CN106708009A - Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering - Google Patents
Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering Download PDFInfo
- Publication number
- CN106708009A CN106708009A CN201611055922.8A CN201611055922A CN106708009A CN 106708009 A CN106708009 A CN 106708009A CN 201611055922 A CN201611055922 A CN 201611055922A CN 106708009 A CN106708009 A CN 106708009A
- Authority
- CN
- China
- Prior art keywords
- data
- dynamic positioning
- support vector
- vector machine
- wavelet
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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
- G05B23/0281—Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/245—Classification techniques relating to the decision surface
- G06F18/2451—Classification techniques relating to the decision surface linear, e.g. hyperplane
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Optimization (AREA)
- Automation & Control Theory (AREA)
- Mathematical Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Algebra (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention provides a ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering. The method comprises the following steps: collecting data of a dynamic positioning measurement sensor; carrying out wavelet filtering; carrying out pretreatment; carrying out data feature extraction; with the data obtained after data feature extraction being as input feature vector, setting threshold values and feature data labels, and establishing a training set; selecting a radial basis kernel function; finding support vector, solving hyper-plane coefficient, establishing an optimal classification hyperplane and obtaining a support vector machine classification model; processing the collected real-time data of the measurement sensor, then, inputting the processed data to the support vector machine classification model and judging whether a fault occurs and which kind of fault occurs through a classification decision value; and carrying out support vector machine clustering on different faults to differentiate occurrence time and occurrence frequency of the different faults. The method can be used for fault diagnosis of an electric gyrocompass, underwater sound, a tension rope and a DGPS in a ship dynamic positioning measurement system.
Description
Technical field
The present invention relates to a kind of dynamic positioning of vessels measuring system method for diagnosing faults, specifically one kind is based on
The dynamic positioning of vessels measuring system method for diagnosing faults of SVMs cluster.
Background technology
Dynamic positioning system is a kind of control system of closed loop, and its function is by the hydrodynamic force system being automatically or manually controlled
System, makes ship in its operation, can be kept under the job area of regulation and environmental condition its accommodation and bow to；In time to ship
It is extremely important for the navigation operational security of ship that the measuring system of oceangoing ship dynamic positioning carries out fault diagnosis.
Machine learning method is one of focus of current fault diagnosis field research, and SVMs clustering method is one
Plant very effective method for diagnosing faults.
The content of the invention
Sensor generation is more in can effectively solve the problem that dynamic positioning vessel running it is an object of the invention to provide one kind
The diagnosis problem of failure is planted, for the dynamic positioning of vessels based on SVMs cluster that the treatment of failure provides reference measures system
System Multiple faults diagnosis approach.
The object of the present invention is achieved like this：
Step one, is gathered the canonical parameter data of dynamic positioning measurement sensor, is sampled by dynamic positioning measuring system
Measure the bow of ship to and position coordinate data, obtain containing noisy sensing data；
Step 2, to carrying out wavelet filtering containing noisy sensing data；
Data after wavelet filtering are pre-processed by step 3；
Step 4, data characteristics is extracted；
Step 5, the data after data characteristics is extracted set as input feature value according to the species of sensor
Threshold value and characteristic label are put, training set is set up；
Step 6, selects Radial basis kernel function；
Step 7, optimization problem is solved according to training set, finds out supporting vector, solves hyperplane coefficient, sets up optimal
Optimal Separating Hyperplane, obtains support vector cassification model；
Step 8, the real time data of the measurement sensor that will be collected is processed according to step 2 to step 4, treatment
Support vector cassification model is input to afterwards, judges whether to break down by the categorised decision value for obtaining, which kind of class occurs
Type failure, realizes fault diagnosis and the classification of sensor.
The present invention can also include：
1st, different faults are supported with vector machine cluster.
2nd, the wavelet filtering is, using the method for wavelet field threshold filter, to specifically include：
(1) small echo and the wavelet decomposition number of plies are selected, noisy sensing data time series will be contained fast using Mallat
The short-cut counting method carries out orthogonal wavelet transformation；
(2) threshold process is carried out to the wavelet coefficient that decomposition is obtained, chooses soft threshold method and processed；
(3) will be reconstructed through the wavelet coefficient of threshold process, obtain the estimate of the sensing data after denoising.
3rd, the pretreatment is expressed as：
Wherein, X and X' are respectively initial data and pretreated data；XmaxAnd XminRespectively in initial data most
Big value and minimum value；BupperAnd BlowerRespectively expect the upper bound and the floor value of pre-processed results.
