CN110134088A - A kind of adaptive quality forecasting procedure based on increment support vector regression - Google Patents
A kind of adaptive quality forecasting procedure based on increment support vector regression Download PDFInfo
- Publication number
- CN110134088A CN110134088A CN201910422208.5A CN201910422208A CN110134088A CN 110134088 A CN110134088 A CN 110134088A CN 201910422208 A CN201910422208 A CN 201910422208A CN 110134088 A CN110134088 A CN 110134088A
- Authority
- CN
- China
- Prior art keywords
- new
- model
- support vector
- vector regression
- sample
- 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 42
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 21
- 238000004519 manufacturing process Methods 0.000 claims abstract description 12
- 230000006870 function Effects 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 19
- 238000013507 mapping Methods 0.000 claims description 4
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 14
- 238000012986 modification Methods 0.000 abstract description 5
- 230000004048 modification Effects 0.000 abstract description 5
- 230000004069 differentiation Effects 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 5
- 239000001273 butane Substances 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- IJDNQMDRQITEOD-UHFFFAOYSA-N n-butane Chemical compound CCCC IJDNQMDRQITEOD-UHFFFAOYSA-N 0.000 description 4
- OFBQJSOFQDEBGM-UHFFFAOYSA-N n-pentane Natural products CCCCC OFBQJSOFQDEBGM-UHFFFAOYSA-N 0.000 description 4
- ATUOYWHBWRKTHZ-UHFFFAOYSA-N Propane Chemical compound CCC ATUOYWHBWRKTHZ-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000017105 transposition Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 239000003054 catalyst Substances 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 239000007789 gas Substances 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000001294 propane Substances 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 239000000126 substance Substances 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]
- G05B19/41885—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 modeling, simulation of the manufacturing system
-
- 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
- G05B2219/32—Operator till task planning
- G05B2219/32339—Object oriented modeling, design, analysis, implementation, simulation language
-
- 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)
- Manufacturing & Machinery (AREA)
- General Engineering & Computer Science (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of adaptive quality forecasting procedures based on increment support vector regression, this method is to adapt to multi-state production process complicated and changeable, on the basis of original support vector regression model, utilize the KKT condition of its own, realize that the differentiation to newly-increased sample learns, incremental learning is carried out to the sample with information, keeps model constant to without the sample of new information, model modification frequency can be reduced while guaranteeing model generalization ability;The present invention can not only significant surface to non-linear in actual industrial process, the problem of being constantly updated also directed to time-varying characteristics in the process to model, and improve model modification efficiency to a certain extent, to achieve the purpose that adaptive quality is forecast.
Description
Technical field
The invention belongs to industrial stokehold and hard measurement fields, are related to a kind of based on the adaptive of increment support vector regression
Answer quality prediction method.
Background technique
There is many can not or be difficult to variable measured directly in industrial process, these variables are usually close with product quality
Cut phase is closed, and is the parameter that must be monitored, soft-measuring technique is a kind of effective ways for solving the problems, such as this, at present in industry
It is used widely in the process.In recent years, the soft-measuring modeling method based on data-driven has obtained the common concern of people.Its
In, support vector regression is based on structural risk minimization, preferably solves small sample, non-linear and local minimum etc. and asks
Topic, has been widely used in multiple research fields.Current support vector regression learning algorithm is offline, batch-type mostly
, and in most of industrial process, due to the change of process environment, the aging of platform instrument equipment, raw material feeds variation
And the various factors such as catalyst activity degeneration, process physical chemical characteristic is among continuous variation, thus it runs work
Condition changes frequent occurrence, generally has the characteristics that dynamic time-varying.The model that off-line training obtains in this way cannot be guaranteed to be applicable in for a long time,
Therefore there is important theory significance and use value to the research of support vector regression on-line study method.
For correct trace-through state, need to carry out timely adaptive updates and correction to soft-sensing model.For work
The time-varying characteristics of industry process, therefore there is expert to propose sliding window support vector regression, local weighted support vector regression etc..If
Without distinction newly sample is increased to each and carry out incremental learning, is bound to cause model frequency similar to this mode of sliding window
Numerous update realizes that the differentiation to newly-increased sample learns, to a certain degree the present invention is based on the KKT condition of support vector regression model
On reduce model modification frequency, improve model modification efficiency.
Summary of the invention
For current industrial process time-varying characteristics, the invention proposes a kind of based on the adaptive of increment support vector regression
Quality prediction method, this method select the sample that those include enough new informations based on the KKT condition of support vector regression
Incremental learning is carried out, model modification frequency can be reduced while guaranteeing model generalization ability, to realize industrial process
Adaptive quality forecast.
Specific technical solution of the present invention is as follows: (1) collecting training dataset P (x, y) the ∈ R in history industrial processn×m,
Indicate that number of samples, m indicate variable number including input data x and quality variable y, n, R indicates set of real numbers;
(2) standardize the training dataset P being collected into step (1): turning to mean value is 0, the new data set that variance is 1And to standardized new data setIt establishes support vector regression model and obtains a part of supporting vector SVs, which can
It is solved by following formula;
Wherein, w is normal vector, and b is displacement item,For nonlinear mapping function, xiFor input data, wTFor normal vector
Transposition, yiFor quality variable value, C is penalty factor, ξ and ξ*It is relaxation factor, ε is tolerable error, and i is sample index.
Lagrange function is introduced, and formula (1) is changed to convex quadratic programming problem:
WhereinFor antithesis parameter, K (xi, xj)=φ (xi)Tφ(xi) it is known as kernel function, φ (xi) it is non-linear reflect
Function is penetrated, selects gaussian kernel function here:
K(xi, x) and=exp (- gamma | | xi-x||2) (8)
Wherein, gamma indicates kernel functional parameter;
Thus decision function can be constructed
Define sample coefficient deviationWith boundary function h (xi)=f (xi)-yi.Then according to coefficients deviation value and
Boundary function value is divided training dataset P for following three subsets based on the KKT condition of the optimization above problem:
Wherein, E is error supporting vector collection, and S is boundary supporting vector collection, and R is remaining sample set.
(3) by collected real time data standardization, and the support vector regression model pair obtained using step (2)
New samples carry out prediction of quality and export result
(4) as the true output y for obtaining quality variable ynewWhen, it calculatesh(xnew) it is that prediction misses
Difference, xnewFor new input data;
(5) pass through h (xnew) further judge whether newly-increased data set meets KKT condition and examine | h (xnew) | it is whether big
In ε, then show to be unsatisfactory for KKT condition if it is greater than ε, indicates that "current" model can not describe newly-increased data set well, need
Incremental learning is carried out to newly-increased data set, newly-increased data set is added in SVs more new model;Show to meet KKT if being less than ε
Condition then continues with old model prediction;
(6) it when there is the acquisition of new training dataset, repeats step (3) to step (5), realizes the adaptive matter of industrial process
Amount forecast.
Compared with prior art, the invention has the advantages that: be directed to process time-varying characteristics, the present invention proposes one kind
Adaptive modeling method based on increment support vector regression.By selectively adding or deleting to training sample, mould is realized
Type online updating, effectively to track the variation of process operating condition.Compared to traditional sliding window method, require to build again every time
Mould improves the update efficiency of model to a certain extent.
Detailed description of the invention
Fig. 1 is online adaptive modeling flow chart;
Fig. 2 is debutanizing tower flow chart;
Fig. 3 is the prediction output result figure of sliding window support vector regression;
Fig. 4 is the prediction output result figure of increment support vector regression;
Specific embodiment
With reference to embodiment to the present invention is based on the adaptive industrial Quality Forecastings of increment support vector regression
Method is described in further detail.
A kind of adaptive quality forecasting procedure based on increment support vector regression, wherein the increment supporting vector
After the process of the adaptive quality forecasting procedure of recurrence is as shown in Figure 1, obtain new training dataset in line, new data set is detected
Whether the KKT condition of existing model is met, if not satisfied, then carrying out incremental learning training dataset information.Specific step is as follows:
(1) training dataset P (x, y) the ∈ R in history industrial process is collectedn×m, including input data x and quality
Variable y, n indicate that number of samples, m indicate variable number, and R indicates set of real numbers;
(2) standardize the training dataset P being collected into step (1): turning to mean value is 0, the new data set that variance is 1And to standardized new data setIt establishes support vector regression model and obtains a part of supporting vector SVs, which can
It is solved by following formula;
Wherein, w is normal vector, and b is displacement item,For nonlinear mapping function, xiFor input data, wTFor normal vector
Transposition, yiFor quality variable value, C is penalty factor, ξ and ξ*It is relaxation factor, ε is tolerable error, and i is sample index.
Lagrange function is introduced, and formula (1) is changed to convex quadratic programming problem:
WhereinFor antithesis parameter, K (xi, xj)=φ (xi)Tφ(xi) it is known as kernel function, φ (xi) it is non-linear reflect
Function is penetrated, selects gaussian kernel function here:
K(xi, x) and=exp (- gamma | | xi-x||2) (13)
Wherein, gamma indicates kernel functional parameter;
Thus decision function can be constructed
Define sample coefficient deviationWith boundary function h (xi)=f (xi)-yi.Then according to coefficients deviation value and
Boundary function value is divided training dataset P for following three subsets based on the KKT condition of the optimization above problem:
Wherein, E is error supporting vector collection, and S is boundary supporting vector collection, and R is remaining sample set.
(3) by collected real time data standardization, and the support vector regression model pair obtained using step (2)
New samples carry out prediction of quality and export result
(4) as the true output y for obtaining quality variable ynewWhen, it calculatesh(xnew) it is that prediction misses
Difference, xnewFor new input data;
(5) pass through h (xnew) further judge whether newly-increased data set meets KKT condition and examine | h (xnew) | it is whether big
In ε, then show to be unsatisfactory for KKT condition if it is greater than ε, indicates that "current" model can not describe newly-increased data set well, need
Incremental learning is carried out to newly-increased data set, newly-increased data set is added in SVs more new model;Show to meet KKT if being less than ε
Condition then continues with old model prediction.The purpose for the arrangement is that in order to allow model to possess the ability of adaptive updates, it is new when encountering
Status information or process when significant change occurs, model can timely update the ginseng of model according to new information
Number;Generally speaking, the present invention is directed to process time-varying characteristics, proposes a kind of adaptive modeling side based on increment support vector regression
Method.By selectively being added or deleted to training sample, implementation model online updating, effectively to track the change of process operating condition
Change.It compared to traditional sliding window method, requires to model again every time, improves the update efficiency of model to a certain extent.
(6) it when there is the acquisition of new training dataset, repeats step (3) to step (5), realizes the adaptive matter of industrial process
Amount forecast.
In addition, root-mean-square error (RMSE) carries out quantitative assessment to estimated performance, expression formula is as follows:
Wherein, yiIt is quality variable value,It is the prediction output of model, Nts indicates the number of on-line testing sample.
Embodiment
Illustrate the performance of increment support vector regression model below in conjunction with a specific debutanizing tower example.Debutanization
Tower is a common normal industry process platform for being used for soft sensor modeling proof of algorithm.Debutanizing tower is refining process
In one of device, flow chart is as shown in Fig. 2, the purpose of the device is the mistake in order to remove propane and butane in naphtha gases
Journey debutanizing tower, the butane content of tower bottom are a highly important key indexes, in order to improve the control quality of debutanizing tower
It needs to establish soft-sensing model for tower bottom butane content.
Table 1 gives for selected 7 auxiliary variables of Key Quality variable butane content, and has carried out centainly to it
Variable description, respectively tower top temperature, tower top pressure, return flow, next stage flow, the temperature of sensitive plate, column bottom temperature
And tower bottom pressure.
Table 1: input variable explanation
For the process, continuous constant duration acquires 2394 samples.1000 initial samples constitute original
Training dataset, remaining 1394 samples are as test data set.
Initial support vector regression model is established to initial training data set first, and 1394 data are learned online
It practises, wherein there is 1148 data to take part in incremental learning.For the variation characteristic of tracking mode, verify of the invention adaptive soft
Measurement method compared sliding window support vector regression method, as shown in Figure 3,4 respectively.Wherein, sliding window is dimensioned to
1000, window step length is set as 1;For the method for the present invention, step sizes are also 1.
Although can be seen that the effect of the two is not much different on the whole by Fig. 3 and Fig. 4, can preferably track pre-
Survey trend, but tracking effect will be more near the biggish sample point of some variation amplitudes for increment support vector regression of the invention
It is good, and the performance of sliding window support vector regression performance is poor.Table 2 gives the prediction effect of two methods and calculates the time
Compare, as can be seen from the table, method of the invention has preferable performance on RMSE, returns compared to sliding window supporting vector
RMSE is returned to reduce 0.0357.But from renewal time, compared to sliding window method, renewal time of the invention is
The former 1/6th or so.Generally speaking, method of the invention in prediction error regardless of still on renewal time, suffering from
Certain advantage.
The prediction effect of 2 two methods of table and calculating time
Method | Sliding window support vector regression | Increment support vector regression |
RMSE | 0.1383 | 0.1026 |
Renewal time (s) | 37.5 | 6.5 |
Claims (1)
1. a kind of adaptive quality forecasting procedure based on increment support vector regression, which is characterized in that the adaptive quality
Forecasting procedure the following steps are included:
(1) training dataset P (x, y) the ∈ R in history industrial process is collectedn×m, including input data x and quality variable
Y, n indicate that number of samples, m indicate variable number, and R indicates set of real numbers;
(2) standardize the training dataset P being collected into step (1): turning to mean value is 0, the new data set that variance is 1And
To standardized new data setIt establishes support vector regression model and obtains a part of supporting vector SVs, which can pass through
Following formula solves;
Wherein, w is normal vector, and b is displacement item,For nonlinear mapping function, xiFor input data, wTFor turning for normal vector
It sets, yiFor quality variable value, C is penalty factor, ξ and ξ*It is relaxation factor, ε is tolerable error, and i is sample index.
Lagrange function is introduced, and formula (1) is changed to convex quadratic programming problem:
Wherein αi,For antithesis parameter, K (xi, xj)=φ (xi)Tφ(xi) it is known as kernel function, φ (xi) it is Nonlinear Mapping letter
Number, selects gaussian kernel function here:
K(xi, x) and=exp (- gamma | | xi-x||2) (3)
Wherein, gamma indicates kernel functional parameter;
Thus decision function can be constructed
Define sample coefficient deviationWith boundary function h (xi)=f (xi)-yi.Then according to coefficients deviation value and boundary
Functional value is divided training dataset P for following three subsets based on the KKT condition of the optimization above problem:
Wherein, E is error supporting vector collection, and S is boundary supporting vector collection, and R is remaining sample set.
(3) by collected real time data standardization, and the support vector regression model obtained using step (2) is to new sample
This progress prediction of quality simultaneously exports result
(4) as the true output y for obtaining quality variable ynewWhen, it calculatesh(xnew) it is prediction error,
xnewFor new input data;
(5) pass through h (xnew) further judge whether newly-increased data set meets KKT condition and examine | h (xnew) | whether it is greater than ε,
Then show to be unsatisfactory for KKT condition if it is greater than ε, indicates that "current" model can not describe newly-increased data set well, need to new
Increase data set and carry out incremental learning, newly-increased data set is added in SVs more new model;Show to meet KKT condition if being less than ε
Then continue with old model prediction;
(6) it when there is the acquisition of new training dataset, repeats step (3) to step (5), realizes that the adaptive quality of industrial process is pre-
Report.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910422208.5A CN110134088A (en) | 2019-05-21 | 2019-05-21 | A kind of adaptive quality forecasting procedure based on increment support vector regression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910422208.5A CN110134088A (en) | 2019-05-21 | 2019-05-21 | A kind of adaptive quality forecasting procedure based on increment support vector regression |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110134088A true CN110134088A (en) | 2019-08-16 |
Family
ID=67571689
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910422208.5A Pending CN110134088A (en) | 2019-05-21 | 2019-05-21 | A kind of adaptive quality forecasting procedure based on increment support vector regression |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110134088A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110866218A (en) * | 2019-11-14 | 2020-03-06 | 北京理工大学 | Hysteresis system compensation method and system |
CN111523710A (en) * | 2020-04-10 | 2020-08-11 | 三峡大学 | Power equipment temperature prediction method based on PSO-LSSVM online learning |
CN113688908A (en) * | 2021-08-25 | 2021-11-23 | 江南大学 | Bluetooth signal indoor propagation model correction method based on online epsilon type twin support vector regression machine |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090193936A1 (en) * | 2008-02-01 | 2009-08-06 | Honeywell International, Inc. | Methods and apparatus for an oxygen furnace quality control system |
CN101520856A (en) * | 2009-04-10 | 2009-09-02 | 东南大学 | Adaptive modeling method for support vector regression based on KKT condition and nearest neighbor method |
CN102693451A (en) * | 2012-06-14 | 2012-09-26 | 东北电力大学 | Method for predicting ammonia process flue gas desulphurization efficiency based on multiple parameters |
CN105139093A (en) * | 2015-09-07 | 2015-12-09 | 河海大学 | Method for forecasting flood based on Boosting algorithm and support vector machine |
