CN109857090A - A kind of equalizing reservoir brake apparatus health evaluation system and method - Google Patents
A kind of equalizing reservoir brake apparatus health evaluation system and method Download PDFInfo
- Publication number
- CN109857090A CN109857090A CN201910163474.0A CN201910163474A CN109857090A CN 109857090 A CN109857090 A CN 109857090A CN 201910163474 A CN201910163474 A CN 201910163474A CN 109857090 A CN109857090 A CN 109857090A
- Authority
- CN
- China
- Prior art keywords
- data
- equalizing reservoir
- brake apparatus
- pressure
- health
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Valves And Accessory Devices For Braking Systems (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses a kind of equalizing reservoir brake apparatus health evaluation system and method, system includes current sensor, pressure sensor and data processing module;Successively by multiple normal and known fault type equalizing reservoir brake apparatus access systems, it follows the steps below respectively and repeatedly tests and record test data: equalization reservoir target pressure being arranged by data processing module, and is sent to electronic brake control unit;Electronic brake control unit is based on equalization reservoir target pressure and observed pressure, corresponding control signal is exported to filling air-valve and exhaust valve, it carries out filling wind or air draft to control equalizing reservoir, and record the data of each sensor acquisition in test process, therefrom extract characteristic, training health evaluation model;Equalizing reservoir brake apparatus to be measured access health evaluation system is tested, the characteristic in test process is obtained, inputs trained health evaluation model, obtain its health evaluating result.The present invention can accurately assess the health status of equalizing reservoir brake apparatus.
Description
Technical field
The present invention relates to a kind of equalizing reservoir brake apparatus health evaluation system and methods.
Background technique
Equalizing reservoir brake apparatus is directly affected as a key subsystem in train braking system, performance quality
The performance of train braking.Since existing train fault diagnostic system is largely dependent upon expert for the diagnosis of failure
System and a large amount of experiences, and diagnostic method is single, is monitored using unitary variant, so be very restricted.Existing diagnosis
The initial failure of the unpredictable equalizing reservoir brake apparatus of system, if but initial failure it is untreated, as time goes by, failure
It will run down, unpredictable loss occurs in some time later.Equalizing reservoir brake apparatus packet on existing train
It includes air supply system, equalizing reservoir, electronic brake control unit (EBCU), fill air-valve, exhaust valve, pressure sensor and air hose
Road;The air supply system is used to provide wind regime to equalizing reservoir;Air-valve and the exhaust valve of filling is for controlling equalizing reservoir progress
Fill wind and air draft;The pressure sensor is for detecting equalizing reservoir pressure.The initial failure of existing equalizing reservoir brake apparatus
Detection method has the following deficiencies: that (1) solenoid valve is one as the executive component of equalizing reservoir brake apparatus and is easily damaged
Device, but existing method is difficult to realize the monitoring of solenoid valve state, because the turn-on time of high-speed electromagnetic valve is generally on the left side 10ms
The right side, signal are difficult to collect in the equalizing reservoir brake apparatus on existing train;(2) pressure sensor is as equalizing reservoir
The acquisition equipment of pressure, in systems, common failure have biasing and drift, and existing train diagnostic system is using monotropic
Amount monitoring, when sensor failure, monitor control index can be greater than its given threshold, and this threshold value monitoring means heavy dependence is artificial
Setting, efficiency are lower;(3) pipeline leakage seriously relies on empirical data, and depends on hardware redundancy.
Therefore, it is necessary to provide a kind of initial failure detection method of new equalizing reservoir brake apparatus.
Summary of the invention
Technical problem solved by the invention is that provide a kind of equalizing reservoir brake apparatus in view of the deficiencies of the prior art strong
Health assessment system and method can accurately assess equalizing reservoir brake apparatus health status.
