CN112651087A - Train motor fault detection method based on distributed estimation - Google Patents

Train motor fault detection method based on distributed estimation Download PDF

Info

Publication number
CN112651087A
CN112651087A CN202011023709.5A CN202011023709A CN112651087A CN 112651087 A CN112651087 A CN 112651087A CN 202011023709 A CN202011023709 A CN 202011023709A CN 112651087 A CN112651087 A CN 112651087A
Authority
CN
China
Prior art keywords
degradation
fault detection
equation
state
distributed
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
Application number
CN202011023709.5A
Other languages
Chinese (zh)
Inventor
程超
王威珺
王艳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Changchun University of Technology
Original Assignee
Changchun University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Changchun University of Technology filed Critical Changchun University of Technology
Priority to CN202011023709.5A priority Critical patent/CN112651087A/en
Publication of CN112651087A publication Critical patent/CN112651087A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/17—Mechanical parametric or variational design
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/34—Testing dynamo-electric machines
    • G01R31/343—Testing dynamo-electric machines in operation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/10—Geometric CAD
    • G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2111/00—Details relating to CAD techniques
    • G06F2111/02—CAD in a network environment, e.g. collaborative CAD or distributed simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/02—Reliability analysis or reliability optimisation; Failure analysis, e.g. worst case scenario performance, failure mode and effects analysis [FMEA]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/04—Ageing analysis or optimisation against ageing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/08—Thermal analysis or thermal optimisation

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Geometry (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Computer Hardware Design (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The invention discloses a train motor fault detection method based on distributed estimation, and particularly relates to the field of fault detection. The method improves the consistency problem of diagnosis of a plurality of single sensors of the existing high-speed train fault detection system and the problem of state estimation deviation caused by system state degradation. The method comprises the steps of constructing a distributed topology and state estimation model for the position structures of a plurality of sensors of a high-speed train traction motor system and collected data, modeling state degradation of the system, simulating real-time change characteristics of the system, setting a fault detection threshold value by combining statistical characteristics of a residual error generator on the basis of state estimation, and further achieving detection.

