CN110187696A - Single order servomechanism sensor fault diagnosis method and system based on dynamic trend - Google Patents
Single order servomechanism sensor fault diagnosis method and system based on dynamic trend Download PDFInfo
- Publication number
- CN110187696A CN110187696A CN201910405560.8A CN201910405560A CN110187696A CN 110187696 A CN110187696 A CN 110187696A CN 201910405560 A CN201910405560 A CN 201910405560A CN 110187696 A CN110187696 A CN 110187696A
- Authority
- CN
- China
- Prior art keywords
- data
- fault
- failure
- window
- additivity
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0275—Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
- G05B23/0278—Qualitative, e.g. if-then rules; Fuzzy logic; Lookup tables; Symptomatic search; FMEA
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Quality & Reliability (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
Single order servomechanism sensor fault diagnosis method and system based on dynamic trend.The present invention relates to single order following control system sensor fault diagnosis methods, comprising the following steps: realizes multiplying property malfunctioning module, the without reason switching between barrier module and additivity malfunctioning module by analog mechanical switch, acquires experimental data；On the basis of analyzing normal and failure system dynamic characteristic, malfunction monitoring, Fault Estimation and fault reconstruction static models are established, propose the diagnostic model scaling method and diagnostic process of application on site；The validity of the method for diagnosing faults and the high-precision of diagnosis are demonstrated according to " complex process system innovation experiment porch " in-circuit emulation.It the composite can be widely applied in the sensor on-line fault diagnosis of single order following control system.
Description
Technical field
The present invention relates to a kind of sensor fault diagnosis system and method more particularly to general single order follow-up control systems
System sensor fault diagnosis method and its system propose the single order servomechanism sensor fault real-time diagnosis based on dynamic trend
Method belongs to data-driven field.
Background technique
In recent years, fault diagnosis technology has obtained developing on a large scale very much, is widely used in different field.Existing failure is examined
Disconnected technology can be divided into three categories.Third class method be based on data-driven method, independent of the analytic modell analytical model of control object,
Fault type, such as wavelet analysis method, Kalman filtering and Hilbert transform are analyzed using real time data online processing；Or
Sum up fault signature according to a large amount of known samples to making inferences, for example, neural network, bayesian theory, genetic algorithm,
Pattern-recognition and Principal Component Analysis.Method for diagnosing faults currently based on data-driven is research hotspot, not only avoids base
In the analytical model algorithm the shortcomings that, such as: due to the scale effect of complication system, the mechanism model of whole system can not be established,
There is significant limitations in physical fault diagnosis；And the deficiency based on qualitative empirical method is compensated for, and such as: it can only qualitatively judge
Failure mode cannot provide more guidances to consequent malfunction processing.Fault diagnosis compared with open cycle system, in closed-loop system
It can be declined, and sensor is perceptual signal, obtains information and transmit the critical elements of information, therefore studies a kind of be applicable in
In the strong real-time of servomechanism, independent of a large amount of test sets, and the black box fault diagnosis side of quantitative analysis can be realized
Method is necessary.
Summary of the invention
Goal of the invention: to solve sensor fault monitoring, Fault Estimation and event common in single order following control system
The problems such as barrier separation, propose a kind of single order servomechanism sensor fault real-time diagnosis method based on dynamic trend.It avoids
System model establishes inaccurate defect, while also avoiding the influence of environmental disturbances and the other component part characteristics of system,
Improve the diagnosis success rate of diagnostic system.
Technical solution: in order to solve the above technical problems, the present invention provides a kind of single order servo antrol based on dynamic trend
System sensor fault diagnosis method and system, " the complex process system innovation experiment porch " for holding water tank based on four, the platform
Server is connected to the slave computer PLC of experimental facilities by OPC mechanics of communication, in conjunction with the real-time of configuration Wincc monitoring interface
Monitor online dynamic data.
