CN104102773B - A kind of equipment fault early-warning and state monitoring method - Google Patents
A kind of equipment fault early-warning and state monitoring method Download PDFInfo
- Publication number
- CN104102773B CN104102773B CN201410322115.2A CN201410322115A CN104102773B CN 104102773 B CN104102773 B CN 104102773B CN 201410322115 A CN201410322115 A CN 201410322115A CN 104102773 B CN104102773 B CN 104102773B
- Authority
- CN
- China
- Prior art keywords
- data
- parameter
- model
- residual
- equipment
- 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.)
- Active
Links
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The present invention relates to equipment fault monitoring technical field, a kind of equipment fault early-warning and state monitoring method are particularly disclosed.The equipment fault early-warning and state monitoring method, it is characterised in that：Including setting up two processes of model and moving model；The step of setting up model process carries out data prediction work to training data first to obtain training data, then chooses dot-blur pattern using nonparametric learning algorithm, trains residual generator, obtains each parameter threshold residual value；The step of moving model, first to obtain real time data, data prediction work is carried out to real time data, then calculates each parameter residual error of real time data, and whether analysis residual error is normal to judge equipment state, and further positions failure cause.The characteristics of present invention both has the versatility based on data-driven method, robustness and strong adaptive ability, the deficiency that early warning result is difficult to analysis interpretation is turn avoid, the introducing additionally, due to nonparametric learning algorithm causes that the degree of accuracy of fault pre-alarming and efficiency are improved.
Description
（One）Technical field
The present invention relates to equipment fault monitoring technical field, more particularly to a kind of equipment fault early-warning and status monitoring side
Method.
（Two）Background technology
The possibility omen that equipment fault early-warning and status monitoring are obtained according to equipment moving law or observation is true in equipment
Before just breaking down, the unusual condition of timely HERALD equipment takes appropriate measures, so as to farthest reduce equipment event
Loss caused by barrier.Scale and complexity with apparatus and engineering control system increasingly increase, to ensure to produce
The safety and steady of journey, timely and effectively monitors and diagnoses process exception and just seem particularly urgent by reliable condition monitoring technology
With it is important.
Existing equipment fault early-warning technology is broadly divided into three major types：Method, Knowledge based engineering side based on mechanism model
Method and the method based on data-driven.
Method based on mechanism model is development earliest also the most deep fault pre-alarming and state monitoring method, and it is main
Including two stages：（1）Residual error produces the stage：Accurate Mathematical Modeling is set up by equipment operation mechanism defeated come estimating system
Go out, and it is compared with actual measured value, obtain residual error, the model that this stage builds is called residual generator；（2）Residual error is commented
The valency stage：Residual error is analyzed so that whether determination process breaks down, and further identification of defective type.Such method and control
Theory processed is combined closely, and mainly realizes residual error sequence using parameter Estimation, state estimation and the specific method of the class of equivalent space three
The structure of row, wherein method for estimating state are the most commonly used, can be used observer or Kalman filter to realize.
Knowledge based engineering method is main based on the enlightening Heuristics of associated specialist and operating personnel, qualitative or fixed
Annexation, fault propagation pattern during amount description between each unit etc., after there is abnormal sign in equipment by reasoning,
The inferential capabilities of the mode simulation process expert in monitoring such as deduction, so as to be automatically performed equipment fault early-warning and monitoring of equipment.
Such method has stronger dependence to expertise without accurate Mathematical Modeling, and conventional method mainly includes expert
System, failure decision tree, digraph, fuzzy logic etc..
Method based on data-driven is by the internal information founding mathematical models in mining process data and expression process
State, according to model come effective monitoring of implementation process.With the extensive use of smart instrumentation and computer memory technical, sea
The process data of amount is able to effectively monitor, collect and store, and such method is based on such mass data, in monitoring
With in warning algorithm it can be divided into based on signal transacting, rough set, machine learning, information fusion and multivariate statistics that this is five big again
Class algorithm, wherein machine learning algorithm are to develop branch the most active in theory and practice, and it includes Bayes classifier,
Neutral net, SVMs, k nearest neighbor algorithms, clustering algorithm, principal component analysis scheduling algorithm.
Monitoring method based on mechanism model can combine physical knowledge with monitoring system, by analyze residual error come
The mode for carrying out fault pre-alarming is more favorable to the understanding of professional, but because most mechanism models are simplified linear system
System, therefore when in face of the non-linear, free degree is higher and during Multivariable Coupling complication system, its using effect is unsatisfactory；
In addition, mechanism model is set up to complication system may pay huge cost；Furthermore, the noise shadow in actual industrial process
Ring, change of environmental factor etc. all improves the risk of model failure.Above reason all causes the monitoring side based on mechanism model
Method Detection results are not good, and range of application is not wide.
Knowledge based engineering monitoring method uses qualitatively model realization early warning and monitoring, when monitored object is relatively simple,
When process knowledge and knowhow are more sufficient, its performance is more excellent.But should be noted that the early warning of such method is accurate
Spend has very strong dependence to the abundant degree of expertise in knowledge base and the height of expertise level；Meanwhile, part
Expert's practical operation experience is difficult to be described with a kind of rational Formal Representation mode, and also having when system is complex can
The problems such as " conflict resolution ", " multiple shot array " can occur；In addition, the versatility of this kind of method is poor, and priori is complete
Property is typically difficult to ensure that.
It is pre- that fault pre-alarming and Condition Monitoring Technology based on data-driven directly set up failure by the historical data of system
Alert model, it is not necessary to know the accurate mechanism model of system, therefore its versatility and adaptive ability are all stronger.But due to this
The internal structure and mechanistic information of class method and indefinite system, so the analysis and explanation to early warning result are then relatively stranded
It is difficult；In addition, the algorithm based on data-driven such as machine learning is all mainly to be applied to fault diagnosis, and the failure for being based on data is pre-
Alert technology is also in the starting stage, and reliable and effective method is also relatively fewer；Furthermore, due to data volume is larger and based on data drive
The time complexity of dynamic algorithm is all universal higher, so the efficiency for how improving monitoring algorithm is also problem demanding prompt solution.