CN106127341A (en) * | 2016-06-24 | 2016-11-16 | 北京市地铁运营有限公司地铁运营技术研发中心 | A kind of urban track traffic newly-built circuit energy consumption Calculating model |
CN107180278A (en) * | 2017-05-27 | 2017-09-19 | 重庆大学 | A kind of real-time passenger flow forecasting of track traffic |
-
2019
- 2019-05-21 CN CN201910422208.5A patent/CN110134088A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20090193936A1 (en) * | 2008-02-01 | 2009-08-06 | Honeywell International, Inc. | Methods and apparatus for an oxygen furnace quality control system |
CN101520856A (en) * | 2009-04-10 | 2009-09-02 | 东南大学 | Adaptive modeling method for support vector regression based on KKT condition and nearest neighbor method |
CN102693451A (en) * | 2012-06-14 | 2012-09-26 | 东北电力大学 | Method for predicting ammonia process flue gas desulphurization efficiency based on multiple parameters |
CN105139093A (en) * | 2015-09-07 | 2015-12-09 | 河海大学 | Method for forecasting flood based on Boosting algorithm and support vector machine |
CN106127341A (en) * | 2016-06-24 | 2016-11-16 | 北京市地铁运营有限公司地铁运营技术研发中心 | A kind of urban track traffic newly-built circuit energy consumption Calculating model |
CN107180278A (en) * | 2017-05-27 | 2017-09-19 | 重庆大学 | A kind of real-time passenger flow forecasting of track traffic |
Non-Patent Citations (2)
Title |
---|
HUI WANG: "Online SVM regression algorithm-based adaptive inverse control", 《NEUROCOMPUTING》 * |
王平: "一种基于增量式SVR学习的在线自适应建模方法", 《化工学报》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110866218A (en) * | 2019-11-14 | 2020-03-06 | 北京理工大学 | Hysteresis system compensation method and system |
CN111523710A (en) * | 2020-04-10 | 2020-08-11 | 三峡大学 | Power equipment temperature prediction method based on PSO-LSSVM online learning |
CN113688908A (en) * | 2021-08-25 | 2021-11-23 | 江南大学 | Bluetooth signal indoor propagation model correction method based on online epsilon type twin support vector regression machine |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103927412B (en) | Instant learning debutanizing tower soft-measuring modeling method based on gauss hybrid models | |
US20200118015A1 (en) | Fuzzy curve analysis based soft sensor modeling method using time difference Gaussian process regression | |
CN109840362B (en) | Multi-objective optimization-based integrated just-in-time learning industrial process soft measurement modeling method | |
CN110134088A (en) | A kind of adaptive quality forecasting procedure based on increment support vector regression | |
Wu et al. | A just-in-time fine-tuning framework for deep learning of SAE in adaptive data-driven modeling of time-varying industrial processes | |
CN106845796B (en) | One kind is hydrocracked flow product quality on-line prediction method | |
CN104765346A (en) | Full-process modeling method for oil refining process | |
CN107403196B (en) | Method for predicting concentration of butane at bottom of debutanizer by instant learning modeling based on spectral clustering analysis | |
CN109389238B (en) | Ridge regression-based short-term load prediction method and device | |
CN113420500B (en) | Intelligent atmospheric and vacuum system | |
CN108334943A (en) | The semi-supervised soft-measuring modeling method of industrial process based on Active Learning neural network model | |
CN107464017A (en) | Based on the adaptive soft-sensor Forecasting Methodology with time difference Bayesian network | |
CN108595892A (en) | Soft-measuring modeling method based on time difference model | |
CN113012766A (en) | Self-adaptive soft measurement modeling method based on online selective integration | |
CN110263488A (en) | A kind of industrial Mooney Viscosity of Rubber Mix flexible measurement method based on integrated instant learning | |
Li et al. | A two-layer fuzzy synthetic strategy for operational performance assessment of an industrial hydrocracking process | |
CN110046377A (en) | A kind of selective ensemble instant learning soft-measuring modeling method based on isomery similarity | |
CN110083065B (en) | Self-adaptive soft measurement method based on flow type variational Bayesian supervised factor analysis | |
CN102621953B (en) | Automatic online quality monitoring and prediction model updating method for rubber hardness | |
Yuan et al. | A comparative study of adaptive soft sensors for quality prediction in an industrial refining hydrocracking process | |
CN112949836A (en) | Method for carrying out regression prediction on-line migration learning on time-varying distribution data | |
CN101201331A (en) | Soft measuring method for on-line determining petroleum naphtha quality index on top of primary tower | |
Zhu | System identification for process control: recent experience and outlook | |
Macias-Hernandez et al. | Soft sensor for predicting crude oil distillation side streams using evolving takagi-sugeno fuzzy models | |
CN109960146A (en) | The method for improving soft measuring instrument model prediction accuracy |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190816 |