CN101937207B - Intelligent visualized monitoring and diagnosing method of mechanical equipment state - Google Patents
Intelligent visualized monitoring and diagnosing method of mechanical equipment state Download PDFInfo
- Publication number
- CN101937207B CN101937207B CN201010263849XA CN201010263849A CN101937207B CN 101937207 B CN101937207 B CN 101937207B CN 201010263849X A CN201010263849X A CN 201010263849XA CN 201010263849 A CN201010263849 A CN 201010263849A CN 101937207 B CN101937207 B CN 101937207B
- Authority
- CN
- China
- Prior art keywords
- equipment state
- dictionary
- state
- equipment
- sparse
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to an intelligent visual monitoring and diagnosing method of a mechanical equipment state in the technical field of mechanical engineering, comprising the following steps of: constructing an original equipment state dictionary; establishing an original equipment state distribution chart; monitoring the sparse representation of data on the basis of solving equipment of the equipment state dictionary; extracting a sparse characteristic value from the sparse representation of the equipment detection data; projecting the sparse characteristic value on the equipment state distribution chart; recognizing an equipment state according to a projected position of the sparse characteristic value on the equipment state distribution chart; detecting whether a new equipment state appears; and when the new equipment state is detected, updating the equipment state dictionary and the equipment state distribution chart. The invention can be used for visually displaying the equipment state directly and clearly, reduces the requirement for professional quality of equipment maintenance staff and can overcome the interference of an equipment diagnosing process by work condition variation.
Description
Technical field
What the present invention relates to is a kind of method of mechanical engineering technical field, specifically is a kind of intelligent visual monitoring and diagnosis method of plant equipment state.
Background technology
Plant maintenance based on state is a kind of emerging plant maintenance technology, and it implements appropriate maintenance on the basis of identification equipment state, not only can effectively avoid the generation of equipment failure, can also reduce maintenance cost simultaneously.Therefore, the plant maintenance based on state is widely used in the plant equipment area of maintenance.The general path of equipment state identification is a sensor installation on institute's care equipment, gathers corresponding signal, obtains status information of equipment through the analysis to signal, combines the diagnostic knowledge of priori to confirm equipment state on this basis.Wherein, extensively pressure, flow, vibration, electric current, strain etc. are arranged by the signal of gathering and analyze.Visual spectral analysis technologies such as the spectrogram of traditional device state identification method through analyzing acquired signal, time-frequency figure come the diagnostic device state.But this mode is higher to plant maintenance personnel specialty competency profiling.Therefore, the equipment state intelligent diagnosing method of robotization is also extensively adopted.But intelligent diagnosing method often lacks people's dirigibility again.In addition, in actual Diagnosis Application, the fault sample signal of equipment obtains the comparison difficulty, makes the limited use of intelligentized diagnostic method, and the equipment failure of the unknown is easy to generate the situation of wrong diagnosis.Therefore, a kind ofly can combine the automatic intelligent diagnostic method, and it is ready to appear to make full use of people's diagnostic mode of dirigibility in diagnostic procedure.If can manifest a kind of plant equipment state scattergram visually; And can the current signal that collects be located on the plant equipment state scattergram; Make the people can be from figure readout equipment state directly perceived and change procedure thereof; That is: realize the visual of plant equipment status monitoring and diagnosis, this is undoubtedly a kind of good diagnostic mode.
Retrieval through to existing document finds that the Chinese patent publication number is: CN101699359A, name is called: the method for visualizing of fault state monitoring, the variation that this technology is come the marking equipment state through the position difference of composite character vector in two dimensional surface.Its deficiency is: characteristic parameter is necessary for discrete state value, and being not suitable for data mode is the seasonal effect in time series waveform signal; And the semanteme of resultant vector position is not clear, still needs the maintainer according to professional knowledge judgment device state.
Find through retrieval again; Lei Wenping has proposed to realize the visual diagnosis to rotating machinery fault based on the self-organization mapping in document " the visual research (master thesis) with Forecasting Methodology of rotating machinery fault "; But the method only is the simple application of self-organization mapping type neural network in fault diagnosis; Do not have further to inquire into the problem of signal analysis and feature extraction in the waveform signal, can not resist working condition and change interference diagnostic result.
Also find through retrieval; Abhinav Saxena diagnoses rotating machinery fault based on self-organization mapping type neural network to waveform signal on the basis of the some simple eigenvalues of extraction in document " Fault Diagnosis in Rotating MechanicalSystems Using Self-Organizing Maps (adopting the rotary machinery fault diagnosis of self-organization mapping type neural network) ", but the method can not overcome the influence of working conditions change to acquired signal equally.
Summary of the invention
The objective of the invention is to overcome the above-mentioned deficiency of prior art, a kind of intelligent visual monitoring and diagnosis method of plant equipment state is provided.The present invention can effectively overcome the interference of working conditions change to equipment condition monitoring and diagnosis, simultaneously, can realize the self-perfection of Device Diagnostic ability.
The present invention realizes through following technical scheme, the present invention includes following steps:
The first step, structure original equipment state dictionary, and set up the original equipment state scattergram.
Described structure original equipment state dictionary may further comprise the steps:
1) the historical sample data under the some equipment states of collection adopt the dictionary learning method, and one of study sets up state subgroup dictionary fully the historical sample data under every kind of equipment state respectively;
Described dictionary learning method is: when the position that the statistic structural features in the historical sample data of equipment occurs in data block has consistance, then adopt the dictionary learning method based on standard sparse coding model; Otherwise, adopt based on the time constant sparse coding model the dictionary learning method.
2) allly set up fully that the state subgroup dictionary constitutes original equipment state dictionary.
The described original equipment state scattergram of setting up may further comprise the steps:
1) based on the original equipment state dictionary of setting up, adopts sparse method for solving, obtain the rarefaction representation of all historical sample data of equipment;
Described sparse method for solving is: when the position that the statistic structural features in the historical sample data of equipment occurs in data block has consistance, then adopt the sparse method for solving based on standard sparse coding model; Otherwise, adopt based on the time constant sparse coding model sparse method for solving.
2) extract the sparse features value in the rarefaction representation of all historical sample data of slave unit;
Described sparse features value; Be the activation number of times of basis function in rarefaction representation in the equipment state dictionary; Or the quadratic sum of the coefficient of basis function in rarefaction representation in the equipment state dictionary, or in the equipment state dictionary coefficient of basis function in rarefaction representation absolute value with.
3) go out U type matrix diagram (U-Matrix) through training self-organization mapping type neural network configuration; And with the RGB color node is played up according to the distance measure of node on sparse features value under the different plant equipment states and the U type matrix diagram; And the distinct device state according to the sparse features value representative of historical sample data marks the node on the U type matrix diagram, and this U type matrix diagram is exactly the original equipment state scattergram.
Second step; In real time the plant equipment state is monitored; Obtain the Monitoring Data of plant equipment; When the appearance position of statistic structural features in data block of the Monitoring Data of plant equipment has consistance, then adopt the rarefaction representation that obtains current Monitoring Data based on the sparse method for solving of standard sparse coding model; Otherwise, then adopt based on the time constant sparse coding model sparse method for solving obtain the rarefaction representation of current Monitoring Data.
The 3rd step; Adopt the method for the first step; From the rarefaction representation of current Monitoring Data, extract current sparse features value, and, obtain the optimal match point (BMU of this current sparse features value on the equipment state distribution plan according to self-organization mapping type neural net method; Best Matching Unit), thus current sparse features value is mapped on the equipment state distribution plan.
In the 4th step, when the optimal match point that obtains when the 3rd step had been marked equipment state, current equipment state was exactly the equipment state that optimal match point marks, and returns for second step, continues equipment state is monitored and discerned; Otherwise, carry out the unknown device state detection process, when new equipment state occurring, carried out for the 5th step, otherwise, returned for second step, continue the plant equipment state is monitored and discerned.
Described unknown device state detection process may further comprise the steps:
1) calculates the optimal match point of current sparse features value on equipment state figure and the minimum quantization error of current sparse features value, when minimum quantization error is greater than or equal to threshold value θ
MQE, then carry out 2); Otherwise, assert new equipment state not occur, and the equipment state that node marked close with the optimal match point color confirmed as the current device state;
2) carry out manual work and confirm, when manual work is confirmed new equipment state not occur, to threshold value θ
MQECarry out correcting process, otherwise, assert new equipment state to occur.
Described correcting process is meant that the minimum quantization error with the optimal match point of current sparse features value on equipment state figure is made as threshold value θ
MQE
The 5th step; Gather the sample data under the new equipment state, carry out equipment state dictionary update processing and equipment state distribution plan update processing, the equipment state dictionary that obtains upgrading and the equipment state distribution plan of renewal; And returned for second step, continue the plant equipment state is monitored and discerned.
Described equipment state dictionary update processing; Be: the method that adopts first step structure original equipment state dictionary; The sub-dictionary of facility for study new state the sample data under the plant equipment new state; And will represent to add the equipment state dictionary in the sub-dictionary of equipment new state, realize the renewal of equipment state dictionary.
Described equipment state distribution plan update processing; Be: adopt the first step to set up the method for original equipment state scattergram; Incorporate the sample data under the plant equipment new state into the historical sample data; Again find the solution the rarefaction representation of current all historical sample data, and therefrom extract the sparse features value identical, through the U type matrix diagram of training self-organization mapping type neural network configuration to make new advances with the first step; New equipment state occupies an independent cluster on equipment state figure, thereby realizes the renewal to the equipment state distribution plan.
Compared with prior art, the invention has the beneficial effects as follows: the state to the monitoring plant equipment shows simple, intuitive, thereby can give full play to the dirigibility of people in equipment state diagnosis identification; Variation to working condition has certain anti-interference, simultaneously, can detect unknown equipment state; Update mechanism through apparatus for establishing state dictionary can realize the self-perfection to the equipment state diagnosis capability.
Description of drawings
Fig. 1 is the process flow diagram of original equipment state dictionary of the present invention and the foundation of original equipment state scattergram.
Fig. 2 is the process flow diagram that present device state dictionary and equipment state distribution plan are used in the plant equipment status monitoring.
Fig. 3 is the process flow diagram that present device state dictionary and equipment state distribution plan upgrade.
Fig. 4 is the mapping process of embodiment centre bearer vibration signal on the bearing state distribution plan.
Fig. 5 is the bearing state distribution plan that the vibration signal under normal based on bearing among the embodiment, inner ring fault and the ball malfunction is set up;
Wherein: (a) be U type matrix diagram; (b) be the mark after three kinds of bearing state distribution plans.
Vibration signal the projection on initial bearing state scattergram of Fig. 6 for monitoring under three kinds of different bearing states among the embodiment;
Wherein: be to decide under the load working condition (a); (b) be under the variable load operating mode.
Fig. 7 is with the new bearing state of threshold decision---outer ring fault among the embodiment.
Fig. 8 upgrades among the embodiment the bearing state distribution plan after the Fault Diagnosis ability of outer ring;
Wherein: (a) be U type matrix diagram; (b) be the mark after four kinds of bearing state distribution plans.
Embodiment
Below in conjunction with accompanying drawing embodiments of the invention are elaborated: present embodiment provided concrete embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment being to implement under the prerequisite with technical scheme of the present invention.
Embodiment
Present embodiment is the application of the present invention in the bearing state monitoring and fault diagnosis, and institute's monitoring bearing is the deep groove ball bearing on the motor output shaft.Acceleration transducer is installed in and is used to gather vibration signal on the casing directly over the output shaft.The bearing state of being monitored comprises: normal, inner ring fault, ball fault and outer ring fault; Part wherein is normal, the bear vibration monitor signal under inner ring fault and the ball malfunction is known, and is used to set up initial bearing status word allusion quotation and initial bearing state scattergram as the historical sample data.Initial bearing status word allusion quotation of setting up and initial bearing state scattergram are used for the further monitoring and diagnosis of bearing state and use; When monitoring new bearing state, carry out the renewal of bearing state dictionary and bearing state distribution plan.
Present embodiment may further comprise the steps:
(1) as shown in Figure 1, the vibration signal under, inner ring fault normal according to the bearing of having grasped and the ball malfunction is constructed initial bearing status word allusion quotation and is set up initial bearing state scattergram.
At first, the initial bearing status word allusion quotation of structure:
1) bearing is nonserviceabled down, and damage of the bearing can cause bearing in rolling process, producing improper impact, in the vibration signal that collects, occurs violent shock pulse accordingly.In to Vibration Signal in Its Processing, need vibration signal is divided into the time series of certain-length, this makes shock pulse be split to the optional position in the time series, that is: make the appearance position of shock pulse in time series be lack of consistency.Therefore, from vibration signal study sub-dictionary should adopt based on the time constant sparse coding model the dictionary learning method.
The time constant sparse coding model make basis function can appear at the seasonal effect in time series optional position through convolution algorithm:
Wherein: matrix
Be called dictionary, each row d in the dictionary
kBe basis function, ∈ is Gauss's residual noise.For given data set X=[x
1, x
2..., x
N], suppose that basis function obeys priori and evenly distribute, calculate simultaneously basis function and rarefaction representation maximum a posteriori probability (maximum a posterior MAP) is estimated as:
Wherein: s
KiBe basis function d
kAt signal x
iIn sparse activation; C is that a constraint constant is used to the d that prevents to find the solution
kBecome too big, simultaneously s
K, iToo little; β is the optimization penalty factor; || ||
1Be sparsely to estimate.Objective function (2) is protruding though be not the protruding optimization of associating to single variable.Therefore, fixing constant protruding optimization about another variable of variable can approach and find the solution; Intersection is found the solution these two protruding optimization problems and is carried out dictionary study; After objective function (2) convergence, the basis function of finding the solution constitutes dictionary D; And then based on the dictionary D that learns, i.e. the constant protruding optimization of finding the solution s of d in the fixed target function (2) can be obtained the rarefaction representation of importing data.
For three kinds of known bearing states: normal, inner ring fault and ball fault, the vibration signal from every kind of bearing state under respectively according to formula (2) and maximum a posteriori probability estimation the carrying out study that single shaft holds the state subgroup dictionary shown in the formula (3).The vibratory impulse that makes discovery from observation continues about 80 sampled points in gathering vibration signal, so the basis function length setting is 80 in each sub-dictionary, and the basis function number is made as 10.Under the constant sparse coding model, each basis function possibly appear at any point in the time series, so each sub-dictionary is actually the ultra complete dictionary of 10x when considering.
2) respectively under normal, inner ring fault of study and the ball malfunction behind the sub-dictionary, the initial bearing status word allusion quotation
that comprises these three kinds of bearing states is set up according to following formula:
Wherein: D
N, D
IF, D
BFBe respectively the sub-dictionary of from the vibration signal of normal, inner ring fault, ball malfunction, learning.
Secondly, set up initial bearing state scattergram:
1) according to the bearing state dictionary of setting up
based on formula (2); Constant sparse method for solving when promptly adopting, the rarefaction representation of finding the solution historical data under normal, inner ring fault and the ball malfunction respectively;
2) under normal, inner ring fault and ball malfunction, extract sparse features the rarefaction representation of historical data respectively; At this, sparse features is defined as a vector, and wherein each is the quadratic sum of nonzero coefficient in the sparse activation of a basis function:
Wherein: f
iFor from vibration signal x
iRarefaction representation in the sparse features vector that extracts, s
K, iBe basis function d
kAt signal x
iIn sparse activation;
3) with the self-organization mapping type neural network of sparse features vector training under normal, inner ring fault of extracting and the ball malfunction, the U type matrix diagram that computing is exported in the back is accomplished in training, shown in Fig. 5 (a); Because the bearing state of each sparse features vector is known, the bearing state of each node representative can mark according to the bearing state that projects to the sparse features vector on this node on the U type matrix diagram; Shown in Fig. 5 (b), the bearing state of each node representative uses N (normal condition), IF (inner ring fault), BF (ball fault) to mark respectively; The bearing state that does not mark the node representative of bearing state can be with reference to the bearing state that closes on the node representative close with its distance value.
So far, initial bearing status word allusion quotation is set up with initial bearing state scattergram and is finished.Bearing state dictionary of being set up and bearing state distribution plan have formed normal, the inner ring fault of institute's monitoring bearing and ball Fault Diagnosis ability.
(2) as shown in Figure 2, use bearing state dictionary and bearing state distribution plan bearing state is monitored application.In to the bearing state observation process, the rarefaction representation that constant sparse method for solving is found the solution the bearing vibration signal that monitors when adopting according to () is said based on the bearing state dictionary of setting up
.
(3) from the rarefaction representation of the bearing vibration signal found the solution, extract the sparse features vector according to method shown in formula (5) in (), thereby and find the solution this current sparse features value according to self-organization mapping type neural net method and current sparse features value is mapped on the bearing state distribution plan at the optimal match point on the bearing state distribution plan.
(4) confirm current bearing state according to the mapping position of the optimal match point of current sparse features value on the bearing state distribution plan.As shown in Figure 5, if the optimal match point that current sparse features value is mapped on the bearing state distribution plan has N, IF or BF label, then current bearing state is the bearing state that is identified on this optimal match point; If do not mark any label on the optimal match point, then carry out unknown bearing state and detect.The bearing vibration signal that monitors under shown in Figure 6 for deciding under load and the variable load working condition respectively, normal to bearing respectively, inner ring fault and the ball malfunction is carried out projection result on the bearing state distribution plan.The big more expression of intranodal sexangle area has more vibration signal sample to be projected to this node among Fig. 6.Can see: deciding under the load-up condition, the vibration signal under three kinds of bearing states is projected to respectively on the node of respective markers; Under the variable load condition; Bearing vibration signal under the normal condition is deviation slightly, but still around flag node, in the case; Be no more than in minimum quantization error under the situation of threshold range, can learn that bearing state is normal with reference to its flag node that centers on " N ".
Fig. 4 has schematically described the bearing vibration signal of gathering, and through finding the solution its rarefaction representation, extracts the sparse features vector, is mapped to the process on the bearing state distribution plan at last.
In the actual monitoring process; Collect respectively that bearing is normal, the bearing vibration signal under the inner ring fault, ball fault, outer ring malfunction; With (three) said step application bearing state dictionary bearing vibration signal is carried out projection according to (two) on the bearing state distribution plan respectively, minimum quantization error (MQE) variation tendency is as shown in Figure 7.Actual conditions are in the bearing monitoring process, to find that minimum quantization error (MQE) continues to surpass pre-set threshold parameter θ
MQE=0.5176, confirm through artificial, find to occur new bearing state---outer ring fault.In the case, the collecting part bearing vibration signal is carried out the renewal of bearing state dictionary and bearing state distribution plan according to step shown in Figure 3.
(5) renewal process of bearing state dictionary and bearing state distribution plan:
Upgrade the bearing state dictionary: said according to (one), based on the time constant sparse coding model, estimate the sub-dictionary of study bearing outer ring malfunction the vibration signal under the bearing outer ring malfunction according to formula (2) and the maximum a posteriori probability shown in the formula (3); The basis function length setting is 80 in the sub-dictionary, and the basis function number is made as 10; The sub-dictionary of learning out of bearing outer ring malfunction is labeled as D
OFThe sub-dictionary of bearing outer ring malfunction with new study upgrades former bearing status word allusion quotation according to following formula:
Upgrade the bearing state distribution plan: the vibration signal under the bearing outer ring fault of new monitoring makes to incorporate into original historical sample data that comprise normal, inner ring fault and ball fault; Then according to the process of setting up of (one) initial bearing constitutional diagram, carry out the finding the solution of rarefaction representation of current all historical sample data based on the bearing state dictionary
after upgrading; Carry out the extraction of sparse features vector according to formula (5); Train the bearing state distribution plan after the output of self-organization mapping type neural network is upgraded as shown in Figure 8 at last.Can see that the bearing state figure after the renewal is through having realized bearing outer ring Fault Diagnosis ability.In the actual monitoring process, confirm to realize perfect to the bearing state diagnosis capability through the unknown bearing state of continuous detection and through artificial.
Present embodiment can the historical sample data are represented bearing state on two-dimentional bearing state distribution plan, show intuitively; Bearing is being carried out in the real-time observation process,, can represent current bearing state simply through the vibration signal that monitors is carried out projection on the bearing state distribution plan.With respect to spectral analysis methods such as traditional spectrogram, time-frequency figure; The present embodiment method is clear and easy to understand to the demonstration of bearing state; Reduced the requirement to plant maintenance personnel specialty quality, the interference table to the working condition in the bearing operational process reveals certain anti-interference simultaneously.Present embodiment can be realized the detection to unknown bearing state in to the bearing state observation process, and through to the renewal of bearing state dictionary and bearing state figure realization perfect to the bearing state diagnosis capability.
Claims (6)
1. the intelligent visual monitoring and diagnosis method of a plant equipment state is characterized in that, may further comprise the steps:
The first step, structure original equipment state dictionary, and set up the original equipment state scattergram;
Described structure original equipment state dictionary may further comprise the steps:
1) the historical sample data under the some equipment states of collection adopt the dictionary learning method, and one of study sets up state subgroup dictionary fully the historical sample data under every kind of equipment state respectively;
2) allly set up fully that the state subgroup dictionary constitutes original equipment state dictionary;
The described original equipment state scattergram of setting up may further comprise the steps:
1) based on the original equipment state dictionary of setting up, adopts sparse method for solving, obtain the rarefaction representation of all historical sample data of equipment;
2) extract the sparse features value in the rarefaction representation of all historical sample data of slave unit;
3) go out U type matrix diagram through training self-organization mapping type neural network configuration; And with the RGB color node is played up according to the distance measure of node on sparse features value under the different plant equipment states and the U type matrix diagram; And the distinct device state according to the sparse features value representative of historical sample data marks the node on the U type matrix diagram, and this U type matrix diagram is exactly the original equipment state scattergram;
Second step; In real time the plant equipment state is monitored; Obtain the Monitoring Data of plant equipment; Be divided into the time series of certain-length when Monitoring Data after, shock pulse wherein or incident still can come the seasonal effect in time series same position, adopt the rarefaction representation that obtains current Monitoring Data based on the sparse method for solving of standard sparse coding model; Otherwise, adopt based on the time constant sparse coding model sparse method for solving obtain the rarefaction representation of current Monitoring Data;
The 3rd step; Adopt the method for the first step; From the rarefaction representation of current Monitoring Data, extract current sparse features value; And, obtain the optimal match point of this current sparse features value on the equipment state distribution plan, thereby current sparse features value is mapped on the equipment state distribution plan according to self-organization mapping type neural net method;
In the 4th step, when the optimal match point that obtains when the 3rd step had been marked equipment state, current equipment state was exactly the equipment state that optimal match point marks, and returns for second step, continues the plant equipment state is detected and discerns; Otherwise, carry out the unknown device state detection process, when new equipment state occurring, carried out for the 5th step, otherwise, returned for second step, continue the plant equipment state is monitored and discerned;
Described unknown device state detection process may further comprise the steps:
1) calculates the optimal match point of current sparse features value on equipment state figure and the minimum quantization error of current sparse features value, when minimum quantization error is greater than or equal to threshold value θ
MQE, then carry out 2); Otherwise, assert new equipment state not occur, and the equipment state that node marked close with the optimal match point color confirmed as the current device state;
2) carry out manual work and confirm, when manual work is confirmed new equipment state not occur, to threshold value θ
MQECarry out correcting process, otherwise, assert new equipment state to occur;
The 5th step; Gather the sample data under the new equipment state, carry out equipment state dictionary update processing and equipment state distribution plan update processing, the equipment state dictionary that obtains upgrading and the equipment state distribution plan of renewal; And returned for second step, continue the plant equipment state is monitored and discerned.
2. the intelligent visual monitoring and diagnosis method of plant equipment state according to claim 1; It is characterized in that; Described dictionary learning method; Be: be divided into the time series of certain-length when Monitoring Data after, shock pulse wherein or incident still can come the seasonal effect in time series same position, adopt the dictionary learning method based on standard sparse coding model; Otherwise, adopt based on the time constant sparse coding model the dictionary learning method.
3. the intelligent visual monitoring and diagnosis method of plant equipment state according to claim 1; It is characterized in that; Sparse features value described in the 3rd step; Be the activation number of times of basis function in rarefaction representation in the equipment state dictionary, or the quadratic sum of the coefficient of basis function in rarefaction representation in the equipment state dictionary, or in the equipment state dictionary coefficient of basis function in rarefaction representation absolute value with.
4. the intelligent visual monitoring and diagnosis method of plant equipment state according to claim 1 is characterized in that, described correcting process is meant that the minimum quantization error with the optimal match point of current sparse features value on equipment state figure is made as threshold value θ
MQE
5. the intelligent visual monitoring and diagnosis method of plant equipment state according to claim 1; It is characterized in that; Equipment state dictionary update processing described in the 5th step is: adopt the method for first step structure original equipment state dictionary, the sub-dictionary of facility for study new state the sample data under the plant equipment new state; And will represent to add the equipment state dictionary in the sub-dictionary of equipment new state, realize the renewal of equipment state dictionary.
6. the intelligent visual monitoring and diagnosis method of plant equipment state according to claim 1; It is characterized in that the equipment state distribution plan update processing described in the 5th step is: adopt the first step to set up the method for original equipment state scattergram; Incorporate the sample data under the plant equipment new state into the historical sample data; Again find the solution the rarefaction representation of current all historical sample data, and therefrom extract the sparse features value identical, through the U type matrix diagram of training self-organization mapping type neural network configuration to make new advances with the first step; New equipment state occupies an independent cluster on equipment state figure, thereby realizes the renewal to the equipment state distribution plan.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010263849XA CN101937207B (en) | 2010-08-27 | 2010-08-27 | Intelligent visualized monitoring and diagnosing method of mechanical equipment state |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010263849XA CN101937207B (en) | 2010-08-27 | 2010-08-27 | Intelligent visualized monitoring and diagnosing method of mechanical equipment state |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101937207A CN101937207A (en) | 2011-01-05 |
| CN101937207B true CN101937207B (en) | 2012-09-05 |
Family
ID=43390589
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010263849XA Expired - Fee Related CN101937207B (en) | 2010-08-27 | 2010-08-27 | Intelligent visualized monitoring and diagnosing method of mechanical equipment state |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101937207B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102789197B (en) * | 2011-05-19 | 2016-06-22 | 科德数控股份有限公司 | The expression system of Digit Control Machine Tool machining information and expression thereof |
| CN103309347B (en) * | 2013-06-05 | 2015-11-18 | 浙江大学 | A kind of multiple operating modes process method for supervising based on rarefaction representation |
| US10996662B2 (en) * | 2014-05-20 | 2021-05-04 | Toshiba Mitsubishi-Electric Industrial Systems Corporation | Manufacturing equipment diagnosis support system |
| WO2016040082A1 (en) * | 2014-09-10 | 2016-03-17 | Siemens Aktiengesellschaft | Gas turbine sensor failure detection utilizing a sparse coding methodology |
| CN104614166B (en) * | 2015-01-29 | 2017-04-19 | 西北工业大学 | Method for identifying failure state of rotor vibration signal of aircraft engine |
| KR102042368B1 (en) * | 2016-02-03 | 2019-11-07 | 도시바 미쓰비시덴키 산교시스템 가부시키가이샤 | Manufacturing equipment diagnostic support device and manufacturing equipment diagnostic support method |
| CN105973595B (en) * | 2016-04-27 | 2019-03-19 | 清华大学深圳研究生院 | A kind of diagnostic method of rolling bearing fault |
| GB201613318D0 (en) * | 2016-08-02 | 2016-09-14 | Univ Oxford Innovation Ltd | System monitor and method of system monitoring |
| CN107193275A (en) * | 2017-07-01 | 2017-09-22 | 浙江三青环保科技有限公司 | A kind of equipment operation intelligent diagnosis system and diagnostic method |
| CN108544303A (en) * | 2018-03-30 | 2018-09-18 | 上海交通大学 | A kind of main shaft of numerical control machine tool fault diagnosis method and system |
| CN108710891B (en) * | 2018-04-03 | 2020-08-11 | 济南大学 | Steady signal dimension reduction and identification method and system based on perceptual hash |
| CN109141881B (en) * | 2018-07-06 | 2020-03-31 | 东南大学 | Rotary machine health assessment method of deep self-coding network |
| CN108983749B (en) * | 2018-07-10 | 2021-03-30 | 福州大学 | Photovoltaic array fault diagnosis method based on K-SVD training sparse dictionary |
| AU2019375200A1 (en) * | 2018-11-09 | 2021-06-17 | Augury Systems Ltd. | Automated analysis of non-stationary machine performance |
| CN110327043B (en) * | 2019-08-05 | 2022-04-01 | 哈尔滨工业大学 | Event-related potential waveform map solving method based on sparse modeling |
| CN115219912A (en) * | 2022-04-24 | 2022-10-21 | 山东大学 | Early fault diagnosis and safety advanced early warning method and system for energy storage battery |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101699359A (en) * | 2009-10-28 | 2010-04-28 | 上海理工大学 | Method for visualizing fault state monitoring |
-
2010
- 2010-08-27 CN CN201010263849XA patent/CN101937207B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101699359A (en) * | 2009-10-28 | 2010-04-28 | 上海理工大学 | Method for visualizing fault state monitoring |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101937207A (en) | 2011-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101937207B (en) | Intelligent visualized monitoring and diagnosing method of mechanical equipment state | |
| CN106338406B (en) | The on-line monitoring of train traction electric drive system and fault early warning system and method | |
| CN106769052B (en) | A kind of mechanical system rolling bearing intelligent failure diagnosis method based on clustering | |
| CN109800697B (en) | Transformer target detection and appearance defect identification method based on VGG-net style migration | |
| CN103152823B (en) | A kind of wireless indoor location method | |
| CN105003453A (en) | Online monitoring and fault diagnosis system of mine fan | |
| CN108275524A (en) | A kind of elevator maintenance operation monitoring and guiding device based on the assessment of the first multi-view video series of operations | |
| CN108830335A (en) | Photovoltaic plant fault early warning method and system | |
| CN102707713B (en) | Fault diagnosis system and method for automobile safety air bag assembly working procedure | |
| CN115438726A (en) | Device life and fault type prediction method and system based on digital twin technology | |
| CN113673884A (en) | Method for monitoring transformer area data based on digital twin mapping model | |
| CN115840394A (en) | Intelligent building carbon tube control system | |
| CN117406026A (en) | Power distribution network fault detection method suitable for distributed power supply | |
| CN107450428A (en) | A kind of main transformer oil level method for real-time monitoring equivalent based on translation | |
| CN102930408B (en) | A kind of 750kV electric grid secondary equipment state appraisal procedure based on information fusion | |
| CN104731970B (en) | The quality testing and control method of the multi-source heterogeneous data of highway | |
| CN112926656A (en) | Method, system and equipment for predicting state of circulating water pump of nuclear power plant | |
| CN108093210A (en) | A kind of transformer oil level warning system and its alarm method | |
| CN117142009A (en) | Scraper conveyor health state assessment method based on graph rolling network | |
| CN110495896B (en) | Wearable knee joint monitoring device and monitoring method based on GPRS communication | |
| CN110455370B (en) | Flood-prevention drought-resisting remote measuring display system | |
| CN106197759A (en) | A kind of temperature sensor control system | |
| CN116451121A (en) | Switch fault diagnosis method, system, equipment and medium | |
| CN110222814A (en) | The pipe recognition methods again of Ethylene Cracking Furnace Tubes based on embedded DCNN | |
| CN108595006A (en) | A kind of interactive system of the experimental facilities Automatic Control based on remote control |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120905 Termination date: 20210827 |