（Three）The content of the invention
The present invention is in order to make up the deficiencies in the prior art, there is provided one kind is not only suitable for Complex Nonlinear System, is easy to again
The equipment fault early-warning and state monitoring method of professional's analysis and understanding.
The present invention is achieved through the following technical solutions：
A kind of equipment fault early-warning and state monitoring method, it is characterised in that：Including setting up model and moving model two
Process；The step of setting up model process carries out data prediction work first to obtain training data to training data, then using non-
Parameter Learning Algorithm chooses dot-blur pattern, trains residual generator, obtains each parameter threshold residual value；The step of moving model, is for first
Obtain real time data, data prediction work carried out to real time data, then calculate each parameter residual error of real time data, analysis residual error with
Judge whether equipment state is normal, and further position failure cause.
The present invention is based on device data, using nonparametric learning algorithm (Non-Parametric Learning
) and support vector regression Algorithm（SVR）The mode being combined builds the residual generator in traditional mechanisms model, and
Residual error is analyzed to reach the purpose of fault pre-alarming.
More excellent technical scheme of the invention is：
It is described to set up in model process, the key parameter related to equipment safety operation is chosen, and to device history data
Screened, using the history health data under equipment normal operating condition as training data, then training data is deleted
Except invalid data, normalized pretreatment.
It is described to set up in model process, dot-blur pattern is chosen using nonparametric learning algorithm, first in calculating training data
The norm of each observation vector, and calculate the span N of normrange, then according to norm span NrangeIt is divided into h parts,
Again with Nrange/ h picks out several satisfactory observation vectors from training matrix and is added in dot-blur pattern D for step pitch,
Remove the remaining data after dot-blur pattern in training data to be preserved as residual matrix, in case the residual error threshold of training pattern
Used during value.
It is described to set up in model process, then on the basis of the dot-blur pattern of use nonparametric learning algorithm extraction, using non-thread
Property regression algorithm support vector regression train the estimate computation model of equipment, make regression machine be output as certain in dot-blur pattern
One parameter of one observation vector, it is the vectorial other specification to be input into, i.e., be fitted one with the other specification in observation vector
Individual parameter, after obtaining estimate computation model, you can obtain residual generator.
It is described to set up in model process, bring residual matrix into residual generator, you can obtain under normal operating condition, number
According to residual error scope, wherein the bound of each parameter residual error can be used as the threshold residual value of each parameter.
During the moving model, the pretreatment to real time data is that it is normalized, and makes each parameter value
It is all mapped in the interval of [0,1], normalized real time data is then substituted into residual generator, obtains each parameter corresponding
Residual error.
During the moving model, each parameter residual error of real time data is entered with each parameter threshold residual value obtained by training
Row contrast, thinks that equipment occurs abnormal if the bound that real time data residual error has exceeded setting, and the corresponding parameter that transfinites
As equipment may occur abnormal position, so as to further position failure cause.
The present invention is the fault early warning method of data-driven, while being effectively utilized the prior information of particular diagnosis object again
And historical information, compared to the fault early warning method that tradition utilizes residual generator, the present invention need not set up accurate mathematics machine
Reason model, therefore there is very strong versatility to complicated nonlinear system.
The present invention by way of presentation device key parameter actual value and system estimation value residual error, the operation of presentation device
Situation, therefore compared to the general fault early warning method based on data-driven, and with more directly perceived, be more easy to carry out early warning result
The characteristics of analysis and explanation.
The Early-warning Model that the present invention sets up is the nonlinear model based on multi-parameter, has taken into full account the phase between multi-parameter
Mutually influence, and the fault pre-alarming of single parameter is not based on, can more disclose the complicated causality and bar implied between parameter
Part relation, so that more accurately source of early warning failure.
The early warning system that the present invention sets up is calculated by estimate and threshold residual value analysis is constituted, and has effectively formed a base
In the early warning system of dynamic threshold, compared to the early warning system of traditional fixed threshold, with more preferable early warning effect, void is reduced
Report wrong report, effectively increases the accuracy rate of early warning.
The early warning system that the present invention sets up is based on nonparametric learning algorithm, changes model parameter number in traditional learning algorithm
Amount is fixed, the features such as model immobilizes after modeling, therefore with stronger adaptive ability, and be alleviate over-fitting/
The effective means of poor fitting problem.
The estimate computation model set up based on support vector regression that the present invention is constructed, is even unlimited higher-dimension
The nonlinear model of dimension, compared to traditional fitting of a polynomial or neural net regression algorithm etc., with accuracy rate and fortune higher
Calculate speed.
（Four）Brief description of the drawings
The present invention is further illustrated below in conjunction with the accompanying drawings.
Fig. 1 is the schematic flow sheet that model of the present invention sets up process；
Fig. 2 is the algorithm flow chart of nonparametric learning algorithm dot-blur pattern extraction process of the present invention；
Fig. 3 is the structural representation of the support vector regression that the present invention builds；
Fig. 4 is the schematic flow sheet of model running process of the present invention；
Fig. 5 is embodiment of the present invention bearing vibrations estimate and residual error curve synoptic diagram；
Fig. 6 is embodiment of the present invention motor coil Temperature estimate value and residual error curve synoptic diagram；
Fig. 7 is embodiment of the present invention drive end watt temperature estimate and residual error curve synoptic diagram.
（Five）Specific embodiment
Below in conjunction with the accompanying drawings and embodiment, the present invention is further elaborated：
The present invention is a kind of fault pre-alarming and state monitoring method suitable for nonlinear multivariable complication system, and it is based on setting
Standby data, combine nonparametric learning algorithm (Non-Parametric Learning Algorithm) and support vector regression
Machine is analyzed by the residual error to equipment real time data and set with reaching to build the residual generator in traditional mechanisms model
The function of standby status monitoring and failure automatic early-warning.The method mainly includes setting up two processes of model and moving model.
