, whether judge ultimate range
More than range error threshold value, if then obtaining the measuring point a maximum with average m distance and enter next step, if otherwise exit adding
Speed sensor fault diagnoses;Step b.3, using rbf neural network prediction device obtain measuring point a at acceleration transducer average
Estimate, and judge at measuring point a, whether Estimation of Mean value and the error of actual value are more than range error threshold value, if then same respectively
Five groups of vibratory impulse data of step four acceleration transducers of collection simultaneously enter next step, if otherwise exiting fault diagnosis;Step
The rapid average calculating each 5 groups of impact data of four acceleration transducers b.4, respectively, and obtained using rbf neural network prediction device
Five Estimation of Mean values of acceleration transducer at measuring point a, calculate average d of five Estimation of Mean values and actual mean value error,
Judge whether average d is more than range error threshold value, if then judging that at measuring point a, acceleration transducer has fault, if otherwise moving back
It is out of order diagnosis.Wherein, step b.1 in the ultimate range of four measuring point averages be that four measuring point vibration and shock signal averages are maximum
Value and the distance of minimum of a value, average m is the average of four measuring point vibration and shock signal averages.The step b.2 measuring point a of middle acquisition
Vibration and shock signal average is maximum with average m deviation;B.3, step is passed through to make remaining the vibration and shock signal average except measuring point a
The estimate of the input value On-line Estimation measuring point a vibration and shock signal average of the rbf neural network prediction device for measuring point a;Step
B.4 five groups of data of middle four measuring points of synchronous acquisition.
As described above, the present invention can preferably be realized.
Claims (6)
1. the self-diagnosing method of rolling stock bogie detection acceleration transducer is it is characterised in that comprise the following steps:
Step one, according to the feature of same type bearing symmetry distribution detected on rolling stock bogie, bogie is detected and survey
Point is divided into multiple monitoring arrays, and the quantity of quantity detection required with the same bogie bearing type of monitoring array is identical, often
Individual monitoring array corresponding detection one class bearing, each monitoring array includes k and assumes symmetrical measuring point, and each measuring point corresponds to
Bearing is detected, wherein, k is the even number more than or equal to 4 needed for one;
On step 2, respectively detection bearing needed for each, an acceleration transducer is installed;
Step 3, set up the respective rbf neural network prediction of k measuring point acceleration transducer in each monitoring array respectively offline
Device;
Vibratory impulse when vibration and shock signal average threshold value, short trouble when step 4, setting acceleration transducer short trouble
Vibration and shock signal variance threshold values when vibration and shock signal average threshold value, open fault when signal variance threshold value, open fault, with
And the range error threshold value of acceleration transducer;
Step 5, when rolling stock bogie works, realize multiple monitoring arrays patrol and examine and carry out acceleration transducer therefore
Barrier diagnosis;
Set up each in described step 3 offline and monitor the respective rbf neural network prediction of k measuring point acceleration transducer of array
The concrete operation step of device is as follows: obtain the training sample of n each measuring point acceleration transducer of moment, according to training sample,
Adjust weight matrix according to gradient descent method, after reaching setting accuracy, represent the pre- then device of rbf neutral net of each measuring point
Offline foundation completes;Wherein, the training sample in each measuring point a certain moment includes the input data in this measuring point corresponding moment and defeated
Go out data, the acquisition of training sample is believed by the vibratory impulse that the length taking remaining the measuring point synchronization outer sampling of this measuring point is l
Number average as the input data of the rbf neural network prediction device of this measuring point, and take this measuring point synchronization to play the length of collection
Spend the output data as the rbf neural network prediction device of this measuring point for the average of vibration and shock signal for l.
2. the self-diagnosing method of rolling stock bogie detection acceleration transducer according to claim 1, its feature
It is, in described step 4, the threshold value of acceleration transducer open fault and short trouble sets in the following ways: acceleration
During open sensor fault, vibration and shock signal average threshold value is, vibratory impulse letter during acceleration transducer short trouble
Number average threshold value is, when acceleration transducer open circuit or short trouble, vibration and shock signal variance threshold values are, its
In,For the output maximum of acceleration transducer,,.
3. the self-diagnosing method of rolling stock bogie detection acceleration transducer according to claim 1, its feature
It is, described rolling stock bogie detection measuring point is divided into three monitoring arrays, three monitoring arrays are respectively the monitoring of axle box position
Array, motor position are monitored array and are seized position monitoring array, and it is similar that each monitoring array comprises symmetrical on bogie 4
Detection bearing, corresponding 4 measuring points.
4. the self-diagnosing method of rolling stock bogie detection acceleration transducer according to claim 1, its feature
It is, the bearing of described rolling stock bogie is provided with speed probe, and described acceleration transducer is in speed probe
After output pulse signal, synchronous acquisition monitors the vibration and shock signal of k measuring point in array.
5. the autodiagnosis of the rolling stock bogie detection acceleration transducer according to any one in Claims 1 to 4
Disconnected method is it is characterised in that realize specifically including following steps during the detection of a certain monitoring array in described step 5:
Step 5.1, the synchronous vibration and shock signal obtaining the corresponding k acceleration transducer generation of this monitoring array, and calculate
The average of this k measuring point vibration and shock signal of monitoring array and variance;
Step 5.2, the average of k measuring point and variance are contrasted with the average threshold value setting and variance threshold values, and according to right
Judging that acceleration transducer whether there is open fault or short trouble than result, if there is open fault or short trouble, sending out
Go out acceleration transducer open circuit or short trouble is reported to the police, enter next and monitor array detection, otherwise enter next step and carry out this
Monitoring array acceleration transducer other fault diagnosis;
The distance of step 5.3, k measuring point vibration and shock signal mean-max of calculating and minimum of a value, if this distance is more than setting
Acceleration transducer range error threshold value, then enter next step, otherwise exit the acceleration transducer of this monitoring array
Fault diagnosis and enter next monitoring array detection;
Step 5.4, average m of k measuring point vibration and shock signal average of calculating, and obtain the vibration punching maximum with average m deviation
Hit the measuring point a of signal average, and using except measuring point a remaining vibration and shock signal average as measuring point a rbf neural network prediction
The estimate of the input value On-line Estimation measuring point a vibration and shock signal average of device, if the estimation of measuring point a vibration and shock signal average
The range error threshold value that error between value and actual average is more than the acceleration transducer setting then enters next step, otherwise
Exit the acceleration transducer fault diagnosis of this monitoring array and enter the detection of next monitoring array;
Step 5.5, the multi-group data of k measuring point of synchronous acquisition, and it is equal with reality to calculate the Estimation of Mean value of measuring point a multi-group data
The mean error of value, if mean error is more than the range error threshold value of the acceleration transducer setting, the acceleration of measuring point a passes
There is fault in sensor, otherwise, the acceleration transducer detection of measuring point a is normal.
6. the self-diagnosing method of rolling stock bogie detection acceleration transducer according to claim 5, its feature
It is, judge in described step 5.2 that open fault and short trouble are realized in the following ways: first determine whether vibration and shock signal
Average whether vibration and shock signal average threshold value during more than open fault, and vibration and shock signal when variance is less than open fault
Variance threshold values, if then judging that corresponding measuring point acceleration transducer has open fault, if otherwise judge the vibration punching of certain measuring point
Whether the average hitting signal is less than vibration and shock signal average threshold value during short trouble, and variance is less than vibration punching during short trouble
Hit signal variance threshold value, if then judging that corresponding measuring point acceleration transducer has short trouble.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410635543.0A CN104316729B (en) | 2014-11-13 | 2014-11-13 | Self-diagnosis method of acceleration sensors for locomotive bogie detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410635543.0A CN104316729B (en) | 2014-11-13 | 2014-11-13 | Self-diagnosis method of acceleration sensors for locomotive bogie detection |
Publications (2)
Publication Number | Publication Date |
---|---|
CN104316729A CN104316729A (en) | 2015-01-28 |
CN104316729B true CN104316729B (en) | 2017-01-25 |
Family
ID=52371982
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410635543.0A Active CN104316729B (en) | 2014-11-13 | 2014-11-13 | Self-diagnosis method of acceleration sensors for locomotive bogie detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104316729B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104596564B (en) * | 2015-02-04 | 2017-03-15 | 中国工程物理研究院化工材料研究所 | The system and method that sensor fault judges |
DE102016204734A1 (en) * | 2015-04-09 | 2016-10-13 | Schaeffler Technologies AG & Co. KG | Method for operating an actuator, in particular a Hydrostataktors for actuating an automated friction clutch |
CN105730431B (en) * | 2016-01-29 | 2018-04-20 | 清华大学 | EMU checking cylinder fault monitoring method and fault monitoring system |
US10204461B2 (en) * | 2016-07-19 | 2019-02-12 | GM Global Technology Operations LLC | Detection and reconstruction of sensor faults |
CN106197759B (en) * | 2016-07-22 | 2019-07-02 | 蚌埠大洋传感系统工程有限公司 | A kind of temperature sensor control system |