A kind of single order following control system sensor fault diagnosis method based on dynamic trend, includes the following steps:
(1) research process is modeled by Simulirik；It is built by host computer configuration software SIMATIC WinCC
Vertical man-machine interface, facilitates control experimentation, meets research needs；Establish it is single hold the infrastest platform in kind of liquid level control system half into
Mobile state emulation；
(2) as shown in Figure 1, realizing multiplying property malfunctioning module, without reason barrier module and additivity failure by analog mechanical switch
Switching between module acquires experimental data；And the man-machine interface established by SIMATIC WinCC observes water level change in real time
Change, handles abnormal conditions in time；
(3) discrete data point is handled using Matlab software, fault eigenvalue is studied according to numerical analysis figure；
(4) based on the data processing model of sliding window, by Fault Isolation in window, malfunction monitoring static state mould is constructed
Type；
Data absolute value residual error is a criterion, can be used for boundary value when statistical system breaks down:
E=| A-A*| (1)
Wherein e is absolute value residual error, and A is observable data value, A*For desired data value.System after system jam
Response needs certain control time, and individual data is not enough to reflect variation after making corresponding control action in order to avoid system,
It is particularly important to intercept reasonable data window, sensor itself can also be filtered out using data window and measure noise to fault diagnosis
Interference.
When system communication cycle be T, data point sum be N, choose data window size t (data point n), then have:
N=t/T (2)
Freshly harvested dynamic data filling into window and is rejected by earliest data using stack architecture, if sharing m number
According to window, then:
M=N-n+1 (3)
The window residual sum of data is added to obtain by data point absolute value residual errors all in the window, if EiFor i-th of window
Mouth residual sum, then:
A in formulaiFor i-th of actual data point, Ai *For i-th of expected data point.
Assuming that normal data residual sum meets normal distribution, the maximum residul difference and calculating l group number of l group normal data are acquired
According to maximum residul difference and average value:
Standard deviation are as follows:
Using 2 σ as the method for removal abnormal data, then residual sum when failure occurs:
Fault monitoring method is that window absolute value residual sum is made comparisons with threshold value, if having exceeded threshold range, is sentenced
Failure has occurred in disconnected system, otherwise system belongs to normal operating condition.
Formula (7) is the static models of malfunction monitoring, and when practical application will be demarcated in conjunction with real system online data.
(5) fault vectors are calculated using adaptive threshold binarization method, constructs Fault Estimation static models；
Fault vectors are calculated using ascent method.
If y'iFor slope:
Wherein (ti,yi), (ti+1,yi+1) it is two consecutive number strong points, sampling period T.
Additivity Fault Estimation method is when the fault type of system is additivity failure, additivity failure deviation a at this time are as follows:
When the symbol of a is timing, system failure value is increment, otherwise is decrement.
Multiplying property Fault Estimation method is when the fault type of system is multiplying property failure, multiplying property failure gain k at this time are as follows:
As k > 1, system failure value is increment；As k < 1, system failure value is decrement.
Formula (9) and (10) are the static models of Fault Estimation.
(6) a curve matching is carried out using data of the least square method to different simulated failure degree, by fitting song
The Monomial coefficient of line carries out quadratic fit and constructs fault reconstruction static models as fault eigenvalue；
After determining that failure occurs, data variation section t the most apparent is chosen, time variable is obtained and integrates as t=[t1,
t2,...,tw], corresponding dynamic response measurement integrates as y=[y1,y2,...,yw], it is obtained in this section of section according to least square method
The fit equation of data dynamic trend:
It is handled using the additivity malfunction of system and the simulation dynamic data of multiplying property malfunction varying strength, it will
Monomial coefficient in formula (11)As the characteristic value EV extracted:
Wherein P is additivity fault eigenvalue, and Q is multiplying property fault eigenvalue.For quantitative two kinds of fault eigenvalues of analysis
Difference, the Monomial coefficient of different faults type is fitted using Linear regression again, it is inclined with additivity failure
Difference obtains additivity fault eigenvalue fit equation using additivity failure Monomial coefficient as dependent variable as independent variable:
P=f (a) (13)
Using the gain of multiplying property failure as independent variable, using multiplying property failure Monomial coefficient as dependent variable, the event of multiplying property is obtained
Hinder characteristic value fit equation:
Q=g (k) (14)
In system jam, the failure strength of additivity He multiplying property is respectively obtained based on Fault Estimation, by additivity failure
Deviation and the gain of multiplying property failure, which substitute into corresponding fit equation, to be obtained accordinglyWith
The Monomial coefficient of measured data is utilized when fault separating method fault diagnosisWithWithIt calculates:
Smaller is considered as effectively as a result, thus judging fault type for additivity failure or multiplying property failure.