4th, the vector machine cluster that is supported to different faults is specifically included：
Fault diagnosis and grouped data are converted into two-dimensional coordinate point (ti,ai)：
Wherein tiIt is the moment that failure occurs,
By the generation of above coordinate points on the two-dimensional coordinate with the time as transverse axis, according to aiIt is worth different display different colours
Point.
The present invention is directed to dynamic positioning of vessels fault diagnosis technology, it is proposed that a kind of failure based on SVMs cluster
Diagnostic method, is used to detect the failure in dynamic positioning of vessels measuring system and is classified and multiple faults problem is gathered
Class, so as to draw what time period easily which kind of failure, can effectively solve the problem that sensor in dynamic positioning vessel running
There is the diagnosis problem of various faults, for the treatment of failure provides reference.The present invention can be used to measure dynamic positioning of vessels
The fault diagnosis of gyro compass, the underwater sound, side tension cords, DGPS in system.
The present invention has the following effects that：
1. can be reduced by wavelet filtering or even reject signal noise completely, it is ensured that some other characteristics of signal are not received
Influence.
2. effectively can in the process of running there are various faults to measuring system by SVMs clustering method to ask
Topic is classified and is clustered, to distinguish the time of origin point of different faults and the frequency occurs, so as to what time period easily to draw
There is which kind of failure, sensor occurs the diagnosis problem of various faults in can effectively solve the problem that dynamic positioning vessel running, is
The treatment of failure provides reference.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention；
Fig. 2 is binary-tree support vector machine multicategory classification workflow diagram in the present invention.
Specific embodiment
The invention will be further described for citing below in conjunction with the accompanying drawings.
With reference to Fig. 1, the present invention is specifically realized according to the following steps：
First, measuring system measurement and output data：
The Three Degree Of Freedom Mathematical Modeling of dynamic positioning vessel is set up, the parameters such as wind, wave, stream, thrust are set, and are entered with MATLAB
Row emulation；By every 0.1 second of measuring system one secondary data of collection, the bow of sampled measurements output ship to and position coordinate data.
2nd, wavelet filtering：
Due to being mingled with noise signal in sensor measurement signal, can be reduced by wavelet filtering and even reject letter completely
Number noise, it is ensured that some other characteristics of signal are unaffected.
Using the method for wavelet field threshold filter, the larger coefficient of Main Basiss amplitude produces this basic by signal of interest
Assuming that to filter.
(1) suitable small echo and the wavelet decomposition number of plies are selected, noisy sensor output time series will be contained and utilized
Mallat fast algorithms carry out orthogonal wavelet transformation.
(2) threshold process is carried out to the wavelet coefficient that decomposition is obtained.Soft threshold method is chosen to be processed.
(3) will be reconstructed through the wavelet coefficient of threshold process, obtain the estimate of the primary signal after denoising, the party
What method was obtained is the optimal estimation value of primary signal.
3rd, data prediction：
It is in order to avoid data deviation in SVMs input data space that certain pretreatment is carried out to input data
Excessive influence classification results, while can also avoid the excessive data of introducing or feature and increase computation complexity.
Original training data refers to change with time, and the bow exported in the measuring system of ship is to the sampling of east northeast coordinate
Value.
Wherein, X and X' are respectively original training data and pretreated data；XmaxAnd XminRespectively in initial data
Maximum and minimum value；BupperAnd BlowerRespectively expect the upper bound and the floor value of pre-processed results.
It is certain it is noted that needing to take identical pretreatment side to training data and test data during pretreatment
Method, so can just cause that classification results have uniformity to improve confidence level.
4th, data characteristics is extracted：
In order to improve the accuracy rate of diagnosis of SVMs, according to the characteristics of 3 kinds of sensor fault types, for train and
The sensing data feature of test will no longer be simple original property value, but data are handled as follows：
It is assumed that
X=[x1,x2,…,xm]=[x (t0),x(t0+τ),…,x(t0+(m-1)τ)]
It is sensor in t0To t0The initial data measured in+(m-1) the τ time periods, wherein τ is the sampling time；
X'=[x1',x'2,…,x'm]=[x'(t0),x'(t0+τ),…,x'(t0+(m-1)τ)]
It is X data by the data that are obtained after pretreatment.
Order
Z=[z1,z2,…,zm-1]=[x'(t0+τ)-x'(t0),…,x'(t0+(m-1)τ)-x'(t0+(m-2)τ)]
It is the characteristic vector extracted from primitive attribute.The dimension of Z fewer than the dimension of X and X' 1, each feature of Z-direction amount
Value represents the difference between adjacent two data, and the size and number feature of this difference determines whether sensor punching occurs
Hit, deviation and output valve stuck at fault.In the case of different faults, by difference process sensing data feature difference very
Substantially, thus with difference process after characteristic as SVMs input feature value, it will greatly enhance support
The rate of correct diagnosis of vector machine.
In order to further improve the rate of correct diagnosis of SVMs, feature can be further extracted.Using translation window method
Collection difference data, every 10 in the difference data that will be obtained increase a sampled data newly and then reject most as one group, often
An early sampled data, and using every group of variance of data as new extraction feature, reconfigure out characteristic vector：
Z'=[z1',z'2,…,z'm-11]
Wherein,
The characteristic vector Z' that will be finally given is used as the input feature vector for training or testing.
5th, training set is set up：
Data after data characteristics is extracted define different threshold values as input feature value according to different sensors
With characteristic label.
Set up training set T={ (z1',y1),…,(z'm-11,ym-11), wherein zi' be by data processing and carry out feature
Input feature value after extraction, yi∈ { -1 ,+1 }, i=1 ..., m-11.When the training set of SVMs is set up, Ying Gen
Rational label is chosen according to different sensors and its actual application environment feature, the correspondence biography in deviation threshold allowed band
Sensor training characteristics data label is -1；It is+1 beyond the character pair data label of deviation threshold allowed band.
6th, kernel function is selected：
Selection radial direction base (RBF) kernel function, data can be mapped to high-dimensional feature space by it from low-dimensional feature space, especially
It is suitable for data distribution and classification problem with non-linear relation.
7th, train：
Optimization problem is solved according to training set, supporting vector is found out, hyperplane coefficient is solved, optimal classification is set up super flat
Face, obtains disaggregated model, and training terminates.
Libsvm tool boxes are used in MATLAB, is instructed with svmtrain function pairs training sample and respective labels
Practice, obtain training pattern model.
8th, test failure diagnosis and classification：
The data that will be collected by being input to SVMs after same data processing and characteristic processing,
Libsvm tool boxes are used in MATLAB, fault diagnosis is carried out with svmpredict function pairs test sample, obtained classification and determine
Plan value.Judge whether to break down by the categorised decision value for obtaining, which kind of type fault occurs, realize that the failure of sensor is examined
Disconnected and classification.
It is analyzed mainly for deviation, impact, the constant three kinds of failures of output valve.In order to be diagnosed to be the biography of this 3 type
At this moment sensor failure will, it is necessary to the data that sensor is measured when normal are classified with data when there occurs failure
It is accomplished by constructing 3 support vector machine classifiers using Binomial Trees.SVM1 graders are by impulse fault and other all classes
Differentiation is not made；SVM2 graders are that deviation fault and other all categories are made into differentiation；SVM3 graders are by output valve
Data of the stuck at fault with sensor when normal make differentiation, as shown in Figure 2：
When the training set of SVMs is set up for gyro compass, SVM1And SVM2In deviation threshold be 2 °, i.e. deviation
Characteristic label is+1 when data are more than 2 °, is otherwise -1；SVM3In deviation threshold be 0.1 °.
When the training set of SVMs is set up for DGPS, due to the positioning precision of DGPS generally reach 1m with
Interior, from sensor fault and control angle, the threshold value that set allows in SVM1 and SVM2 is 3m, and threshold value is 0.1 in SVM3, i.e., just
Sequence label of the difference less than 0.1 is -1, and other are 1.
Training data interval τ is 0.1s.
9th, different faults are supported with vector machine cluster：
Occur because measuring system may in the process of running have various faults, can enter by same fault type
Row cluster, to distinguish the time of origin point of different faults and the frequency occurs.
According to the diagnosis and classification of previous step, fault moment and correspondence fault type are drawn.Count above-mentioned data and carry out
Data processing.Specific processing method is as follows：
Above-mentioned data are converted into two-dimensional coordinate point (ti,ai)：
Wherein tiIt is the moment that failure occurs,
By the generation of above coordinate points on the two-dimensional coordinate with the time as transverse axis, according to aiIt is worth different display different colours
Point.
Different fault types are represented with different colours point, the time that different failures occur so is can be clearly seen that
Point.Above coordinate points are supported vector machine cluster, with libsvm tool boxes, according to institute's diagnostic measurement system sensor
Feature, selects suitable parameter penalty coefficient C and gaussian kernel function q to be clustered.
It is supported after vector machine cluster, the frequency that can be occurred according to the dense degree failure judgement of trouble point, from
And what time period easily drawing which kind of failure, sensor generation is more in can effectively solve the problem that dynamic positioning vessel running
The diagnosis problem of failure is planted, for the treatment of failure provides reference.
Claims (5)
1. it is a kind of based on SVMs cluster dynamic positioning of vessels measuring system Multiple faults diagnosis approach, it is characterized in that：
Step one, gathers the canonical parameter data of dynamic positioning measurement sensor, by dynamic positioning measuring system sampled measurements
The bow of ship to and position coordinate data, obtain containing noisy sensing data；
Step 2, to carrying out wavelet filtering containing noisy sensing data；
Data after wavelet filtering are pre-processed by step 3；
Step 4, data characteristics is extracted；
Step 5, the data after data characteristics is extracted set threshold as input feature value according to the species of sensor
Value and characteristic label, set up training set；
Step 6, selects Radial basis kernel function；
Step 7, optimization problem is solved according to training set, finds out supporting vector, solves hyperplane coefficient, sets up optimal classification
Hyperplane, obtains support vector cassification model；
Step 8, the real time data of the measurement sensor that will be collected is processed, after treatment according to step 2 to step 4
Support vector cassification model is input to, is judged whether to break down by the categorised decision value for obtaining, which kind of type event is occurred
Barrier, realizes fault diagnosis and the classification of sensor.
2. a kind of dynamic positioning of vessels measuring system multiple faults based on SVMs cluster according to claim 1 is examined
Disconnected method, it is characterized in that：Also include being supported different faults vector machine cluster.
3. a kind of dynamic positioning of vessels measuring system multiple faults based on SVMs cluster according to claim 1 is examined
Disconnected method, it is characterized in that：The wavelet filtering is, using the method for wavelet field threshold filter, to specifically include：
(1) small echo and the wavelet decomposition number of plies are selected, noisy sensing data time series will be contained and calculated quickly soon using Mallat
Method carries out orthogonal wavelet transformation；
(2) threshold process is carried out to the wavelet coefficient that decomposition is obtained, chooses soft threshold method and processed；
(3) will be reconstructed through the wavelet coefficient of threshold process, obtain the estimate of the sensing data after denoising.
4. a kind of dynamic positioning of vessels measuring system multiple faults based on SVMs cluster according to claim 1 is examined
Disconnected method, it is characterized in that：The pretreatment is expressed as：
Wherein, X and X' are respectively initial data and pretreated data；XmaxAnd XminMaximum respectively in initial data
And minimum value；BupperAnd BlowerRespectively expect the upper bound and the floor value of pre-processed results.
5. a kind of dynamic positioning of vessels measuring system multiple faults based on SVMs cluster according to claim 2 is examined
Disconnected method, it is characterized in that：The vector machine cluster that is supported to different faults is specifically included：
Fault diagnosis and grouped data are converted into two-dimensional coordinate point (ti,ai)：
Wherein tiIt is the moment that failure occurs,
By the generation of above coordinate points on the two-dimensional coordinate with the time as transverse axis, according to aiIt is worth the point of different display different colours；
With libsvm tool boxes, according to the characteristics of institute's diagnostic measurement system sensor, selection parameter penalty coefficient C and gaussian kernel function
Q is clustered, and is clustered by same fault type, distinguishes the time of origin point of different faults and the frequency occurs；
It is supported after vector machine cluster, what when is the frequency that the dense degree failure judgement according to trouble point occurs draw
Between section easily there is which kind of failure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611055922.8A CN106708009A (en) | 2016-11-25 | 2016-11-25 | Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611055922.8A CN106708009A (en) | 2016-11-25 | 2016-11-25 | Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106708009A true CN106708009A (en) | 2017-05-24 |
Family
ID=58934078
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611055922.8A Pending CN106708009A (en) | 2016-11-25 | 2016-11-25 | Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106708009A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109254577A (en) * | 2018-08-08 | 2019-01-22 | 佛山科学技术学院 | A kind of intelligence manufacture procedure fault classification method and device based on deep learning |
| CN109635864A (en) * | 2018-12-06 | 2019-04-16 | 佛山科学技术学院 | A kind of fault tolerant control method and device based on data |
| CN110583548A (en) * | 2019-10-28 | 2019-12-20 | 海南省民德海洋发展有限公司 | Anchoring system of culture facility |
| CN112560951A (en) * | 2020-12-15 | 2021-03-26 | 哈尔滨工程大学 | Dynamic positioning ship multi-sensor fusion method under multiplicative noise |
| CN114386451A (en) * | 2021-12-03 | 2022-04-22 | 中铁第一勘察设计院集团有限公司 | Contact net dropper fault diagnosis and alarm method based on sensor information perception |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102944418A (en) * | 2012-12-11 | 2013-02-27 | 东南大学 | Wind turbine generator group blade fault diagnosis method |
| CN104442924A (en) * | 2014-11-05 | 2015-03-25 | 杭州南车城市轨道交通车辆有限公司 | All-weather high speed railway vehicle-mounted obstacle detection system and method |
-
2016
- 2016-11-25 CN CN201611055922.8A patent/CN106708009A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102944418A (en) * | 2012-12-11 | 2013-02-27 | 东南大学 | Wind turbine generator group blade fault diagnosis method |
| CN104442924A (en) * | 2014-11-05 | 2015-03-25 | 杭州南车城市轨道交通车辆有限公司 | All-weather high speed railway vehicle-mounted obstacle detection system and method |
Non-Patent Citations (3)
| Title |
|---|
| 周林成等: "基于二代小波变换的信号去噪及其软测量建模", 《计算机与应用化学》 * |
| 宁继鹏: "船舶动力定位容错控制方法研究", 《中国博士学位论文全文数据库工程科技II辑》 * |
| 阎晓娜等: "基于支持向量机的改进高斯核函数聚类算法研究", 《现代电子技术》 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109254577A (en) * | 2018-08-08 | 2019-01-22 | 佛山科学技术学院 | A kind of intelligence manufacture procedure fault classification method and device based on deep learning |
| CN109635864A (en) * | 2018-12-06 | 2019-04-16 | 佛山科学技术学院 | A kind of fault tolerant control method and device based on data |
| CN110583548A (en) * | 2019-10-28 | 2019-12-20 | 海南省民德海洋发展有限公司 | Anchoring system of culture facility |
| CN112560951A (en) * | 2020-12-15 | 2021-03-26 | 哈尔滨工程大学 | Dynamic positioning ship multi-sensor fusion method under multiplicative noise |
| CN114386451A (en) * | 2021-12-03 | 2022-04-22 | 中铁第一勘察设计院集团有限公司 | Contact net dropper fault diagnosis and alarm method based on sensor information perception |
| CN114386451B (en) * | 2021-12-03 | 2024-05-03 | 中铁第一勘察设计院集团有限公司 | Contact net hanger fault diagnosis alarm method based on sensor information perception |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106708009A (en) | Ship dynamic positioning measurement system multiple-fault diagnosis method based on support vector machine clustering | |
| CN112084974B (en) | Multi-label rolling bearing fault diagnosis method based on meta-learning | |
| CN111830408A (en) | Motor fault diagnosis system and method based on edge calculation and deep learning | |
| CN109299705A (en) | Rotary machinery fault diagnosis method based on one-dimensional depth residual error convolutional neural networks | |
| CN110514957A (en) | Substation's automatic detecting method and platform | |
| CN105629958B (en) | A kind of batch process method for diagnosing faults based on sub-period MPCA SVM | |
| CN106447040A (en) | Method for evaluating the health state of mechanical equipment based on heterogeneous multi-sensor data fusion | |
| CN112629851B (en) | Offshore wind turbine gearbox fault diagnosis method based on data enhancement method and image recognition | |
| CN101738998B (en) | System and method for monitoring industrial process based on local discriminatory analysis | |
| CN106017876A (en) | Wheel set bearing fault diagnosis method based on equally-weighted local feature sparse filter network | |
| CN112486097A (en) | Cutting platform wear state monitoring method and system based on model algorithm and readable storage medium | |
| CN102928435A (en) | Aircraft skin damage identification method and device based on image and ultrasound information fusion | |
| CN110440148B (en) | Method, device and system for classifying and identifying leakage acoustic signals | |
| CN110866448A (en) | Flutter signal analysis method based on convolutional neural network and short-time Fourier transform | |
| CN102855485B (en) | The automatic testing method of one grow wheat heading | |
| CN106874950A (en) | A kind of method for identifying and classifying of transient power quality recorder data | |
| CN116625683A (en) | Wind turbine generator system bearing fault identification method, system and device and electronic equipment | |
| CN106339720A (en) | Automobile engine failure detection method | |
| CN114330430A (en) | Elevator fault judgment method and system based on big data characteristic analysis | |
| CN113919389A (en) | GIS fault diagnosis method and system based on voiceprint imaging | |
| CN116991146A (en) | Control method and system of ultrasonic cleaning device | |
| CN106815835A (en) | Damnification recognition method and device | |
| CN110530631A (en) | A kind of gear list type fault detection method based on hybrid classifer | |
| CN115236272A (en) | Gas sensor fault diagnosis method and device under multi-working condition and storage medium | |
| CN115931318A (en) | Fault intelligent diagnosis method, device, equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170524 |