US10611545B2 (en) * | 2016-09-19 | 2020-04-07 | Emerson Climate Technologie—Transportation Solutions ApS | Container shock detection system |
CN109029815A (en) * | 2018-08-13 | 2018-12-18 | 成都旸谷信息技术有限公司 | Rolling stock key structure Fast Monitoring system |
CN109596860B (en) * | 2018-11-05 | 2020-12-08 | 中国电子产品可靠性与环境试验研究所((工业和信息化部电子第五研究所)(中国赛宝实验室)) | Accelerometer fault diagnosis method, device, circuit and computer equipment |
JP7215131B2 (en) * | 2018-12-12 | 2023-01-31 | 株式会社オートネットワーク技術研究所 | Determination device, determination program, determination method, and neural network model generation method |
CN109753741B (en) * | 2019-01-11 | 2023-04-07 | 中国铁路总公司 | Locomotive mechanical part performance evolution identification method based on data driving |
FR3092368B1 (en) * | 2019-02-05 | 2021-01-08 | Safran Aircraft Engines | Monitoring the health of at least two vibration sensors of a twin-barrel turbomachine |
CN111444574B (en) * | 2020-03-25 | 2023-08-22 | 成都运达科技股份有限公司 | Sensor layout optimization method based on dynamics analysis |
CN113776760B (en) * | 2020-06-09 | 2023-06-27 | 成都运达科技股份有限公司 | Train wheel set out-of-round fault monitoring method and system based on whole-axis vibration analysis |
CN113696934B (en) * | 2021-09-07 | 2023-05-23 | 广州电力机车有限公司 | Locomotive speed abnormal fault processing method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07225244A (en) * | 1994-02-10 | 1995-08-22 | Toyota Motor Corp | Detector for fault of acceleration sensor |
CN103674582A (en) * | 2013-11-29 | 2014-03-26 | 南车青岛四方机车车辆股份有限公司 | Method and device for judging movement stability of railway vehicle bogie |
CN103852271A (en) * | 2012-12-01 | 2014-06-11 | 南车青岛四方机车车辆股份有限公司 | High-speed train running gear fault diagnosis and remote monitoring system based on Internet of Things |
-
2014
- 2014-11-13 CN CN201410635543.0A patent/CN104316729B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07225244A (en) * | 1994-02-10 | 1995-08-22 | Toyota Motor Corp | Detector for fault of acceleration sensor |
CN103852271A (en) * | 2012-12-01 | 2014-06-11 | 南车青岛四方机车车辆股份有限公司 | High-speed train running gear fault diagnosis and remote monitoring system based on Internet of Things |
CN103674582A (en) * | 2013-11-29 | 2014-03-26 | 南车青岛四方机车车辆股份有限公司 | Method and device for judging movement stability of railway vehicle bogie |
Non-Patent Citations (1)
Title |
---|
基于小波包和RBF神经网络的压电加速度传感器故障诊断;杜菲等;《井冈山大学学报》;20130531;第34卷(第3期);第54-57页 * |
Also Published As
Publication number | Publication date |
---|---|
CN104316729A (en) | 2015-01-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104316729B (en) | Self-diagnosis method of acceleration sensors for locomotive bogie detection | |
KR101829645B1 (en) | Device and method for monitoring railway vehicle and method for monitoring using the same | |
CN110704801B (en) | Bridge cluster structure operation safety intelligent monitoring and rapid detection complete method | |
CN104015672B (en) | For the method for registration of vehicle related data, sensor device and vehicle | |
CN106447040B (en) | Mechanical equipment health state evaluation method based on Heterogeneous Multi-Sensor Data fusion | |
CN109142524A (en) | A kind of track damage detecting method, device and equipment | |
CN105424364A (en) | Diagnostic method and device of train bearing failure | |
CN103674234B (en) | State early warning method and system for abnormal vibration of wind generating set | |
CN104297003B (en) | The fault monitoring method of the bogie rotary part based on dynamic alert threshold value | |
CN105403420B (en) | A kind of bogie method for diagnosing faults for merging multisensor | |
CN107036815A (en) | The method and data processing equipment of Severity are carried out to bearing defect | |
US9676402B2 (en) | System for monitoring the operating conditions of a train | |
CN208125377U (en) | A kind of motor rolling bearing failure diagnosis system | |
CN109489931B (en) | Abnormal impact real-time detection method | |
CN105222879B (en) | The long-range self-diagnosable system and its self-diagnosing method of complicated dynamic weighing sensing network | |
CN104266841A (en) | Fault diagnosis device and diagnosis method of diesel engine | |
CN108107365A (en) | A kind of battery pack structure health status on-line monitoring and fault prediction method and device | |
CN106569160A (en) | Output voltage sensor fault diagnosing method of auxiliary current transformer and fault tolerance control method | |
CN110606105A (en) | Safety monitoring system for walking part of unmanned subway vehicle | |
CN110171442B (en) | Detection system and detection method for wheel flat scars | |
CN104271428B (en) | Method for investigating Wheel Rail Contact | |
CN108528475B (en) | A kind of track transition fault alarm method based on multi-level fusion | |
CN110220726A (en) | Detection system, the detection method of wheel out of round | |
CN103743477A (en) | Mechanical failure detecting and diagnosing method and apparatus | |
JP2020008293A (en) | Structure deterioration diagnostic 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 | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |