CN111914889A - Rectifying tower abnormal state identification method based on brief kernel principal component analysis - Google Patents
Rectifying tower abnormal state identification method based on brief kernel principal component analysis Download PDFInfo
- Publication number
- CN111914889A CN111914889A CN202010570653.9A CN202010570653A CN111914889A CN 111914889 A CN111914889 A CN 111914889A CN 202010570653 A CN202010570653 A CN 202010570653A CN 111914889 A CN111914889 A CN 111914889A
- Authority
- CN
- China
- Prior art keywords
- matrix
- data
- kernel
- rectifying tower
- vectors
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 43
- 238000000513 principal component analysis Methods 0.000 title claims abstract description 19
- 239000013598 vector Substances 0.000 claims abstract description 62
- 238000004364 calculation method Methods 0.000 claims abstract description 15
- 238000012544 monitoring process Methods 0.000 claims abstract description 8
- 238000001514 detection method Methods 0.000 claims abstract description 6
- 238000012216 screening Methods 0.000 claims abstract description 5
- 239000011159 matrix material Substances 0.000 claims description 38
- 238000005070 sampling Methods 0.000 claims description 34
- 239000007788 liquid Substances 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000010992 reflux Methods 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 238000010606 normalization Methods 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims 1
- 238000012549 training Methods 0.000 abstract description 4
- 230000006872 improvement Effects 0.000 abstract description 3
- 238000012824 chemical production Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000001311 chemical methods and process Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- BULVZWIRKLYCBC-UHFFFAOYSA-N phorate Chemical compound CCOP(=S)(OCC)SCSCC BULVZWIRKLYCBC-UHFFFAOYSA-N 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Testing Or Calibration Of Command Recording Devices (AREA)
Abstract
The invention discloses a rectifying tower abnormal state identification method based on simple kernel principal component analysis, and aims to improve the efficiency of detecting the abnormal state of rectifying tower equipment in real time by using a KPCA (kernel principal component analysis) method. In particular, rather than training a KPCA model using as much sample data as possible, a KPCA model is trained using typically representative sample data. The method has the advantages and characteristics that: and part of representative data are screened out through the edge point indexes to train the KPCA model, so that the online calculation amount of the kernel vector is greatly reduced. In addition, although the method has screening partial data, the method does not completely deny unselected data, and the unselected data is considered in online monitoring in a mode of constructing a detection index; finally, the improvement of the method in the computational efficiency is verified through a specific implementation case, and meanwhile, the identification capability of the method for the abnormal state is not reduced.
Description
Technical Field
The invention relates to a method for monitoring the running state of a chemical process, in particular to a rectifying tower abnormal state identification method based on simple kernel principal component analysis.
Background
Because the computer technology and the advanced measuring instrument technology are widely applied to chemical production, data information such as temperature, pressure, flow and the like in the chemical production process can be measured and stored in real time, and the mass sample data provides a solid data base for the current chemical intelligent manufacturing. In recent ten years, the detection of abnormal states occurring in the operation of chemical processes by using sampling data has received more and more attention in the field of safe chemical production. After ten years of development, a set of abnormal state identification methods based on multivariate Analysis algorithms such as Principal Component Analysis (PCA) and Partial Least Squares (PLS) has been established for monitoring abnormal states in chemical process. The core of these data-driven methodology techniques lies in the mining of the underlying characteristics of the sampled data. In other words, the data-driven models that are built are all intended to extract features that are hidden in the sampled data.
The rectifying tower equipment is a mass transfer and heat transfer device which is widely applied in petrochemical production and mainly used for realizing the purpose of separating substances by utilizing different volatility of each component in a mixture. Therefore, the real-time detection of whether the rectifying tower equipment has the abnormal operation state has important significance for the whole chemical production. The temperature of each layer of tower plate in the rectifying tower, the reflux flow and other data information can reflect whether the rectifying tower equipment works on the expected working condition or not. In consideration of the non-linear variation characteristic of the sampling data of the rectifying tower apparatus, an abnormal state identification method based on a Kernel PCA (Kernel PCA, abbreviation: KPCA) model may be used. The KPCA method projects original data to a high-dimensional space through a recessive nonlinear mapping function, so that nonlinear features can be directly subjected to feature analysis and extraction in the high-dimensional space according to the idea of the PCA method. Although the KPCA method can well process the nonlinear characteristics of the sampling data of the rectifying tower and achieve the purpose of identifying the abnormal state by monitoring the change of the nonlinear characteristics, the defects of the KPCA method technology are also obvious.
