CN113190792A - Ethylene cracking furnace operation state monitoring method based on neighbor local abnormal factors - Google Patents
Ethylene cracking furnace operation state monitoring method based on neighbor local abnormal factors Download PDFInfo
- Publication number
- CN113190792A CN113190792A CN202110440247.5A CN202110440247A CN113190792A CN 113190792 A CN113190792 A CN 113190792A CN 202110440247 A CN202110440247 A CN 202110440247A CN 113190792 A CN113190792 A CN 113190792A
- Authority
- CN
- China
- Prior art keywords
- cracking furnace
- ethylene cracking
- matrix
- vector
- data
- 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
- VGGSQFUCUMXWEO-UHFFFAOYSA-N Ethene Chemical compound C=C VGGSQFUCUMXWEO-UHFFFAOYSA-N 0.000 title claims abstract description 74
- 239000005977 Ethylene Substances 0.000 title claims abstract description 74
- 238000005336 cracking Methods 0.000 title claims abstract description 72
- 230000002159 abnormal effect Effects 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 title claims abstract description 21
- 238000012544 monitoring process Methods 0.000 title claims abstract description 17
- 239000011159 matrix material Substances 0.000 claims abstract description 37
- 230000005856 abnormality Effects 0.000 claims abstract description 7
- 239000013598 vector Substances 0.000 claims description 62
- 238000005070 sampling Methods 0.000 claims description 22
- 238000010606 normalization Methods 0.000 claims description 6
- 238000007599 discharging Methods 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 230000005855 radiation Effects 0.000 claims description 4
- 238000010790 dilution Methods 0.000 claims description 3
- 239000012895 dilution Substances 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000006386 neutralization reaction Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 5
- 238000004458 analytical method Methods 0.000 abstract description 4
- 239000007789 gas Substances 0.000 description 9
- 238000004519 manufacturing process Methods 0.000 description 6
- 238000009826 distribution Methods 0.000 description 2
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 239000002994 raw material Substances 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 239000006227 byproduct Substances 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000001311 chemical methods and process Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000010779 crude oil Substances 0.000 description 1
- -1 ethylene, propylene Chemical group 0.000 description 1
- 239000003345 natural gas Substances 0.000 description 1
- 238000011112 process operation Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 239000000243 solution Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- 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]
Abstract
The invention discloses an ethylene cracking furnace running state monitoring method based on a neighbor local abnormal factor, which realizes real-time characteristic analysis on sampled data of an ethylene cracking furnace on the premise of considering unsteady state running and time-varying characteristics of the ethylene cracking furnace, thereby effectively monitoring the running state of the ethylene cracking furnace. The method solves the problem of unsteady state change of the sampled data of the ethylene cracking furnace by a technical means of constructing the local abnormal factor of the neighbor in real time, maximizes the difference between the online new sampled data and the reference data set of the neighbor in real time, and judges whether the new sampled data reflects the abnormality of the ethylene cracking furnace. In addition, the method of the invention realizes the real-time update of the monitoring model of the running state of the ethylene cracking furnace by introducing the normal latest sample data into the normal data matrix in real time and deleting the oldest sample data.
Description
Technical Field
The invention relates to a method for monitoring the running state of a chemical process, in particular to a method for monitoring the running state of an ethylene cracking furnace based on a neighbor local abnormal factor.
Background
The ethylene cracking furnace is used for processing cracking gas and is of the types of double radiation chambers, single radiation chambers and millisecond furnaces. The ethylene cracking furnace is the core equipment of an ethylene production device, and mainly has the functions of processing various raw materials such as natural gas, refinery gas, crude oil, naphtha and the like into cracking gas, providing the cracking gas for other ethylene devices, and finally processing the cracking gas into ethylene, propylene and various byproducts. The production capacity and the technical height of the ethylene cracking furnace directly determine the production scale, the yield and the product quality of the whole set of ethylene equipment, so that the ethylene cracking furnace plays a leading role in the ethylene production equipment and even the whole set of petrochemical production.