To achieve the goals above, technical solution provided by the present invention are as follows:
A kind of equalizing reservoir brake apparatus health evaluation system, including the first current sensor, the second current sensor,
One pressure sensor, second pressure sensor, data collecting card and data processing module;
First current sensor and the second current sensor are respectively used to fill wind in detection equalizing reservoir brake apparatus
Valve electric current and exhaust valve electric current;First pressure sensor and with second pressure sensor be respectively used to detection equalizing reservoir braking dress
Set the total wind pressure and equalizing reservoir pressure of middle air supply system output;
First current sensor, the second current sensor, first pressure sensor and second pressure sensor letter
Number output end is connected to data collecting card, and data collecting card (passing through USB) is connected to data processing module;
Firstly, multiple normal and known fault type equalizing reservoir brake apparatus are successively accessed into health evaluation system,
Repeatedly test (simulation train operation operating condition, such as release of brake behavior) is followed the steps below respectively and records test data;
1) equalization reservoir target pressure, and the electricity being sent in equalizing reservoir brake apparatus are arranged by data processing module
Sub- brak control unit;The equalization reservoir target pressure and the second pressure that electronic brake control unit is arranged based on data processing module
The equalizing reservoir pressure that force snesor detects exports corresponding control signal to air-valve and exhaust valve is filled, to control balanced wind
Cylinder carries out filling wind or air draft to reach equalization reservoir target pressure;
2) record test data: record test process in the first current sensor acquisition fill air-valve electric current (Icv) data,
The exhaust valve current data (Idcv) of second current sensor acquisition, total wind pressure (Pmr) of first pressure sensor acquisition count
According to original pressure in equalizing reservoir pressure (Pser) data and equalizing reservoir brake apparatus of the acquisition of, second pressure sensor
Pressure (Per) data of force snesor acquisition;
Then, characteristic is extracted from the test data of above-mentioned normal and known fault type equalizing reservoir brake apparatus
According to for training health evaluation model;
Further according to sample training collection training health evaluation model;
Finally, equalizing reservoir brake apparatus to be measured is accessed health evaluation system, by step 1)~3) repeatedly tested
And test data is recorded, characteristic is extracted from the test data, inputs trained health evaluation model, obtains its health
Assessment result.
Further, described to fill air-valve and exhaust valve is MAC 130B model solenoid valve.
Further, the data collecting card is USB4222 model high-speed data acquisition card;The data collecting card is adopted
Sample frequency is set as 5KHz to guarantee the accurate acquisition of electromagnetic valve driving current.
Further, the data processing module is computer.
Further, the computer is communicated by UDP communication protocol with electronic brake control unit.
A kind of equalizing reservoir brake apparatus health evaluation system carries out equalizing reservoir brake apparatus health using above system
Assessment, comprising the following steps:
Firstly, successively by multiple normal and known fault type (including fault level and defective device) equalizing reservoir systems
Dynamic device accesses health evaluation system, follows the steps below repeatedly test respectively and (simulates the different operating condition of train, such as
Release of brake behavior) and record test data;
1) equalization reservoir target pressure, and the electricity being sent in equalizing reservoir brake apparatus are arranged by data processing module
Sub- brak control unit;The equalization reservoir target pressure and the second pressure that electronic brake control unit is arranged based on data processing module
The equalizing reservoir pressure that force snesor detects exports corresponding control signal to air-valve and exhaust valve is filled, to control balanced wind
Cylinder carries out filling wind or air draft to reach equalization reservoir target pressure;
2) record test data: the acquisition of the first current sensor fills air-valve current data, second in record test process
The exhaust valve current data of current sensor acquisition, total wind pressure force data of first pressure sensor acquisition, second pressure sensing
The number pressure of original pressure sensor acquisition in the equalizing reservoir pressure data and equalizing reservoir brake apparatus of device acquisition
According to;
Then, characteristic is extracted from the test data of above-mentioned normal and known fault type equalizing reservoir brake apparatus
According to for training health evaluation model;
Finally, equalizing reservoir brake apparatus to be measured is accessed health evaluation system, by step 1)~3) repeatedly tested
And test data is recorded, characteristic is extracted from the test data, inputs trained health evaluation model, obtains its health
Assessment result.Health evaluating result may include whether equalizing reservoir brake apparatus to be measured is normal, if abnormal, identify its failure etc.
Grade and defective device.
Further, the characteristic is two-dimensional time window data, and horizontal direction is the characteristic value of sometime window, is hung down
Histogram is to the timing distribution for a certain feature;Sometime the characteristic value of window includes the first current sensor in the time window,
Original pressure sensing in two current sensors, first pressure sensor, second pressure sensor and equalizing reservoir brake apparatus
One of average value, variance, maximum value, minimum value, median and Differential Characteristics of data of device acquisition are a variety of.
Further, the health evaluation model includes CNN model and XGBoost model two parts;Characteristic is defeated
Enter CNN model, obtains health evaluating value A;Characteristic is unfolded to the one-dimensional characteristic vector to be formed, inputs XGBoost model, obtains
To health evaluating value B;Last comprehensive health assessment value A and health evaluating value B, obtains final health evaluating value.The health is commented
Estimating model is integrated model CNN-XGBoost, combines convolutional neural networks model (CNN) feature learning Nonlinear Mapping energy
The advantage of the stronger advantage of power and limit gradient lift scheme (XGBoost) processing high dimensional feature, improves health to a certain degree
The accuracy and robustness of assessment.