Formula (15) is the static models of fault reconstruction.
(7) the diagnostic model scaling method and diagnostic process of application on site are proposed；
The flow chart of sensing system on-line proving and fault diagnosis in practical engineering application as shown in Figure 2.It is every in system
It is secondary to put into operation again and re-start corresponding calibration after Breakdown Maintenance, to adapt to system performance.Fault diagnosis is divided into the progress of three steps,
It is real time fail monitoring, Fault Estimation and fault reconstruction respectively.
(8) validity of the method for diagnosing faults is demonstrated according to " complex process system innovation experiment porch " in-circuit emulation
With the high-precision of diagnosis.
The utility model has the advantages that the present invention has the advantage that in terms of existing technologies
1, this method is suitable for the strong real-time of servomechanism, independent of a large amount of test sets, and can be realized quantitative
The black box method for diagnosing faults of analysis avoids system model and establishes inaccurate defect, at the same also avoid environmental disturbances and
The influence of the other component part characteristics of system, improves the diagnosis success rate of diagnostic system；
2, data source is abundant, based on generated historical data and analog simulation data in practical engineering application, and
Data sample is sufficiently large；
3, this system, which avoids the additional equipment of increase, can be realized on-line fault diagnosis technology, can preferably be integrated into
In original control system, production cost is advantageously reduced, improves the safety of automation engineering.
Detailed description of the invention
Fig. 1 is the single loop control system block diagram in troubleshooting step in the present invention (2)；
Fig. 2 is on-line proving and Troubleshooting Flowchart in troubleshooting step in the present invention (7)；
Fig. 3 is the fault reconstruction schematic illustration in the present invention in specific implementation step 3；
Fig. 4 is the quadratic fit curve of additivity failure deviation a < 0 in the present invention in specific implementation step 3；
Fig. 5 is the quadratic fit curve of additivity failure deviation a > 0 in the present invention in specific implementation step 3；
Fig. 6 is the quadratic fit curve of multiplying property failure gain k < 1 in the present invention in specific implementation step 3；
Fig. 7 is the quadratic fit curve of multiplying property failure gain k > 1 in the present invention in specific implementation step 3；
Window residual sum when Fig. 8 is the additivity failure deviation+1.0cm in the present invention in specific implementation step 4；
Window residual sum when Fig. 9 is the additivity failure deviation -0.1cm in the present invention in specific implementation step 4；
Window residual sum when Figure 10 is the multiplying property failure gain 1.02 in the present invention in specific implementation step 4；
Liquid level change rate when Figure 11 is the additivity failure deviation+1.5cm in the present invention in specific implementation step 5；
Liquid level change rate when Figure 12 is the multiplying property failure 0.80 in the present invention in specific implementation step 5.
Specific embodiment
Online event provided by the invention is introduced so that certain single order is servo-actuated tank level control system ultrasonic sensor as an example below
Hinder the specific implementation step of diagnostic method.
Process as shown in connection with fig. 2, steps are as follows:
1, final expectation level value is set in experiment as 10cm, operation hours is set as 1000s in single experiment, sampling
Cycle Ts=0.5s.
Servomechanism sets input slope as the ramp signal of 0.01cm/s, and zero moment starts to input the letter that initial value is zero
Number.The varying strength failure in 20% range of setting value is added in the selection 500s moment, analyzes the liquid level in the case of different faults
Changing rule.
2, t=25s is set as data window size, then each window number strong point n are as follows:
N=t/T=50
Primary experiment has N=2001 data, shares m data window:
M=N-n+1=1952
According to the method for adding window absolute value residual sum, the maximized window residual error of data under multiple groups normal non-fault state is acquired
With for set EMax:
EMax=[E1max,E2max,...,E15max]
=[5.9134,6.2123,5.1452,7.4221,7.2116,
3.7570,5.3467,4.6190,6.1961,5.7457,
5.9872,6.0425,5.0343,6.6345,4.9857]
The average value and standard deviation of this 10 groups of data are obtained according to formula (5), (6):
Determine that fault threshold is E by formula (7)*=7.6244.