The data-driven method technology has a common characteristic that the more training data, the better. In other words, KPCA methods should use as much sample data as possible when training the model. However, when the off-line modeling stage of KPCA is completed and feature transformation is performed on new sampled data, the on-line computation amount is directly proportional to the data amount used in the off-line modeling stage. In addition, when the kernel vector is calculated for each sample data on line, nonlinear exponential operation is involved. If the sampling interval is short, the online identification of the abnormal state of the rectifying tower by using KPCA faces a serious real-time efficiency problem. In other words, when the KPCA method is using the current sampling time to perform operation and abnormal state detection, the next newly sampled sample data already appears. Therefore, how to improve the online computing efficiency of the KPCA method is a very interesting problem, especially in the identification of abnormal states of the rectification tower equipment with short sampling intervals.
Disclosure of Invention
The invention aims to solve the main technical problems that: how to improve the efficiency of detecting the abnormal state of the rectifying tower equipment in real time by using a KPCA method. The method realizes the improvement of the online identification efficiency by reducing the data volume during offline modeling. Simply speaking, rather than training a KPCA model using as many sample data as possible, a KPCA model is trained using typically representative sample data.
The technical scheme adopted by the method for solving the problems is as follows: a rectifying tower abnormal state identification method based on simplified kernel principal component analysis comprises the following steps:
step (1): by utilizing a measuring instrument arranged on the rectifying tower equipment, N sample data x are collected when the rectifying tower is in a normal operation state1，x2，…，xNWherein the sample data x at the ith sampling timet∈Rm×1The method is composed of m sampling data, and specifically comprises the following steps: column bottom liquid level, column bottom pressure, column bottom product flow, feed temperature, top reflux flow, condenser liquid level, and temperature of each layer of trays, i ∈ {1, 2, …, N }.
Step (2): for N sample data x1，x2，…，xNPerforming normalization to obtain N m × 1-dimensional data vectors
And (3): for N data vectors
Screening of edge points is performed, thereby retaining n data vectors z1，z2，…，znEstablishing a kernel principal component analysis model, wherein N is less than N, specificallyThe implementation of (2) is as follows.
Step (3.1): the initialization i is 1.
Step (3.2): according to the Euclidean distance
From N data vectors
In is
Searching out k data vectors with minimum Euclidean distance and recording the k data vectors
Where i ∈ {1, 2, …, n }.
Step (3.3): calculated according to the formula shown below
Normal vector f ofi。
Where c is 1, 2, …, k,
representation calculation
And
the euclidean distance between them.
Step (3.4): the edge point index ζ is calculated according to the formula shown belowi：
In the above formula, θciIs binary number, and the value law is as follows:
wherein the superscript T represents the transpose of a matrix or vector.
Step (3.5): judging whether the condition i is less than N; if yes, after i is set to i +1, returning to the step (3.2); if not, according to the value, the edge point index zeta is subjected to1，ζ2，…，ζNPerforming descending order arrangement, and recording the data vectors corresponding to the maximum n edge point indexes as z1，z2，…，zn。
And (4): using z1，z2，…，znA kernel principal component analysis model is established, and the specific implementation process is as follows.
Step (4.1): computing a kernel matrix K ∈ R according to the formula shown belown×nRow a, column b element K (a, b):
where, as the kernel parameter, it is generally desirable to be 100m, a ∈ {1, 2, …, n }, b ∈ {1, 2, …, n }, R ∈ {1, 2, …, n }, and R ∈ {1, 2, …, n }, wheren×nA matrix of real numbers representing dimensions n x n.
Step (4.2): the core matrix K is subjected to centralized processing according to the formula shown below to obtain a matrix
Wherein, the matrix IIn∈Rn×nAll elements in (1).
Step (4.3):solving the matrix
All the eigenvectors corresponding to the eigenvalues are arranged in descending order according to the size to obtain lambda1≥λ2≥…≥λnAnd a characteristic value λ1，λ2，…，λnRespectively corresponding feature vectors are p1，p2，…，pn。
Step (4.4): calculating a mean of the eigenvalues
And determining a characteristic value lambda1，λ2，…，λnMiddle to large mean value
The number of (d) is recorded as parameter d.