Further, the CNN model and XGBoost model are write using Python.
Research and development of the invention are by project of national nature science fund project 61672539,61672537,61803394,61873353
Part is provided to support.
The utility model has the advantages that
The present invention makes full use of full-scale investigation platform, soft on the basis of analyzing equalizing reservoir brake apparatus working principle
The convenience of part hardware and software platform design and big data, has built equalizing reservoir brake apparatus health evaluating full-scale investigation platform, constructs
Health evaluation model, has the following characteristics that
1, the full-scale investigation platform that the present invention is built has equalizing reservoir control loop design and data acquisition device, can
Alleviate process to simulate train braking well, existing equalizing reservoir brake apparatus monitoring on opposite train, which are added electromagnetism
Valve current monitoring function precisely acquires electromagnetic valve driving current using USB4222 high-speed data card, has good monitoring function
Energy;
2, the software platform that the present invention is built has the health status and data point of overall monitor equalizing reservoir brake apparatus
The function of analysis and the health evaluation model building based on CNN-XGBoost calls the analysis of Python data by process, significantly
Reduce the engineering development period, so that health evaluation model is built easily updated, expansion is greatly reinforced.Health evaluation model structure without
The physical model that must consider equalizing reservoir brake apparatus, only need to pass through data mining by analyzing a large amount of historical datas
New model can be constructed, so that the transplanting and update of implementation model, greatly shorten equalizing reservoir brake apparatus health evaluation system
With the R&D cycle of method.
The present invention can accurately assess equalizing reservoir brake apparatus health status, enhance the safety of the following locomotive operation braking
Property.The present invention improves the intelligence of equalizing reservoir brake apparatus health evaluating, and improving engineering research staff and service personnel makes
Convenience and high efficiency.
Detailed description of the invention
Fig. 1 equalizing reservoir brake apparatus health evaluation system entire block diagram
Fig. 2 health evaluating main window
Fig. 3 administrator logins window
Fig. 4 data monitoring window
The data introduction of Fig. 5 data analysis and Feature Engineering window
The data de-noising of Fig. 6 analysis and Feature Engineering window
The data prediction of Fig. 7 data analysis and Feature Engineering window
The analysis of Fig. 8 data and the feature of Feature Engineering window are shown
The data preparation of Fig. 9 health evaluation model window
The model of Figure 10 health evaluation model window selects
The model training of Figure 11 health evaluation model window and assessment
Figure 12 equalizing reservoir brake apparatus health evaluation model flow chart
Figure 13 CNN-XGBoost schematic diagram
Specific embodiment
In order to illustrate more clearly of Hardware platform design and Design of Software Platform of the invention and system schema, below with reference to
The present invention is further illustrated for the drawings and specific embodiments.
As shown in Figure 1, the invention discloses a kind of equalizing reservoir brake apparatus health evaluation system and method, including hardware
Design and Design of Software Platform.Hardware components mainly include air supply system, equalizing reservoir control loop and data acquisition device.Its
Middle air supply system includes air compressor machine, total cylinder (total wind air reservoir), pneumatic triple piece and air pipe line, by adjusting pneumatic three
Pressure reducing valve in part can be exported the air pressure of total cylinder is relatively stable to equalizing reservoir, to guarantee the stability of wind regime;It is balanced
Cylinder control loop by equalizing reservoir, EBCU, solenoid valve, pressure sensor, current sensor, data collecting card and computer,
Control loop uses the used PWM technology of train and PID closed-loop control, and data acquisition device uses USB4222 high-speed collection card,
For the real-time acquisition of electromagnetic valve current value in equalizing reservoir brake apparatus and equalizing reservoir pressure value, and store data in number
According in library, sample frequency is set as 5KHz to guarantee the accurate acquisition of electromagnetic valve driving current.Electronic brake control unit will
The analog signals of sensor acquisition are converted into digital quantity signal and are transferred to control panel, complete equalizing reservoir using control panel program
Pressure control.Software section mainly uses C# to write the health evaluation system interface of equalizing reservoir brake apparatus, designs health and comments
Estimate main window, data monitoring window, data analysis and Feature Engineering window and health evaluation model window;It is write using Python
Signature analysis and health evaluating mould model construct interface by process and are embedded into health evaluating interface, and assessed value uses percentage
System, the higher characterization equalizing reservoir brake apparatus operating status of score value is the better, and the present invention effectively realizes equalizing reservoir brake apparatus
Experiment simulation and data acquisition, quickly succinct can be patterned data analysis and health evaluating.