3, fault reconstruction schematic diagram as shown in Figure 3 obtains the fitting function curve under different degrees of malfunction, as schemed institute
Show:
As additivity failure deviation a < 0, quadratic fit curve such as Fig. 4, corresponding fitting function are as follows:
Q (a)=- 0.0206a2- 0.1001a+0.0087, and have r2=0.9978, RMSE=0.001862
As additivity failure deviation a > 0, quadratic fit curve such as Fig. 5, corresponding fitting function are as follows:
Q (a)=0.0390a2- 0.2055a+0.0195, and have r2=0.992, RMSE=0.00777
As multiplying property failure gain k < 1, quadratic fit curve such as Fig. 6, corresponding fitting function are as follows:
P (k)=1.9352k2- 4.8900k+2.9569, and have r2=0.996, RMSE=0.002712
As multiplying property failure gain k > 1, quadratic fit curve such as Fig. 7, corresponding fitting function are as follows:
P (k)=- 1.4981k2+ 2.3195k-0.8284, and have r2=0.9933, RMSE=0.00207
From the point of view of the above fitting result, r2Close to 1, RMSE close to 0, illustrate data to the fitting degree of model very
It is good.
4, as shown in figure 8, being monitored using malfunction monitoring static models to failure；Fault detection sensitivity marginal testing
The missing inspection situation of real time data is determined, as shown in Figure 9 and Figure 10, when additivity failure deviation is in setting value in following control system
1% in the range of when, malfunction monitoring is likely to occur missing inspection situation when the gain of multiplying property failure is in 2% range of setting value.Together
When malfunction monitoring is carried out to situation of the multiple groups failure size outside 2% range of setting value, malfunction monitoring does not occur missing inspection situation.
5, by taking additivity failure+1.5cm and multiplying property failure 0.80 occur for servomechanism as an example, change rate curve such as Figure 11,12
It is shown.
When fault type is additivity failure, failure deviation a is can be obtained in maximum rate of change as shown in Figure 11(+1.5):
a(+1.5)=Ty'm=0.5s × 3cm/s=1.5cm
When fault type is multiplying property failure, failure gain k is can be obtained in maximum rate of change as shown in Figure 12(0.80):
1 Fault Estimation verification result of table
As shown in table 1, it in 5% range for adding desired value, chooses multiple groups additivity failure and multiplying property failure is tested, lead to
Fault Estimation method is crossed to determine the fault vectors of real time data, judges the precision of Fault Estimation.
6、
2 fault reconstruction verification result of table
As shown in table 2, when following control system sensor occur failure size deviate expectation level value 1.5% with
When interior, it cannot separate.For stabilization system, failure in the range of deviateing the 1.5% of expectation level value, cannot equally be separated.
Fault diagnosis summarize the servomechanism sensor fault diagnosis based on dynamic trend failure strength effective range be
1.5%, Fault Estimation precision is 3.0%；Demonstrate the validity of method and the high-precision of diagnosis.
The various embodiments described above are merely to illustrate the present invention, wherein the structure of each component, connection type and manufacture craft etc. are all
It can be varied, all equivalents and improvement carried out based on the technical solution of the present invention should not exclude
Except protection scope of the present invention.
Claims (3)
1. single order following control system sensor fault diagnosis system, it is characterised in that:
" the complex process system innovation experiment porch " for holding water tank based on four, the platform are serviced OPC by OPC mechanics of communication
Slave computer PLC, the Matlab/Simulink Virtual Controller that device is connected to experimental facilities can carry out liquid level by PID arithmetic
Control, in conjunction with configuration Wincc monitoring interface real time monitoring and Matlab in Workspace online dynamic data carry out it is real-time
Diagnostic model is analyzed and established to data.By calling Simulink function to write Program Generating M file dynamic link to Matlab
The verifying of method for diagnosing faults is realized in software.
