CN102589890A - Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences - Google Patents
Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences Download PDFInfo
- Publication number
- CN102589890A CN102589890A CN2012100507549A CN201210050754A CN102589890A CN 102589890 A CN102589890 A CN 102589890A CN 2012100507549 A CN2012100507549 A CN 2012100507549A CN 201210050754 A CN201210050754 A CN 201210050754A CN 102589890 A CN102589890 A CN 102589890A
- Authority
- CN
- China
- Prior art keywords
- cpn
- steam turbine
- independent
- neutral nets
- fault
- 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
Images
Abstract
The invention relates to an integrated fault diagnostic method of a steam turbine based on a CPN (counter propagation network) and D-S (Dempster-Shafer) evidences. The integrated fault diagnostic method utilizes the advantages of a CPN and a D-S evidence theory and takes steam turbine vibration state parameters collected by different sensors as independent data samples aiming at the steam turbine of a power plant; and after being subjected to the characteristic extraction and processing, and the parameter normalization processing, the independent data samples are input into respectively independent CPN to be trained, so that each independent CPN can form a non-linear mapping of a fault symptom to a fault mode. And then, a D-S evidence theory method is used for carrying out further data fusion diagnosis on a diagnosed result of each CPN, so as to improve the accuracy and the reliability of the diagnosed result. Therefore, more accurate diagnosis and analysis on the current operation state of the steam turbine is implemented.
Description
Technical field
The present invention relates to a kind of Steam Turbine Fault Diagnosis, the integrated method for diagnosing faults of more particularly to a kind of steam turbine based on CPN networks and D-S evidences.
Background technology
Turbo-generator Set is the capital equipment of power-generating enterprise, and steam turbine constantly produces a large amount of different information in the process of running, and these information reflect the running status of steam turbine in terms of different.Because steam turbine structure is more complicated, running environment is more special, unit failure situation is inevitable, and failure mode is various, there are many fault characteristic signals, including temperature, pressure, amplitude, voltage, electric current, flow, power etc., and the fault signature wherein included with vibration signal is most, it more can rapidly, directly reflect the running status of unit, also be easier monitored and analyze.Therefore, the research of Steam Turbine Fault Diagnosis is carried out, for finding failure cause and position, the operational reliability of unit is improved, ensures that unit safety, economical operation are significant.
Information fusion technology be by from different purposes, different time, different spaces information, automatically analyzed by computer technology under certain criterion and comprehensive, unified feature representation information is formed, so that system obtains the technology than single piece of information source more accurate, more complete estimation and judgement.It provides first and then reliable method for the information processing and decision problem for solving the information age.In multisensor syste, the general principle of information fusion is just as the process of human brain integrated treatment information, by various sensors provide in the time or spatially redundancy or complementary information, according to certain criterion optimum organization, produce the reasonable description to observing environment.Information fusion technology is on the basis of integrated treatment is carried out to the information of multisensor, to realize real-time state monitoring, the mutation forecasting even fault diagnosis of signal and the alarm to equipment applied to equipment fault diagnosis.In fault diagnosis field, information fusion process can be divided into data Layer, characteristic layer, three levels of decision-making level, and these three levels are represented to the abstract of the different depths of initial data respectively.Compared with traditional fault diagnosis technology, information fusion technology has higher diagnosis accuracy and confidence level.
Evidence theory in information fusion technology is a kind of reasoning method under uncertainty, is proposed first by Dempster, and is further grown up by Shafer, forms a set of mathematical theory on evidence, thus also known as D-S evidence theory.D-S evidence theory is mainly according to reliability function computing, and it is a kind of data fusion method for solving uncertain problem.D-S evidence theory does not need prior information, and describes uncertain information using the method for interval estimation, solves on probabilistic method for expressing.Therefore, D-S evidence theory has the ability of very strong processing uncertain information, do not know that with uncertain aspect and accurately in terms of reflection evidence-gathering there is very big flexibility distinguishing, expression and synthesis for uncertain information provide strong method, decision level information fusion is particularly adapted to, is widely applied in fields such as pattern-recognition, fault diagnosis, problem prediction, expert systems.
Counterpropagation network（Counter-propagation Network, abbreviation CPN）It is a kind of novel feature mapping network of rising in recent years, it can overcome BP neural network conventional to be at present absorbed in local minimum point, slow and poor astringency the defect of pace of learning.CPN neutral nets get up Kohonen Feature Mappings network and the basic competitive type network integrations of Grossberg, have played respective speciality, it is adaptable to fault diagnosis, pattern classification, function approximation, statistical analysis and data compression etc..