Therefore, the operation of the ethylene cracking furnace in the expected state (namely, the normal state) plays an important role in the whole ethylene production process, and the real-time monitoring of the operation state of the ethylene cracking furnace is essential. However, conventional monitoring of the operating conditions of ethylene cracking furnaces is accomplished by monitoring changes in one or a few key technical parameters. The monitoring of the change of the key variable according to the univariate thought can only monitor whether a single technical parameter is within an allowable change range in real time, and cannot monitor the running state of the ethylene cracking furnace in real time from the perspective of the whole system.
In addition, in the actual operation of the ethylene cracking furnace, the state of the ethylene cracking furnace is influenced by the fluctuation of the factors such as raw material feeding, steam feeding and the like, so that the operation of the ethylene cracking furnace is unstable. The data collected under this condition are all non-steady state sample data. The traditional data-driven approach assumes a gaussian distribution, which is not true in sample data for ethylene cracking furnaces. Therefore, data-driven monitoring of the operating conditions of ethylene cracking furnaces is known to be a relatively difficult problem. Fortunately, due to the promotion of industrial big data at the present stage, a plurality of measuring instruments and computer-aided systems are installed on the ethylene cracking furnace object, and the sample data can be measured in real time at short intervals. The method lays a full data foundation for implementing data-driven monitoring of the operation state of the ethylene cracking furnace.
In the case of sufficient sampled data volume of the ethylene cracking furnace, the unsteady nature of the process operation causes the gaussian distribution assumption to remain false. However, considering that the abnormal operation state of the ethylene cracking furnace is unknown, and any number of abnormalities or any variable of abnormalities may indicate that the operation state of the ethylene cracking furnace is abnormal. Therefore, the characteristic analysis of the sample data of the ethylene cracking furnace should be carried out on line in real time. In addition, the unsteady state characteristic of the sampled data is further considered, the historical sample data of the ethylene cracking furnace cannot be taken as a whole, and the sample data under the normal operation state of the ethylene cracking furnace can be updated in real time.
Disclosure of Invention
The invention aims to solve the main technical problems that: on the premise of considering the unsteady state operation and the time-varying characteristic of the ethylene cracking furnace, how to perform real-time characteristic analysis on the sampled data of the ethylene cracking furnace is carried out, so that the operation state of the ethylene cracking furnace is effectively monitored. Specifically, the method constructs a near-neighbor local abnormal factor for online sampling data of the ethylene cracking furnace in real time, and monitors whether the running state of the ethylene cracking furnace is abnormal or not by judging whether the abnormal factor is over-limit or not. And then, the sample data in the normal operation state is updated to the reference data set in real time, so that the self-adaptive updating characteristic of the model is ensured.
The technical scheme adopted by the method for solving the problems is as follows: a method for monitoring the running state of an ethylene cracking furnace based on a neighbor local abnormal factor comprises the following steps.
Step (1): obtaining sample data x of n sampling moments of an ethylene cracking furnace in a normal operation state1,x2,…,xnAnd composing it into a data matrix X ═ X of dimension n × m1,x2,…,xn]T(ii) a Wherein, the sample data x of the ith sampling timei∈R33×1The 33 measurement data in the process sequentially comprise the feeding pressure of 6 furnace tube radiation chambers, the feeding flow of 6 furnace tubes, the dilution steam flow of 6 furnace tubes, the discharging temperature of 6 furnace tubes, the cross temperature of 6 furnace tubes, the average discharging temperature of 1 ethylene cracking furnace, the furnace wall gas flow of 1 ethylene cracking furnace and the bottom gas flow of 1 ethylene cracking furnace, i belongs to {1, 2, …, n }, R } in turn33×1A real number vector of 33 × 1 dimensions is represented, R represents a real number set, and the upper symbol T represents a matrix or a transpose of a vector.
Step (2): according to the formulaFor column vectors z in the data matrix X, respectively1,z2,…,z33Performing normalization to obtain normalized data matrixWherein z iskAndrespectively represent X andcolumn vector of the kth column, k ∈ {1, 2, …, 33}, μkAnd deltakRespectively representing column vectors zkMean and standard deviation of all elements in (a).
And (3): determining the normal upper limit value Q of the adjacent local abnormal factor according to the following steps (3.1) to (3.6)lim。
Step (3.1): the initialization i is 1.
Step (3.2): is provided withThen, X is addediThe row vector of the ith row in the matrix is deleted, thereby obtaining a matrix X consisting of n-1 row vectorsi∈R(n-1)×33(ii) a Wherein R is(n-1)×33A real number matrix representing (n-1) × 33 dimensions.
Step (3.3): calculating XiMiddle row vector and row vector yiDistance between, then XiIs in with yiThe N row vectors with the minimum distance between them form a neighbor reference matrixWherein R isN×33Representing a matrix of real numbers of Nx 33 dimensions, yiTo representThe row vector of the ith row.
Step (3.4): solving generalized eigenvalue problemMedium maximum eigenvalue lambdaiCorresponding feature vector alphaiThen according to the formulaUpdating the feature vector alphai。
Step (3.5): according to the formulaCalculating a nearest neighbor local anomaly factor QiThen, judging whether the condition i is less than n; if yes, setting i to i +1, and then returning to the step (3.2); if not, obtaining Q1,Q2,…,Qn。
Step (3.6): will Q1,Q2,…,QnThe maximum value in (1) is recorded as the normal upper limit value Q of the nearest local abnormality factorlim。
And (4): at the latest sampling time t, sample data x of the ethylene cracking furnace is collectedt∈R1×33According to the formulaFor xtEach element in the data is normalized to obtain normalized online data vectorWherein x ist(k) Andrespectively represent xtAndthe kth element in (1).
And (5): computingMiddle row vector and number of linesData vectorThe distance between the twoNeutralization ofThe N row vectors with the minimum distance between them form a neighbor reference matrix
And (6): solving generalized eigenvalue problemMedium maximum eigenvalue lambdatCorresponding feature vector ptThen according to the formulaUpdating feature vector pt。
And (7): according to the formulaCalculating the nearest neighbor local anomaly factor Q of the latest sampling time ttThen, whether or not the condition Q is satisfied is judgedt≤Qlim(ii) a If yes, executing step (8); if not, executing step (10).
And (8): according to the formula shown belowRow vector y of middle and last n-1 rows2,y3,…,ynAndafter merging into a matrix Y, step (9) is performed.
And (9): according to the formulaRespectively to column vectors in matrix YPerforming normalization to obtain normalized matrixAccording to the formulaAndthe mean values mu are updated separately1,μ2,…,μ33And standard deviation delta1,δ2,…,δ33Then, set upAnd returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment; wherein the content of the first and second substances,and η k represent Y andthe column vector of the k-th column,andrespectively represent column vectorsMean and standard deviation of all elements in (a).
Step (10): returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment until the adjacent local abnormal factors of the continuous 6 latest sampling moments are obtained, and judging whether the 6 adjacent local abnormal factors are all larger than Qlim(ii) a If yes, triggering an abnormal alarm; if not, the operation state of the ethylene cracking furnace is normal, and the step (8) is executed.
In the above implementation steps, step (3.4) and step (6) both involve the problem of generalized eigenvalues with similar calculations. The generalized eigenvalue problem solving is actually a key core process for implementing the construction of the nearest neighbor local anomaly factor. Taking step (3.4) as an example, the so-called neighbor local feature factor is intended to pass through a feature vector αiTo quantize the row vector yiWith respect to neighbor reference matrixDegree of abnormality of (d).
Simply put, it is the neighbor reference matrixAlpha warpiThe range of change is minimized after projective transformation, and the line vector y is simultaneously enablediAlpha warpiThe projective transformation is as maximum as possible. By taking the basic idea of discriminant analysis algorithm as a reference, an objective function can be constructed as follows:
without loss of generality, the denominator in the objective function can be setThus, the above equation can be equivalently transformed into a maximization problem with constraints as shown below:
the solution of equation (c) above can use the classical lambertian multiplier method, i.e. construct the lambertian function L as shown below.
Then, solve L relative to alphaiAnd set equal to 0, the following equation is obtained:
and then converting into a generalized eigenvalue problem in step (3.4):if the two sides of the equation of the generalized eigenvalue problem are multiplied by the left sideThen can obtainThus, λiEqual to the maximum target in the above formula (c), i.e. the generalized eigenvalue problemThe largest eigenvalue needs to be calculated.
By carrying out the steps described above, the advantages of the method of the invention are presented below.
The method solves the problem of unsteady state change of the sampled data of the ethylene cracking furnace by a technical means of constructing a neighbor local abnormal factor, maximizes the difference between the online new sampled data and a neighbor reference data set in real time, and judges whether the new sampled data reflects the abnormality of the ethylene cracking furnace. In addition, the method of the invention realizes the real-time update of the monitoring model of the running state of the ethylene cracking furnace by introducing the normal latest sample data into the normal data matrix in real time and deleting the oldest sample data.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention.
FIG. 2 is a schematic view of an ethylene cracking furnace.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
As shown in figure 1, the invention discloses an ethylene cracking furnace operation state monitoring method based on a neighbor local abnormal factor. The following describes a specific embodiment of the method of the present invention in conjunction with a specific application example.
The ethylene cracking furnace shown in fig. 2 is a tap device for producing ethylene in a certain chemical industry in China, and the device comprises 6 furnace tubes, wherein 5 variables which can be measured by each furnace tube specifically comprise: feed pressure, feed flow, dilution steam flow, exit temperature, and crossover temperature of the radiant chamber. Thus, there were 30 measured variables for 6 tubes. In addition, three measurable variables are provided, specifically furnace wall gas flow of the ethylene cracking furnace, bottom gas flow of the ethylene cracking furnace and average discharging temperature of the ethylene cracking furnace. In summary, in this embodiment, 33 measurement data can be acquired at each sampling time, and a 33 × 1 dimensional sample data is formed.
Step (1): acquiring sample data x of n sampling moments of an ethylene cracking furnace in a normal operation state1,x2,…,xnAnd recording it as a data matrix X ═ X in n × m dimensions1,x2,…,xn]T。
Step (2): according to the formulaFor column vectors z in the data matrix X, respectively1,z2,…,z33Performing normalization to obtain normalized data matrix
And (3): determining the normal upper limit value Q of the adjacent local abnormal factor according to the steps (3.1) to (3.6)lim。
And (4): at the latest sampling time t, sample data x of the ethylene cracking furnace is collectedt∈R1×33According to the formulaFor xtEach element in the data is normalized to obtain normalized online data vectorWherein x ist(k) Andrespectively represent xtAndthe kth element in (1).
And (5): computingMiddle row vector and on-line data vectorThe distance between the twoNeutralization ofThe N row vectors with the minimum distance between them form a neighbor reference matrix
In step (5), calculatingMiddle ith row vector yiAnd online data vectorThe specific embodiments of the distance between are:
and (6): solving generalized eigenvalue problemMedium maximum eigenvalue lambdatCorresponding feature vector ptThen according to the formulaUpdating feature vector pt。
And (7): according to the formulaCalculating the nearest neighbor local anomaly factor Q of the latest sampling time ttThen, whether or not the condition Q is satisfied is judgedt≤Qlim(ii) a If yes, executing step (8); if not, executing step (10).
And (8): according to the formulaRow vector y of middle and last n-1 rows2,y3,…,ynAndare combined into a matrix Y1Then, step (9) is performed.
And (9): according to the formulaAre respectively aligned with matrix Y1Column vector ofPerforming normalization to obtain normalized matrixAccording to the formulaAndthe mean values mu are updated separatelykAnd standard deviation deltakThen, set upAnd returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment.
Step (10): returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment until the adjacent local abnormal factors of the continuous 6 latest sampling moments are obtained, and judging whether the 6 adjacent local abnormal factors are all larger than Qlim(ii) a If yes, triggering an abnormal alarm; if not, the operation state of the ethylene cracking furnace is normal, and the step (8) is executed.
Claims (1)
1. A method for monitoring the running state of an ethylene cracking furnace based on neighbor local abnormal factors is characterized by comprising the following steps:
step (1): acquiring sample data x of n sampling moments of an ethylene cracking furnace in a normal operation state1,x2,…,xnAnd composing it into a data matrix X ═ X in n × m dimensions1,x2,…,xn]T(ii) a Wherein, the sample data x of the ith sampling timei∈R33×1The 33 measurement data specifically comprise the feeding pressure of 6 furnace tube radiation chambers, the feeding flow of 6 furnace tubes, the dilution steam flow of 6 furnace tubes, the discharging temperature of 6 furnace tubes, the crossing temperature of 6 furnace tubes, the average discharging temperature of 1 ethylene cracking furnace, the furnace wall gas flow of 1 ethylene cracking furnace and the bottom gas flow of 1 ethylene cracking furnace, i belongs to {1, 2, …, n }, R } of33×1Representing a real number vector of 33 x 1 dimensions, R representing a real number set, and the upper label T representing a matrix or a transpose of a vector;
step (2): according to the formulaFor column vectors z in the data matrix X, respectively1,z2,…,z33Performing normalization to obtain normalized data matrixWherein z iskAndrespectively represent X andcolumn vector of the kth column, k ∈ {1, 2, …, 33}, μkAnd deltakRespectively representing column vectors zkMean and standard deviation of all elements in (a);
and (3): determining the normal upper limit value Q of the adjacent local abnormal factor according to the following steps (3.1) to (3.6)lim;
Step (3.1): initializing i to 1;
step (3.2): is provided withThen, X is addediThe row vector of the ith row in the matrix is deleted, thereby obtaining a matrix X consisting of n-1 row vectorsi∈R(n-1)×33(ii) a Wherein R is(n-1)×33Is represented by (n)-1) x 33 dimensional real matrix;
step (3.3): calculating XiMiddle row vector and row vector yiDistance between, then XiIs in with yiThe N row vectors with the minimum distance between them form a neighbor reference matrixWherein R isN×33Representing a matrix of real numbers of Nx 33 dimensions, yiTo representA row vector of the ith row;
step (3.4): solving generalized eigenvalue problemMedium maximum eigenvalue lambdaiCorresponding feature vector alphaiThen according to the formulaUpdating the feature vector alphai;
Step (3.5): according to the formulaCalculating a nearest neighbor local anomaly factor QiThen, judging whether the condition i is less than n; if yes, setting i to i +1, and then returning to the step (3.2); if not, obtaining Q1,Q2,…,Qn;
Step (3.6): will Q1,Q2,…,QnThe maximum value in (1) is recorded as the normal upper limit value Q of the nearest local abnormality factorlim;
And (4): at the latest sampling time t, sample data x of the ethylene cracking furnace is collectedt∈R1×33According to the formulaFor xtEach ofCarrying out standardization processing on each element to obtain an online data vector after standardization processingWherein x ist(k) Andrespectively represent xtAndthe kth element in (1);
and (5): computingMiddle row vector and on-line data vectorThe distance between the twoNeutralization ofThe N row vectors with the minimum distance between them form a neighbor reference matrix
And (6): solving generalized eigenvalue problemMedium maximum eigenvalue lambdatCorresponding feature vector ptThen according to the formulaUpdating feature vector pt;
And (7): according to the formulaCalculating the nearest neighbor local anomaly factor Q of the latest sampling time ttThen, whether or not the condition Q is satisfied is judgedt≤Qlim(ii) a If yes, executing step (8); if not, executing the step (10);
and (8): according to the formula shown belowRow vector y of middle and last n-1 rows2,y3,…,ynAndafter the matrix Y is combined, the step (9) is executed;
and (9): according to the formulaAre respectively aligned with matrix Y1Column vector ofPerforming normalization to obtain normalized matrixAccording to the formulaAndthe mean values mu are updated separately1,μ2,…,μ33And standard deviation delta1,δ2,…,δ33Then, set upAnd returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment; wherein the content of the first and second substances,and ηkRespectively represent Y andthe column vector of the k-th column,andrespectively represent column vectorsMean and standard deviation of all elements in (a);
step (10): returning to the step (4) to continue to monitor the running state of the ethylene cracking furnace at the latest sampling moment until the adjacent local abnormal factors of the continuous 6 latest sampling moments are obtained, and judging whether the 6 adjacent local abnormal factors are all larger than Qlim(ii) a If yes, triggering an abnormal alarm; if not, the operation state of the ethylene cracking furnace is normal, and the step (8) is executed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110440247.5A CN113190792B (en) | 2021-04-18 | 2021-04-18 | Ethylene cracking furnace running state monitoring method based on neighbor local abnormal factors |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110440247.5A CN113190792B (en) | 2021-04-18 | 2021-04-18 | Ethylene cracking furnace running state monitoring method based on neighbor local abnormal factors |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113190792A true CN113190792A (en) | 2021-07-30 |
CN113190792B CN113190792B (en) | 2023-10-24 |
Family
ID=76978113
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110440247.5A Active CN113190792B (en) | 2021-04-18 | 2021-04-18 | Ethylene cracking furnace running state monitoring method based on neighbor local abnormal factors |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113190792B (en) |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105518654A (en) * | 2013-08-23 | 2016-04-20 | 应用材料公司 | K-nearest neighbor-based method and system to provide multi-variate analysis on tool process data |
JP2016099863A (en) * | 2014-11-25 | 2016-05-30 | 日本電信電話株式会社 | Abnormal value detection device and operation method thereof |
CN110362608A (en) * | 2019-06-11 | 2019-10-22 | 广东工业大学 | Energy consumption method for detecting abnormality based on rain flow method and local outlier factor |
CN110865260A (en) * | 2019-11-29 | 2020-03-06 | 南京信息工程大学 | Method for monitoring and evaluating MOV actual state based on outlier detection |
CN111913460A (en) * | 2019-05-20 | 2020-11-10 | 宁波大学 | Fault monitoring method based on sequence correlation local preserving projection algorithm |
CN111914206A (en) * | 2019-05-20 | 2020-11-10 | 宁波大学 | Process monitoring method based on novel dynamic neighbor preserving embedding algorithm |
CN111914887A (en) * | 2020-06-13 | 2020-11-10 | 宁波大学 | Novel multi-mode chemical process abnormal state detection method |
-
2021
- 2021-04-18 CN CN202110440247.5A patent/CN113190792B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105518654A (en) * | 2013-08-23 | 2016-04-20 | 应用材料公司 | K-nearest neighbor-based method and system to provide multi-variate analysis on tool process data |
JP2016099863A (en) * | 2014-11-25 | 2016-05-30 | 日本電信電話株式会社 | Abnormal value detection device and operation method thereof |
CN111913460A (en) * | 2019-05-20 | 2020-11-10 | 宁波大学 | Fault monitoring method based on sequence correlation local preserving projection algorithm |