Step (4.5): calculating a kernel matrix J epsilon R according to the formula shown belowN×nRow i and column b element J (i, b):
wherein i belongs to {1, 2, …, n }, and b belongs to {1, 2, …, n }.
Step (4.6): the J is subjected to centering treatment according to the formula shown below to obtain
In the above formula, matrix IIN∈RN×nWherein all elements are 1, RN×nA matrix of real numbers representing dimensions N × N.
Step (4.7): according to the formula
And
separately calculating a principal score matrix S1And the secondary score matrix S2Then according to formula A ═ S1 TS1)-1Calculating a matrix A, wherein a primary transformation matrix P1＝[p1，p2，…，pd]Sub-transformation matrix P2＝[pd+1，pd+2，…，pn]。
Step (4.8): according to the formula D ═ diag { S ═ D1AS1 TQ and diag { S }2S2 TCalculating index vectors D and Q respectively, wherein diag { } represents an operation of converting a matrix diagonal element in braces into a vector.
Step (4.9): estimating the upper control limit D of the index vectors D and Q by using a kernel density estimation methodlimAnd Qlim。
As can be seen from steps (4.5) to (4.8), although some data vectors are screened out by using the edge point indexes to solve the feature value vectors of the KPCA, all the sampling data are used when the index vectors are calculated; on one hand, the sample data of the on-line calculation kernel vector is greatly reduced, which is why the method is called as the simplified kernel principal component analysis; on the other hand, all the collected sample data are assigned to use when determining the control upper limit.
The off-line modeling process of the method is completely finished, and then the real-time monitoring of the running state of the multi-rectifying-tower equipment is continuously carried out by utilizing on-line sampling data, so that the abnormal state is timely identified.
And (5): at the latest sampling time t, a data vector x consisting of m sampling data is obtained by measuring with a measuring instrument arranged on the rectifying tower equipmentt∈Rm×1And carrying out the same standardization processing as the step (2) to obtain a new data vector
And (6): the kernel vector k is calculated according to the formula shown belowt∈R1×nThe b-th element k in (1)t(b)：
In the above formula, b is ∈ {1, 2, …, n }, R1×nRepresenting a real number vector of dimension 1 xn.
As can be seen from the formula, the method can obtain the kernel vector by only calculating n times aiming at each data measured on line; in contrast, the conventional KPCA method requires N times of calculation to obtain a kernel vector; therefore, the online calculation amount of the method is obviously reduced.
And (7): kernel vector k is normalized according to the formula shown belowtImplementing a centering treatment to obtain
In the above formula, vector IIt∈R1×nAll elements in 1, matrix IIn∈Rn×nAll elements in (1).
And (8): according to the formula
And
calculating the principal score vectors s respectively1And the sub-score vector s2Then, according to formula Dt＝s1As1 TAnd Qt＝s2s2 TCalculating an identification index DtAnd Qt。
Step (ii) of(9): judging whether the conditions are met: dt≤DlimAnd Qt≤Qlim(ii) a If so, the running state of the rectifying tower at the current sampling moment is normal, and the step (5) is returned to continue to implement the abnormal state identification of the next latest sampling moment; if not, step (10) is performed to decide whether an abnormal state is identified.
Step (10): returning to the step (5) to continue to implement the abnormal state identification of the sample data at the next latest sampling moment, and if the detection indexes of 3 continuous sampling moments do not meet the judgment condition in the step (9), enabling the rectifying tower to enter an abnormal working state; otherwise, returning to the step (5) to continue to implement the abnormal state identification of the next latest sampling moment.
The advantages and features of the method of the present invention are shown below.
Firstly, the method aims to improve the calculation efficiency of KPCA when the KPCA is used for the online abnormal state recognition of the rectifying tower, and screens out part of representative data through edge point indexes to train a KPCA model, thereby greatly reducing the online calculation amount of kernel vectors; secondly, although the method has screening partial data, the method does not totally reject unselected data, and considers the unselected data into online abnormal state identification in a mode of constructing identification indexes; finally, the improvement of the method in the computational efficiency is verified through a specific implementation case, and meanwhile, the identification capability of the method for the abnormal state is not reduced.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention.
Fig. 2 is a schematic view of a real view and a schematic structural diagram of a rectifying tower device.
FIG. 3 is a comparison of the ability of the method of the present invention to identify various abnormal conditions of the rectification column.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
Referring to fig. 1, the invention discloses a method for identifying an abnormal state of a rectifying tower based on a simplified kernel principal component analysis, and a specific embodiment of the method is described below with reference to a specific application example.
As can be seen from the real view of the rectifying tower in FIG. 2, the rectifying tower is not only a single rectifying device, but also a matched reboiler at the bottom and a condenser at the top. As can be seen from the schematic structural diagram in fig. 2, the rectification column apparatus measuring instrument includes: flow meter, temperature instrument, three types of liquid level meter, there are 17 corresponding measured variables, specifically include: column bottom liquid level, column bottom pressure, column bottom product flow, feed temperature, top reflux flow, condenser liquid level, and temperature of each layer of trays (10 layers in total).
Step (1): by using a measuring instrument installed on the rectifying tower equipment, 1000 sample data x are acquired when the rectifying tower is in a normal operation state1，x2，…，x1000。
Step (2): for 1000 sample data x1，x2，…，x1000Performing normalization to obtain 1000 data vectors of 17 × 1 dimension
And (3): the 1000 data vectors are processed according to the steps (3.1) to (3.5)
The screening of the edge points is carried out, so that n is reserved as 400 data vectors z1，z2，…，z400And a kernel principal component analysis model is established, so that the number of samples is greatly reduced.
Due to the great reduction of the number of samples, the calculation time required for calculating the kernel vector on line (formula (r)) is reduced in proportion. Because the original KPCA needs to call the formula (1000) for each online measurement sample, the method of the invention only needs to call 400 times. Obviously, the method of the invention has high calculation efficiency.
And (4): using z1，z2，…，z400Establishing kernel principal component analysis model, and reserving principal and subordinate transformation matrix P1And P2Matrix A, and upper control limit DlimAnd Qlim。
After the off-line modeling stage is completed, the on-line state monitoring of the rectifying tower can be continuously carried out according to the following steps, and sample data at each new sampling moment is required to be utilized.
And (5): at the latest sampling time t, a data vector x consisting of m sampling data is obtained by measuring with a measuring instrument arranged on the rectifying tower equipmentt∈R17×1And carrying out the same standardization processing as the step (2) to obtain a new data vector
And (6): calculating a kernel vector k according to the formula [ + ]t∈R1×400The ith element k int(i)。
And (7): according to the formula, n is a nuclear vector ktImplementing a centering treatment to obtain
And (8): according to the formula
And
calculating the principal score vectors s respectively1And the sub-score vector s2Then, according to formula Dt＝s1As1 TAnd Qt＝s2s2 TCalculating an identification index DtAnd Qt。
And (9): judging whether the conditions are met: dt≤DlimAnd Qt≤Qlim(ii) a If so, the running state of the rectifying tower at the current sampling moment is normal, and the step (5) is returned to continue to implement the abnormal state identification of the next latest sampling moment; if not, step (10) is performed to decide whether an abnormal state is identified.
Step (10): returning to the step (5) to continue to implement the abnormal state identification of the sample data at the next latest sampling moment, and if the detection indexes of 3 continuous sampling moments do not meet the judgment condition in the step (9), enabling the rectifying tower to enter an abnormal working state; otherwise, returning to the step (5) to continue to implement the abnormal state identification of the next latest sampling moment.
Finally, in order to verify that the method of the invention does not cause negative influence on the state monitoring result while improving the calculation efficiency, the identification probability of the method of the invention and the traditional KPCA is contrastively analyzed aiming at the identification problems of the abnormal states of the rectifying tower, such as the abnormal temperature of condensed water, the abnormal temperature of a reboiler and the viscous state of a reflux valve. As can be seen from fig. 3, the method of the present invention is almost different from the conventional KPCA method. However, it has been analyzed before that the calculation efficiency of the method of the present invention is reduced from the original calculation that needs 1000 times to the present calculation that needs 400 times, and the calculation time can be obviously reduced.
Claims (2)
1. A rectifying tower abnormal state identification method based on simplified kernel principal component analysis is characterized by comprising the following steps:
firstly, the off-line modeling stage comprises the following steps (1) to (4);
step (1): by utilizing a measuring instrument arranged on the rectifying tower equipment, N sample data x are collected when the rectifying tower is in a normal operation state1，x2，…，xNWherein the sample data x at the ith sampling timei∈Rm×1The method is composed of m sampling data, and specifically comprises the following steps: the liquid level of the tower kettle, the pressure of the tower kettle, the flow rate of a product at the bottom of the tower kettle, the feeding flow rate, the feeding temperature, the top reflux flow rate, the liquid level of a condenser and the temperature of each layer of tower plates, wherein i belongs to {1, 2, …, N };
step (2): for N sample data x1，x2，…，xNPerforming normalization to obtain N m × 1-dimensional data vectors
And (3): for N data vectors
Screening of edge points is performed, thereby retaining n data vectors z1，z2，…，znEstablishing a kernel principal component analysis model, wherein N is less than N, and the specific implementation process is shown in the steps (3.1) to (3.5);
step (3.1): initializing i to 1;
step (3.2): according to the Euclidean distance
From N data vectors
In is
Searching out k data vectors with minimum Euclidean distance and recording the k data vectors
Where i ∈ {1, 2, …, n };
step (3.3): calculated according to the formula shown below
Normal vector f ofi；
Where c is 1, 2, …, k,
representation calculation
And
the upper label T represents the transposition of a matrix or a vector;
step (3.4): the edge point index ζ is calculated according to the formula shown belowi：
In the above formula, θciIs binary number, and the value law is as follows:
step (3.5): judging whether the condition i is less than N; if yes, after i is set to i +1, returning to the step (3.2); if not, according to the value, the edge point index zeta is subjected to1，ζ2，…，ζNPerforming descending order arrangement, and recording the data vectors corresponding to the maximum n edge point indexes as z1，z2，…，zn；
And (4): using z1，z2，…，znEstablishing kernel principal component analysis model, and reserving principal and subordinate transformation matrix P1And P2Matrix A, and upper control limit DlimAnd Qlim；
Secondly, after the off-line modeling stage is finished, the on-line state monitoring of the rectifying tower can be continuously carried out according to the following steps;
and (5): at the latest sampling time t, a data vector x consisting of m sampling data is obtained by measuring with a measuring instrument arranged on the rectifying tower equipmentt∈Rm×1And carrying out the same standardization processing as the step (2) to obtain a new data vector
And (6): the kernel vector k is calculated according to the formula shown belowt∈R1×nThe b-th element k in (1)t(b)：
In the above formula, b is ∈ {1, 2, …, n }, R1×nReal number vectors of 1 Xn dimension are expressed as kernel parameters;
and (7): kernel vector k is normalized according to the formula shown belowtImplementing a centering treatment to obtain
In the above formula, vector IIt∈R1×nAll elements in 1, matrix IIn∈Rn×nAll elements in (A) are 1;
and (8): according to the formula
And
calculating the principal score vectors s respectively1And the sub-score vector s2Then, according to formula Dt＝s1As1 TAnd Qt＝s2s2 TCalculating an identification index DtAnd Qt；
And (9): judging whether the conditions are met: dt≤DlimAnd Qt≤Qlim(ii) a If so, the running state of the rectifying tower at the current sampling moment is normal, and the step (5) is returned to continue to implement the abnormal state identification of the next latest sampling moment; if notThen step (10) is performed to decide whether an abnormal condition is identified;
step (10): returning to the step (5) to continue to implement the abnormal state identification of the sample data at the next latest sampling moment, and if the detection indexes of 3 continuous sampling moments do not meet the judgment condition in the step (9), enabling the rectifying tower to enter an abnormal working state; otherwise, returning to the step (5) to continue to implement the abnormal state identification of the next latest sampling moment.