Fig. 1 is the flow chart that the present invention sets up model, and whole modeling process is mainly included the following steps that：
Step 1：Obtain training data
The key parameter related to equipment safety operation is chosen, and device history data is screened, it is normal with equipment
History health data under running status is used as training data.Need in the monitoring parameter and index for equipment, choose with
They are modeled with testing equipment operation conditions by the related key parameter of equipment safety operation.Assuming that certain equipment have n this
The monitoring parameter of sample, then this n index at a time detecting may make up an observation vector for n dimensions：
。
Therefore the historical data of the equipment is that can be considered the data observation matrix being made up of above-mentioned observation vector, is then needed
Observation vector in data observing matrix screened, get rid of the unusual part in historical data, system will covered just
The data of normal running status are used as model training data.Common Exception Type mainly includes Sudden Anomalies, beyond limit value, frequency
It is abnormal etc., should meet claimed below by the training data after screening：(1):Cover equipment as complete as possible normally to run
Situation；(2) each observation vector represents a normal operating condition of equipment；(3) each ginseng in each observation vector
Numerical value should be the sampled value of synchronization.
Step 2：Data prediction work is carried out to training data
Need to carry out training data deletion invalid data, the pretreatment measure such as normalization.Due in data acquisition
Problem that may be present, the original training data for initially obtaining there may be the invalid datas such as sky data, it is necessary to will be comprising invalid
The observation vector of data is deleted.Further, since the dimension of device model relevant parameter is different, and different parameters data are absolute
Value differs greatly, to ensure correctly to weigh the distance between different observation vectors using nonlinear operator, it is necessary to parameters
Measured value be normalized according to respective extreme value.Can be using the linear normalization mode as shown in 1 formula, to each parameter
Data are pre-processed, and it is interval interior that each parameter value is all mapped into [0 1]：
。
Step 3：Dot-blur pattern is chosen using nonparametric learning algorithm
The observation vector of enormous amount is may included due to training data, therefore in order to greatly improve model training and fortune
Capable efficiency, the present invention automatically extracts dot-blur pattern using nonparametric learning algorithm from training data, and in dot-blur pattern
On the basis of carry out the training of residual generator.
Model parameter quantity is fixed during nonparametric learning algorithm changes traditional learning algorithm, is fixed once model after modeling
Constant the features such as, the number of its model parameter changes with the size of training data, and model also can be according to phase after modeling
The difference of the data answered and change, therefore nonparametric learning algorithm has a stronger adaptive ability, and is effectively to alleviate
The means of over-fitting/poor fitting problem.
The dot-blur pattern obtained by training data is typical nonparametric learning algorithm, and dot-blur pattern should be covered as far as possible
Complete equipment normal operation, therefore take following method to extract dot-blur pattern：Each in training data is calculated first
The norm of observation vector, and calculate the span N of normrange, then according to norm span NrangeIt is divided into h parts, then with
Nrange/ h picks out several satisfactory observation vectors from training matrix and is added in dot-blur pattern D for step pitch, specifically
Dot-blur pattern extracting method algorithm flow chart shown in Figure 2, wherein m is the number of observation vector in training matrix, and δ is
One small positive number is used for controlling satisfactory observation vector quantity.The dot-blur pattern constructed using the method, will can train
Representational observation vector is selected into dot-blur pattern in data, has not only ensured not repeat typing but also can preferably overlay device is being just
Normal working space.
Remove the remaining data after dot-blur pattern in training data to be preserved as residual matrix, in case training pattern
Threshold residual value when use.
Step 4：Training residual generator
On the basis of the dot-blur pattern for being extracted using nonparametric learning algorithm before, using nonlinear regression algo support to
Amount regression machine trains the estimate computation model of equipment, the model to be used for the normal output of each parameter in estimating system.With biography
The estimate computation model of system is different, and the model is based on data, without setting up accurate mathematics mechanism model.
The cardinal principle of non-linear support vector regression of the present invention is, by introducing kernel function, will be former empty
Between vector nonlinear be mapped to a feature space, in this feature space Central Plains, problem is a linear separability, and can be asked
The optimal interface of this problem is solved, this interface is a non-linear interface in former space, support vector regression
Essence is to solve for following optimization problem, and the wherein C in object function is penalty factor, ξi,ξi *Respectively slack variable is upper
Lower limit,。
The structural representation of support vector regression can be expressed as the form in Fig. 3, and the wherein input of model is left side
Multiple parameter values, centre by kernel function K (x, xi) conversion, finally construct regression function output estimation value.From Fig. 3
It can be seen that, support vector regression model is a model for multiple input single output, therefore in training estimate computation model
When, regression machine should be made to be output as a parameter of a certain observation vector in dot-blur pattern, it is input into other ginsengs for the vector
Number, i.e., be fitted a parameter with the other specification in observation vector.So, an estimate computation model for n dimensions is trained, is needed
Train n support vector regression one-to-one with parameters.
After estimate computation model is obtained using the training of dot-blur pattern data, you can obtain residual generator, residual error
Computing formula be：Residual error=measured data-estimate.
Step 5：Obtain each parameter threshold residual value
Because residual matrix is to remove the training data after dot-blur pattern, therefore observation vector therein also represent and set
Each standby normal operating condition.Residual matrix is substituted into residual generator, you can obtain normal operating condition under, data it is residual
Difference scope, wherein the bound of each parameter residual error can be used as the threshold residual value of each parameter.When certain parameter residual error of real time data
More than the parameter threshold residual value when, then it is assumed that equipment occurs abnormal.
Fig. 4 is the flow chart of moving model of the present invention, is mainly included the following steps that：
Step 1：Obtain real time data
The data that real time data is monitored on-line when being equipment real time execution, it is consistent with the observation vector in training data
Also it is made up of the observation of the n key parameter related to equipment safety operation：
。
Step 2：Data prediction work is carried out to real time data
Real time data is needed also exist for be normalized, each parameter value is all mapped in the interval of [0 1],
The used formula of normalization is as shown in 1 formula, it should be noted that the x in formulaiminAnd ximaxIt is corresponding when being normalized with training data
Value used by parameter is identical.