Description

Train motor fault detection method based on distributed estimation
Technical Field
The invention relates to a fault detection method for multi-sensor and state correction, which realizes fault detection based on a state estimation model of a target system. The method can be particularly applied to the following aspects: firstly, for a traction motor system monitored by a distributed sensor, the method can realize the state estimation of the motor system. Secondly, the degradation phenomenon exists in most mechanical systems, the internal performance of the system is reduced along with the prolonging of the running time, and the method can simulate the actual degradation state through modeling. And finally, combining the system state and the degradation state to realize the fault detection function of the traction motor system.
Background
With the development of science and the progress of technology, mechanical systems are larger and more complex, and once accidents happen to the large systems, huge casualties and property losses are caused. Therefore, the safety and reliability of complex mechanical systems are very critical factors to be considered. The fault detection technology is an important method for improving the system reliability and reducing the accident risk. Currently, in order to improve the reliability of related systems, fault detection usually employs a redundancy concept to monitor, locate and identify faults. The basic idea of hardware redundancy is to use the same components with the same input signals to compare the repeated output signals to make diagnostic decisions by means of limit checking and majority voting, among other methods, which is reliable but expensive and takes up more space. Diagnostic methods that analyze redundancy are less costly than hardware redundancy methods, but are more challenging due to environmental noise, unavoidable modeling errors, and complexity of system dynamics and control architecture.
At present, for a data-based fault detection method widely applied to fault detection of a traction system, the most important step is to extract characteristic information of fault data, but for a complex mechanical system, a fault once caused can cause a result which is difficult to measure, and the method is an indispensable process for timely repair and periodic maintenance of the complex mechanical system. This directly results in the shortage of the amount of invalid data, and if only a small amount of sample data or data with insignificant characteristic signals is used for research, the accuracy of fault characteristic extraction cannot be ensured, thereby directly influencing the effect of system fault detection. In addition, in an actual system, data acquired by nodes at different positions in the same monitoring area are not completely the same, which indirectly affects the consistency of information acquisition and reduces the diagnosis precision, i.e., the state estimation method does not fully consider the influence of adjacent nodes and cannot ensure the consistency of node states. Finally, the consideration of the system state degradation problem of the traction motor is deficient, the state degradation is an inevitable reality, if the degradation problem is not fully considered, the false alarm rate of fault detection is increased, and the diagnosis accuracy is reduced. In summary, a new method is needed to consider the actual operation characteristics of the traction system as much as possible, improve the state estimation accuracy of the traction motor system and further improve the fault detection performance.
Disclosure of Invention
The invention aims to detect the fault of a traction system of a high-speed train, and realize distributed modeling and degradation process modeling by state estimation and degradation simulation of an actual system so as to further realize fault detection of the whole system.
The invention specifically comprises the following steps:
step 1: the method comprises the steps of collecting measurement data values of a system in an industrial process in a normal operation state, wherein the measurement data values comprise measurement values of a plurality of sensor nodes.
Step 2: and constructing an actual sensor topology according to the physical structure of the actual traction motor system.
And step 3: and constructing a distributed state estimation model by taking the topological structure as a support. Firstly, considering the influence of a sensor in the process of acquiring information, processing a system equation to obtain an equivalent system equation:
Figure RE-GDA0002959567010000021
secondly, constructing a distributed state estimation model:
Figure RE-GDA0002959567010000022
the filter parameter H to be obtained needs to be obtained on the premise of minimum mean square error. The principle is equivalent to a conditional extremum, and the obtained optimal solution H is represented as:
Figure RE-GDA0002959567010000023
the distributed state estimation curve and the error covariance curve are shown in fig. 6 and 7.
And 4, step 4: and establishing an operation state degradation model based on the actual traction motor data. The method specifically comprises the following steps: taking the degradation process driven by the Brownian motion as an example, a nonlinear process model is taken as a basic degradation process model to carry out parameter estimation, and finally, a state correction factor is obtained.