2. The method for identifying the abnormal state of the rectifying tower based on the abbreviated nuclear principal component analysis as claimed in claim 1, wherein the specific implementation process of the step (4) is as follows:
step (4.1): computing a kernel matrix K ∈ R according to the formula shown belown×nRow a, column b element K (a, b):
wherein, as the kernel parameter, a belongs to {1, 2, …, n }, b belongs to {1, 2, …, n }, R belongs to {1, 2, …, n }, andn×na matrix of real numbers representing dimensions n × n;
step (4.2): the core matrix K is subjected to centralized processing according to the formula shown below to obtain a matrix
Wherein, the matrix IIn∈Rn×nAll elements in (A) are 1;
step (4.3): solving the matrix
All the eigenvectors corresponding to the eigenvalues are arranged in descending order according to the size to obtain lambda1≥λ2≥…≥λnAnd a characteristic value λ1，λ2，…，λnRespectively corresponding feature vectors are p1，p2，…，pnAnd the feature vector meets the length requirement:
step (4.4): calculating a mean of the eigenvalues
And determining a characteristic value lambda1，λ2，…，λnMiddle to large mean value
The number of (d) is recorded as parameter d;
step (4.5): calculating a kernel matrix J epsilon R according to the formula shown belowN×nRow i and column b element J (i, b):
wherein i belongs to {1, 2, …, n }, and b belongs to {1, 2, …, n };
step (4.6): the J is subjected to centering treatment according to the formula shown below to obtain
In the above formula, matrix IIN∈RN×nWherein all elements are 1, RN×nA real number matrix representing dimensions N × N;
step (4.7): according to the formula
And
separately calculating a principal score matrix S1And the secondary score matrix S2Then according to formula A ═ S1 TS1)-1Calculating a matrix A, wherein a primary transformation matrix P1＝[p1，p2，…，pd]Sub-transformation matrix P2＝[pd+1，pd+2，…，pn]；
Step (4.8): according to the formula D ═ diag { S ═ D1AS1 TQ and diag { S }2S2 TCalculating index vectors D and Q respectively, wherein diag { } represents operation of converting matrix diagonal elements in braces into vectors;
step (4.9): estimating the upper control limit D of the index vectors D and Q by using a kernel density estimation methodlimAnd Qlim。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010570653.9A CN111914889A (en) | 2020-06-13 | 2020-06-13 | Rectifying tower abnormal state identification method based on brief kernel principal component analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010570653.9A CN111914889A (en) | 2020-06-13 | 2020-06-13 | Rectifying tower abnormal state identification method based on brief kernel principal component analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111914889A true CN111914889A (en) | 2020-11-10 |
Family
ID=73226196
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010570653.9A Pending CN111914889A (en) | 2020-06-13 | 2020-06-13 | Rectifying tower abnormal state identification method based on brief kernel principal component analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111914889A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113065583A (en) * | 2021-03-13 | 2021-07-02 | 宁波大学科学技术学院 | Rectification process abnormity monitoring method based on online nonlinear discriminant feature analysis |
| CN113688895A (en) * | 2021-08-19 | 2021-11-23 | 广东工业大学 | Method and system for detecting abnormal firing zone of ceramic roller kiln based on simplified KECA |
| CN113935408A (en) * | 2021-09-22 | 2022-01-14 | 宁波大学科学技术学院 | Wind power generation equipment state monitoring method based on simplified nuclear principal component network |
| CN114970375A (en) * | 2022-07-29 | 2022-08-30 | 山东飞扬化工有限公司 | Rectification process monitoring method based on real-time sampling data |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050238200A1 (en) * | 2004-04-27 | 2005-10-27 | Rakesh Gupta | Simultaneous localization and mapping using multiple view feature descriptors |
| CN101308385A (en) * | 2008-07-11 | 2008-11-19 | 东北大学 | Non-linear procedure failure testing method based on two-dimensional dynamic kernel principal component analysis |
| CN109740641A (en) * | 2018-12-18 | 2019-05-10 | 清华四川能源互联网研究院 | A kind of area user identification zone method clustered based on principal component analysis and k-means |
| CN110009020A (en) * | 2019-03-12 | 2019-07-12 | 宁波大学 | A kind of non-linear process monitoring method based on multicore Principal Component Analysis Model |
| JP2020027342A (en) * | 2018-08-09 | 2020-02-20 | 富士電機株式会社 | Information processing apparatus, monitoring device and information processing method |
-
2020
- 2020-06-13 CN CN202010570653.9A patent/CN111914889A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20050238200A1 (en) * | 2004-04-27 | 2005-10-27 | Rakesh Gupta | Simultaneous localization and mapping using multiple view feature descriptors |
| CN101308385A (en) * | 2008-07-11 | 2008-11-19 | 东北大学 | Non-linear procedure failure testing method based on two-dimensional dynamic kernel principal component analysis |
| JP2020027342A (en) * | 2018-08-09 | 2020-02-20 | 富士電機株式会社 | Information processing apparatus, monitoring device and information processing method |
| CN109740641A (en) * | 2018-12-18 | 2019-05-10 | 清华四川能源互联网研究院 | A kind of area user identification zone method clustered based on principal component analysis and k-means |
| CN110009020A (en) * | 2019-03-12 | 2019-07-12 | 宁波大学 | A kind of non-linear process monitoring method based on multicore Principal Component Analysis Model |
Non-Patent Citations (1)
| Title |
|---|
| MANIA NAVI ET.AL.: "Sensor fault detection and isolation of an industrial gas turbine using partial adaptive KPCA", 《JOURNAL OF PROCESS CONTROL》, pages 37 - 48 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113065583A (en) * | 2021-03-13 | 2021-07-02 | 宁波大学科学技术学院 | Rectification process abnormity monitoring method based on online nonlinear discriminant feature analysis |
| CN113065583B (en) * | 2021-03-13 | 2023-11-14 | 宁波大学科学技术学院 | Rectification process abnormity monitoring method based on online nonlinear discriminant feature analysis |
| CN113688895A (en) * | 2021-08-19 | 2021-11-23 | 广东工业大学 | Method and system for detecting abnormal firing zone of ceramic roller kiln based on simplified KECA |
| CN113935408A (en) * | 2021-09-22 | 2022-01-14 | 宁波大学科学技术学院 | Wind power generation equipment state monitoring method based on simplified nuclear principal component network |
| CN114970375A (en) * | 2022-07-29 | 2022-08-30 | 山东飞扬化工有限公司 | Rectification process monitoring method based on real-time sampling data |
| CN114970375B (en) * | 2022-07-29 | 2022-11-04 | 山东飞扬化工有限公司 | Rectification process monitoring method based on real-time sampling data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111914889A (en) | Rectifying tower abnormal state identification method based on brief kernel principal component analysis | |
| CN108803520B (en) | Dynamic process monitoring method based on variable nonlinear autocorrelation rejection | |
| CN108508865B (en) | A kind of fault detection method based on distributing OSC-PLS regression model | |
| CN109240274B (en) | Industrial process fault diagnosis method based on high-order correlation | |
| CN110009020B (en) | Nonlinear process monitoring method based on multi-core principal component analysis model | |
| CN112904810B (en) | Process industry nonlinear process monitoring method based on effective feature selection | |
| CN109409425B (en) | Fault type identification method based on neighbor component analysis | |
| CN107153409A (en) | A kind of nongausian process monitoring method based on missing variable modeling thinking | |
| CN105868164B (en) | A kind of soft-measuring modeling method based on the linear dynamic system model for having supervision | |
| CN108345284A (en) | A kind of quality dependent failure detection method becoming gauge block based on two | |
| CN114757269A (en) | Complex process refined fault detection method based on local subspace-neighborhood preserving embedding | |
| CN108388234A (en) | A kind of fault monitoring method dividing changeable gauge block pca model based on correlation | |
| CN109144028A (en) | A kind of rectifying column efficiency deterioration detecting | |
| CN112580693A (en) | Petrochemical process fault diagnosis method based on self-help resampling neighborhood preserving embedding | |
| CN110427019B (en) | Industrial process fault classification method and control device based on multivariate discriminant analysis | |
| CN114879628A (en) | Multi-mode industrial process fault diagnosis method based on local maximum mean difference resistance | |
| CN111914471A (en) | Rectification tower fault detection method based on rapid nuclear independent component analysis | |
| CN111912638B (en) | Rectifying tower fault diagnosis method for online fault source identification | |
| CN114611067A (en) | Chemical process slow-change fault detection method based on typical variable dissimilarity analysis | |
| CN111914886B (en) | Nonlinear chemical process monitoring method based on online brief kernel learning | |
| CN111915121B (en) | Chemical process fault detection method based on generalized typical variable analysis | |
| CN115985411A (en) | Soft measurement method for polymer melt index based on Gaussian process regression model | |
| CN113156908B (en) | Multi-working-condition industrial process monitoring method and system with mechanism and data combined fusion | |
| CN111914214B (en) | PTA production process soft measurement method based on reduced KPLS model | |
| CN111913461B (en) | Distributed chemical process monitoring method based on regularization GCCA model |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240218 Address after: 518000 1002, Building A, Zhiyun Industrial Park, No. 13, Huaxing Road, Henglang Community, Longhua District, Shenzhen, Guangdong Province Applicant after: Shenzhen Wanzhida Technology Co.,Ltd. Country or region after: China Address before: 315211 Ningbo University, 818 Fenghua Road, Jiangbei District, Ningbo, Zhejiang Applicant before: Ningbo University Country or region before: China |
|
| TA01 | Transfer of patent application right |