Software section includes health evaluation system Platform Designing and health evaluation model.
The health evaluation system platform includes health evaluating main window, data monitoring window, data analysis and feature work
Journey window and health evaluation model window, between window can free switching, function expansibility is strong;EBCU operating system be used for together with
Operating system in EBCU.
As shown in Fig. 2, health evaluating main window include the sub- window of sensing data, the sub- window of characteristic, the sub- window of assessment models,
The sub- window of health evaluating, EBCU monitor sub- window and the sub- window of System Analysis Report, interrelated, intelligence display equalizing reservoir between sub- window
Health evaluating state;It clicks more buttons in any sub- window to go to such as Fig. 3 administrator's login system, gives and weigh if passing through
Limit, otherwise only shows main interface health status, no change model permission;The sub- window real-time display sensing data of sensing data,
Click more button connection data monitor window mouths;The sub- window refresh timing of data characteristics shows the characteristic in certain a period of time,
Click more button connection data analyses and Feature Engineering window;Assessment models show sub- window synchronous healthy assessment models window
Types of models and model parameter;The sub- window of health evaluating includes health evaluating instruction disk, health evaluating history and warning information, is clicked
More button system connection health evaluation model windows;
It includes EBCU running state monitoring and the control of cylinder pressure that EBCU, which monitors sub- window, and it is anti-in real time to be conducive to monitoring effect
Feedback;The sub- window real-time display system CPU usage of System Analysis Report, current time and equalizing reservoir operation report;EBCU monitoring
Sub- window sends pressure instruction to EBCU by UDP communication protocol;It is bound by data, by digital input, number in EBCU
The signal real-time display of output board, analog input, power panel, control panel is measured in monitoring sub- window, convenient for directly observation control
Effect processed, and monitor whether EBCU is abnormal, and record failure can realize fault simulation in real time;
As shown in figure 4, data monitoring window utilizes DynamicDataDisplay dynamic link library real-time reception and figure
Change the data of display current monitor sensor;By the range of normal value (or fault threshold) of setting sensor, Preliminary detection is passed
Sensor abnormal conditions (detect obvious abnormal data), and carry out frequency of abnormity statistics;By connecting database realizing monitoring data
Storage, for data analysis, Feature Engineering and model foundation;Local data can be obtained by file reading, is realized additional
The extension of data;
The data analysis carries out mathematical statistics and data mining to historical data with Feature Engineering window, completes feature
It extracts and constructs and save into database.As viewed in figures 5-8, data analysis and Feature Engineering window include data introduction, number
Four parts are shown according to denoising, data prediction, feature, the common data analysis for completing historical data and Feature Engineering;Data
Introduce include file selection, file path show, data statistics shows and remarks be written, file select document source in
The csv file that database stores data passes through data conversion storage;Data de-noising includes algorithms selection and filter effect visualization display,
Denoising Algorithm may be selected to be median filtering or mean filter, can customize setting filter window parameter and shows filter effect, leads to
Reading and writing data is crossed to save into database;Data prediction includes time window feature extraction, other feature extractions, data normalization
It is combined with feature;Time window feature extraction obtains the average value of window data, variance, maximum by customized setting window size
It is value, minimum value, median, one or more in difference value tag;In order to effectively realize the identification of operational mode, operation is increased
Pattern feature, other features are operational mode feature;History data set is clustered using K-Means first, is obtained not
With the Sub Data Set of operational mode;Then for the Sub Data Set of different operational modes increase a searching value of the column since 1 ing with
One arranges the identifier constant of operational mode for identification;Then the Sub Data Set of different operational modes is pressed into sample time order weight
Group can recognize which kind of operational mode of the data processing of any time;Data normalization may be selected to be min-max standardization, z-
One in score standardization and the conversion of log function;Feature combination using original time window data+time window characteristic+its
Time window and the feature of extraction are overlapped by the combination of his feature, and obtained expansion time window is before retaining raw information
It puts, reconstructs training set data, increase derived character, improve the accuracy of health state evaluation;The visualization of feature display module
Initial data and characteristic are conducive to the effect of contrast characteristic's processing, can be reserved for characteristic to local;Wherein, for equal
Weigh cylinder multi-state data, and the feature of extraction is simultaneously attached in original time window data by extraction time window feature, constitutes extension
Time window data, thus the characterization ability of lifting feature, the accuracy of the health evaluating preferably promoted.