The content of the invention
The present invention be directed to the important sex chromosome mosaicism of fault diagnosis of the steam turbine comprising vibration signal, propose a kind of integrated method for diagnosing faults of the steam turbine based on CPN networks and D-S evidences, using turbine rotor vibration analogue experiment installation as test platform, it regard the Steam Turbine Vibration state parameter collected from different sensors as independent data sample, using CPN neutral nets and the advantage of D-S evidence theory, realize and steam turbine current operating conditions are carried out more accurately to diagnose and analyze.
The technical scheme is that：A kind of integrated method for diagnosing faults of steam turbine based on CPN networks and D-S evidences, method is comprised the following specific steps that：
1）It regard the Steam Turbine Vibration state parameter collected from sensor on turbine rotor vibration simulator stand as independent data sample, each independent CPN neutral nets are input to after the processing of feature extraction, processing and parameter normalization to be trained so that each independent CPN neutral nets can form failure symptom to the Nonlinear Mapping of fault mode；
2）Carried out using each independent CPN neutral nets after fault diagnosis, the output valve of CPN neutral nets be normalized using formula below, be converted into should CPN neutral nets various malfunctions basic probability assignment BPA,
Wherein
Represent the
Individual evidence in CPN networks
Individual output result；
Represent
Individual evidence is to stateBPA；
3）Further data fusion is carried out to the diagnostic results of each failure CPN neutral nets according to the fusion rule of D-S evidence theory, try to achieve its it is orthogonal and
, the basic probability assignment of each status fault state, obtains fault diagnosis result after as merging.
The beneficial effects of the present invention are：Steam turbine integrated method for diagnosing faults of the invention based on CPN networks and D-S evidences, utilize CPN neutral nets and the advantage of D-S evidence theory, for power plant steam turbine, it regard the Steam Turbine Vibration state parameter collected from different sensors as independent data sample, each independent CPN neutral nets are inputted after the processing of feature extraction, processing and parameter normalization to be trained so that each independent CPN neutral nets can form failure symptom to the Nonlinear Mapping of fault mode.Occur simultaneously because each independent fault signature and fault mode are present, further data fusion diagnosis is carried out to the diagnostic result of each CPN neutral net using D-S evidence theory method, steam turbine current operating conditions are carried out more accurately to diagnose and analyze so as to realize.
Brief description of the drawings
Fig. 1 is vibration faults of turbine rotor simulator stand structural representation of the present invention；
Fig. 2 is the steam turbine integrated fault diagnosis model figure of the invention based on CPN networks and D-S evidence theory；
Fig. 3 is the topology diagram of CPN neutral nets of the present invention.
Embodiment
In actual motion, the common failure of steam turbine has following several：Rotor unbalance, rotor misalignment, oil whirl and oil whip, fulcrum bearing loosen, rotor impact and rub and axle crack fault etc..Wherein, rotor unbalance is the most common failure of steam turbine, the reason for causing steam turbine synchronous vibration may have original quality imbalance, rotor thermal unbalance, rotor thermal bending, rotary part to come off and rotor part fouling etc., and these reasons will all cause the imbalance fault of rotor.Under certain working condition, rotor-bearing system also occurs that the problem of oil whirl and oil whip, and high vibration of the rotor journal in oil film will directly result in the damage of element part.
Steam turbine integrated method for diagnosing faults of the invention based on CPN neutral nets and D-S evidence theory is using turbine rotor vibration analogue experiment installation as test platform, by changing rotor speed, axis rigidity, mass unbalance, the friction of bearing and impact condition and the form of shaft coupling, the test platform can effectively reproduce the common vibration fault produced by steam turbine.Test platform includes six eddy current displacement sensors, a photoelectric sensor and two magnetoelectric velocity transducers, the real time data of testing stand during for gathering fault simulation experiment.The present invention is simulated on test platform to the most common failure of steam turbine, while being sampled to whole boosting velocity procedure, to provide multisensor in fault data not in the same time.Need the diverse location in rotating shaft that multiple sensors are installed in order to obtain required fault-signal, when every kind of malfunction test, while the signal for choosing sensor carries out convergence analysis.
The present invention illustrates embodiment by taking oil whirl and oil whip failure as an example.