2. single order following control system sensor fault diagnosis method comprising following steps:
1) follow-up signal input process is modeled by Simulirik；Pass through host computer configuration software SIMATIC WinCC
Man-machine interface is established, control experimentation is facilitated, meets research needs；Establish single appearance liquid level control system half infrastest platform in kind
Carry out dynamic simulation；
2) multiplying property malfunctioning module, the without reason switching between barrier module and additivity malfunctioning module are realized by analog mechanical switch,
Acquire experimental data；And the man-machine interface established by SIMATIC WinCC observes the variation of servomechanism liquid level in real time, locates in time
Manage abnormal conditions；
3) based on the data processing model of sliding window, by Fault Isolation in window, malfunction monitoring static models are constructed:
4) fault vectors are calculated using adaptive threshold binarization method, construct Fault Estimation static models:
5) after determining that failure occurs, data variation section t the most apparent is chosen, time variable is obtained and integrates as t=[t1,
t2,...,tw], corresponding dynamic response measurement integrates as y=[y1,y2,...,yw], it is obtained in this section of section according to least square method
The fit equation of data dynamic trend:
It, will be in formula (12) using the additivity malfunction of system and the simulation Dynamic Data Processing of multiplying property malfunction varying strength
Monomial coefficientAs the characteristic value EV extracted:
Wherein P is additivity fault eigenvalue, and Q is multiplying property fault eigenvalue.For the area of quantitative two kinds of fault eigenvalues of analysis
Not, the Monomial coefficient of different faults type is fitted using Linear regression again, it is inclined with additivity failure respectively
The gain of difference and multiplying property failure is as independent variable, and respectively using respective Monomial coefficient as dependent variable, it is special to obtain additivity failure
Value indicative and multiplying property fault eigenvalue static models:
In system jam, the failure strength of additivity He multiplying property is respectively obtained based on Fault Estimation, by additivity failure deviation
It substitutes into corresponding fit equation and is obtained accordingly with the failure gain of multiplying propertyWith
The Monomial coefficient of measured data is utilized when fault location static models fault diagnosisWithWithIt calculates:
Smaller is considered as effectively as a result, thus judging fault type for additivity failure or multiplying property failure.
3. single order following control system sensor fault diagnosis method as claimed in claim 2, it is characterised in that: the step 3)
The determination method of middle failure determination threshold value:
Data absolute value residual error is a criterion, can be used for boundary value when statistical system breaks down:
E=| A-A*| (8)
Wherein e is absolute value residual error, and A is observable data value, A*For desired data value.System response after system jam
Certain control time is needed, individual data is not enough to reflect variation after making corresponding control action in order to avoid system, interception
Reasonable data window is particularly important, can also filter out sensor itself using data window and measures noise and do to fault diagnosis
It disturbs.
When system communication cycle be T, data point sum be N, choose data window size t (data point n), then have:
N=t/T (9)
Freshly harvested dynamic data filling into window and is rejected by earliest data using stack architecture, if shared m data window
Mouthful, then:
M=N-n+1 (10)
The window residual sum of data is added to obtain by data point absolute value residual errors all in the window, if EiFor i-th of window residual error
With, then:
A in formulaiFor i-th of actual data point, Ai *For i-th of expected data point.
Assuming that normal data residual sum meets normal distribution, acquire l group normal data maximum residul difference and, calculating l group data are most
Big residual sum average value:
Standard deviation are as follows:
Using 2 σ as the method for removal abnormal data, then residual sum when failure occurs:
Wherein E*For malfunction monitoring threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910405560.8A CN110187696A (en) | 2019-05-16 | 2019-05-16 | Single order servomechanism sensor fault diagnosis method and system based on dynamic trend |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910405560.8A CN110187696A (en) | 2019-05-16 | 2019-05-16 | Single order servomechanism sensor fault diagnosis method and system based on dynamic trend |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110187696A true CN110187696A (en) | 2019-08-30 |
Family
ID=67716450
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910405560.8A Pending CN110187696A (en) | 2019-05-16 | 2019-05-16 | Single order servomechanism sensor fault diagnosis method and system based on dynamic trend |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110187696A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110609479A (en) * | 2019-10-23 | 2019-12-24 | 中国科学院工程热物理研究所 | Gas turbine sensor fault-tolerant control method based on linear variable parameter model |
| CN110687885A (en) * | 2019-09-09 | 2020-01-14 | 中国计量大学 | Fault diagnosis method and system for regulating valve of first-order constant value control system |
| CN111931349A (en) * | 2020-07-15 | 2020-11-13 | 中南大学 | Multiplicative fault diagnosis method of CSCH (continuous stirred tank reactor) based on dynamic principal component analysis |
| CN113295266A (en) * | 2021-05-07 | 2021-08-24 | 重庆川仪自动化股份有限公司 | Stress wave sensor fault processing method |
| CN114578789A (en) * | 2022-03-04 | 2022-06-03 | 中国计量大学 | Cascade constant value control system regulating valve fault diagnosis method based on data driving |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1734342A1 (en) * | 2005-06-16 | 2006-12-20 | Hitachi, Ltd. | Fault diagnosis apparatus for sensors used in a system |
| CN103558029A (en) * | 2013-10-22 | 2014-02-05 | 重庆建设摩托车股份有限公司 | Abnormal engine sound fault on-line diagnostic system and diagnostic method |
| CN104360679A (en) * | 2014-10-21 | 2015-02-18 | 南京航空航天大学 | Train suspension system fault diagnosis and fault-tolerant control method based on dynamic actuator |
| CN104596780A (en) * | 2015-02-12 | 2015-05-06 | 清华大学 | Diagnosis method for sensor faults of motor train unit braking system |
| CN106647286A (en) * | 2017-02-16 | 2017-05-10 | 吉林大学 | Early failure removal testing method for carrying industrial robot |
| CN107566353A (en) * | 2017-08-21 | 2018-01-09 | 浙江大学 | A kind of industrial control system safety experiment platform for encrypted master experimental study |
| JP2018125979A (en) * | 2017-02-01 | 2018-08-09 | 株式会社デンソー | Sensor fault diagnosis device |
| JP2018173948A (en) * | 2017-03-31 | 2018-11-08 | 西日本電信電話株式会社 | Malfunction diagnosis device, malfunction diagnosis method, and computer program |
-
2019
- 2019-05-16 CN CN201910405560.8A patent/CN110187696A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1734342A1 (en) * | 2005-06-16 | 2006-12-20 | Hitachi, Ltd. | Fault diagnosis apparatus for sensors used in a system |
| CN103558029A (en) * | 2013-10-22 | 2014-02-05 | 重庆建设摩托车股份有限公司 | Abnormal engine sound fault on-line diagnostic system and diagnostic method |
| CN104360679A (en) * | 2014-10-21 | 2015-02-18 | 南京航空航天大学 | Train suspension system fault diagnosis and fault-tolerant control method based on dynamic actuator |
| CN104596780A (en) * | 2015-02-12 | 2015-05-06 | 清华大学 | Diagnosis method for sensor faults of motor train unit braking system |
| JP2018125979A (en) * | 2017-02-01 | 2018-08-09 | 株式会社デンソー | Sensor fault diagnosis device |
| CN106647286A (en) * | 2017-02-16 | 2017-05-10 | 吉林大学 | Early failure removal testing method for carrying industrial robot |
| JP2018173948A (en) * | 2017-03-31 | 2018-11-08 | 西日本電信電話株式会社 | Malfunction diagnosis device, malfunction diagnosis method, and computer program |
| CN107566353A (en) * | 2017-08-21 | 2018-01-09 | 浙江大学 | A kind of industrial control system safety experiment platform for encrypted master experimental study |
Non-Patent Citations (4)
| Title |
|---|