The health evaluation model window returns the characteristic of historical data by model selection and parameter setting
Return prediction, constructs health evaluation model, and be synchronized to health evaluating main window, realize online health evaluating;The main window of health evaluating
Mouth provides user-centered interface, develops conducive to the health evaluating of equalizing reservoir brake apparatus;As described in Fig. 9-11, health evaluation model
Window includes data preparation, model selection and model training and assessment;Data preparation include file selection, file address show,
Characteristic introduction and remarks write-in, the characteristic that the file of file selection is saved from data analysis with Feature Engineering window
File, data characteristics data introduction show the mathematical statistics information of characteristic;Model selection includes that algorithms selection and parameter are selected
It selects;Model training and assessment display model training process and verifying collection result;Model selection calls Python to write by process
Health evaluating algorithm routine, algorithms selection Integrated Algorithm CNN-XGBoost, provide parameter interface realize algorithm parameter setting,
And existing model importing and new model store function are provided, flexibility greatly reinforces;
As shown in figure 12, health evaluating algorithm routine is reduced significantly using the convenience of Python data processing exploitation and is based on
The engineering developme of the equalizing reservoir brake apparatus health evaluating of CNN-XGBoost, algorithm is portable strong, and the development cycle is short,
Model construction is simple, is substituted by model, greatly strengthens equalizing reservoir brake apparatus health evaluating accuracy;It is being sensed
On the basis of the preliminary abnormality detection of device, when sensor values are not in normal interval, fault warning is issued, but when sensor detects
In normal interval (fault type, fault level is artificially marked according to the fault data of existing device in historical data in value
Data), training have Monitor assessment models, then online according to sensing data assess current operating conditions it is whether good
It is good.
As shown in figure 13, the present invention proposes to establish the health based on data-driven using the method for integrated CNN-XGBoost
Assessment models;CNN-XGBoost models coupling CNN model and the powerful feature learning of XGBoost model, significantly improve
The accuracy of weighing apparatus cylinder brake apparatus health evaluating;CNN model is a kind of efficient neural network structure, passes through convolution operation
Feature is continually strengthened with pondization operation, there is very strong non-linear mapping capability;XGBoost model is a kind of efficient Assembled tree
Model has very strong characteristic strengthening important by boosting integrated approach, constantly reduction error of fitting;The CNN of use
Model structure is to be followed successively by input layer, convolutional layer, full articulamentum and output layer.Input layer data is 2D time window data, horizontal
Direction is the characteristic value at a certain moment, and vertical direction is the timing distribution of a certain feature;Convolutional layer is the 1*1 convolution of multichannel
Core, input sample are mapped as multiple Feature Mappings, i.e., the different features learnt by different parameters by multiple convolution kernels
Mapping;The feature of full articulamentum connection convolutional layer mapping, then commented by the health that non-linear function regression is fitted to obtain output layer
Valuation A;For the XGBoost model used to integrate tree-model, output is the 1D feature vector of 2D time window data expansion, is passed through
Boosting integrates more trees, constantly reduction error of fitting, stops iteration when error of fitting tends to convergence, obtains health evaluating
Value B;Finally health evaluating value A and health evaluating value B is averaging, obtains the health evaluating value of target.
The present invention effectively realizes the simulated experiment and data acquisition of equalizing reservoir brake apparatus, being capable of quickly succinct progress
Graphics data analysis and health evaluating.
Claims (8)
1. a kind of equalizing reservoir brake apparatus health evaluation system, which is characterized in that including the first current sensor, the second electric current
Sensor, first pressure sensor, second pressure sensor, data collecting card and data processing module;
First current sensor and the second current sensor are respectively used to fill air-valve electricity in detection equalizing reservoir brake apparatus
Stream and exhaust valve electric current;First pressure sensor and with second pressure sensor be respectively used to detection equalizing reservoir brake apparatus in
The total wind pressure and equalizing reservoir pressure of air supply system output;
First current sensor, the second current sensor, the signal of first pressure sensor and second pressure sensor are defeated
Outlet is connected to data collecting card, and data collecting card is connected to data processing module;
Firstly, multiple normal and known fault type equalizing reservoir brake apparatus are successively accessed health evaluation system, respectively
It follows the steps below and repeatedly tests and record test data;
1) equalization reservoir target pressure, and the electronics system being sent in equalizing reservoir brake apparatus are arranged by data processing module
Dynamic control unit;The equalization reservoir target pressure and second pressure biography that electronic brake control unit is arranged based on data processing module
The equalizing reservoir pressure that sensor detects, exports corresponding control signal to filling air-valve and exhaust valve, with control equalizing reservoir into
Row fills wind or air draft to reach equalization reservoir target pressure;
2) record test data: the acquisition of the first current sensor fills air-valve current data, the second electric current in record test process
The exhaust valve current data of sensor acquisition, total wind pressure force data, the second pressure sensor of first pressure sensor acquisition are adopted
The pressure data of original pressure sensor acquisition in the equalizing reservoir pressure data and equalizing reservoir brake apparatus of collection;
Then, characteristic is extracted from the test data of above-mentioned normal and known fault type equalizing reservoir brake apparatus,
For training health evaluation model;
Finally, equalizing reservoir brake apparatus to be measured is accessed health evaluation system, by step 1)~2) is repeatedly tested and remembered
Test data is recorded, characteristic is extracted from the test data, inputs trained health evaluation model, obtain its health evaluating
As a result.
2. equalizing reservoir brake apparatus health evaluation system according to claim 1, which is characterized in that the data acquisition
Card is the high-speed data acquisition card of model USB4222;The sample frequency of the data collecting card is set as 5KHz.
3. equalizing reservoir brake apparatus health evaluation system according to claim 1, which is characterized in that the data processing
Module is computer.
4. equalizing reservoir brake apparatus health evaluation system according to claim 3, which is characterized in that the computer is logical
UDP communication protocol is crossed to be communicated with electronic brake control unit.
5. equalizing reservoir brake apparatus health evaluation system according to claim 1, which is characterized in that use claim
Equalizing reservoir brake apparatus health evaluation system described in any one of 1~4, which is characterized in that carried out using above system equal
Weigh cylinder brake apparatus health evaluating, comprising the following steps:
Firstly, multiple normal and known fault type equalizing reservoir brake apparatus are successively accessed health evaluation system, respectively
It follows the steps below and repeatedly tests and record test data;
1) equalization reservoir target pressure, and the electronics system being sent in equalizing reservoir brake apparatus are arranged by data processing module
Dynamic control unit;The equalization reservoir target pressure and second pressure biography that electronic brake control unit is arranged based on data processing module
The equalizing reservoir pressure that sensor detects, exports corresponding control signal to filling air-valve and exhaust valve, with control equalizing reservoir into
Row fills wind or air draft to reach equalization reservoir target pressure;
2) record test data: the acquisition of the first current sensor fills air-valve current data, the second electric current in record test process
The exhaust valve current data of sensor acquisition, total wind pressure force data, the second pressure sensor of first pressure sensor acquisition are adopted
The pressure data of original pressure sensor acquisition in the equalizing reservoir pressure data and equalizing reservoir brake apparatus of collection;
Then, characteristic is extracted from the test data of above-mentioned normal and known fault type equalizing reservoir brake apparatus,
For training health evaluation model;
Finally, equalizing reservoir brake apparatus to be measured is accessed health evaluation system, by step 1)~2) is repeatedly tested and remembered
Test data is recorded, characteristic is extracted from the test data, inputs trained health evaluation model, obtain its health evaluating
As a result.
6. equalizing reservoir brake apparatus health evaluation system according to claim 5, which is characterized in that the characteristic
For two-dimensional time window data, horizontal direction is the characteristic value of sometime window, and vertical direction is the timing distribution of a certain feature;Certain
The characteristic value of one time window includes the first current sensor in the time window, the second current sensor, first pressure sensor,
The average value of the data of original pressure sensor acquisition, variance, maximum in two pressure sensors and equalizing reservoir brake apparatus
One of value, minimum value, median and Differential Characteristics are a variety of.
7. equalizing reservoir brake apparatus health evaluation system according to claim 6, which is characterized in that the health evaluating
Model includes CNN model and XGBoost model two parts;Characteristic is inputted into CNN model, obtains health evaluating value A;It will be special
The one-dimensional characteristic vector to be formed is unfolded in sign data, inputs XGBoost model, obtains health evaluating value B;Last comprehensive health assessment
Value A and health evaluating value B, obtains final health evaluating value.