When steam turbine axle journal is rotated in bearing shell, the gap between bearing shell and axle journal can form oil film, and the hydrodynamic pressure of oil film makes axle journal have bearing capacity.Once the bearing capacity of oil film reaches balance with external applied load, axle journal is just in equilbrium position, and when rotating shaft is by certain external disturbance, bearing film will not only produce along the elastic restoring force of offset direction to keep the balance with external load, and produce one perpendicular to the tangential unstability component of offset direction to drive rotor formation and rotor direction of rotation identical whirling motion.Because the eddy velocity of axle journal is close to the half of rotating speed, often it is referred to as " half-speed vortex ".After generation whirling motion, no matter amplitude size, rotor all loses stability, i.e., so-called rotor unstability.Under certain condition, though rotor unstability, axle journal may the only whirling motion in the range of a very little, i.e. whirling motion amplitude very little, may be still stable from the perspective of machine run.Whirling motion angular speed increases with the raising of working speed, but always it is about equal to the half of velocity of rotation, when rotating shaft rotating speed is increased to slightly above after 2 times of first critical speed, first critical speed of the whirling motion angular speed of half-speed vortex just with rotating shaft coincides, so as to produce resonance, strong covibration is shown as.Orbit of shaft center suddenly becomes the irregular curve of diffusion simultaneously, if continuing to improve rotating speed, the frequency of vortex motion of rotor keeps constant, consistently equal to the first critical speed of rotor, and this phenomenon is oil whip.
Due to being limited by limit speed, when doing oil whip experiment, this testing stand needs that rotor-support-foundation system first critical speed is down into below 4000rpm by way of increasing wheel disc, to be observed that oil whip phenomenon near 9000rpm.Vibration faults of turbine rotor simulator stand is as shown in figure 1, motor 1, bearing I 4, bearing II 7, the bearing IV 16 of bearing III 9 are fixed on pedestal 17.There is shaft coupling 2 to connect between motor 1 and bearing I 4, there is shaft coupling 8 to connect between bearing II 7 and bearing III 9, the testing stand uses twin spans rotor, from one wheel disc of installing at whirling motion bearing block 2/3 in oil whirl rotating shaft, it is 5 wheel disc A, 6 wheel disc B, 11 wheel disc C respectively equipped with three wheel discs on rotor.Along axial arranged a photoelectric sensor 3 and two current vortex sensors 10,12 of rotating shaft, wherein photoelectric sensor 3 is located at the right-hand member of shaft coupling 2 between motor 1 and bearing I 4, rotating speed and phase for measuring rotor；Current vortex sensor 10 is is arranged vertically and current vortex sensor 12 is arranged horizontally, the amplitude for rotor.X, Y-direction eddy current sensor are installed on eddy current sensor support, and distance of the adjustment probe to rotor surface on request；The output signal line of preamplifier is connected on vialog, and vialog is connected on host computer.During experiment, open needle valve lubricating cup 15, until oil drips from bearing, firing test motor, gradually steps up rotating speed, about in 3000~4000rpm, generation whirling motion, if whirling motion does not occur immediately, axle is lightly lifted in whirling motion bearing end using nylon preload rod, continues to raise rotating speed until oil whip failure occurs for testing stand.Excessive to prevent from refueling, testing stand has an oil return opening 14 on the side of bearing IV, and unnecessary oil is flowed into oil-containing groove 13.
The present invention is test platform for vibration faults of turbine rotor analogue experiment installation, it regard the Steam Turbine Vibration state parameter collected from different sensors as independent data sample, each independent CPN neutral nets are inputted after the processing of feature extraction, processing and parameter normalization to be trained so that each independent CPN neutral nets can form failure symptom to the Nonlinear Mapping of fault mode.Because these vibration fault states are taken from same simulator stand, so each there is common factor in independent fault signature and fault mode, further data fusion is carried out to the diagnostic result of each CPN neutral net using D-S evidence theory method, steam turbine current operating conditions are carried out more accurately to diagnose and analyze so as to realize.Fig. 2 show the integrated fault diagnosis model of steam turbine based on CPN neutral nets and D-S evidence theory.
The implementation process of specific fault diagnosis algorithm of the invention is described below：
, it is necessary to be trained to CPN neutral nets before application CPN neutral nets carry out Steam Turbine Fault Diagnosis.The topological structure of CPN neutral nets by CPN neutral nets symbol as shown in figure 3, set as follows：If the input vector of CPN networks is
, the output vector of competition layer is, the reality output vector of output layer is
, target output vector is
, whereinThe respectively neuron number of CPN neural network input layers, competition layer and output layer,
, wherein
Represent the number of CPN neural network failure samples.It is by the connection weight vector of input layer to competition layer
；It is by the connection weight vector of competition layer to output layer
。
The learning process of CPN neutral nets is as follows：
（1）Data prediction：By all input patterns
According to formula（1）It is normalized, and weight vector will be connected
With
Assign the random value in [0,1].
(1)
（2）CPN neural metwork trainings：
1）Input layer to competition layer learns without teacher's type：
By connection weight vector
According to formula（2）It is normalized
,
,
；
(2)
By
Individual input patternNetwork input layer is supplied to, then according to formula（3）Calculate each neuron in competition layer weighting input and：
(3)
According to formula（4）Try to achieve connection weight vectorIn with
Closest vector
,
(4)
And by its corresponding neuron
Output
It is set as 1, the output of remaining competition layer neuron
It is set as 0；Weight vector will finally be connected
According to formula（5）It is modified, and by connection weight
Again normalize.
(5)
Wherein
For learning rate,
。
2）Competition layer has the study of teacher's type to output layer：
According to formula（6）To correct competition layer to the connection weight vector of output layer
。
(6)
Wherein
For learning rate,
。
Due to having determined competition layer neuron in the study of input layer to competition layer
1 is output as, and other neurons are output as 0, so only needing to correct neuron
Corresponding connection weight vector, institute's above formula can be with abbreviation into formula（7）,
(7)
The weighting input of each neuron of output layer is tried to achieve, and is translated into the real output value such as formula of output layer neuron（8）It is shown.It can similarly simplify such as formula（9）In form
(8)
(9)
3）The repetition training of CPN networks：
Will
Individual input pattern is all supplied to CPN neutral nets learn without teacher's type and has the study of teacher's type, completes the training of a CPN neutral net.Make again
, by input pattern
E-learning is re-supplied to, untilOr untill network error E is less than predetermined error.Wherein
For study total degree set in advance.
(10)
In the present invention, CPN neural network training process algorithms are improved.Due in standard CPN neutral nets, if the triumph neuron repeatedly trained is identical, then algorithm is only adjusted to the corresponding connection weight of the neuron, so that the information record of multiple input patterns is in same neuron, this can cause the chaotic situation of the information of record, be unfavorable for improving the training effect of CPN neutral nets.In order to avoid such case, it is necessary to artificially intervene neuron, make the information record of fault mode in different neurons, so as to improve CPN neural metwork training effects.
According to formula（3）Calculate
, in theseAccording to formula（4）The maximum weighting input of selection and
, as connect weight vectorIn with
Closest vector.If
's
（T is activated number of times for the competition layer neuron）, then will
Corresponding neuron causes as competition layer triumph neuron
's
；If, then select to remove
Outer maximum weighting input and
If,
's
, then willCorresponding neuron causes as competition layer triumph neuron
's, otherwise successively by weighting input and
Order from big to small finds competition layer triumph neuron.Adjusted by such algorithm, can be by the information record of fault mode in different neurons.
After application CPN neutral nets carry out fault diagnosis, diagnostic result is normalized, fusion diagnosis is further carried out by D-S evidence theory, to improve the accuracy and reliability of diagnostic result.
D-S evidence theory is mainly according to reliability function computing, and it is a kind of data fusion method for solving uncertain problem.D-S evidence theory does not need prior information, and describes uncertain information using the method for interval estimation, solves on probabilistic method for expressing.
The following is the relating basic concepts of D-S evidence theory：
1）The basic conception of identification framework：
If
For finite aggregate,
For natural number,
In haveIndividual various separate possibility answers on certain proposition or it is assumed that then
It is shared
Individual subset, wherein all use of all subsets
To represent.The then finite aggregate
For identification framework FD（Frame of Discement）.
2）The basic conception of probability distribution function：
If
For identification framework, proposition in field all by
Subset represent, then set function
Meet
(11)
Then claim
It is
On basic probability assignment BPA（Basic Probability Assignment）；,
For
Elementary probability number BPN（Basic Probability Number）.
3）The basic conception of belief function：
If
For identification framework,
For framework
On basic probability assignment, then belief function（Belief Function）
For
, whereinRepresent
All subsets,
Also known as lower limit function, is represented pair
Proposition is genuine trusting degree.
