CN109472070A - A kind of flexible measurement method of the ESN furnace operation variable based on PLS - Google Patents
A kind of flexible measurement method of the ESN furnace operation variable based on PLS Download PDFInfo
- Publication number
- CN109472070A CN109472070A CN201811262632.XA CN201811262632A CN109472070A CN 109472070 A CN109472070 A CN 109472070A CN 201811262632 A CN201811262632 A CN 201811262632A CN 109472070 A CN109472070 A CN 109472070A
- Authority
- CN
- China
- Prior art keywords
- state network
- echo state
- heating furnace
- network model
- operation variable
- 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
- 238000000691 measurement method Methods 0.000 title claims abstract description 11
- 238000010438 heat treatment Methods 0.000 claims abstract description 43
- 238000010606 normalization Methods 0.000 claims abstract description 9
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000000354 decomposition reaction Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 210000002569 neuron Anatomy 0.000 claims description 6
- 238000004519 manufacturing process Methods 0.000 abstract description 12
- 238000000034 method Methods 0.000 description 33
- 230000008569 process Effects 0.000 description 11
- 238000013528 artificial neural network Methods 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000007246 mechanism Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012549 training Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010304 firing Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- 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
-
- 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
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/30—Circuit design
- G06F30/32—Circuit design at the digital level
- G06F30/333—Design for testability [DFT], e.g. scan chain or built-in self-test [BIST]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/08—Thermal analysis or thermal optimisation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Computational Mathematics (AREA)
- Computer Hardware Design (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Databases & Information Systems (AREA)
- Algebra (AREA)
- Evolutionary Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Operations Research (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Feedback Control In General (AREA)
Abstract
The flexible measurement method of the invention discloses a kind of ESN furnace operation variable based on PLS, comprising: obtain heating furnace data, the heating furnace data include heating furnace pressure, inlet amount, temperature;The heating furnace data are normalized using normalization formula;The parameters value of echo state network is initialized;Echo state network model is optimized using partial least squares algorithm;Furnace operation variable is predicted according to the echo state network model after optimization.Technical solution provided by the invention solves the Problems of Multiple Synteny of traditional echo state network algorithm, optimize echo state network model accordingly, the precision and stability that can be improved echo state network model is widely used among the modeling of furnace operation variable.Therefore, technical solution provided by the invention can be realized that performance variable is effectively predicted among production process to heating furnace, to improve the thermal efficiency of heating furnace.
Description
Technical field
The present invention relates to heating furnace technical field more particularly to a kind of soft surveys of the ESN furnace operation variable based on PLS
Amount method.
Background technique
Modern industry process becomes increasingly complex, and causes the modeling of modern industry process also more and more difficult.Since operation passes through
The accumulation tested reduces many unnecessary losses, improves production efficiency, so that the modeling to performance variable in industrial production
It becomes more and more important.Veteran operator can rapidly and accurately adjusting parameter value so that entire industrial process begins
Good state is kept eventually.
Common three kinds of modeling methods are as follows: what modelling by mechanism, data-driven modeling, mechanism were combined with data-driven mixes
Build mould jointly.Modelling by mechanism is suitable for the object with mathematical models, and data-driven modeling is suitable for object model and is difficult to obtain
The case where a large amount of process datas is obtained and has, the method that mechanism is combined with data-driven is suitable for object accurate model and is difficult to
It obtains and there is the case where historical empirical data.
However, the demand to experienced operator is also increasing, huge manpower with the increasingly complexity of technique
Cost reduces the profit of enterprise.In order to reduce cost of labor, more accurate operation model is needed to carry out automatic adjusting parameter, so that
Model adjusted can bring optimal operation, so as to replace veteran operator, realize reduce manually at
This purpose.
Summary of the invention
To solve limitation and defect of the existing technology, the present invention provides a kind of ESN furnace operation change based on PLS
The flexible measurement method of amount, comprising:
Heating furnace data are obtained, the heating furnace data include heating furnace pressure, inlet amount, temperature;
The heating furnace data are normalized using normalization formula, the normalization formula are as follows:
Wherein,
YminAnd YmaxIt is the minimum value and maximum value of output mode vector Y respectively;
The parameters value of echo state network is initialized;
The state and output of echo state network are obtained, calculation formula is as follows:
X (t+1)=f (Win×u(t+1)+Wbackx(t)) (3)
Y (t+1)=fout×(Wout× (u (t+1), x (t+1))) (4)
Wherein, WinAnd WbackFor the parameter value of initialization, u (t+1) is current input, and x (t+1) is current time deposit
The state in pond, x (t) is the state of previous moment reserve pool, and as t=0, the initialization value of x (t) is 0;F (g) is in reserve pool
The activation function of portion's neuron, foutIndicate the activation primitive of output layer neuron;
Using offset minimum binary to the two-way principal component decomposition of progress is output and input, calculation formula is as follows:
Wherein, N is sample size, qi∈RkAnd oi∈RkFor score vector, Q and O are score matrix, pi∈RkAnd Si∈Rk
For load vector, P and S are load matrix, EUAnd ETFor resolution error matrix;
Echo state network model is optimized according to the result of the two-way principal component decomposition;
Furnace operation variable is predicted according to the echo state network model after optimization.
Optionally, the step that the result according to the two-way principal component decomposition optimizes echo state network model
Suddenly include:
Following parameter is initialized according to unit variance and zero-mean:
It is calculated according to returningIn riOn recurrence weight obtain input matrix Win, input weight is converted as unit
Matrix obtains score vector qi:
It is calculated according to returningIn riOn recurrence weight obtain load vector Si, according to the load vector SiIt obtains
Export score vector ri:
Work as riWhen convergence, according to EUIn qiOn recurrence weight obtain load vector pi:
Obtain the regression coefficient of internal links model:
Echo state network model is optimized according to the regression coefficient.
Optionally, further includes:
To ETReturn calculating and obtains ETRegressand value
According to the regressand valueBy θ1q1、θ2q2、L、θiqiIt is expressed asL、Linear combination;
The discreet value of output is obtained according to the linear combination are as follows:
Regression equation of the y about x is obtained by renormalization.
The present invention have it is following the utility model has the advantages that
The flexible measurement method of ESN furnace operation variable provided by the invention based on PLS, comprising: obtain heating furnace number
According to the heating furnace data include heating furnace pressure, inlet amount, temperature;Using normalization formula to the heating furnace data into
Row normalized;The parameters value of echo state network is initialized;Using partial least squares algorithm to echo shape
State network model optimizes;Furnace operation variable is predicted according to the echo state network model after optimization.This
The technical solution that invention provides solves the Problems of Multiple Synteny of traditional echo state network algorithm, optimizes echo shape accordingly
State network model can be improved the precision and stability of echo state network model, be widely used in furnace operation variable
Among modeling.Therefore, technical solution provided by the invention can be realized that performance variable has among production process to heating furnace
Effect prediction, to improve the thermal efficiency of heating furnace.
Detailed description of the invention
Fig. 1 is the structural schematic diagram for the oil plant heating under reduced pressure furnace that the embodiment of the present invention one provides.
Fig. 2 is the flow diagram for the echo state network model that the embodiment of the present invention one provides.
Fig. 3 is the extensive course prediction distribution of results figure that the embodiment of the present invention one provides.
Fig. 4 is the error schematic diagram for the echo state network model that the embodiment of the present invention one provides.
Specific embodiment
To make those skilled in the art more fully understand technical solution of the present invention, the present invention is mentioned with reference to the accompanying drawing
The flexible measurement method of the ESN furnace operation variable based on PLS supplied is described in detail.
Embodiment one
Fig. 1 is the structural schematic diagram for the oil plant heating under reduced pressure furnace that the embodiment of the present invention one provides.As shown in Figure 1, industrial
Tubular heater is a complicated difficult to obtain the process industry object of accurate model, but operator makes among industrial process
Optimum operation data are can to obtain.Therefore, the present embodiment is using method of the mechanism in conjunction with data-driven to producing
Performance variable in journey is modeled.Be fitted based on historical sample data mapping relations between input and output data to
The soft-sensing model for establishing furnace operation variable, which greatly simplifies the processes of modeling.Currently, there are many kinds of based on data into
The method, such as artificial neural network, fuzzy reasoning, data mining, Association Rule Analysis etc. of row modeling.These are based on data
In modeling method, it is used widely due to its powerful non-linear mapping capability based on the modeling method of artificial neural network.
Any continuous nonlinear function can the methods of employment artificial neural networks be fitted, the modeling of furnace operation variable
It is a complicated process industry problem, data are that have the time series data of Continuous Nonlinear, and artificial mind can be used
Method through network is modeled.
Artificial neural network has received good effect in the industrial flow modeled based on data, but with
Process industry becomes increasingly complex, and corresponding data variable and data volume are also more and more huger, traditional artificial neural network into
When row modeling, because the full connection performance of each layer network in its algorithm makes calculation amount extremely huge, the speed of modeling is caused to be got over
To be more difficult to meet the needs of people.Echo state network algorithm is a kind of learning algorithm of profound level, and echo state network is calculated
Method can be used for handling industrial process time series data.The nuclear structure of echo state network algorithm is the deposit generated at random
Pond can efficiently solve the modeling problem of dynamic time sequence data.The neuron of Random sparseness connection is taken as hidden layer to be formed
Reserve pool, echo state network realized using the reserve pool input higher-dimension and non-linear expression.Echo state network utilizes
Least-squares algorithm calculates the weight between reserve pool and output layer, to avoid falling into locally optimal solution.The present embodiment is directed to
Be a tubular heater, production process data has the feature of time series, therefore can use echo state network
To carry out soft sensor modeling to the performance variable of production process.
Although echo state network model or improved echo state network model are successfully applied in process industry
In the processing of data, but still have limitation.Echo state network model is limited in that least-squares algorithm cannot
Handle the multicollinearity between hidden layer node output.In order to overcome this disadvantage, the present embodiment proposes a kind of by echo
State network and offset minimum binary (offset minimum binary-echo state network) in conjunction with method to carry out hard measurement to performance variable
Modeling.Offset minimum binary-echo state network model replaces traditional echo state network algorithm using partial least squares algorithm
In least-squares algorithm echo state network is trained, it is that may be present in modeling to solve echo state network
Problems of Multiple Synteny.
The present embodiment obtains data and goes forward side by side line number Data preprocess.Specifically, heating furnace of the present embodiment to collection in worksite
Missing data present in data, abnormal data and noise data are handled, and I sample { (X is finally obtainedi, Yi) | i=1,
2 ..., I }, wherein Xi=[xi1, xi2..., xin]∈RnRepresent i-th of input sample, xinRepresent i-th of input sample Xi?
N element.Referring to table 1, n element respectively corresponds pressure, inlet amount, temperature etc. in heating furnace production, Yi∈ R represents output
Vector-air door aperture.Technical solution provided in this embodiment solves the multiple total of traditional echo state network algorithm
Linear problem optimizes echo state network model accordingly, can be improved the precision and stability of echo state network model, extensively
Among modeling applied to furnace operation variable.
1 input/output variable of table
The present embodiment models furnace operation variable-air door aperture.Specifically, the present embodiment uses inclined
Least square-echo state network model carries out hard measurement to furnace operation variable, and input variable parameter is input to training
Among model later, to obtain neural network forecast value, i.e. the predicted value of air door aperture.Therefore, provided in this embodiment
Technical solution can be realized that performance variable is effectively predicted among production process to heating furnace, to improve the thermal effect of heating furnace
Rate.
It is provided in this embodiment partially minimum for the multicollinearity feature of the hiding node layer of echo state network output
Two multiply-echo state network soft-sensing model, for predicting the variation of furnace operation variable.The present embodiment uses minimum two partially
Multiply method instead of the least-squares algorithm in echo state network, solves the multicollinearity of traditional echo state network algorithm
Problem optimizes echo state network model accordingly.The present embodiment may be implemented to heating furnace performance variable in process of production
Be effectively predicted, to improve the thermal efficiency of heating furnace.
Fig. 2 is the flow diagram for the echo state network model that the embodiment of the present invention one provides.As shown in Fig. 2, this reality
The flexible measurement method for applying the ESN furnace operation variable based on PLS of example offer includes: to obtain heating furnace data, the heating
Furnace data include heating furnace pressure, inlet amount, temperature;Place is normalized to the heating furnace data using normalization formula
Reason;The parameters value of echo state network is initialized;Using partial least squares algorithm to echo state network model
It optimizes;Furnace operation variable is predicted according to the echo state network model after optimization.
The present embodiment obtains training sample (X, Y) and it is normalized, and eliminates influence of the dimension to model.Its
In, shown in normalization process such as formula (1) and formula (2):
Wherein,
Ymin and Ymax is the minimum value and maximum value of output mode vector Y respectively.
The present embodiment initializes echo state network, randomly chooses the init state of echo state network, and one
As be 0, i.e. x (0)=0.Training set (u (t), t=1,2, K, K) passes through weight matrix WinIt is added to reserve pool.
Offset minimum binary-echo state network algorithm is specifically described in the present embodiment.Echo state network is obtained first
The state and output of network, calculation formula are as follows:
X (t+l)=f (Win×u(t+l)+Wbackx(t)) (3)
Y (t+l)=fout×(Wout× (u (t+l), x (t+l))) (4)
Wherein, WinAnd WbackFor the parameter value of initialization, u (t+1) is current input, and x (t+1) is current time deposit
The state in pond, x (t) is the state of previous moment reserve pool, and as t=0, the initialization value of x (t) is 0;F (g) is in reserve pool
The activation function of portion's neuron, foutIndicate the activation primitive of output layer neuron.
In order to avoid the multicollinearity in least square method, the present embodiment is using offset minimum binary to outputting and inputting
Two-way principal component decomposition is carried out, calculation formula is as follows:
Wherein, N is sample size, qi∈RkAnd oi∈RkFor score vector, Q and O are score matrix, pi∈RkAnd si∈Rk
For load vector, P and S are load matrix, EUAnd ETFor resolution error matrix.Technical solution provided in this embodiment solves biography
The Problems of Multiple Synteny of the echo state network algorithm of system optimizes echo state network model accordingly, can be improved echo shape
The precision and stability of state network model is widely used among the modeling of furnace operation variable.
Among the technical solution that embodiment provides, the quantity k of latent variable*It is determined by cross-validation method.The present embodiment
Using unit variance and zero-mean, following parameter is initialized:
It is calculated according to returningIn riOn recurrence weight obtain input matrix Win, the present embodiment by input weight turn
It turns to unit matrix and obtains score vector qi:
It is calculated according to returningIn riOn recurrence weight obtain load vector Si, the present embodiment according to it is described load to
Measure SiObtain output score vector ri:
Recurring formula 8 and formula 9, until riConvergence.Work as riWhen convergence, the present embodiment is according to EUIn qiOn recurrence weight
Obtain load vector pi:
So far, the present embodiment can obtain the regression coefficient of internal links model:
θi=ri Tqi (11)
To ETRecurrence calculating is carried out, the present embodiment can obtain ETRegressand valueAccording to the regressand valueBy θ1q1、
θ2q2、L、θiqiIt is expressed asL、Linear combination.In conclusion the present embodiment is according to the linear combination
Obtain the discreet value of output are as follows:
Finally, the present embodiment obtains regression equation of the y about x by renormalization, according to the echo state after optimization
Network model predicts furnace operation variable.
In order to verify the validity of the above method, the above method is applied to the soft survey of furnace operation variable by the present embodiment
Among amount modeling, for improving the precision of performance variable modeling and the thermal efficiency of heating furnace.Heating furnace be modern process industry it
In widely used equipment, the height of the thermal efficiency of operation directly affects the height of economic benefit, and the height of the thermal efficiency by
To the influence of factors, such as furnace design, status of equipment, firing optimization, technological operation, operating load etc..Wherein, good
Good operation is the important channel for improving the thermal efficiency.The aperture of air door often by as performance variable, relevant pressure,
The factors such as flow, temperature are as detection target and control object, furnace operation variable is modeled and be predicted.If operation
Variable remains at optimum state, then heating furnace can remain optimal efficiency levels.
Fig. 3 is the extensive course prediction distribution of results figure that the embodiment of the present invention one provides, and Fig. 4 is that the embodiment of the present invention one mentions
The error schematic diagram of the echo state network model of confession.As shown in Figure 3 and Figure 4, the present embodiment is by test data come to partially most
Small two multiply-echo state network method tested, further verify the validity for the technical solution that the present embodiment proposes.
2 effect of table compares
As shown in table 2, offset minimum binary-echo state network model average relative error (ARE) and average root-mean-square
Error (ARMSE) is respectively less than traditional echo state network model, shows offset minimum binary-echo state network provided in this embodiment
Network method is more accurate.Therefore, the present embodiment is by offset minimum binary-echo state network method and traditional echo state network
Method is compared, and shows that offset minimum binary-echo state network furnace operation variable soft-sensing model can be improved essence
Degree improves the thermal efficiency of heating furnace to improve the quality of product with robustness well.
The flexible measurement method of ESN furnace operation variable provided in this embodiment based on PLS, comprising: obtain heating furnace
Data, the heating furnace data include heating furnace pressure, inlet amount, temperature;Using normalization formula to the heating furnace data
It is normalized;The parameters value of echo state network is initialized;Using partial least squares algorithm to echo
State network model optimizes;Furnace operation variable is predicted according to the echo state network model after optimization.
Technical solution provided in this embodiment solves the Problems of Multiple Synteny of traditional echo state network algorithm, optimizes back accordingly
Sound state network model, can be improved the precision and stability of echo state network model, be widely used in furnace operation change
Among the modeling of amount.Therefore, technical solution provided in this embodiment can be realized operates change to heating furnace among production process
Amount is effectively predicted, to improve the thermal efficiency of heating furnace.
It is understood that the principle that embodiment of above is intended to be merely illustrative of the present and the exemplary implementation that uses
Mode, however the present invention is not limited thereto.For those skilled in the art, essence of the invention is not being departed from
In the case where mind and essence, various changes and modifications can be made therein, these variations and modifications are also considered as protection scope of the present invention.
Claims (3)
1. a kind of flexible measurement method of the ESN furnace operation variable based on PLS characterized by comprising
Heating furnace data are obtained, the heating furnace data include heating furnace pressure, inlet amount, temperature;
The heating furnace data are normalized using normalization formula, the normalization formula are as follows:
Wherein,
YminAnd YmaxIt is the minimum value and maximum value of output mode vector Y respectively;
The parameters value of echo state network is initialized;
The state and output of echo state network are obtained, calculation formula is as follows:
X (t+1)=f (Win×u(t+1)+Wbockx(t)) (3)
Y (t+1)=fout×(Wout× (u (t+1), x (t+1))) (4)
Wherein, WinAnd WbackFor the parameter value of initialization, u (t+1) is current input, and x (t+1) is current time reserve pool
State, x (t) is the state of previous moment reserve pool, and as t=0, the initialization value of x (t) is 0;F (g) is mind inside reserve pool
Activation function through member, foutIndicate the activation primitive of output layer neuron;
Using offset minimum binary to the two-way principal component decomposition of progress is output and input, calculation formula is as follows:
Wherein, N is sample size, qi∈RkAnd oi∈RkFor score vector, Q and O are score matrix, pi∈RkAnd si∈RkIt is negative
Vector is carried, P and S are load matrix, EUAnd ETFor resolution error matrix;
Echo state network model is optimized according to the result of the two-way principal component decomposition;
Furnace operation variable is predicted according to the echo state network model after optimization.
2. the flexible measurement method of the ESN furnace operation variable according to claim 1 based on PLS, which is characterized in that institute
Stating the step of optimizing according to the result of the two-way principal component decomposition to echo state network model includes:
Following parameter is initialized according to unit variance and zero-mean:
It is calculated according to returningIn riOn recurrence weight obtain input matrix Win, unit matrix is converted by input weight
Obtain score vector qi:
It is calculated according to returningIn riOn recurrence weight obtain load vector Si, according to the load vector SiIt is exported
Score vector ri:
Work as riWhen convergence, according to EUIn qiOn recurrence weight obtain load vector pi:
Obtain the regression coefficient of internal links model:
θi=ri Tqi (11)
Echo state network model is optimized according to the regression coefficient.
3. the flexible measurement method of the ESN furnace operation variable according to claim 2 based on PLS, which is characterized in that also
Include:
To ETReturn calculating and obtains ETRegressand value
According to the regressand valueBy θ1q1、θ2q2、L、θiqiIt is expressed as Linear combination;
The discreet value of output is obtained according to the linear combination are as follows:
Regression equation of the y about x is obtained by renormalization.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811262632.XA CN109472070A (en) | 2018-10-27 | 2018-10-27 | A kind of flexible measurement method of the ESN furnace operation variable based on PLS |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811262632.XA CN109472070A (en) | 2018-10-27 | 2018-10-27 | A kind of flexible measurement method of the ESN furnace operation variable based on PLS |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109472070A true CN109472070A (en) | 2019-03-15 |
Family
ID=65666037
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811262632.XA Pending CN109472070A (en) | 2018-10-27 | 2018-10-27 | A kind of flexible measurement method of the ESN furnace operation variable based on PLS |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109472070A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112364527A (en) * | 2020-12-02 | 2021-02-12 | 岳文琦 | Debutanizer soft measurement modeling method based on ALIESN online learning algorithm |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030191728A1 (en) * | 2002-03-27 | 2003-10-09 | Kulkarni Bhaskar Dattatray | Performance of artificial neural network models in the presence of instrumental noise and measurement errors |
CN105978732A (en) * | 2016-06-27 | 2016-09-28 | 华北理工大学 | Method and system for optimizing parameters of minimum complexity echo state network based on particle swarm |
CN107299170A (en) * | 2017-08-10 | 2017-10-27 | 东北大学 | A kind of blast-melted quality robust flexible measurement method |
CN108197380A (en) * | 2017-12-29 | 2018-06-22 | 南京林业大学 | Gauss based on offset minimum binary returns soft-measuring modeling method |
-
2018
- 2018-10-27 CN CN201811262632.XA patent/CN109472070A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030191728A1 (en) * | 2002-03-27 | 2003-10-09 | Kulkarni Bhaskar Dattatray | Performance of artificial neural network models in the presence of instrumental noise and measurement errors |
CN105978732A (en) * | 2016-06-27 | 2016-09-28 | 华北理工大学 | Method and system for optimizing parameters of minimum complexity echo state network based on particle swarm |
CN107299170A (en) * | 2017-08-10 | 2017-10-27 | 东北大学 | A kind of blast-melted quality robust flexible measurement method |
CN108197380A (en) * | 2017-12-29 | 2018-06-22 | 南京林业大学 | Gauss based on offset minimum binary returns soft-measuring modeling method |
Non-Patent Citations (2)
Title |
---|
李瑞祥: "基于改进回声状态神经网络的出水总磷软测量研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
贺彦林: "前馈神经网络结构设计研究及其复杂化工过程建模应用", 《中国博士学位论文全文数据库信息科技辑》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112364527A (en) * | 2020-12-02 | 2021-02-12 | 岳文琦 | Debutanizer soft measurement modeling method based on ALIESN online learning algorithm |
CN112364527B (en) * | 2020-12-02 | 2021-09-07 | 岳文琦 | Debutanizer soft measurement modeling method based on ALIESN online learning algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109002686B (en) | Multi-grade chemical process soft measurement modeling method capable of automatically generating samples | |
CN108932567B (en) | Convolutional neural network-based multi-energy-consumption index prediction method for cement sintering process | |
CN108764540B (en) | Water supply network pressure prediction method based on parallel LSTM series DNN | |
Lu et al. | Generalized radial basis function neural network based on an improved dynamic particle swarm optimization and AdaBoost algorithm | |
CN111474965B (en) | Fuzzy neural network-based method for predicting and controlling water level of series water delivery channel | |
CN108764568B (en) | Data prediction model tuning method and device based on LSTM network | |
CN107704962B (en) | Steam flow interval prediction method based on incomplete training data set | |
CN112818595B (en) | Correction method and system for digital twin model data of evaporation area of thermal power plant | |
CN110046408A (en) | Bounded-but-unknown uncertainty optimization method based on BP neural network differential and interval analysis | |
CN110097929A (en) | A kind of blast furnace molten iron silicon content on-line prediction method | |
CN112712189A (en) | Heat supply demand load prediction method | |
CN110018675A (en) | Nonlinear system modeling method based on LWDNN-ARX model | |
Budak et al. | Accuracy improvement of point data reduction with sampling-based methods by Fuzzy logic-based decision-making | |
CN114707712A (en) | Method for predicting requirement of generator set spare parts | |
CN114492191A (en) | Heat station equipment residual life evaluation method based on DBN-SVR | |
CN111210089A (en) | Stock price prediction method of gated cyclic unit neural network based on Kalman filtering | |
CN111754034A (en) | Time sequence prediction method based on chaos optimization neural network model | |
CN111311577A (en) | Intelligent water seepage detection method based on generation of confrontation network and reinforcement learning | |
CN116992759A (en) | Freeze-thawing concrete strength evaluation method and device based on interpretable neural network | |
Mao et al. | Coating thickness modeling and prediction for hot-dip galvanized steel strip based on GA-BP neural network | |
CN109472070A (en) | A kind of flexible measurement method of the ESN furnace operation variable based on PLS | |
CN110795891B (en) | Method for learning operating parameters of trailing suction hopper dredger | |
CN110909492A (en) | Sewage treatment process soft measurement method based on extreme gradient lifting algorithm | |
CN108446506B (en) | Uncertain system modeling method based on interval feedback neural network | |
CN111414927A (en) | Method for evaluating seawater quality |
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: 20190315 |