Step 3：Calculate each parameter residual error of real time data
Real time data after normalization is substituted into residual generator, you can obtain the corresponding residual error of each parameter.Should be noted
Be that residual generator is made up of n estimate computation model, the input of each computation model is a reality for n-1 dimensions
When data observation vector, calculating is a remaining estimate for parameter.Residual error is by real time data and the difference table of estimate
Show.
Step 4：Analysis residual error simultaneously carries out early warning to failure
The step is residual error evaluation phase, and wind is whether there is to the operation that the residual error of each parameter is analyzed to judge equipment
Danger, if need to carry out fault pre-alarming.Specific method：By each parameter residual error of real time data and each parameter residual error obtained by training
Threshold value is contrasted, and thinks that equipment occurs if the bound that real time data residual error has exceeded setting abnormal, and corresponding super
Limit parameter is equipment may occur abnormal position, so as to further position failure cause.
Embodiment：
The present embodiment is monitoring object with the primary air fan of northern certain thermal power plant 1# units, and primary air fan is power plant's weight
The subsidiary engine equipment wanted, its complex structure, influence factor is more, it is difficult to set up accurate mathematical mechanism model, and easily sends out multiple malfunctions, symbol
The characteristics of closing the present invention targeted nonlinear multivariable system.By elaborating for the present embodiment, the present invention is further illustrated
Implementation process.
The embodiment of the present invention is as follows to the fault pre-alarming of certain power plant's primary air fan equipment and the implementation steps of status monitoring：
1. equipment fault early-warning and condition detecting system modeling process
（1）Obtain training data
The key parameter related to the primary air fan safe operation has 28, including real hair power (MW), fan outlet pressure
Power (kPa), bearing x to vibration (mm/s) etc., therefore the equipment observation vector be 28 dimensions vector：
。
The historical data of acquisition is the equipment 1 day on 2 1st, 2013 data of half a year of August in 2012, rejects history number
Abnormal data in, remaining normal historical data is the training data of model.
（2）Data prediction work is carried out to training data
Training data to picking out carries out deleting the pretreatment measure such as invalid data and normalization.Enter line according to 1 formula
Property normalization mode, by each parameter value be all mapped to [0 1] interval in.
（3）Dot-blur pattern is chosen using nonparametric learning algorithm
The extraction of dot-blur pattern in dot-blur pattern, this example is automatically extracted from training data using nonparametric learning algorithm
Method algorithm flow chart shown in Figure 2.
（4）Training residual generator
Because the equipment that this example is related to has 28 key parameters, it is therefore desirable to train 28 support vector regressions to divide
The estimate of these parameters is not calculated, and the wherein input of each regression machine is 27 dimensional vectors, is output as 1 dimensional vector, supporting vector
The structure of regression machine is as shown in Figure 3.After estimate computation model is obtained using the training of dot-blur pattern data, you can obtain residual
Difference generator.
（5）Obtain each parameter threshold residual value
Residual matrix is substituted into residual generator, you can obtain under normal operating condition, the residual error scope of data, wherein respectively
The bound of parameter residual error can be used as the threshold residual value of each parameter.
2. equipment fault early-warning and condition detecting system model running process
（1）Collection real time data
The sampling period of real time data is 1 minute, real time data consistent with the observation vector in training data in this example
It is made up of the observation of 28 key parameters.
（2）Data prediction work is carried out to real time data
The real time data of collection is normalized, is mapped in the interval of [0 1], normalize used formula such as 1
Shown in formula, the x in formulaiminAnd ximaxThe value used when being normalized with training data is identical.
（3）Calculate each parameter residual error of real time data
Real time data after normalization is substituted into residual generator, you can obtain the corresponding residual error of each parameter, residual error passes through
Real time data is represented with the difference of estimate.
（4）Analysis residual error simultaneously carries out early warning to failure
Each parameter residual error of real time data is contrasted with each parameter threshold residual value obtained by training, if residual error surpasses in real time
Cross threshold residual value and then send fault pre-alarming, and further identification of defective reason.
This example is carried out with certain thermal power plant primary air fan in August, 2012 historical data of 1 day on 2 1st, 2013
Modeling, and real time data is gathered since on 2 2nd, 2013 carry out status monitoring, Fig. 5 to Fig. 7 illustrates the implementation of this example
Effect, as shown in figure 5, after equipment normally runs a period of time, key parameter bearing x is to the actual value and estimate vibrated
Obvious deviation is there occurs, the residual error curve in Fig. 5 fully indicates the degradation trend of the equipment, it can be seen that bear vibration is residual
Difference gradually increases and has exceeded upper threshold, and so as to trigger early warning, and traditional method for early warning is more than in vibration signal
Alarm can be just triggered during 4.6mm/s, it turns out that, send alarm about 6 days afterwards in the early warning system of present invention design, the equipment
Because x is to the vibration catastrophe failure that excessively acutely something unexpected happened shuts down, the current failure of success prediction of the present invention.Fig. 6 and Fig. 7
It is two other key parameter electrode coil temperature (DEG C) and the operation curve figure of drive end watt temperature (DEG C), it can be seen that the two
The actual value of parameter is not deviated considerably from estimate, their residual error also without departing from threshold residual value, therefore this two
The individual parameter not broken down does not trigger early warning, therefore this example explanation source of early warning event that can not only succeed of the invention
Barrier, it is also possible to help quick positioning failure Producing reason.