4）The basic conception of likelihood function：
If
For
On belief function, then likelihood function
For
, wherein,
Also known as upper limit function or function can not be refuted.Due to
Expression pair
For genuine trusting degree, so
Mean that pair
It is true（I.e.
It is false）Trusting degree, therefore likelihood function
Expression pair
For non-false trusting degree.
5）The basic conception of burnt member：
If
, then claimFor belief function
Jiao's member, the unions of all burnt members are referred to as
Core.
After the basic conception of above D-S evidence theory is established, the rule of evidence combined effect can be reflected by the fusion rule of D-S evidence theory.For same thing, according to different sources of evidence, it can set
With
It is same identification framework
On two belief functions, simultaneously
With
It is its corresponding BPA function, if they are separate, corresponding belief function
WithIt is also separate.According to
With
A new BPA function can be calculated
, corresponding belief function
It can be passed through according to the definition of belief function
To try to achieve.If burnt member is respectively
With
, and
,
,
, then
(12)
Wherein,
.Formula（12）In, if
, then
It is defined as a BPA function；If
, then it is assumed that in the absence of orthogonal and
, i.e., BPA can not be combined.
And the compositional rule of many belief functions is as follows：If
For same identification frameworkOn belief function,
It is its corresponding elementary probability number, if
In the presence of and elementary probability number be
, then
(13)
Wherein
。
The reliability of data and the local diagnosis result of single evidence is only obtained according to existing sensor to construct the basic probability function of D-S evidence theory global diagnosis, the evidence that one of sensor is provided could be enabled to be merged with the evidence that other sensors are obtained, that is, the diagnostic result of CPN networks is converted into Evidence Reasoning Model.
Generally, the acquisition of Basic Probability As-signment relies on expertise, it is believed that the output valve of certain failure is big in CPN neutral net output vectors, then the probability that corresponding failure occurs is big.So the output valve of CPN neutral nets is used formula（14）Be normalized, be converted into should network various malfunctions basic probability assignment BPA, then according to the fusion rule of D-S evidence theory try to achieve its it is orthogonal and
, the BPA of each state after as merging.
(14)
Formula（14）In,
Represent the
Individual evidence in CPN networks
Individual output result；
Represent
Individual evidence is to state
BPA.
When application D-S evidence theory carries out fusion diagnosis, using the result of the Steam Turbine Fault Diagnosis of synchronization as evidence body, merged in D-S Fusion Modules, obtain the fusion results judged current state, it is final to submit judging module, carry out the condition adjudgement of steam turbine failure.Through above-mentioned processing, whole observation judging process have passed through the fusion process from data-feature-knowledge, can be effectively prevented from the erroneous judgement of single differentiation or problem of failing to judge, reduce the error rate of fault diagnosis, it is ensured that the accuracy of fault diagnosis.
The present invention is by taking the oil whip failure of turbine rotor vibration simulation experiment platform as an example, the testing stand according to Fig. 1, exemplified by the frequency spectrum data for 1. 2. locating steam turbine oil whip vibration failure as measuring point with current vortex sensor 10 using photoelectric sensor 3 as measuring point, these data will by 0-0.39f, 0.4-0.49f, 0.5f, 0.51-0.59f, 1f, 2f, 3f, 3-5f,>5f（F is speed）Deng the amplitude components energy on 9 different frequency ranges as fault characteristic value, steam turbine oil whip vibration fault sample as shown in table 1.
Table 1
Fault data at two measuring points is diagnosed with CPN neutral nets, the fault diagnosis result for obtaining correspondence measuring point is as shown in table 2.
Table 2
According to formula（14）Data in table 2 are normalized, basic probability assignment BPA can be obtained as shown in table 3.Wherein defective space touches mill for axial direction
, misalign
, turbine being utilized
, it is uneven
, oil whirl
, oil whip
.Because this six kinds of malfunctions are separate, so only needing to calculate the basic probability assignment of following six kinds of independent failures.
Table 3
Above-mentioned basic probability function is a kind of possibility description of failure, it is impossible to further accurate to judge, it is necessary to carry out evidence merging to above-mentioned basic probability function.Step is as follows：
1）Calculate conflict weights
=1-[0.0235*(0.0072+0.0536+0.1263+0.0269+0.7610)+0.0047*(0.0250+0.0536+0.1263+0.0269+0.7610)+0.0883*(0.0250+0.0072+0.1263+0.0269+0.7610)+0.1366*(0.0250+0.0072+0.0536+0.0269+0.7610)+0.0227*(0.0250+0.0072+0.0536+0.1263+0.7610)+0.7242*(0.0250+0.0072+0.0536+0.1263+0.0269)]=0.5743
2）The basic probability function of each possible judgement is after merging：
=0.0235×0.025/0.5743≈0.0010
=0.1366×0.1263/0.5743≈0.0300
The output result of CPN neutral nets is after D-S evidence theory processing, and the comparison for obtaining final fusion results with CPN Neural Network Diagnosis result before merging is as shown in table 4.
Table 4
To the diagnostic result of oil whip failure it is respectively 0.8714 and 0.9031 before can be seen that fusion from above-mentioned result of calculation, the observed result rises to 0.9596 after fusion, and the diagnostic value of other failures is then significantly declined before and after merging.As can be seen that the diagnostic result of two evidence bodies to 1. 2. being obtained with measuring point from measuring point is merged, it further enhancing to criterion
Support so that the differentiation of each proposition is more obvious, is conducive to the differentiation to true proposition, realizes more accurate and reliable fault diagnosis.
Claims (1)
1. the integrated method for diagnosing faults of a kind of steam turbine based on CPN networks and D-S evidences, it is characterised in that method is comprised the following specific steps that：
1）It regard the Steam Turbine Vibration state parameter collected from sensor on turbine rotor vibration simulator stand as independent data sample, each independent CPN neutral nets are input to after the processing of feature extraction, processing and parameter normalization to be trained so that each independent CPN neutral nets can form failure symptom to the Nonlinear Mapping of fault mode；
2）Carried out using each independent CPN neutral nets after fault diagnosis, the output valve of CPN neutral nets be normalized using formula below, be converted into should CPN neutral nets various malfunctions basic probability assignment BPA,
Wherein
Represent the
Individual evidence in CPN networks
Individual output result；RepresentIndividual evidence is to state
BPA；
3）Further data fusion is carried out to the diagnostic results of each failure CPN neutral nets according to the fusion rule of D-S evidence theory, try to achieve its it is orthogonal and
, the basic probability assignment of each status fault state, obtains fault diagnosis result after as merging.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100507549A CN102589890A (en) | 2012-03-01 | 2012-03-01 | Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100507549A CN102589890A (en) | 2012-03-01 | 2012-03-01 | Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102589890A true CN102589890A (en) | 2012-07-18 |
Family
ID=46478836
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100507549A Pending CN102589890A (en) | 2012-03-01 | 2012-03-01 | Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102589890A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102928231A (en) * | 2012-11-13 | 2013-02-13 | 上海电力学院 | Equipment fault diagnosis method based on D-S (Dempster-Shafer) evidence theory |
| CN103207070A (en) * | 2013-04-09 | 2013-07-17 | 苏州经贸职业技术学院 | Method for diagnosing failure of rotary machine by fusing oil and vibration |
| CN103557884A (en) * | 2013-09-27 | 2014-02-05 | 杭州银江智慧城市技术集团有限公司 | Multi-sensor data fusion early warning method for monitoring electric transmission line tower |
| CN106447040A (en) * | 2016-09-30 | 2017-02-22 | 湖南科技大学 | Method for evaluating the health state of mechanical equipment based on heterogeneous multi-sensor data fusion |
| CN106503643A (en) * | 2016-10-18 | 2017-03-15 | 上海电力学院 | Tumble detection method for human body |
| CN107729920A (en) * | 2017-09-18 | 2018-02-23 | 江苏海事职业技术学院 | A kind of method for estimating state combined based on BP neural network with D S evidence theories |
| CN108318249A (en) * | 2018-01-24 | 2018-07-24 | 广东石油化工学院 | A kind of method for diagnosing faults of bearing in rotating machinery |
| CN108628968A (en) * | 2018-04-24 | 2018-10-09 | 哈尔滨汽轮机厂有限责任公司 | A kind of steam turbine measuring point historical data base method for building up |
| CN109583036A (en) * | 2018-11-05 | 2019-04-05 | 中国航空工业集团公司西安飞机设计研究所 | A kind of distribution method of the fault detection rate of integrated failure |
| CN109948636A (en) * | 2017-12-21 | 2019-06-28 | 北京京东尚科信息技术有限公司 | Data fusion method and device |
| CN111082402A (en) * | 2019-12-31 | 2020-04-28 | 西安理工大学 | Prediction method for cascading failure sequence of power transmission network |
| CN113033600A (en) * | 2021-02-02 | 2021-06-25 | 湖南科技大学 | Rotor misalignment state identification method based on improved D-S evidence fusion |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1389710A (en) * | 2002-07-18 | 2003-01-08 | 上海交通大学 | Multiple-sensor and multiple-object information fusing method |
| US20030103667A1 (en) * | 2001-12-05 | 2003-06-05 | New Mexico Technical Research Foundation | Neural network model for compressing/decompressing image/acoustic data files |
| CN101539241A (en) * | 2009-05-07 | 2009-09-23 | 北京航空航天大学 | Hierarchical multi-source data fusion method for pipeline linkage monitoring network |
| US7613360B2 (en) * | 2006-02-01 | 2009-11-03 | Honeywell International Inc | Multi-spectral fusion for video surveillance |
| CN101713776A (en) * | 2009-11-13 | 2010-05-26 | 长春迪瑞实业有限公司 | Neural network-based method for identifying and classifying visible components in urine |
| CN102175282A (en) * | 2011-01-24 | 2011-09-07 | 长春工业大学 | Method for diagnosing fault of centrifugal air compressor based on information fusion |
-
2012
- 2012-03-01 CN CN2012100507549A patent/CN102589890A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20030103667A1 (en) * | 2001-12-05 | 2003-06-05 | New Mexico Technical Research Foundation | Neural network model for compressing/decompressing image/acoustic data files |
| CN1389710A (en) * | 2002-07-18 | 2003-01-08 | 上海交通大学 | Multiple-sensor and multiple-object information fusing method |
| US7613360B2 (en) * | 2006-02-01 | 2009-11-03 | Honeywell International Inc | Multi-spectral fusion for video surveillance |
| CN101539241A (en) * | 2009-05-07 | 2009-09-23 | 北京航空航天大学 | Hierarchical multi-source data fusion method for pipeline linkage monitoring network |
| CN101713776A (en) * | 2009-11-13 | 2010-05-26 | 长春迪瑞实业有限公司 | Neural network-based method for identifying and classifying visible components in urine |
| CN102175282A (en) * | 2011-01-24 | 2011-09-07 | 长春工业大学 | Method for diagnosing fault of centrifugal air compressor based on information fusion |
Non-Patent Citations (2)
| Title |
|---|
| 彭文季等: "基于对向传播神经网络的水电机组振动故障诊断研究", 《西安理工大学学报》, vol. 22, no. 4, 31 December 2006 (2006-12-31), pages 366 * |
| 徐春梅等: "基于改进D-S的汽轮机组集成故障诊断研究", 《系统仿真学报》, vol. 23, no. 10, 31 October 2011 (2011-10-31), pages 2191 - 2193 * |
Cited By (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102928231A (en) * | 2012-11-13 | 2013-02-13 | 上海电力学院 | Equipment fault diagnosis method based on D-S (Dempster-Shafer) evidence theory |
| CN103207070A (en) * | 2013-04-09 | 2013-07-17 | 苏州经贸职业技术学院 | Method for diagnosing failure of rotary machine by fusing oil and vibration |
| CN103207070B (en) * | 2013-04-09 | 2015-08-26 | 苏州经贸职业技术学院 | The rotary machinery fault diagnosis method that fluid merges mutually with vibration |
| CN103557884A (en) * | 2013-09-27 | 2014-02-05 | 杭州银江智慧城市技术集团有限公司 | Multi-sensor data fusion early warning method for monitoring electric transmission line tower |
| CN103557884B (en) * | 2013-09-27 | 2016-06-29 | 杭州银江智慧城市技术集团有限公司 | A kind of Fusion method for early warning of electric power line pole tower monitoring |
| CN106447040B (en) * | 2016-09-30 | 2018-11-23 | 湖南科技大学 | Mechanical equipment health state evaluation method based on Heterogeneous Multi-Sensor Data fusion |
| CN106447040A (en) * | 2016-09-30 | 2017-02-22 | 湖南科技大学 | Method for evaluating the health state of mechanical equipment based on heterogeneous multi-sensor data fusion |
| CN106503643A (en) * | 2016-10-18 | 2017-03-15 | 上海电力学院 | Tumble detection method for human body |
| CN106503643B (en) * | 2016-10-18 | 2019-06-28 | 上海电力学院 | Tumble detection method for human body |
| CN107729920A (en) * | 2017-09-18 | 2018-02-23 | 江苏海事职业技术学院 | A kind of method for estimating state combined based on BP neural network with D S evidence theories |
| CN109948636A (en) * | 2017-12-21 | 2019-06-28 | 北京京东尚科信息技术有限公司 | Data fusion method and device |
| CN108318249A (en) * | 2018-01-24 | 2018-07-24 | 广东石油化工学院 | A kind of method for diagnosing faults of bearing in rotating machinery |
| CN108318249B (en) * | 2018-01-24 | 2020-04-17 | 广东石油化工学院 | Fault diagnosis method for rotary mechanical bearing |
| CN108628968A (en) * | 2018-04-24 | 2018-10-09 | 哈尔滨汽轮机厂有限责任公司 | A kind of steam turbine measuring point historical data base method for building up |
| CN108628968B (en) * | 2018-04-24 | 2022-06-21 | 哈尔滨汽轮机厂有限责任公司 | Turbine measuring point historical database establishing method |
| CN109583036A (en) * | 2018-11-05 | 2019-04-05 | 中国航空工业集团公司西安飞机设计研究所 | A kind of distribution method of the fault detection rate of integrated failure |
| CN111082402A (en) * | 2019-12-31 | 2020-04-28 | 西安理工大学 | Prediction method for cascading failure sequence of power transmission network |
| CN111082402B (en) * | 2019-12-31 | 2022-01-07 | 西安理工大学 | Prediction method for cascading failure sequence of power transmission network |
| CN113033600A (en) * | 2021-02-02 | 2021-06-25 | 湖南科技大学 | Rotor misalignment state identification method based on improved D-S evidence fusion |
| CN113033600B (en) * | 2021-02-02 | 2022-05-27 | 湖南科技大学 | Rotor misalignment state identification method based on improved D-S evidence fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102589890A (en) | Integrated fault diagnostic method of steam turbine based on CPN (counter-propagation network) and D-S (dempster-shafer) evidences | |
| CN110276416B (en) | Rolling bearing fault prediction method | |
| CN107003663B (en) | The monitoring of device with movable part | |
| Xue et al. | A fuzzy system of operation safety assessment using multimodel linkage and multistage collaboration for in-wheel motor | |
| CN104200396B (en) | A kind of wind turbine component fault early warning method | |
| CN104850889B (en) | Airplane rotation actuator drive unit adaptive fault detection, isolation and confidences assessment method | |
| CN113092115B (en) | Digital twin model construction method of digital-analog combined drive full-life rolling bearing | |
| CN110308002A (en) | A kind of municipal rail train suspension method for diagnosing faults based on ground detection | |
| Xu et al. | A bearing fault diagnosis method without fault data in new working condition combined dynamic model with deep learning | |
| CN109583092A (en) | A kind of intelligent machine diagnosis method for system fault of multi-level multi-mode feature extraction | |
| CN105487009A (en) | Motor fault diagnosis method based on k-means RBF neural network algorithm | |
| CN106762343A (en) | The diagnostic method of the hydraulic generator set thrust bearing failure based on online data | |
| CN102609764A (en) | CPN neural network-based fault diagnosis method for stream-turbine generator set | |
| Shang et al. | Fault diagnosis method of rolling bearing based on deep belief network | |
| Zhang et al. | Vibration-based structural damage detection via phase-based motion estimation using convolutional neural networks | |
| CN112765890A (en) | Dynamic domain adaptive network-based multi-working-condition rotating machine residual life prediction method | |
| Guo et al. | Reconstruction domain adaptation transfer network for partial transfer learning of machinery fault diagnostics | |
| Chen et al. | Rotor fault diagnosis system based on sGA-based individual neural networks | |
| Zhou et al. | Structural health monitoring of offshore wind power structures based on genetic algorithm optimization and uncertain analytic hierarchy process | |
| Liu et al. | Intelligent cross-condition fault recognition of rolling bearings based on normalized resampled characteristic power and self-organizing map | |
| Rahimilarki et al. | Time-series deep learning fault detection with the application of wind turbine benchmark | |
| Yu et al. | A new method for quantitative estimation of rolling bearings under variable working conditions | |
| CN103245491A (en) | Rotor system fault diagnosis method based on confirmed learning theory | |
| Xing et al. | Detection of magnitude and position of rotor aerodynamic imbalance of wind turbines using Convolutional Neural Network | |
| Du et al. | Research on the application of artificial intelligence method in automobile engine fault diagnosis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20120718 |