| MARIUSZ BUCIAKOWSKI: "uator fault diagnosis and fault-tolerant control: Application to the quadruple-tank process", 《JOURNAL OF PHYSICS: CONFERENCE SERIES》 * |
| N. TUDOROIU: "Fault detection and diagnosis of the valve actuators in HVAC systems, using frequency analysis", 《2005 INTERNATIONAL CONFERENCE ON INDUSTRIAL ELECTRONICS AND CONTROL APPLICATIONS》 * |
| 于金海: "海洋平台执行器故障最优抗扰自修复容错控制", 《计算机仿真》 * |
| 陆云松 等: "基于定性仿真和模糊知识的离心式压缩机", 《自动化学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110687885A (en) * | 2019-09-09 | 2020-01-14 | 中国计量大学 | Fault diagnosis method and system for regulating valve of first-order constant value control system |
| CN110609479A (en) * | 2019-10-23 | 2019-12-24 | 中国科学院工程热物理研究所 | Gas turbine sensor fault-tolerant control method based on linear variable parameter model |
| CN110609479B (en) * | 2019-10-23 | 2022-08-26 | 中国科学院工程热物理研究所 | Gas turbine sensor fault-tolerant control method based on linear variable parameter model |
| CN111931349A (en) * | 2020-07-15 | 2020-11-13 | 中南大学 | Multiplicative fault diagnosis method of CSCH (continuous stirred tank reactor) based on dynamic principal component analysis |
| CN111931349B (en) * | 2020-07-15 | 2022-03-18 | 中南大学 | Multiplicative fault diagnosis method of CSCH (continuous stirred tank reactor) based on dynamic principal component analysis |
| CN113295266A (en) * | 2021-05-07 | 2021-08-24 | 重庆川仪自动化股份有限公司 | Stress wave sensor fault processing method |
| CN113295266B (en) * | 2021-05-07 | 2022-09-16 | 重庆川仪自动化股份有限公司 | Stress wave sensor fault processing method |
| CN114578789A (en) * | 2022-03-04 | 2022-06-03 | 中国计量大学 | Cascade constant value control system regulating valve fault diagnosis method based on data driving |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110187696A (en) | Single order servomechanism sensor fault diagnosis method and system based on dynamic trend | |
| Xu et al. | PHM-oriented integrated fusion prognostics for aircraft engines based on sensor data | |
| Controllers | Other titles published in this Series: Supervision and Control for Industrial Processes | |
| EP1982301B1 (en) | Method of condition monitoring | |
| CN104573850B (en) | A kind of Power Plant Equipment state evaluating method | |
| CN105046402B (en) | A kind of state evaluating method applied to secondary equipment of intelligent converting station | |
| CN105629958B (en) | A kind of batch process method for diagnosing faults based on sub-period MPCA SVM | |
| CN104756029B (en) | A kind of system of the parts group of monitoring device | |
| CN104504248B (en) | A kind of fault diagnosis modeling method analyzed based on design data | |
| Mohanty et al. | Bayesian statistic based multivariate Gaussian process approach for offline/online fatigue crack growth prediction | |
| CN103499921A (en) | Fault diagnosis method for variable structure fuzzy system sensor and application thereof in flight control system | |
| CN107065834B (en) | The method for diagnosing faults of concentrator in hydrometallurgy process | |
| CN111665066B (en) | Equipment fault self-adaptive upper and lower early warning boundary generation method based on convolutional neural network | |
| CN109675935A (en) | A kind of IPCA operation of rolling on-line fault diagnosis method becoming control limit | |
| JP2000259223A (en) | Plant monitoring device | |
| CN108398637A (en) | A kind of method for diagnosing faults of non-linear Mechatronic Systems | |
| WO2021114320A1 (en) | Wastewater treatment process fault monitoring method using oica-rnn fusion model | |
| CN109976311A (en) | Single order fixed set point control system sensor fault diagnosis method and diagnostic system | |
| CN111860839A (en) | Shore bridge fault monitoring method based on multi-signal fusion and Adam optimization algorithm | |
| CN105678423A (en) | Fault diagnosis system sensor optimal configuration method based on quantitative structure model | |
| CN105137354A (en) | Motor fault detection method based on nerve network | |
| EP1752898A2 (en) | Exception analysis for multimissions | |
| Calderano et al. | An enhanced aircraft engine gas path diagnostic method based on upper and lower singleton type-2 fuzzy logic system | |
| CN206833239U (en) | A kind of thermal power plant's control system fault detection system based on data-driven | |
| EP1724717A2 (en) | Real-time spatio-temporal coherence estimation for autonomous mode identification and invariance tracking |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190830 |