Claims (5)
1. a kind of equipment fault early-warning and state monitoring method, it is characterised in that：Including setting up two mistakes of model and moving model
Journey；The step of setting up model process carries out data prediction work to training data first to obtain training data, then using non-ginseng
Mathematics practises algorithm picks dot-blur pattern, trains residual generator, obtains each parameter threshold residual value；The step of moving model is first to obtain
Real time data is taken, data prediction work is carried out to real time data, then calculate each parameter residual error of real time data, analyze residual error to sentence
Whether disconnected equipment state is normal, and further positions failure cause；It is described to set up in model process, using nonparametric learning algorithm
Dot-blur pattern is chosen, the norm of each observation vector in training data is calculated first, and calculate the span N of normrange, so
Afterwards according to norm span NrangeIt is divided into h parts, then with Nrange/ h picks out several from training matrix and conforms to for step pitch
The observation vector asked is added in dot-blur pattern D, and the remaining data after dot-blur pattern is removed in training data as remaining square
Battle array is preserved, in case being used during the threshold residual value of training pattern；It is described to set up in model process, learning to calculate using nonparametric
On the basis of the dot-blur pattern that method is extracted, mould is calculated using the estimate of nonlinear regression algo support vector regression training equipment
Type, makes regression machine be output as a parameter of a certain observation vector in dot-blur pattern, and it is the vectorial other specification to be input into, i.e.,
A parameter is fitted with the other specification in observation vector, after obtaining estimate computation model, you can obtain residual error and produce
Device.
2. equipment fault early-warning according to claim 1 and state monitoring method, it is characterised in that：It is described to set up model mistake
Cheng Zhong, chooses the key parameter related to equipment safety operation, and device history data is screened, and is normally run with equipment
Then history health data under state carries out deletion invalid data, normalized pre- place as training data to training data
Reason.
3. equipment fault early-warning according to claim 1 and state monitoring method, it is characterised in that：It is described to set up model mistake
Cheng Zhong, brings residual matrix into residual generator, you can obtain under normal operating condition, the residual error scope of data, wherein each ginseng
The bound of number residual error is the threshold residual value as each parameter.
4. equipment fault early-warning according to claim 1 and state monitoring method, it is characterised in that：The moving model mistake
Cheng Zhong, the pretreatment to real time data is that it is normalized, and each parameter value is all mapped to the interval of [0,1]
It is interior, normalized real time data is then substituted into residual generator, obtain the corresponding residual error of each parameter.
5. equipment fault early-warning according to claim 1 and state monitoring method, it is characterised in that：The moving model mistake
Cheng Zhong, each parameter residual error of real time data is contrasted with each parameter threshold residual value obtained by training, if real time data is residual
Difference has exceeded the bound of setting, and then to think that equipment occurs abnormal, and the corresponding parameter as equipment that transfinites may occur it is abnormal
Position, so as to further position failure cause.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410322115.2A CN104102773B (en) | 2014-07-05 | 2014-07-05 | A kind of equipment fault early-warning and state monitoring method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410322115.2A CN104102773B (en) | 2014-07-05 | 2014-07-05 | A kind of equipment fault early-warning and state monitoring method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104102773A CN104102773A (en) | 2014-10-15 |
| CN104102773B true CN104102773B (en) | 2017-06-06 |
Family
ID=51670924
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410322115.2A Active CN104102773B (en) | 2014-07-05 | 2014-07-05 | A kind of equipment fault early-warning and state monitoring method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104102773B (en) |
Families Citing this family (91)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105302848B (en) * | 2014-10-11 | 2018-11-13 | 山东鲁能软件技术有限公司 | A kind of assessed value calibration method of device intelligence early warning system |
| CN104317778A (en) * | 2014-10-30 | 2015-01-28 | 国家电网公司 | Massive monitoring data based substation equipment fault diagnosis method |
| CN104462846B (en) * | 2014-12-22 | 2017-11-10 | 山东鲁能软件技术有限公司 | A kind of equipment fault intelligent diagnosing method based on SVMs |
| CN104742895B (en) * | 2015-02-09 | 2018-01-05 | 中国计量学院 | A kind of car air braking system fault detection method based on analytic modell analytical model |
| CN104935464B (en) * | 2015-06-12 | 2018-07-06 | 北京奇虎科技有限公司 | The fault early warning method and device of a kind of web station system |
| CN105424224B (en) * | 2015-12-14 | 2018-04-06 | 国家电网公司 | A kind of disconnecting switch state monitoring method and device |
| CN105574284B (en) * | 2015-12-29 | 2019-06-14 | 山东鲁能软件技术有限公司 | A kind of Fault Diagnosis for Electrical Equipment method based on trend feature point |
| CN105548764B (en) * | 2015-12-29 | 2018-11-06 | 山东鲁能软件技术有限公司 | A kind of Fault Diagnosis for Electrical Equipment method |
| CN105721194B (en) * | 2016-01-13 | 2017-11-21 | 广州衡昊数据科技有限公司 | Mobile network potential faults intelligent positioning system |
| CN105928611A (en) * | 2016-04-14 | 2016-09-07 | 中国神华能源股份有限公司 | Fault early-warning system and method of mechanical device |
| KR101827108B1 (en) * | 2016-05-04 | 2018-02-07 | 두산중공업 주식회사 | Plant fault detection learning method and system |
| CN106649727B (en) * | 2016-12-23 | 2019-12-24 | 南京航空航天大学 | Database construction method for fault detection of unmanned aerial vehicle flight control system |
| CN106709662B (en) * | 2016-12-30 | 2021-07-02 | 山东鲁能软件技术有限公司 | Power equipment operation condition division method |
| JP6848656B2 (en) * | 2017-04-28 | 2021-03-24 | 横河電機株式会社 | Display device, display method and program |
| CN107505884B (en) * | 2017-07-31 | 2021-03-12 | 新奥泛能网络科技股份有限公司 | Diagnosis method of universal energy station equipment, cloud server and system |
| CN107701468B (en) * | 2017-09-27 | 2019-07-05 | 郑州大学 | A kind of online integrated monitoring of mixed-flow pump and device |
| WO2019071438A1 (en) * | 2017-10-10 | 2019-04-18 | 西门子公司 | Method and apparatus for monitoring state of device in process industry and medium |
| CN109766138A (en) | 2017-11-08 | 2019-05-17 | 广东欧珀移动通信有限公司 | Application program prediction model is established, preloads method, apparatus, medium and terminal |
| CN109784127B (en) * | 2017-11-10 | 2023-08-01 | 郑州大学 | Equipment health state early warning method and system thereof |
| CN109814937A (en) * | 2017-11-20 | 2019-05-28 | 广东欧珀移动通信有限公司 | Application program prediction model is established, preloads method, apparatus, medium and terminal |
| CN109814936A (en) * | 2017-11-20 | 2019-05-28 | 广东欧珀移动通信有限公司 | Application program prediction model is established, preloads method, apparatus, medium and terminal |
| CN107967489A (en) * | 2017-11-29 | 2018-04-27 | 中国科学院空间应用工程与技术中心 | A kind of method for detecting abnormality and system |
| CN108037387B (en) * | 2017-12-05 | 2020-06-19 | 北京能源集团有限责任公司 | Equipment fault analysis method and device based on clustering |
| CN108133326A (en) * | 2017-12-22 | 2018-06-08 | 华润电力（菏泽）有限公司 | A kind of status early warning method and system based on thermal power generating equipment |
| CN108199795B (en) * | 2017-12-29 | 2019-05-10 | 北京百分点信息科技有限公司 | A kind of monitoring method and device of equipment state |
| CN108108852B (en) * | 2018-01-05 | 2020-04-21 | 广东电科院能源技术有限责任公司 | Thermal power generating unit short-term reliability evaluation method and device based on fault early warning technology |
| CN108168548B (en) * | 2018-02-13 | 2022-03-15 | 南京师范大学 | Pedestrian inertial navigation system and method assisted by machine learning algorithm and model |
| CN108460207A (en) * | 2018-02-28 | 2018-08-28 | 上海华电电力发展有限公司 | A kind of fault early warning method of the generating set based on operation data model |
| CN108460144B (en) * | 2018-03-14 | 2021-11-12 | 西安华光信息技术有限责任公司 | Coal equipment fault early warning system and method based on machine learning |
| CN108446734A (en) * | 2018-03-20 | 2018-08-24 | 中科边缘智慧信息科技（苏州）有限公司 | Disk failure automatic prediction method based on artificial intelligence |
| CN108664696B (en) * | 2018-04-02 | 2023-04-07 | 国家计算机网络与信息安全管理中心 | Method and device for evaluating running state of water chiller |
| CN108596229B (en) * | 2018-04-13 | 2021-09-10 | 北京华电智慧科技产业有限公司 | Method and system for monitoring and diagnosing online abnormity |
| CN109344976A (en) * | 2018-08-24 | 2019-02-15 | 华能国际电力股份有限公司海门电厂 | A kind of electrical system operating status intelligent analysis method based on Keras |
| CN110968069B (en) * | 2018-09-28 | 2022-01-25 | 新疆金风科技股份有限公司 | Fault prediction method of wind generating set, corresponding device and electronic equipment |
| CN109660419B (en) * | 2018-10-08 | 2022-06-17 | 平安科技（深圳）有限公司 | Method, device, equipment and storage medium for predicting abnormity of network equipment |
| CN109524139B (en) * | 2018-10-23 | 2023-06-13 | 中核核电运行管理有限公司 | Real-time equipment performance monitoring method based on equipment working condition change |
| CN109586239B (en) * | 2018-12-10 | 2020-03-31 | 国网四川省电力公司电力科学研究院 | Real-time diagnosis and fault early warning method for intelligent substation |
| CN109657982B (en) * | 2018-12-20 | 2022-02-11 | 三一重能有限公司 | Fault early warning method and device |
| CN109726463A (en) * | 2018-12-25 | 2019-05-07 | 中铁隧道局集团有限公司 | A kind of shield TBM fault early warning method based on SVM algorithm |
| CN110046182A (en) * | 2019-03-21 | 2019-07-23 | 华能澜沧江水电股份有限公司 | A kind of huge hydroelectric power plant's intelligent alarm threshold setting method and system |
| JP7044175B2 (en) * | 2019-03-26 | 2022-03-30 | 東芝三菱電機産業システム株式会社 | Abnormality judgment support device |
| CN110173453A (en) * | 2019-04-04 | 2019-08-27 | 上海发电设备成套设计研究院有限责任公司 | A kind of online assessment method of power plant pressure fan state |
| CN110046146A (en) * | 2019-04-16 | 2019-07-23 | 中国联合网络通信集团有限公司 | The monitoring method and device of industrial equipment based on mobile edge calculations |
| CN110059968A (en) * | 2019-04-23 | 2019-07-26 | 深圳市华星光电技术有限公司 | Process data monitoring method and process data monitoring system |
| CN110322049B (en) * | 2019-06-03 | 2023-06-09 | 浙江图灵软件技术有限公司 | Public security big data early warning method |
| CN110298455B (en) * | 2019-06-28 | 2023-06-02 | 西安因联信息科技有限公司 | Mechanical equipment fault intelligent early warning method based on multivariate estimation prediction |
| CN110533294B (en) * | 2019-07-30 | 2024-04-16 | 中国核电工程有限公司 | Nuclear power plant operation fault alarm method based on artificial intelligence technology |
| CN110414154B (en) * | 2019-07-31 | 2022-09-30 | 北京天泽智云科技有限公司 | Fan component temperature abnormity detection and alarm method with double measuring points |
| CN111913443A (en) * | 2019-08-24 | 2020-11-10 | 南京鸿雁讯通信息科技有限公司 | Industrial equipment fault early warning method based on similarity |
| CN110501154B (en) * | 2019-09-05 | 2020-12-29 | 国网河北省电力有限公司电力科学研究院 | GIS equipment fault detection and positioning method based on MOSVR and boxplot analysis |
| CN110689203A (en) * | 2019-09-30 | 2020-01-14 | 国网山东省电力公司电力科学研究院 | Self-encoder-based primary fan fault early warning method for thermal power plant |
| CN110764474B (en) * | 2019-10-16 | 2023-01-31 | 上海电气集团股份有限公司 | Method and system for monitoring running state of equipment |
| CN112747416B (en) * | 2019-10-31 | 2022-04-05 | 北京国双科技有限公司 | Energy consumption prediction method and device for air conditioning system |
| CN110991666B (en) * | 2019-11-25 | 2023-09-15 | 远景智能国际私人投资有限公司 | Fault detection method, training device, training equipment and training equipment for model, and storage medium |
| CN110794799A (en) * | 2019-11-28 | 2020-02-14 | 桂林电子科技大学 | Big data system with fault diagnosis function applied to industrial production |
| CN110806743A (en) * | 2019-12-05 | 2020-02-18 | 成都天玙兴科技有限公司 | Equipment fault detection and early warning system and method based on artificial intelligence |
| CN113027696B (en) * | 2019-12-24 | 2022-11-15 | 新疆金风科技股份有限公司 | Fault diagnosis method and device of hydraulic variable pitch system |
| CN111275331A (en) * | 2020-01-20 | 2020-06-12 | 张家口卷烟厂有限责任公司 | Monitoring method and device for production system |
| CN111563685B (en) * | 2020-05-09 | 2022-03-08 | 国网江苏省电力有限公司 | Power generation equipment state early warning method based on auto-associative kernel regression algorithm |
| CN111814849B (en) * | 2020-06-22 | 2024-02-06 | 浙江大学 | DA-RNN-based wind turbine generator set key component fault early warning method |
| CN111814848B (en) * | 2020-06-22 | 2024-04-09 | 浙江大学 | Self-adaptive early warning strategy design method for temperature faults of wind turbine generator |
| CN111783486B (en) * | 2020-06-24 | 2023-08-22 | 中国工商银行股份有限公司 | Maintenance early warning method and device for card reader equipment |
| CN111767275B (en) * | 2020-06-28 | 2024-04-19 | 北京林克富华技术开发有限公司 | Data processing method and device and data processing system |
| CN112067335A (en) * | 2020-09-03 | 2020-12-11 | 华能国际电力股份有限公司玉环电厂 | Power plant blower fault early warning method based on multivariate state estimation |
| CN112036089A (en) * | 2020-09-03 | 2020-12-04 | 华能国际电力股份有限公司玉环电厂 | Coal mill fault early warning method based on DPC-MND and multivariate state estimation |
| CN112231849B (en) * | 2020-11-09 | 2023-03-10 | 北京国信会视科技有限公司 | Axle temperature fault detection method based on NEST and SPRT fusion algorithm |
| CN112629585A (en) * | 2020-12-02 | 2021-04-09 | 三门核电有限公司 | Equipment on-line monitoring method and device based on multi-dimensional parameter estimation |
| CN112633561A (en) * | 2020-12-09 | 2021-04-09 | 北京名道恒通信息技术有限公司 | Production risk intelligent prediction early warning method based on industrial big data |
| CN112611971B (en) * | 2020-12-23 | 2021-10-15 | 东北大学 | Networked direct current motor anomaly detection method based on data driving |
| CN112686389A (en) * | 2020-12-25 | 2021-04-20 | 中能融安(北京)科技有限公司 | Estimation method and estimation device for optimal value of equipment parameter |
| CN112668200B (en) * | 2021-01-06 | 2023-08-29 | 西安理工大学 | Special equipment safety analysis method and system |
| CN113011656B (en) * | 2021-03-22 | 2022-08-02 | 内蒙古电力(集团)有限责任公司内蒙古电力科学研究院分公司 | Power station auxiliary machine fault early warning method and system |
| CN112948163B (en) * | 2021-03-26 | 2023-09-19 | 中国航空无线电电子研究所 | Method for evaluating influence of equipment on functional failure based on BP neural network |
| CN113219939A (en) * | 2021-04-07 | 2021-08-06 | 山东润一智能科技有限公司 | Equipment fault prediction method and system based on residual autoregression |
| CN113420954A (en) * | 2021-05-08 | 2021-09-21 | 中国电建集团华东勘测设计研究院有限公司 | Engineering management service informatization method based on mechanism model |
| CN113110402B (en) * | 2021-05-24 | 2022-04-01 | 浙江大学 | Knowledge and data driven large-scale industrial system distributed state monitoring method |
| CN113657454B (en) * | 2021-07-23 | 2024-02-23 | 杭州安脉盛智能技术有限公司 | Nuclear power rotating machinery state monitoring method based on autoregressive BiGRU |
| CN113777908B (en) * | 2021-08-09 | 2023-11-21 | 杭州集益科技有限公司 | Control signal optimization method for multiple redundancy uncertain measurement parameters |
| CN113433917A (en) * | 2021-08-11 | 2021-09-24 | 内蒙古京隆发电有限责任公司 | DCS (distributed control system) for power plant and control method |
| CN113673600B (en) * | 2021-08-23 | 2024-02-02 | 中海石油气电集团有限责任公司 | Industrial signal abnormality early warning method, system, storage medium and computing device |
| CN113743531A (en) * | 2021-09-14 | 2021-12-03 | 华润电力技术研究院有限公司 | Equipment working condition early warning method and related device |
| CN113915153A (en) * | 2021-09-30 | 2022-01-11 | 山东浪潮通软信息科技有限公司 | Method, system, equipment and storage medium for detecting abnormality of mine ventilator |
| CN113984114B (en) * | 2021-10-18 | 2022-12-06 | 大连理工大学 | Method for diagnosing abnormality of underwater structure of ocean floating platform |
| CN113878214B (en) * | 2021-12-08 | 2022-03-25 | 苏芯物联技术(南京)有限公司 | Welding quality real-time detection method and system based on LSTM and residual distribution |
| CN114152880B (en) * | 2022-02-08 | 2022-04-12 | 湖南大学 | Soft package battery sensor fault online detection method |
| CN115059634B (en) * | 2022-06-14 | 2023-04-18 | 重庆英科铸数网络科技有限公司 | Fault diagnosis method and device for air blowing equipment and electronic equipment |
| CN115169650B (en) * | 2022-06-20 | 2022-11-22 | 四川观想科技股份有限公司 | Equipment health prediction method for big data analysis |
| CN115795999B (en) * | 2022-10-26 | 2023-08-01 | 国网新源控股有限公司 | Early warning method for abnormal performance of long-term service pumped storage unit |
| CN116049654B (en) * | 2023-02-07 | 2023-10-13 | 北京奥优石化机械有限公司 | Safety monitoring and early warning method and system for coal preparation equipment |
| CN116861164A (en) * | 2023-05-08 | 2023-10-10 | 华电电力科学研究院有限公司 | Turbine operation fault monitoring system |
| CN116384980B (en) * | 2023-05-25 | 2023-08-25 | 杭州青橄榄网络技术有限公司 | Repair reporting method and system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010011918A2 (en) * | 2008-07-24 | 2010-01-28 | University Of Cincinnati | Methods for prognosing mechanical systems |
| CN101930486B (en) * | 2010-07-12 | 2012-10-17 | 沈阳工业大学 | Device and method for predicating fan load index of wind powder plant |
| CN102331772B (en) * | 2011-03-30 | 2013-03-27 | 浙江省电力试验研究院 | Method for carrying out early warning of abnormal superheated steam temperature and fault diagnosis on direct current megawatt unit |
| CN103324834B (en) * | 2013-04-07 | 2016-08-17 | 北京航空航天大学 | Mechatronic Systems under a kind of completely cut off data qualification and critical component life-span prediction method thereof |
| CN103488091A (en) * | 2013-09-27 | 2014-01-01 | 上海交通大学 | Data-driving control process monitoring method based on dynamic component analysis |
-
2014
- 2014-07-05 CN CN201410322115.2A patent/CN104102773B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN104102773A (en) | 2014-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104102773B (en) | A kind of equipment fault early-warning and state monitoring method | |
| CN107301884B (en) | A kind of hybrid nuclear power station method for diagnosing faults | |
| CN103983453B (en) | A kind of executing agency of aero-engine and the differentiating method of sensor fault diagnosis | |
| CN104298225B (en) | Chemical process unusual service condition causality inference pattern is modeled and graphical representation method | |
| CN110738274A (en) | nuclear power device fault diagnosis method based on data driving | |
| KR101065767B1 (en) | Online early fault detection and diagnostic method for plant operation | |
| RU2313815C2 (en) | Device and method for controlling technical plant, which contains a set of systems, in particular, electric power plant | |
| CN107358347A (en) | Equipment cluster health state evaluation method based on industrial big data | |
| CN107835964A (en) | Control situation and the reasoning on control | |
| CN107291991B (en) | Early defect early warning method for wind turbine generator based on dynamic network sign | |
| JP2009053938A (en) | Equipment diagnosing system and equipment-diagnosing method on the basis of multiple model | |
| CN103324155A (en) | System monitoring | |
| Li et al. | Framework and case study of cognitive maintenance in Industry 4.0 | |
| CN116756909A (en) | Early warning diagnosis system of thermal power plant based on data model and mechanism model | |
| CN116483054A (en) | Industrial robot running state monitoring and early warning system and method | |
| CN117252051A (en) | Cable tunnel monitoring and early warning method and system based on digital twinning | |
| Gu et al. | Real-time novelty detection of an industrial gas turbine using performance deviation model and extreme function theory | |
| Nikitin et al. | Mechatronic modules diagnosis by use of fuzzy sets | |
| Hu et al. | Mutual information-based feature disentangled network for anomaly detection under variable working conditions | |
| CN114019935A (en) | Real-time detection and diagnosis system based on industrial Internet of things equipment | |
| CN104216397B (en) | Failure recognition and detection method for intelligent drive axle system | |
| CN113673600A (en) | Industrial signal abnormity early warning method, system, storage medium and computing equipment | |
| CN108092802A (en) | The numerical prediction maintenance system and method for ocean nuclear power platform nuclear power unit | |
| CN105279553B (en) | A kind of height adds to water system fault degree recognition methods | |
| Zhang | Comparison of data-driven and model-based methodologies of wind turbine fault detection with SCADA data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| ASS | Succession or assignment of patent right |
Owner name: ELECTRIC POWER RESEARCH INSTITUTE, STATE GRID SHAN Effective date: 20141117 |
|
| C41 | Transfer of patent application or patent right or utility model | ||
| C53 | Correction of patent for invention or patent application | ||
| CB03 | Change of inventor or designer information |
Inventor after: Ding Shugeng Inventor after: Xu Yang Inventor after: Li Haibin Inventor after: An Baijing Inventor before: Xu Yang Inventor before: Li Haibin Inventor before: An Baijing |
|
| COR | Change of bibliographic data |
Free format text: CORRECT: INVENTOR; FROM: XU YANG LI HAIBIN AN BAIJING TO: DING SHUGENG XU YANG LI HAIBIN AN BAIJING |
|
| TA01 | Transfer of patent application right |
Effective date of registration: 20141117 Address after: Xinluo Avenue high tech Zone of Ji'nan City, Shandong province 250101 silver bearing No. 2008 building B block 5 layer Applicant after: Shandong Luneng Software Technology Co., Ltd. Applicant after: Electric Power Research Institute of State Grid Shandong Electric Power Company Address before: Xinluo Avenue high tech Zone of Ji'nan City, Shandong province 250101 silver bearing No. 2008 building B block 5 layer Applicant before: Shandong Luneng Software Technology Co., Ltd. |
|
| C53 | Correction of patent for invention or patent application | ||
| CB03 | Change of inventor or designer information |
Inventor after: Ding Shugeng Inventor after: Xu Yang Inventor after: Li Haibin Inventor after: An Baijing Inventor before: Ding Shugeng Inventor before: Xu Yang Inventor before: Li Haibin Inventor before: An Baijing |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 250101 5th floor, block B, Yinhe building, 2008 Xinluo street, high tech Zone, Jinan City, Shandong Province Patentee after: Shandong luruan Digital Technology Co.,Ltd. Patentee after: ELECTRIC POWER RESEARCH INSTITUTE OF STATE GRID SHANDONG ELECTRIC POWER Co. Address before: 250101 5th floor, block B, Yinhe building, 2008 Xinluo street, high tech Zone, Jinan City, Shandong Province Patentee before: SHANDONG LUNENG SOFTWARE TECHNOLOGY Co.,Ltd. Patentee before: ELECTRIC POWER RESEARCH INSTITUTE OF STATE GRID SHANDONG ELECTRIC POWER Co. |