CN111914206A (en) * | 2019-05-20 | 2020-11-10 | 宁波大学 | Process monitoring method based on novel dynamic neighbor preserving embedding algorithm |
CN110362608A (en) * | 2019-06-11 | 2019-10-22 | 广东工业大学 | Energy consumption method for detecting abnormality based on rain flow method and local outlier factor |
CN110865260A (en) * | 2019-11-29 | 2020-03-06 | 南京信息工程大学 | Method for monitoring and evaluating MOV actual state based on outlier detection |
CN111914887A (en) * | 2020-06-13 | 2020-11-10 | 宁波大学 | Novel multi-mode chemical process abnormal state detection method |
Non-Patent Citations (5)
Title |
---|
HEHE MA 等: "Fault Detection and Identification Based on the Neighborhood Standardized Local Outlier Factor Method", INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH, pages 2389 - 2402 * |
余立苹 等: "基于高维数据流的异常检测算法", 计算机工程, vol. 44, no. 1, pages 51 - 55 * |
吕志坚 等: "状态监测技术在裂解炉引风机故障诊断中的应用", 中国设备工程, pages 160 - 161 * |
尹新亮: "基于偏离特征的离群点挖掘方法研究", 中国优秀硕士学位论文全文数据库信息科技辑, pages 138 - 1119 * |
郭小芳;李锋;王卫东;: "基于k-近邻的多元时间序列局部异常检测", 江苏科技大学学报(自然科学版), vol. 26, no. 5, pages 505 - 508 * |
Also Published As
Publication number | Publication date |
---|---|
CN113190792B (en) | 2023-10-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Jiang et al. | Data-driven batch-end quality modeling and monitoring based on optimized sparse partial least squares | |
Kourti et al. | Experiences with industrial applications of projection methods for multivariate statistical process control | |
CN108062565B (en) | Double-principal element-dynamic core principal element analysis fault diagnosis method based on chemical engineering TE process | |
Said et al. | Machine learning technique for data-driven fault detection of nonlinear processes | |
CN101446831B (en) | Decentralized process monitoring method | |
CN108508865B (en) | A kind of fault detection method based on distributing OSC-PLS regression model | |
CN112036089A (en) | Coal mill fault early warning method based on DPC-MND and multivariate state estimation | |
Martin et al. | Manufacturing performance enhancement through multivariate statistical process control | |
CN108469805A (en) | A kind of distributing dynamic process monitoring method based on dynamic optimal selection | |
CN110942258B (en) | Performance-driven industrial process anomaly monitoring method | |
MacGregor et al. | Analysis, monitoring and fault diagnosis of industrial processes using multivariate statistical projection methods | |
CN108919755A (en) | A kind of distributed fault detection method based on muti-piece Nonlinear and crossing relational model | |
Wise et al. | Process chemometrics | |
CN112149054B (en) | Construction and application of orthogonal neighborhood preserving embedding model based on time sequence expansion | |
CN113190792B (en) | Ethylene cracking furnace running state monitoring method based on neighbor local abnormal factors | |
CN112288126B (en) | Sampling data abnormal change online monitoring and diagnosing method | |
CN110033175B (en) | Soft measurement method based on integrated multi-core partial least square regression model | |
Zhou et al. | Process monitoring based on classification tree and discriminant analysis | |
JP6953812B2 (en) | Anomaly detection device and anomaly detection system | |
CN113253682B (en) | Nonlinear chemical process fault detection method | |
CN113191614B (en) | Polypropylene product quality real-time monitoring method based on slow characteristic regression analysis | |
CN114936520A (en) | Steam turbine fault early warning analysis method based on improved MSET | |
CN114841621A (en) | Method for detecting abnormal operation state of synthetic ammonia process based on online neighbor embedding index | |
CN108388224A (en) | A kind of Desalting and Dewatering from Crude Oil process operation method for monitoring state based near infrared spectrum | |
CN113191616B (en) | Polypropylene product quality abnormity detection method based on double-layer correlation characteristic analysis |
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 |