And 5: and calculating a fault detection index of the measured data, and performing fault detection analysis on the data. The method specifically comprises the following steps: a residual generator is constructed and statistical properties given to the residual generator are calculated. Based on the characteristic, under the derivation of the Chebyshev inequality, a fault detection threshold value under the premise of a certain confidence coefficient is given, so that fault detection is carried out.
The beneficial effects of the invention specifically comprise:
firstly, the neighbor node information is fully considered, the state estimation of the target node is optimized, a parameter analysis solution of the optimal state estimation of the multi-sensor system is given, and the consistency of the multi-sensor monitoring of the target system is realized. Compared with single sensor monitoring, the robustness is stronger, and the stability is better; secondly, the noise problem and the initial information problem of the actual system are unavoidable, the method corrects the initial information in an unbiased constraint mode, and processes the colored measurement noise in a differential construction mode, so that the conventional assumption of white noise is avoided; finally, considering the degradation process of the actual system, the method combines the non-stationary process to establish a non-linear degradation model for representing the degradation trend of the system state. In order to improve the parameter estimation problem of the degradation model, a degradation increment needs to be constructed, the statistical characteristic of the degradation increment is calculated, and a parameter analysis solution of the degradation process can be calculated by combining a maximum likelihood estimation method. In order to verify the effectiveness of the improved high-speed train traction motor fault detection system, monitoring data of the high-speed train traction motor system is used as a simulation case for verification. The method comprises the steps of firstly collecting 600 observation data of a traction motor system of the high-speed train, wherein the first 300 observation data are normal operation data, and the last 300 observation data comprise fault data exceeding a threshold value. The failure decision after selecting the threshold is shown in fig. 9.
Drawings
FIG. 1 is a flow chart of a high speed train traction system fault detection;
FIG. 2 is a diagram of a distributed topology architecture;
FIG. 3 is a flow diagram of distributed state estimation;
FIG. 4 is a state degradation model building flow diagram;
FIG. 5 is a flow chart of fault detection threshold generation;
FIG. 6 is a graph of distributed state estimation;
FIG. 7 is a plot of a distributed state estimation error covariance trace;
FIG. 8 is a state estimation diagram after state correction;
fig. 9 is a fault detection decision diagram.
Detailed Description
Embodiments of the invention are further described below with reference to the following figures and specific examples:
as shown in fig. 1, the method for detecting the fault of the traction system of the high-speed train specifically comprises the following steps:
step 1, collecting measurement data values in the normal operation state of a system in an industrial process, wherein the measurement data values comprise measurement values of a plurality of sensor nodes.
Step 2: and constructing an actual sensor topology according to the physical structure of the actual traction motor system. Taking a monitoring system with 6 nodes as an example, the actual acquisition environment can be abstracted into a distributed topology in the form of fig. 2. Where circles represent 6 sensor nodes and triangles represent the monitored area.
And step 3: and constructing a distributed state estimation model by taking the topological structure as a support. As shown in fig. 3, firstly, the system equation is processed to obtain an equivalent system equation in consideration of the influence on the sensor in the process of acquiring information. The method specifically comprises the following steps: the invention considers the gain attenuation and the colored measurement noise of the sensor, and constructs the model of the target system into the following form:
Figure RE-GDA0002959567010000031
in order to better accord with the change situation of the actual state, the method considers the influence of the temperature change on the system state, and changes the model of the target system into a bilinear form as follows:
Figure RE-GDA0002959567010000032
where k is the discrete time index, u (k) is the control signal, w (k) is the process noise and has a mean of 0 and a covariance of Q. v (k) is the measurement noise of the ith sensor. λ is the gain attenuation of the ith sensor, A, B, Ci, Di are known matrices of dimensional matching, and N (k) u (k) x (k) is a bilinear term.
Firstly, colored noise needs to be processed in order to realize distributed estimation of a multi-sensor system, a differential form is constructed to replace an original measurement equation, and auxiliary signals are defined as follows:
zi(k)＝yi(k)-ψi(k-1)vi(k-1) (6)
the equivalent system after replacement by the auxiliary signal is:
Figure RE-GDA0002959567010000041
the method gives the following definitions to the distributed state estimation structure according to the system model:
Figure RE-GDA0002959567010000042
wherein
Figure RE-GDA0002959567010000043
Is the ith passEstimating the state of the sensor node, wherein Ni is a set formed by the node i and the neighbor nodes thereof, H (k) is the filter gain of the multi-sensor, and the optimal solution is as follows:
Figure RE-GDA0002959567010000044
the state estimation of the target node is corrected through the information of the neighbor node, and the state estimation method under the distributed monitoring environment is obtained. By passing
Figure RE-GDA0002959567010000045
To satisfy the unbiased property of the state estimation and combine the error function of the state estimation
Figure RE-GDA0002959567010000046
And obtaining a filter gain parameter on the premise of minimum error mean square error. The specific state estimation performance and error covariance trace is shown in fig. 6 and fig. 7.
And 4, step 4: and establishing an operation state degradation model based on the actual traction motor data. The method specifically comprises the following steps: taking the degradation process driven by the Brownian motion as an example, a nonlinear process model is taken as a basic degradation process model to carry out parameter estimation, and finally, a state correction factor is obtained. The specific flow is shown in fig. 4, and in order to better describe the degradation phenomenon existing in the system, the method considers degradation modeling based on a non-stationary non-linear process. Similar to the process of the above model, considering the changes of random impact factors and environmental factors in the degradation model, a nonlinear degradation model of the following process is proposed:
Figure RE-GDA0002959567010000047
wherein, X (t) is a degradation process driven by standard drift Brownian motion B (t), mu (u; theta) is a nonlinear function, J represents the amplitude of random impact, N represents the number of random impact, and u (t) x (t) represents a parameter of environmental factors for degrading the system state. It can be seen that when μ (u; θ) is μ, the degradation process becomes a linear degradation model. The parameters of the model can be obtained by constructing degradation increments and calculating the statistical properties thereof.
The degradation increment can be expressed as:
Figure RE-GDA0002959567010000048
where Δ may represent the effect of the environment (medium) on the state degradation, an average of the two sides of the above equation may be taken:
E[X(i+1)]＝aE[t(i+1)b-t(i)b]-E[Δ] (12)
the covariance can be expressed as:
Figure RE-GDA0002959567010000051
wherein σB(. represents. sigma.)BThe item concerned.
Let Q { { B [ t (i +1) ]]-B[t(i)]}{B[t(j+1)]-B[t(j)]} the variance of the available degradation increments is
Figure RE-GDA0002959567010000052
Carrying out maximum likelihood estimation on the parameters of the degradation process, and constructing a likelihood function:
Figure RE-GDA0002959567010000053
taking logarithm and pair a and sigmaBCalculating a partial derivative:
Figure RE-GDA0002959567010000054
Figure RE-GDA0002959567010000055
a and sigmaBIs brought into a log-likelihood functionAnd calculating the estimated value of b by maximizing the log-likelihood function by a simplex method.
And calculating a correction factor according to the obtained state degradation curve:
Figure RE-GDA0002959567010000056
the state estimation performance curve after state correction is shown in fig. 8.
And 5: as shown in fig. 5, a fault detection index of the measurement data is calculated, and fault detection analysis is performed on the data. The method specifically comprises the following steps: a residual generator is constructed and statistical properties given to the residual generator are calculated. Based on the characteristic, under the derivation of the Chebyshev inequality, a fault detection threshold value under the premise of a certain confidence coefficient is given, so that fault detection is carried out.
The following residual generator needs to be constructed:
Figure RE-GDA0002959567010000057
wherein
Figure RE-GDA0002959567010000058
Therefore, the method comprises the following steps:
Figure RE-GDA0002959567010000061
given a residual generator satisfying the following distribution
Figure RE-GDA0002959567010000062
The final fault detection threshold is given using the chebyshev inequality:
Figure RE-GDA0002959567010000063
in order to verify the effectiveness of the improved high-speed train traction motor fault detection system, monitoring data of the high-speed train traction motor system is used as a simulation case for verification. Firstly, 600 observation data of a high-speed train traction motor system are collected, wherein the first 300 observation data are normal operation data, and the last 300 observation data comprise fault data exceeding a threshold value. All data comes from traction motor stator temperature of the traction motor system of the high speed train itself. Considering that the sampling frequency of an actual data acquisition system is lower, and the method is more prone to the test of the whole life cycle of the system, the 600 screened observation data are selected in an averaging mode, the complete test is carried out from the data in a normal state to the data derived from a degradation state, and the detection result verifies the effectiveness of the invention.
The embodiments of the present invention are described above, but the description is for the convenience of understanding the present invention and is not intended to limit the present invention, and those skilled in the art should make modifications and variations within the spirit and scope of the present invention.

Claims (4)

1. A train motor fault detection method based on distributed estimation is characterized by comprising the following steps:
collecting a plurality of sensor node measured values of a monitoring area in an industrial process for analysis;
constructing an actual sensor topology according to the physical structure of an actual traction motor system, and abstracting an actual acquisition environment into a distributed topology structure by taking a monitoring system with 6 nodes as an example;
thirdly, constructing a distributed state estimation model by taking the topological structure as a support;
establishing an operation state degradation model based on actual traction motor data;
and step five, constructing a residual error generator.
2. The train motor fault detection method based on distributed estimation according to claim 1, wherein the third step is specifically:
firstly, the influence of colored noise on the sensor in the information acquisition process is processed to obtain an equivalent system equation as shown in formula 1:
Figure DEST_PATH_IMAGE001
(1)
considering the influence of the actual temperature on the system state, the bilinear form is considered to obtain an equation as shown in formula 2 in the system model:
Figure DEST_PATH_IMAGE002
(2)
constructing a difference form to replace the original measurement equation, and defining an auxiliary signal as formula 3:
Figure DEST_PATH_IMAGE003
(3)
the equivalent system after substitution by the auxiliary signal is as follows:
Figure DEST_PATH_IMAGE004
(4)
in combination with the actual topology, the distributed state estimation equation is defined as shown in equation 5:
Figure DEST_PATH_IMAGE005
(5)
wherein
Figure DEST_PATH_IMAGE006
Is the state estimation of the ith sensor node, and Ni is the set formed by the node i and the neighbor nodes thereof;
h (k) is the filter gain of the multi-sensor, and the optimal solution is as follows:
Figure DEST_PATH_IMAGE007
(6)
by passing
Figure DEST_PATH_IMAGE008
To satisfy the unbiased property of the state estimation and combine the error function of the state estimation
Figure DEST_PATH_IMAGE009
And obtaining a filter gain parameter on the premise of minimum error mean square error.
3. The train motor fault detection method based on distributed estimation according to claim 1, wherein the fourth step is specifically:
the invention considers the degradation modeling based on the non-stable non-linear process, considers the change of random impact factors and environmental factors in a degradation model, and provides a non-linear degradation model as the formula 7:
Figure DEST_PATH_IMAGE010
(7)
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE011
is a standard drift of Brownian motion
Figure DEST_PATH_IMAGE012
The process of degradation of the drive is such that,
Figure DEST_PATH_IMAGE013
j represents the amplitude of the random impact, N represents the number of times of the random impact, u (t) x (t) represents the environmental factor versus the system stateParameters of state degradation;
the degradation delta can be expressed as equation 8:
Figure DEST_PATH_IMAGE014
(8)
wherein
Figure DEST_PATH_IMAGE015
The influence of the environment (medium) on the state degradation can be represented, and the statistical characteristics of the degradation increment can be represented as formula 9 and formula 10:
Figure DEST_PATH_IMAGE016
(9)
Figure DEST_PATH_IMAGE017
(10)
wherein the content of the first and second substances,
Figure DEST_PATH_IMAGE018
represent and
Figure DEST_PATH_IMAGE019
related item, order
Figure DEST_PATH_IMAGE020
Variance of available degradation increments of
Figure DEST_PATH_IMAGE021
；
The likelihood function is constructed as in equation 11:
Figure DEST_PATH_IMAGE022
(11)
the parameters may be represented by equation 12:
Figure DEST_PATH_IMAGE023
(12)
the correction factor can be expressed as equation 13:
Figure DEST_PATH_IMAGE024
(13)。
4. the train motor fault detection method based on distributed estimation according to claim 1, wherein the step five specifically comprises:
Figure DEST_PATH_IMAGE025
(14)
given a distribution of residual generators that satisfies equation 15:
Figure DEST_PATH_IMAGE026
(15)
the final fault detection threshold is obtained by using the chebyshev inequality, as shown in formula 16:
Figure DEST_PATH_IMAGE027
(16)。
CN202011023709.5A 2020-09-25 2020-09-25 Train motor fault detection method based on distributed estimation Pending CN112651087A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011023709.5A CN112651087A (en) 2020-09-25 2020-09-25 Train motor fault detection method based on distributed estimation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011023709.5A CN112651087A (en) 2020-09-25 2020-09-25 Train motor fault detection method based on distributed estimation

Publications (1)

Publication Number Publication Date
CN112651087A true CN112651087A (en) 2021-04-13

Family

ID=75346571

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011023709.5A Pending CN112651087A (en) 2020-09-25 2020-09-25 Train motor fault detection method based on distributed estimation

Country Status (1)

Country Link
CN (1) CN112651087A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113805056A (en) * 2021-09-18 2021-12-17 中车大连电力牵引研发中心有限公司 Method for intelligently monitoring insulation fault of traction motor on line

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100222897A1 (en) * 2009-03-02 2010-09-02 Toyota Motor Engineering & Manufacturing North America, Inc. Distributed fault diagnosis
CN103124105A (en) * 2012-03-27 2013-05-29 湖南大学 Wireless intelligent sensor network system for monitoring states of intelligent substation devices
CN105460027A (en) * 2016-01-04 2016-04-06 唐智科技湖南发展有限公司 Vehicle-mounted distributed running part fault diagnostic system for urban rail transit train
CN107329083A (en) * 2017-07-28 2017-11-07 南京航空航天大学 For the method for diagnosing faults of bullet train traction electric machine nonlinear system sensor
CN114877926A (en) * 2022-04-25 2022-08-09 江南大学 Sensor fault detection and diagnosis method, medium, electronic device and system

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100222897A1 (en) * 2009-03-02 2010-09-02 Toyota Motor Engineering & Manufacturing North America, Inc. Distributed fault diagnosis
CN103124105A (en) * 2012-03-27 2013-05-29 湖南大学 Wireless intelligent sensor network system for monitoring states of intelligent substation devices
CN105460027A (en) * 2016-01-04 2016-04-06 唐智科技湖南发展有限公司 Vehicle-mounted distributed running part fault diagnostic system for urban rail transit train
CN107329083A (en) * 2017-07-28 2017-11-07 南京航空航天大学 For the method for diagnosing faults of bullet train traction electric machine nonlinear system sensor
CN114877926A (en) * 2022-04-25 2022-08-09 江南大学 Sensor fault detection and diagnosis method, medium, electronic device and system

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
CHAO CHENG 等: ""State-Degradation-Oriented Fault diagnosis for High-Speed Train Running Gears System "", 《SENSORS》 *
张铭 等: ""城市轨道交通设备故障聚类与贝叶斯网络预警"", 《计算机工程与应用》 *
李颖 等: ""基于控制器局域网总线的车载轨道检测***设计研制"", 《中国铁道科学》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113805056A (en) * 2021-09-18 2021-12-17 中车大连电力牵引研发中心有限公司 Method for intelligently monitoring insulation fault of traction motor on line
CN113805056B (en) * 2021-09-18 2023-12-19 中车大连电力牵引研发中心有限公司 Intelligent on-line traction motor insulation fault monitoring method

Similar Documents

Publication Publication Date Title
CN109857094B (en) Two-stage Kalman filtering algorithm-based aeroengine fault diagnosis method
Xu et al. An incorrect data detection method for big data cleaning of machinery condition monitoring
US7415328B2 (en) Hybrid model based fault detection and isolation system
CN112284440B (en) Sensor data deviation self-adaptive correction method
CN110008253B (en) Industrial data association rule mining and abnormal working condition prediction method
Jiang Sensor Fault Detection and Isolation Using System Dynamics Identification Techniques.
Alippi et al. A cognitive fault diagnosis system for distributed sensor networks
CN111460392B (en) Magnetic suspension train and suspension system fault detection method and system thereof
CN107725283A (en) A kind of fan trouble detection method based on depth belief network model
CN102880170A (en) System failure early warning method based on baseline model and Bayesian factor
CN107644148B (en) On-orbit satellite abnormal state monitoring method and system based on multi-parameter association
CN108257365B (en) Industrial alarm design method based on global uncertainty evidence dynamic fusion
CN111103137A (en) Wind turbine gearbox fault diagnosis method based on deep neural network
CN111324110A (en) Fermentation process fault monitoring method based on multiple shrinkage automatic encoders
CN112651087A (en) Train motor fault detection method based on distributed estimation
CN109657329B (en) Sensor fault and structure damage identification method based on improved adaptive particle filtering
CN117668684A (en) Power grid electric energy data anomaly detection method based on big data analysis
Kozionov et al. Wavelet-based sensor validation: Differentiating abrupt sensor faults from system dynamics
CN117454221A (en) Abnormality identification method for multi-sensor fusion of high-speed rail traction braking system
CN116627116A (en) Process industry fault positioning method and system and electronic equipment
CN115358294A (en) Micro fault detection method for high-speed train traction system
CN110045716B (en) Method and system for detecting and diagnosing early fault of closed-loop control system
CN114357667A (en) Engine starting state fault detection method based on RESID recursive structure identification
Sundaram et al. Aircraft engine health monitoring using density modelling and extreme value statistics
Chen et al. A multimode anomaly detection method based on oc-elm for aircraft engine system

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: 20210413