8. equalizing reservoir brake apparatus health evaluation system according to claim 7, which is characterized in that the CNN model
It is write with XGBoost model using Python.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910163474.0A CN109857090B (en) | 2019-03-05 | 2019-03-05 | Health evaluation system and method for balanced air cylinder braking device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910163474.0A CN109857090B (en) | 2019-03-05 | 2019-03-05 | Health evaluation system and method for balanced air cylinder braking device |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109857090A true CN109857090A (en) | 2019-06-07 |
CN109857090B CN109857090B (en) | 2020-12-15 |
Family
ID=66899870
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910163474.0A Active CN109857090B (en) | 2019-03-05 | 2019-03-05 | Health evaluation system and method for balanced air cylinder braking device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109857090B (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110281895A (en) * | 2019-07-10 | 2019-09-27 | 中南大学 | A kind of control method and its control system of equalizing reservoir pressure |
CN111308909A (en) * | 2020-02-18 | 2020-06-19 | 中南大学 | Relay valve fault diagnosis method and device |
CN111399474A (en) * | 2020-02-29 | 2020-07-10 | 中南大学 | Health index-based life prediction method and device for balance control module |
CN112350878A (en) * | 2019-08-07 | 2021-02-09 | 阿里巴巴集团控股有限公司 | Pressure test system |
CN112613224A (en) * | 2020-11-25 | 2021-04-06 | 西人马联合测控(泉州)科技有限公司 | Training method, detection method, device and equipment of bridge structure detection model |
CN114413055A (en) * | 2022-01-20 | 2022-04-29 | 济南轨道交通集团有限公司 | Operation and maintenance device and method for subway tunnel fan interlocking air valve |
CN114684096A (en) * | 2022-03-08 | 2022-07-01 | 中南大学 | Multi-source data fusion-based electromagnetic valve service life prediction device and method |
CN114969990A (en) * | 2022-08-02 | 2022-08-30 | 中国电子科技集团公司第十研究所 | Multi-model fused avionic product health assessment method |
CN116501027A (en) * | 2023-06-29 | 2023-07-28 | 中南大学 | Distributed braking system health assessment method, system, equipment and storage medium |
CN116861369A (en) * | 2023-09-04 | 2023-10-10 | 山东捷瑞数字科技股份有限公司 | Air compressor health assessment method and system for industrial internet heterogeneous data fusion |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102608452A (en) * | 2012-02-24 | 2012-07-25 | 安徽建筑工业学院 | High speed train equipment state and electric energy quality monitoring system and method |
WO2013093207A1 (en) * | 2011-12-23 | 2013-06-27 | Aalto-Korkeakoulusäätiö | A method for detecting faults in a flow network controlled by valves or in a part of it |
CN103364203A (en) * | 2012-03-29 | 2013-10-23 | 通用汽车环球科技运作有限责任公司 | Collaborative vehicle health model |
CN106740798A (en) * | 2016-11-24 | 2017-05-31 | 郑州智辆电子科技有限公司 | Locomotive brake on-line monitoring system |
CN206265032U (en) * | 2016-11-24 | 2017-06-20 | 郑州智辆电子科技有限公司 | Locomotive brake on-line monitoring system |
CN107844113A (en) * | 2017-11-08 | 2018-03-27 | 中车青岛四方机车车辆股份有限公司 | A kind of rail vehicle braking system debugging monitoring system and debugging monitoring method |
CN109344517A (en) * | 2018-10-14 | 2019-02-15 | 毛述春 | A kind of high-voltage isulation method for diagnosing faults of new-energy automobile |
-
2019
- 2019-03-05 CN CN201910163474.0A patent/CN109857090B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013093207A1 (en) * | 2011-12-23 | 2013-06-27 | Aalto-Korkeakoulusäätiö | A method for detecting faults in a flow network controlled by valves or in a part of it |
CN102608452A (en) * | 2012-02-24 | 2012-07-25 | 安徽建筑工业学院 | High speed train equipment state and electric energy quality monitoring system and method |
CN103364203A (en) * | 2012-03-29 | 2013-10-23 | 通用汽车环球科技运作有限责任公司 | Collaborative vehicle health model |
CN106740798A (en) * | 2016-11-24 | 2017-05-31 | 郑州智辆电子科技有限公司 | Locomotive brake on-line monitoring system |
CN206265032U (en) * | 2016-11-24 | 2017-06-20 | 郑州智辆电子科技有限公司 | Locomotive brake on-line monitoring system |
CN107844113A (en) * | 2017-11-08 | 2018-03-27 | 中车青岛四方机车车辆股份有限公司 | A kind of rail vehicle braking system debugging monitoring system and debugging monitoring method |
CN109344517A (en) * | 2018-10-14 | 2019-02-15 | 毛述春 | A kind of high-voltage isulation method for diagnosing faults of new-energy automobile |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110281895A (en) * | 2019-07-10 | 2019-09-27 | 中南大学 | A kind of control method and its control system of equalizing reservoir pressure |
CN112350878A (en) * | 2019-08-07 | 2021-02-09 | 阿里巴巴集团控股有限公司 | Pressure test system |
CN111308909A (en) * | 2020-02-18 | 2020-06-19 | 中南大学 | Relay valve fault diagnosis method and device |
CN111399474A (en) * | 2020-02-29 | 2020-07-10 | 中南大学 | Health index-based life prediction method and device for balance control module |
CN112613224A (en) * | 2020-11-25 | 2021-04-06 | 西人马联合测控(泉州)科技有限公司 | Training method, detection method, device and equipment of bridge structure detection model |
CN114413055B (en) * | 2022-01-20 | 2024-04-16 | 济南轨道交通集团有限公司 | Operation and maintenance device and method for interlocking air valve of subway tunnel fan |
CN114413055A (en) * | 2022-01-20 | 2022-04-29 | 济南轨道交通集团有限公司 | Operation and maintenance device and method for subway tunnel fan interlocking air valve |
CN114684096A (en) * | 2022-03-08 | 2022-07-01 | 中南大学 | Multi-source data fusion-based electromagnetic valve service life prediction device and method |
CN114684096B (en) * | 2022-03-08 | 2023-02-17 | 中南大学 | Multi-source data fusion-based electromagnetic valve service life prediction device and method |
CN114969990B (en) * | 2022-08-02 | 2022-10-14 | 中国电子科技集团公司第十研究所 | Multi-model fused avionic product health assessment method |
CN114969990A (en) * | 2022-08-02 | 2022-08-30 | 中国电子科技集团公司第十研究所 | Multi-model fused avionic product health assessment method |
CN116501027A (en) * | 2023-06-29 | 2023-07-28 | 中南大学 | Distributed braking system health assessment method, system, equipment and storage medium |
CN116501027B (en) * | 2023-06-29 | 2023-10-03 | 中南大学 | Distributed braking system health assessment method, system, equipment and storage medium |
CN116861369A (en) * | 2023-09-04 | 2023-10-10 | 山东捷瑞数字科技股份有限公司 | Air compressor health assessment method and system for industrial internet heterogeneous data fusion |
Also Published As
Publication number | Publication date |
---|---|
CN109857090B (en) | 2020-12-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109857090A (en) | A kind of equalizing reservoir brake apparatus health evaluation system and method | |
CN104712542B (en) | A kind of reciprocating compressor sensitive features based on Internet of Things are extracted and method for diagnosing faults | |
CN107272644B (en) | The DBN network fault diagnosis method of latent oil reciprocating oil pumping unit | |
CN107193271A (en) | Preposition service adapter and trouble-shooter for industrial information physical system | |
CN106338406A (en) | On-line monitoring and fault early-warning system and method for traction electric transmission system of train | |
CN105487530B (en) | The low failure prediction system of diesel engine row's temperature and method | |
CN105629958B (en) | A kind of batch process method for diagnosing faults based on sub-period MPCA SVM | |
CN104777418B (en) | A kind of analog-circuit fault diagnosis method based on depth Boltzmann machine | |
CN104523264B (en) | Electrocardiosignal processing method | |
CN109838696A (en) | Pipeline fault diagnostic method based on convolutional neural networks | |
CN106951197A (en) | A kind of printing equipment remote fault diagnosis method and system based on signal testing | |
CN110469561B (en) | Device, method and system for monitoring state of hydraulic system | |
CN204113701U (en) | A kind of mine fan on-line monitoring and fault diagnosis system | |
CN107462425A (en) | The failure prediction method and failure predication platform of a kind of car kits system | |
CN104471542A (en) | Discriminative hidden kalman filters for classification of streaming sensor data in condition monitoring | |
CN106404441A (en) | Nonlinear similarity index based fault classification and diagnosing method | |
CN112507479B (en) | Oil drilling machine health state assessment method based on manifold learning and softmax | |
CN105571638A (en) | Machinery device fault combination prediction system and method | |
CN110174878A (en) | A kind of unmanned intelligence equipment is healthy and ensures integrated management general-purpose system | |
CN108153987A (en) | A kind of hydraulic pump Multiple faults diagnosis approach based on the learning machine that transfinites | |
CN105626502A (en) | Plunger pump health assessment method based on wavelet packet and Laplacian Eigenmap | |
CN104142680B (en) | Multiple-sensor fault diagnosis system and method based on robust input training neural network | |
CN105678423A (en) | Fault diagnosis system sensor optimal configuration method based on quantitative structure model | |
CN104880217A (en) | Fault sensor information reconstruction method based on measured value association degree | |
CN105403705A (en) | Continuous blood glucose monitoring equipment comprising blood glucose classification function fault detection module |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |