CN110109435A - A kind of on-line monitoring method improving two step Subspace partitions - Google Patents
A kind of on-line monitoring method improving two step Subspace partitions Download PDFInfo
- Publication number
- CN110109435A CN110109435A CN201910432047.8A CN201910432047A CN110109435A CN 110109435 A CN110109435 A CN 110109435A CN 201910432047 A CN201910432047 A CN 201910432047A CN 110109435 A CN110109435 A CN 110109435A
- Authority
- CN
- China
- Prior art keywords
- new
- data
- matrix
- spe
- subspace
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000012544 monitoring process Methods 0.000 title claims abstract description 19
- 238000005192 partition Methods 0.000 title claims abstract description 11
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000005070 sampling Methods 0.000 claims description 8
- 238000011068 loading method Methods 0.000 claims description 7
- 238000004519 manufacturing process Methods 0.000 claims description 7
- 241001269238 Data Species 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims description 3
- 230000007257 malfunction Effects 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000009628 steelmaking Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
- G05B19/4185—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by the network communication
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/30—Nc systems
- G05B2219/31—From computer integrated manufacturing till monitoring
- G05B2219/31088—Network communication between supervisor and cell, machine group
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
- Complex Calculations (AREA)
Abstract
The invention discloses a kind of on-line monitoring methods for improving two step Subspace partitions, include the following steps: step 1, according to the subspace progress dividing subset after division and are standardized according to the spatial character of sample to progress Subspace partition, and to initial data;Step 2 is monitored according to pivotal methods, and is calculated monitoring index and judged whether exception occur.The present invention can more accurately implement to monitor.
Description
Technical field
The invention belongs to automatic industrial process monitoring fields, and in particular to a kind of two step space of improvement of industrial process
The on-line monitoring method of division.
Background technique
The complicated variety of actual production process, which results in the precision of single space monitoring and rapidity, can not expire
Foot requires.Then the monitoring method for dividing subspace is generated under the premise of mass production data, these methods will be originally single
Space be divided into two or more subspaces according to certain feature, and be monitored in every sub-spaces.These monitoring methods
Data are refined and classified, and are all monitored for inhomogeneous data, although real work amount increases many, it is accurate
The advantages of it is very significant, therefore have great advantage in industrial application, and be widely used.
Summary of the invention
It is inaccurate object of the present invention is to be directed to single space monitoring, original space is divided into two sub-spaces, is obtained
A kind of new monitoring method can more accurately be implemented to monitor.This method is primarily based on industrial processes, and sampling is monitored
Otherwise the case where point, divide simultaneously divides data, is then monitored according to pivotal methods, and calculate monitoring to data space
Index judges whether to occur abnormal.
Step of the invention includes:
Step 1, according to the spatial character of sample to carrying out Subspace partition, and to initial data according to division after
Subspace carries out dividing subset and standardizes, and comprises the concrete steps that:
1-1. obtains the data at a certain moment by sampling, and is standardized:
Wherein X ' is the sampled data at a certain moment, and X is the data after standardization, X 'aIt is the average value of data X ',
X’sdIt is the variance of data X '.
Data after standardization are carried out host element extraction by 1-2.:
X=APT+E
Wherein, A is pivot matrix, and E is residual matrix, and P is matrix of loadings, and T is transposition symbol.
1-3. calculates the angle of each data and principal component space:
Wherein θiIt is the angle of i-th of data and principal component space, cos-1It is inverse cosine function, Ui、ViIt is two data squares
Battle array, QAIt is one group of orthonormal basis of main metadata, QiIt is one group of orthonormal basis of i-th of data, ΣiIt is preceding i data
With.
1-4. calculates the angled average value of institute:
Wherein θaIt is all angular averages, Σ (θi) be i angle and.
1-5. is by each angle and θaCompare, if θi< θaThen the data and pivot characteristic spatial simlanty are strong, and will own
Data group synthon space S (A) corresponding data for meeting all conditions is XA;If θi> θaThen the data and residual error feature are empty
Between similitude it is strong, and by all data group synthon space S (E) corresponding datas for meeting all conditions be XEIf θi=θaThen
The data are suitable with two sub-spaces similitudes, are divided into S (A) for convenience of calculating.
The mean value and variance of 1-6. calculating history square prediction error:
E=X (I-PPT)
SPE=eeT
Wherein e is the residual matrix of data, and I is unit matrix, SPE1…SPEiIt is the flat of the 1st i-th moment of moment ... respectively
Side's prediction error, m, v are the mean value and variance of history square prediction error respectively.
Step 2, malfunction monitoring, comprise the concrete steps that:
2-1. sampling obtains the data at some new moment, is obtained according to the modeling process of step 1-1 and 1-2:
Xnew=Anew(Pnew)T+Enew
Wherein XnewIt is new data, AnewIt is new data pivot matrix, E is the residual matrix of new data, and P is new data
Matrix of loadings.
2-2. is by new data XnewSubspace S is obtained according to step 1-3,1-4,1-5new(A)、Snew(E) corresponding data is distinguished
For Xnew A、Xnew E。
2-3. is to sub- space Snew(A) data Xnew AIt chooses the identical R data of order and obtains main procedure vector, it is remaining
J-R data obtain residual vector, calculate the residual matrix and score vector of new data:
tnew=XnewPnew
Wherein enewIt is the residual matrix of new data, tnewIt is the score vector of new data, pnewIt is the load moment of new data
Battle array.
The square prediction error of 2-4. calculating new data:
SPEnew=(enew)(enew)T
2-5. calculates real time information statistic:
(Tnew)2=tnewΛ-1(tnew)T
Wherein Λ-1It isCorresponding eigenvalue cluster at diagonal matrix it is inverse.
2-6. obtains the expression formula abbreviation of real time information statistic:
(Tnew)2=XnewPΛ-1P-1(Xnew)T
2-7. calculatingSPEα:
Wherein FR,αIt is that the F that freedom degree is R and α is distributed;
WhereinIt is that freedom degree isχ2Distribution, m and v are history square prediction error SPE respectivelyαMean value and
Variance.
2-8. comparison result, if meetingSPEnew< SPEαThen production is normal at this time.
2-9. is to sub- space Snew(E) step 2-3 is repeated to step 2-8.
Specific embodiment
The invention will be further described below.
By taking industry steel-making as an example, to make steel furnace temperature as monitoring objective:
Method and step of the invention includes:
Step 1 obtains the spatial character of converter temperature data according to sampling to carrying out Subspace partition, and to original temperature
Degree evidence carries out dividing subset according to the subspace after division and standardizes, and comprises the concrete steps that:
Step 1, according to the host element of temperature data and it is Gaussian Subspace partition is carried out to data, and to original temperature number
According to according to space dividing subset and standardizing, comprise the concrete steps that:
1-1. obtains the temperature data at a certain moment by sampling, and is standardized:
Wherein X ' is the temperature data at a certain moment, and X is the temperature data after standardization, X 'aIt is temperature data X '
Average value, X 'sdIt is the variance of temperature data X '.
Temperature data after standardization is carried out host element extraction by 1-2.:
X=APT+E
Wherein, A is main temperature matrix, and E is temperature residual matrix, and P is matrix of loadings, and T is transposition symbol.
1-3. calculates the angle of each temperature data Yu main temperature data space:
Wherein θiIt is the angle of i-th of temperature data Yu main temperature space, cos-1It is inverse cosine function, Ui、ViIt is two numbers
According to matrix, QAIt is one group of orthonormal basis of main temperature data, QiIt is one group of orthonormal basis of i-th of temperature data, ΣiIt is
The sum of preceding i temperature data.
1-4. calculates the angled average value of institute:
Wherein θaIt is all angular averages, Σ (θi) be i angle and.
1-5. is by the angle and θ of each temperatureaCompare, if θi< θaThen the temperature data is similar to main temperature feature space
Property it is strong, and by all temperature datas for meeting all conditions be combined into subspace S (A) corresponding temperature data be XA;If θi> θa
Then the temperature data and temperature residual error feature space similitude are strong, and all temperature datas for meeting all conditions are combined into
S (E) corresponding data in subspace is XEIf θi=θaThen the temperature data is suitable with two sub-spaces similitudes, draws for convenience of calculating
Enter S (A).
The mean value and variance of 1-6. calculating historical temperature square prediction error:
E=X (I-PPT)
SPE=eeT
Wherein e is the residual matrix of temperature data, and I is unit matrix, SPE1…SPEiIt was the 1st i-th moment of moment ... respectively
Temperature square prediction error, m, v are the mean value and variance of historical temperature square prediction error respectively.
Step 2, malfunction monitoring, comprise the concrete steps that:
2-1. sampling obtains the temperature data at some new moment, is obtained according to the modeling process of step 1-1 and 1-2:
Xnew=Anew(Pnew)T+Enew
Wherein XnewIt is new temperature data, AnewIt is the main temperature matrix of new temperature data, E is that the temperature of new temperature data is residual
Poor matrix, P are the matrixs of loadings of new temperature data.
2-2. is by new temperature data XnewSubspace S is obtained according to step 1-3,1-4,1-5new(A)、Snew(E) corresponding temperature
Data are respectively Xnew A、Xnew E。
2-3. is to sub- space Snew(A) temperature data Xnew AChoose the identical R temperature data of order obtain main temperature to
Amount, remaining J-R temperature data obtain residuals temperatures vector, calculate the temperature residual matrix and score vector of new temperature data:
tnew=XnewPnew
Wherein enewIt is the residual matrix of new temperature data, tnewIt is the score vector of new temperature data, pnewIt is new temperature number
According to matrix of loadings.
2-4. calculates the square prediction error of new temperature data:
SPEnew=(enew)(enew)T
2-5. calculates real time information statistic:
(Tnew)2=tnewΛ-1(tnew)T
Wherein Λ-1It isCorresponding eigenvalue cluster at diagonal matrix it is inverse.
2-6. obtains the expression formula abbreviation of real time information statistic:
(Tnew)2=XnewPΛ-1P-1(Xnew)T
2-7. calculatingSPEα:
Wherein FR,αIt is that the F that freedom degree is R and α is distributed;
WhereinIt is that freedom degree isχ2Distribution, m and v are history square prediction error SPE respectivelyαMean value and
Variance.
2-8. comparison result, if meetingSPEnew< SPEαThen temperature is normal at this time.
2-9. is to sub- space Snew(E) step 2-3 is repeated to step 2-8.
Claims (3)
1. a kind of on-line monitoring method for improving two step Subspace partitions, includes the following steps:
Step 1, according to the spatial character of sample to carrying out Subspace partition, and it is empty according to the son after division to initial data
Between dividing subset and standardize;
Step 2 is monitored according to pivotal methods, and is calculated monitoring index and judged whether exception occur.
2. improving the on-line monitoring method of two step Subspace partitions as described in claim 1, it is characterised in that:
The step 1 is specific as follows:
1-1. obtains the data at a certain moment by sampling, and is standardized:
Wherein X ' is the sampled data at a certain moment, and X is the data after standardization, X 'aIt is the average value of data X ', X 'sd
It is the variance of data X ';
Data after standardization are carried out host element extraction by 1-2.:
X=APT+E
Wherein, A is pivot matrix, and E is residual matrix, and P is matrix of loadings, and T is transposition symbol;
1-3. calculates the angle of each data and principal component space:
Wherein θiIt is the angle of i-th of data and principal component space, cos-1It is inverse cosine function, Ui、ViIt is two data matrixes, QA
It is one group of orthonormal basis of main metadata, QiIt is one group of orthonormal basis of i-th of data, ΣiBe preceding i data and;
1-4. calculates the angled average value of institute:
Wherein θaIt is all angular averages, Σ (θi) be i angle and;
1-5. is by each angle and θaCompare, if θi< θaThen the data and pivot characteristic spatial simlanty are strong, and meet all
Data group synthon space S (A) corresponding data of all conditions is XA;If θi> θaThe then data and residual error feature space phase
It is strong like property, and be X by all data group synthon space S (E) corresponding datas for meeting all conditionsEIf θi=θaThe then number
According to, for convenience calculating cut-in S (A) suitable with two sub-spaces similitudes;
The mean value and variance of 1-6. calculating history square prediction error:
E=X (I-PPT)
SPE=eeT
Wherein e is the residual matrix of data, and I is unit matrix, SPE1…SPEiIt is the square pre- of the 1st i-th moment of moment ... respectively
Error is surveyed, m, v are the mean value and variance of history square prediction error respectively.
3. improving the on-line monitoring method of two step Subspace partitions as claimed in claim 2, it is characterised in that:
The step 2 specifically:
2-1. sampling obtains the data at some new moment, is obtained according to the modeling process of step 1-1 and 1-2:
Xnew=Anew(Pnew)T+Enew
Wherein XnewIt is new data, AnewIt is new data pivot matrix, E is the residual matrix of new data, and P is the load of new data
Matrix;
2-2. is by new data XnewSubspace S is obtained according to step 1-3,1-4,1-5new(A)、Snew(E) corresponding data is respectively
Xnew A、Xnew E;
2-3. is to sub- space Snew(A) data Xnew AIt chooses the identical R data of order and obtains main procedure vector, remaining J-R is a
Data obtain residual vector, calculate the residual matrix and score vector of new data:
tnew=XnewPnew
Wherein enewIt is the residual matrix of new data, tnewIt is the score vector of new data, pnewIt is the matrix of loadings of new data;
The square prediction error of 2-4. calculating new data:
SPEnew=(enew)(enew)T
2-5. calculates real time information statistic:
(Tnew)2=tnewΛ-1(tnew)T
Wherein Λ-1It isCorresponding eigenvalue cluster at diagonal matrix it is inverse;
2-6. obtains the expression formula abbreviation of real time information statistic:
(Tnew)2=XnewPΛ-1P-1(Xnew)T;
2-7. calculatingSPEα:
Wherein FR,αIt is that the F that freedom degree is R and α is distributed;
WhereinIt is that freedom degree isχ2Distribution, m and v are history square prediction error SPE respectivelyαMean value and variance;
2-8. comparison result, if meetingSPEnew< SPEαThen production is normal at this time;
2-9. is to sub- space Snew(E) step 2-3 is repeated to step 2-8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910432047.8A CN110109435B (en) | 2019-05-22 | 2019-05-22 | Online monitoring method for improving two-step subspace division |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910432047.8A CN110109435B (en) | 2019-05-22 | 2019-05-22 | Online monitoring method for improving two-step subspace division |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110109435A true CN110109435A (en) | 2019-08-09 |
CN110109435B CN110109435B (en) | 2020-07-24 |
Family
ID=67491796
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910432047.8A Active CN110109435B (en) | 2019-05-22 | 2019-05-22 | Online monitoring method for improving two-step subspace division |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110109435B (en) |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050060103A1 (en) * | 2003-09-12 | 2005-03-17 | Tokyo Electron Limited | Method and system of diagnosing a processing system using adaptive multivariate analysis |
CN101446827A (en) * | 2008-11-06 | 2009-06-03 | 西安交通大学 | Process fault analysis device of process industry system and method therefor |
CN103197663A (en) * | 2013-03-07 | 2013-07-10 | 北京信息科技大学 | Method and system of failure prediction |
CN103336507A (en) * | 2013-06-24 | 2013-10-02 | 浙江大学 | Statistical modeling and on-line monitoring method based on multimodality collaboration time frame automatic division |
CN103389701A (en) * | 2013-07-15 | 2013-11-13 | 浙江大学 | Plant-level process fault detection and diagnosis method based on distributed data model |
CN103729562A (en) * | 2013-12-31 | 2014-04-16 | 东北大学 | Penicillin fermentation process fault monitoring method based on reconstruction discriminatory analysis |
CN104898646A (en) * | 2015-04-30 | 2015-09-09 | 东北大学 | KPCA-based fused magnesium furnace fault diagnosis method for fault separation and reconstruction |
CN109189028A (en) * | 2018-10-15 | 2019-01-11 | 江南大学 | PCA method for diagnosing faults based on muti-piece information extraction |
-
2019
- 2019-05-22 CN CN201910432047.8A patent/CN110109435B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050060103A1 (en) * | 2003-09-12 | 2005-03-17 | Tokyo Electron Limited | Method and system of diagnosing a processing system using adaptive multivariate analysis |
CN101446827A (en) * | 2008-11-06 | 2009-06-03 | 西安交通大学 | Process fault analysis device of process industry system and method therefor |
CN103197663A (en) * | 2013-03-07 | 2013-07-10 | 北京信息科技大学 | Method and system of failure prediction |
CN103336507A (en) * | 2013-06-24 | 2013-10-02 | 浙江大学 | Statistical modeling and on-line monitoring method based on multimodality collaboration time frame automatic division |
CN103389701A (en) * | 2013-07-15 | 2013-11-13 | 浙江大学 | Plant-level process fault detection and diagnosis method based on distributed data model |
CN103729562A (en) * | 2013-12-31 | 2014-04-16 | 东北大学 | Penicillin fermentation process fault monitoring method based on reconstruction discriminatory analysis |
CN104898646A (en) * | 2015-04-30 | 2015-09-09 | 东北大学 | KPCA-based fused magnesium furnace fault diagnosis method for fault separation and reconstruction |
CN109189028A (en) * | 2018-10-15 | 2019-01-11 | 江南大学 | PCA method for diagnosing faults based on muti-piece information extraction |
Non-Patent Citations (1)
Title |
---|
常玉清,王姝,王福利,谭帅,刘炎: "基于多PCA模型的过程监测方法", 《仪器仪表学报》 * |
Also Published As
Publication number | Publication date |
---|---|
CN110109435B (en) | 2020-07-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112699913B (en) | Method and device for diagnosing abnormal relationship of household transformer in transformer area | |
CN108062565B (en) | Double-principal element-dynamic core principal element analysis fault diagnosis method based on chemical engineering TE process | |
CN108508865B (en) | A kind of fault detection method based on distributing OSC-PLS regression model | |
CN107292023B (en) | Bridge structure state diagnosis method based on narrow-area characteristics of damage index system | |
Wang et al. | A novel multi‐mode data processing method and its application in industrial process monitoring | |
CN106647650B (en) | Distributing Industrial Process Monitoring method based on variable weighting pca model | |
Wang et al. | Multiscale neighborhood normalization-based multiple dynamic PCA monitoring method for batch processes with frequent operations | |
CN102601881B (en) | Method for monitoring on-line quality and updating prediction model of rubber hardness | |
CN108388234A (en) | A kind of fault monitoring method dividing changeable gauge block pca model based on correlation | |
CN113570200A (en) | Power grid operation state monitoring method and system based on multidimensional information | |
CN108469805A (en) | A kind of distributing dynamic process monitoring method based on dynamic optimal selection | |
CN108345294B (en) | A kind of fault detection method based on distributing principal component regression model | |
CN105955214A (en) | Batch process fault detection method based on sample timing sequence and neighborhood similarity information | |
CN113298162A (en) | Bridge health monitoring method and system based on K-means algorithm | |
CA2034492A1 (en) | Method of monitoring the operational state of a system | |
CN114819636A (en) | Industrial production data processing method and system based on SPC detection | |
CN115372748A (en) | Method and system for processing maintenance data of cable | |
CN112989587B (en) | Online analysis method and system for degradation cause of capacitive voltage transformer | |
CN107346122A (en) | Improve the manufacturing process multivariate quality diagnostic classification device of fuzzy support vector machine | |
CN112241832B (en) | Product quality grading evaluation standard design method and system | |
US9043249B2 (en) | Automatic chemical assay classification using a space enhancing proximity | |
CN110109435A (en) | A kind of on-line monitoring method improving two step Subspace partitions | |
CN113609216A (en) | Block chain-based product quality diagnosis method, device, equipment and storage medium | |
CN117521518A (en) | Magnesium alloy heat treatment process optimization method based on machine learning | |
CN112434739A (en) | Chemical process fault diagnosis method of support vector machine based on multi-core learning |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |