CN108241789A - The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR - Google Patents
The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR Download PDFInfo
- Publication number
- CN108241789A CN108241789A CN201810030797.8A CN201810030797A CN108241789A CN 108241789 A CN108241789 A CN 108241789A CN 201810030797 A CN201810030797 A CN 201810030797A CN 108241789 A CN108241789 A CN 108241789A
- Authority
- CN
- China
- Prior art keywords
- awls
- svr
- sample
- parameter
- 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.)
- Withdrawn
Links
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Regulation And Control Of Combustion (AREA)
Abstract
The invention discloses a kind of carbon containing measuring methods of boiler slag and equipment, method based on adaptive weighted AWLS SVR to be:The auxiliary variable and leading variable in certain time in the past are obtained, per hour one group of data, form sample data;Sample data is standardized, pivot analysis is carried out to the sample data after standardization, extracts p principal component of sample data;After p principal component is normalized, as AWLS SVR model training samples, AWLS SVR models are established;Utilize the boiler slag carbon content at the AWLS SVR model prediction T moment of foundation.The AWLS SVR models that the present invention is established by adaptive weighted combination Least square support vector regression method, and it combines pivot analysis and realizes the carbon containing measurement of boiler slag, precision of prediction is high, foundation is provided for boiler combustion process operation optimization, adjustment operating parameter in time, it is energy-saving so as to achieve the purpose that.
Description
Technical field:
The present invention relates to data processing field more particularly to the boiler slag phosphorus content based on adaptive weighted AWLS-SVR
Measuring method and apparatus field.
Background technology:
In the prior art, for the carbon containing measurement of boiler slag, generally use off-line measurement method, using offline inspection stove
During slag phosphorus content method, a slag sample analysis result wants several hours or even longer, makes boiler attendance control lag, and influence
The operating parameter of boiler slag carbon content is more and with time variation, and then can not carry out timely, accurate firing optimization.
Invention content:
The purpose of the present invention is to provide a kind of carbon containing measurings of the boiler slag based on adaptive weighted AWLS-SVR
Method and equipment integrate pivot analysis, multiple output system modeling method of least squares support method and based on time slip-window
Sample set on-line tuning method, realizes the measurement of boiler slag carbon content, and the prediction model based on AWLS-SVR has learning ability
By force, the features such as Generalization Capability is good, boiler slag carbon content that can be preferably in on-line prediction boiler combustion process, precision of prediction is high,
Foundation is provided for boiler combustion process operation optimization, adjusts operating parameter in time, it is energy-saving so as to achieve the purpose that.
The present invention is implemented by following technical solution:
In a first aspect, the present invention provides a kind of carbon containing measuring of boiler slag based on adaptive weighted AWLS-SVR
Method, including:
Step S1 obtains auxiliary variable and leading variable in certain time in the past, per hour one group of data, forms sample
Data, wherein, the auxiliary variable be influenced in boiler combustion process boiler slag phosphorus content simultaneously can on-line measurement parameter, institute
Leading variable is stated as boiler slag phosphorus content;
Step S2 is standardized the sample data, and pivot is carried out to the sample data after standardization
P principal component of the sample data is extracted in analysis;
Step S3 after the p principal component is normalized, as AWLS-SVR model training samples, is established
AWLS-SVR models;
Step S4 utilizes the boiler slag carbon content at the AWLS-SVR model prediction T moment of the foundation.
The carbon containing measuring method of boiler slag provided by the invention based on adaptive weighted AWLS-SVR, technical solution
For:Including obtaining auxiliary variable and leading variable in certain time in the past, one group of data, forms sample data per hour,
In, the auxiliary variable be influenced in boiler combustion process boiler slag phosphorus content simultaneously can on-line measurement parameter, it is described leading
Variable is boiler slag phosphorus content;The sample data is standardized, to the sample data after standardization into
Row pivot analysis extracts p principal component of the sample data;After the p principal component is normalized, as
AWLS-SVR model training samples, establish AWLS-SVR models;Utilize the stove at the AWLS-SVR model prediction T moment of the foundation
Slag phosphorus content.
The carbon containing measuring method of boiler slag provided by the invention based on adaptive weighted AWLS-SVR, by adaptive
The AWLS-SVR models that weighted combination Least square support vector regression method is established, and combine pivot analysis and realize boiler slag
Carbon containing measurement, precision of prediction is high, provides foundation for boiler combustion process operation optimization, adjusts operating parameter in time, so as to reach
To energy-saving purpose.
Further, the method further includes:
Step S5 carries out renormalization processing to the boiler slag carbon content, obtains the boiler slag carbon content prediction at the T moment
Value y* (T), and the combustion process data at the T moment are acquired, it is added in the sample data;
Step S6 calculates the relative error M of the predicted value, and the sample is adjusted using adaptive training sample method of adjustment
Notebook data.
Further, in the step S1, the auxiliary variable includes boiler load parameter, excess air coefficient, primary
It is air quantity parameter, secondary air flow parameter, 6 secondary air register aperture parameters, bellows burner hearth differential pressure parameter, 5 coal-supplying amount parameters, primary
Wind speed parameter, coal pulverizer operation mode parameter, coal-fired industry composition parameters, fineness of pulverized coal parameter and fire box temperature parameter.
Further, the step S2, specially:
If the sample data matrix being standardized is X ∈ Rm×n, wherein m is the length of input sample, and n is defeated
Enter the variable number of sample;
The covariance matrix for defining X is Cov (X)=Rm×n, Orthogonal Decomposition is carried out to Cov (X), is obtained
In formula, V=diag (λ1,λ2,…,λn) it is that n characteristic value of covariance matrix arranges (λ according to descending1,
λ2,…,λn) form diagonal matrix, Um×n=[λ1,λ2,…,λn] it is its corresponding unitization eigenmatrix;
Pivot number P determined by accumulative variance contribution ratio, i.e.,:
Take η>85% obtains preceding P principal component (t1,t2,…,tp) p principal component as the sample data.
Further, in the step S3, the p principal component is normalized, specially:
If the sampled data of a certain auxiliary variable is, wherein X(m,i)To represent the auxiliary
Parameter in variable;
Data after standardization are
Further, in the step S3, the AWLS-SVR models are supported by sectionally weighting strategy and least square
Vector regression method, specially:
Using sectionally weighting strategy, different weighted values, a moment are used to the different time segment data of the training sample
The training sample weights proportion of acquisition is smaller than the training sample weights proportion that the b moment acquires, and when a moment is earlier than the b
It carves;
Wherein, the mathematical method of the sectionally weighting strategy is expressed as follows:
Weights are set as vi:
In formula, weight coefficients of the d for first training sample, 0<d<1;φ is given parameters, φ>0;M is training sample
Collect length;
Based on Least square support vector regression method, according to structural risk minimization principle, regression problem is expressed as
Formula constrained optimization problem:
(i=1,2 ..., m)
It solves above-mentioned optimization problem and is converted into solution linear equation:
In formula,Ω ties up square formation for m,α=[α1,…,αm], y
=[y1,…,ym]T
AWLS-SVR models are obtained according to the calculation formula of boiler slag phosphorus content, wherein, the meter of boiler slag phosphorus content
Calculating formula is:
Radial basis kernel function K (x, x in formulai)=exp (- | | x-xi||2/δ2), wherein δ is kernel functional parameter, and x is waits to ask
Solve variable, xiIt is supporting vector, α and b are to solve for the solution that linear equation obtains, and m is sample number.
Further, it in the step S3, during establishing the AWLS-SVR models, further includes:Utilize K-fold
Cross-validation method and grid data service optimize the AWLS-SVR models.
Further, it is described to optimize the AWLS-SVR models, tool using K-fold cross-validation methods and grid data service
Body is:
A. network coordinate is established, is enabled,, step-length 1;
B. the training sample data are divided into K subset;
C. for each group (C, δ) in grid, using any one subset as test set, remaining K-1 subset is as instruction
Practice collection, test set is predicted after training pattern, the mean square error of statistical test result:
Wherein,For predicted value;YiFor sample actual value;
D. the test set is changed to another subset, then remaining K-1 subset is taken to count square again as training set
Error until after K subset is all once predicted, takes the average value of K group mean square errorsPrediction as the group (C, δ)
Error;
E. parameter combination (C, δ) is replaced, repeats step b to d, training pattern under each parameter combination sequentially in calculating gridAnd compare one by one, the parameter combination of mean error minimum is the optimal parameter combination in grid section.
Further, the step S6, specially:The relative error M of the predicted value is calculated, utilizes adaptive training sample
This method of adjustment adjusts the sample data.
The relative error M of the predicted value is calculated, if M≤δ, deletes the data group at earliest moment in the sample data
To update the sample data;
If M>Nearest deleted data group is added to the sample data, increases sample data length by δ, wherein,
δ is error setting value.
Second aspect, the present embodiment provides a kind of carbon containing measurements of the boiler slag based on adaptive weighted AWLS-SVR to set
It is standby, including:At least one processor, at least one processor and the computer program instructions being stored in the memory,
Realize when the computer program instructions are performed by the processor as described in first aspect based on adaptive weighted
The carbon containing measuring method of boiler slag of AWLS-SVR.
Advantages of the present invention:
Comprehensive pivot analysis, multiple output system modeling method of least squares support method and the sample based on time slip-window
This collection on-line tuning method, realizes the measurement of boiler slag carbon content, the prediction model based on AWLS-SVR have learning ability it is strong,
The features such as Generalization Capability is good, boiler slag carbon content that can be preferably in on-line prediction boiler combustion process, precision of prediction is high, is pot
Stove combustion process operation optimization provides foundation, adjusts operating parameter in time, energy-saving so as to achieve the purpose that.
Description of the drawings:
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention, for those of ordinary skill in the art, without creative efforts, can be with
Other attached drawings are obtained according to these attached drawings.
The carbon containing measurement of a kind of boiler slag based on adaptive weighted AWLS-SVR that Fig. 1 is provided by the embodiment of the present invention
The flow chart of amount method;
The carbon containing measurement of a kind of boiler slag based on adaptive weighted AWLS-SVR that Fig. 2 is provided by the embodiment of the present invention
Measure the structure diagram of equipment.
Specific embodiment:
Below in conjunction with the attached drawing in the embodiment of the present invention, the technical solution in the embodiment of the present invention is carried out clear, complete
Site preparation describes, it is clear that described embodiment is only part of the embodiment of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art are obtained every other without creative efforts
Embodiment shall fall within the protection scope of the present invention.
Embodiment
In a first aspect, Fig. 1 shows a kind of pot based on adaptive weighted AWLS-SVR that the embodiment of the present invention is provided
The flow chart of stove boiler slag carbon content measuring method;As shown in Figure 1, embodiment offer is a kind of based on adaptive weighted AWLS-SVR's
The carbon containing measuring method of boiler slag, including:
Step S1 obtains auxiliary variable and leading variable in certain time in the past, per hour one group of data, forms sample
Data;Auxiliary variable and leading variable in the present embodiment in the past 72 hours of selection;Wherein, the auxiliary variable is in the past 72
Influenced in boiler combustion process in hour boiler slag phosphorus content simultaneously can on-line measurement parameter, the leading variable is in the past 72
Boiler slag phosphorus content in hour.
Step S2 is standardized the sample data, and pivot is carried out to the sample data after standardization
Analysis, extracts p principal component of the sample data, is denoted as (t1,t2,…,tp);
Step S3 after the p principal component is normalized, as AWLS-SVR model training samples, is established
AWLS-SVR models;
Step S4 utilizes the boiler slag carbon content at the AWLS-SVR model prediction T moment of the foundation.
Anemometer tower provided by the invention, technical solution are:Including obtaining auxiliary variable and master in certain time in the past
Variable is led, per hour one group of data, form sample data;The sample data is standardized, to standardization
Sample data afterwards carries out pivot analysis, extracts p principal component of the sample data;Normalizing is carried out to the p principal component
After change processing, as AWLS-SVR model training samples, AWLS-SVR models are established;Utilize the AWLS-SVR models of the foundation
Predict the boiler slag carbon content at T moment.
The carbon containing measuring method of boiler slag provided by the invention based on adaptive weighted AWLS-SVR, by adaptive
The AWLS-SVR models that weighted combination Least square support vector regression method is established, and combine pivot analysis and realize boiler slag
Carbon containing measurement, precision of prediction is high, provides foundation for boiler combustion process operation optimization, adjusts operating parameter in time, so as to reach
To energy-saving purpose.
Preferably, the method further includes:
Step S5 carries out renormalization processing to the boiler slag carbon content, obtains the boiler slag carbon content prediction at the T moment
Value y* (T), and the combustion process data at the T moment are acquired, it is added in the sample data, sample data is protected with sample set S
It deposits, is denoted as S=S ∪ { x(T,1),x(T,2),…,x(T,n),y(T)};
Step S6 calculates the relative error M of the predicted value, and the sample is adjusted using adaptive training sample method of adjustment
Notebook data.
On-line tuning is carried out to sample data based on time slip-window, it can be achieved that optimization to sample data, improves clinker
The measurement accuracy of phosphorus content.
Preferably, in the step S1, the auxiliary variable includes boiler load parameter, excess air coefficient, First air
Measure parameter, secondary air flow parameter, 6 secondary air register aperture parameters (AA~EF), bellows burner hearth differential pressure parameter, 5 coal-supplying amount ginsengs
Number, primary air velocity parameter, coal pulverizer operation mode parameter, coal-fired industry composition parameters, fineness of pulverized coal parameter and fire box temperature ginseng
Number.
Wherein, coal-fired industry composition parameters include generation amount parameter, volatile matter parameter, ash content parameter and water parameters.
Therefore, selecting above 24 parameters in the present embodiment, the auxiliary variable of selection is more, makes sample data more as auxiliary variable
It is abundant, and then make the boiler slag carbon content of prediction more accurate.
Preferably, the step S2 carries out pivot analysis PCA, specially:
If the sample data matrix being standardized is X ∈ Rm×n, wherein m is the length of input sample, and n is defeated
Enter the variable number of sample;
The covariance matrix for defining X is Cov (X)=Rm×n, Orthogonal Decomposition is carried out to Cov (X), is obtained
In formula, V=diag (λ1,λ2,…,λn) it is that n characteristic value of covariance matrix arranges (λ according to descending1,
λ2,…,λn) form diagonal matrix, Um×n=[λ1,λ2,…,λn] it is its corresponding unitization eigenmatrix;
Pivot number P determined by accumulative variance contribution ratio, i.e.,:
Take η>85% obtains preceding P principal component (t1,t2,…,tp) p principal component as the sample data.
Preferably, in the step S3, the p principal component is normalized, specially:
If the sampled data of a certain auxiliary variable is, wherein X(m,i)To represent the auxiliary
Parameter in variable, m are not more than 24;
Data after standardization are
Preferably, in the step S3, the AWLS-SVR models by sectionally weighting strategy and least square support to
The Return Law is measured, specially:
Using sectionally weighting strategy, different weighted values, a moment are used to the different time segment data of the training sample
The training sample weights proportion of acquisition is smaller than the training sample weights proportion that the b moment acquires, and when a moment is earlier than the b
It carves;
Wherein, the mathematical method of the sectionally weighting strategy is expressed as follows:
Weights are set as vi:
In formula, weight coefficients of the d for first training sample, 0<d<1;φ is given parameters, φ>0;M is training sample
Collect length;
Based on Least square support vector regression method, according to structural risk minimization principle, regression problem is expressed as
Formula constrained optimization problem:
(i=1,2 ..., m)
It solves above-mentioned optimization problem and is converted into solution linear equation:
In formula,Ω ties up square formation for m,α=[α1,…,αm], y
=[y1,…,ym]T
AWLS-SVR models are obtained according to the calculation formula of boiler slag phosphorus content, wherein, the meter of boiler slag phosphorus content
Calculating formula is:
Radial basis kernel function K (x, x in formulai)=exp (- | | x-xi||2/δ2), wherein δ is kernel functional parameter, and x is waits to ask
Solve variable, xiIt is supporting vector, α and b are to solve for the solution that linear equation obtains, and m is sample number.
The on-line prediction of boiler slag carbon content is carried out for method through this embodiment, as new measured data constantly adds
Entering into sample set S, old historical data gradually weakens the effect of estimated performance, and the present embodiment uses sectionally weighting strategy,
Different weighted values is used to the different time segment data of sample data, the sample weights proportion of more early acquisition is smaller, and newer
Sample weights proportion it is bigger.
Preferably, it in the step S3, during establishing the AWLS-SVR models, further includes:It is handed over using K-fold
It pitches proof method and grid data service optimizes the AWLS-SVR models.
It is highly preferred that described optimize the AWLS-SVR models, tool using K-fold cross-validation methods and grid data service
Body is:
A. network coordinate is established, enables C=[2-3,27], δ=[2-3,27], step-length 1;
B. the training sample data are divided into K subset;The general values of K are 4~10;To ensure that sample size is long-range
In test capacity.
C. for each group (C, δ) in grid, using any one subset as test set, remaining K-1 subset is as instruction
Practice collection, test set is predicted after training pattern, the mean square error of statistical test result:
Wherein,For predicted value;YiFor sample actual value;
D. the test set is changed to another subset, then remaining K-1 subset is taken to count square again as training set
Error until after K subset is all once predicted, takes the average value of K group mean square errorsPrediction as the group (C, δ)
Error;
E. parameter combination (C, δ) is replaced, repeats step b to d, training pattern under each parameter combination sequentially in calculating gridAnd compare one by one, the parameter combination of mean error minimum is the optimal parameter combination in grid section.
Cross validation and grid search are combined, ginseng is improved with the preferred target of the minimum parameter of parameter mean error
The preferred efficiency of number and accuracy, while evaded influence of the sampling randomness of training sample to model performance.
It is highly preferred that the step S6, specially:The relative error M of the predicted value is calculated, utilizes adaptive training sample
This method of adjustment adjusts the sample data.
The relative error M of the predicted value is calculated, the measured value at T moment with predicted value is compared and calculates its opposite mistake
Difference, i.e.,:
If M≤δ, then the data group at earliest moment in the sample data is deleted to update the sample data;
If M>Nearest deleted data group is added to the sample data, increases sample data length by δ, wherein,
δ is error setting value.
By calculating relative error, screening and optimizing is carried out to sample data, further improves the carbon containing measurement of boiler slag
Accuracy.
Second aspect, with reference to the boiler slag based on adaptive weighted AWLS-SVR of Fig. 2 embodiment of the present invention described
Carbon containing measuring method can be adaptive weighted by being based on
The boiler slag phosphorus content measuring apparatus of AWLS-SVR is realized.Fig. 2 shows bases provided in an embodiment of the present invention
In the hardware architecture diagram of the boiler slag phosphorus content measuring apparatus of adaptive weighted AWLS-SVR.
Boiler slag phosphorus content measuring apparatus based on adaptive weighted AWLS-SVR can include processor 401 and deposit
Contain the memory 402 of computer program instructions.
Specifically, above-mentioned processor 401 can include central processing unit (CPU) or specific integrated circuit
It (Application Specific Integrated Circuit, ASIC) or may be configured to implement implementation of the present invention
One or more integrated circuits of example.
Memory 402 can include the mass storage for data or instruction.For example it is unrestricted, memory
402 may include hard disk drive (Hard Disk Drive, HDD), floppy disk, flash memory, CD, magneto-optic disk, tape or logical
With the combination of universal serial bus (Universal Serial Bus, USB) driver or two or more the above.It is closing
In the case of suitable, memory 402 may include can be removed or the medium of non-removable (or fixed).In a suitable case, it stores
Device 402 can be inside or outside data processing equipment.In a particular embodiment, memory 402 is nonvolatile solid state storage
Device.In a particular embodiment, memory 402 includes read-only memory (ROM).In a suitable case, which can be mask
The ROM of programming, programming ROM (PROM), erasable PROM (EPROM), electric erasable PROM (EEPROM), electrically-alterable ROM
(EAROM) or the combination of flash memory or two or more the above.
Processor 401 is by reading and performing the computer program instructions stored in memory 402, to realize above-mentioned implementation
Any one carbon containing measuring method of boiler slag based on adaptive weighted AWLS-SVR in example.
In one example, the boiler slag phosphorus content measuring apparatus based on adaptive weighted AWLS-SVR may also include logical
Believe interface 403 and bus 410.Wherein, as shown in Fig. 2, processor 401, memory 402, communication interface 403 pass through bus 410
It connects and completes mutual communication.
Communication interface 403 is mainly used for realizing in the embodiment of the present invention between each module, device, unit and/or equipment
Communication.
Bus 410 includes hardware, software or both, by the carbon containing measurement of boiler slag based on adaptive weighted AWLS-SVR
The component of amount equipment is coupled to each other together.For example it is unrestricted, bus may include accelerated graphics port (AGP) or other
Graphics bus, enhancing Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), super transmission (HT) interconnection, Industry Standard Architecture
(ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro- channel architecture (MCA) bus, periphery
Component interconnection (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, video electronic
The combination of Standard Association part (VLB) bus or other suitable buses or two or more the above.Suitable
In the case of, bus 410 may include one or more buses.Although specific bus has been described and illustrated in the embodiment of the present invention,
The present invention considers any suitable bus or interconnection.
In addition, with reference to the carbon containing measuring of the boiler slag based on adaptive weighted AWLS-SVR in above-described embodiment
Method, the embodiment of the present invention can provide a kind of computer readable storage medium to realize.It is stored on the computer readable storage medium
There are computer program instructions;The computer program instructions realize in above-described embodiment when being executed by processor any one be based on
The carbon containing measuring method of boiler slag of adaptive weighted AWLS-SVR.
It should be clear that the invention is not limited in specific configuration described above and shown in figure and processing.
For brevity, it is omitted here the detailed description to known method.In the above-described embodiments, several tools have been described and illustrated
The step of body, is as example.But procedure of the invention is not limited to described and illustrated specific steps, this field
Technical staff can be variously modified, modification and addition or suitable between changing the step after the spirit for understanding the present invention
Sequence.
Above structural frames functional block shown in figure can be implemented as hardware, software, firmware or combination thereof.When
When realizing in hardware, electronic circuit, application-specific integrated circuit (ASIC), appropriate firmware, plug-in unit, function may, for example, be
Card etc..When being realized with software mode, element of the invention is used to perform the program or code segment of required task.Journey
Sequence either code segment can be stored in machine readable media or the data-signal by being carried in carrier wave in transmission medium or
Person's communication links are sent." machine readable media " can include being capable of any medium of storage or transmission information.It is machine readable
The example of medium include electronic circuit, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disk, CD-ROM,
CD, hard disk, fiber medium, radio frequency (RF) link, etc..Code segment can be via the calculating of internet, Intranet etc.
Machine network is downloaded.
It should also be noted that, the exemplary embodiment referred in the present invention, is retouched based on a series of step or device
State certain methods or system.But the present invention is not limited to the sequence of above-mentioned steps, that is to say, that can be according in embodiment
The sequence referred to performs step, may also be distinct from that the sequence in embodiment or several steps are performed simultaneously.
The foregoing is merely a prefered embodiment of the invention, is not intended to limit the invention, it is all the present invention spirit and
Within principle, any modification, equivalent replacement, improvement and so on should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of carbon containing measuring method of boiler slag based on adaptive weighted AWLS-SVR, which is characterized in that including:
Step S1 obtains auxiliary variable and leading variable in certain time in the past, per hour one group of data, forms sample number
According to, wherein, the auxiliary variable be influenced in boiler combustion process boiler slag phosphorus content simultaneously can on-line measurement parameter, it is described
Leading variable is boiler slag phosphorus content;
Step S2 is standardized the sample data, and pivot analysis is carried out to the sample data after standardization,
Extract p principal component of the sample data;
Step S3 after the p principal component is normalized, as AWLS-SVR model training samples, establishes AWLS-
SVR models;
Step S4 utilizes the boiler slag carbon content at the AWLS-SVR model prediction T moment of the foundation.
2. according to the method described in claim 1, it is characterized in that:
The method further includes:
Step S5 carries out renormalization processing to the boiler slag carbon content, obtains the boiler slag carbon content predicted value at the T moment
y* (T), and the combustion process data at the T moment are acquired, it is added in the sample data;
Step S6 calculates the relative error M of the predicted value, and the sample number is adjusted using adaptive training sample method of adjustment
According to.
3. according to the method described in claim 1, it is characterized in that:
In the step S1, the auxiliary variable includes boiler load parameter, excess air coefficient, primary air flow parameter, secondary
Air quantity parameter, 6 secondary air register aperture parameters, bellows burner hearth differential pressure parameter, 5 coal-supplying amount parameters, primary air velocity parameter, coal-grinding
Machine operation mode parameter, coal-fired industry composition parameters, fineness of pulverized coal parameter and fire box temperature parameter.
4. according to the method described in claim 1, it is characterized in that:
The step S2, specially:
If the sample data matrix being standardized is X ∈ Rm×n, wherein m is the length of input sample, and n is input sample
This variable number;
The covariance matrix for defining X is Cov (X)=Rm×n, Orthogonal Decomposition is carried out to Cov (X), is obtained
In formula, V=diag (λ1,λ2,…,λn) it is that n characteristic value of covariance matrix arranges (λ according to descending1,λ2,…,λn)
The diagonal matrix of composition, Um×n=[λ1,λ2,…,λn] it is its corresponding unitization eigenmatrix;
Pivot number P determined by accumulative variance contribution ratio, i.e.,:
Take η>85% obtains preceding P principal component (t1,t2,…,tp) p principal component as the sample data.
5. according to the method described in claim 1, it is characterized in that:
In the step S3, the p principal component is normalized, specially:
If the sampled data of a certain auxiliary variable isWherein X(m,i)To represent the auxiliary variable
In parameter;
Data after standardization are
6. according to the method described in claim 1, it is characterized in that:
In the step S3, the AWLS-SVR models pass through sectionally weighting strategy and Least square support vector regression method, tool
Body is:
Using sectionally weighting strategy, different weighted values is used to the different time segment data of the training sample, a moment acquires
Training sample weights proportion it is smaller than the training sample weights proportion that the b moment acquires, and a moment is earlier than the b moment;
Wherein, the mathematical method of the sectionally weighting strategy is expressed as follows:
Weights are set as vi:
In formula, weight coefficients of the d for first training sample, 0<d<1;φ is given parameters, φ>0;M is long for training sample set
Degree;
Based on Least square support vector regression method, according to structural risk minimization principle, regression problem is expressed as equation about
Beam optimization problem:
It solves above-mentioned optimization problem and is converted into solution linear equation:
In formula,Ω ties up square formation for m,α=[α1,…,αm], y=
[y1,…,ym]T
AWLS-SVR models are obtained according to the calculation formula of boiler slag phosphorus content, wherein, the calculating of boiler slag phosphorus content is public
Formula is:
Radial basis kernel function K (x, x in formulai)=exp (- | | x-xi||2/δ2), wherein δ is kernel functional parameter, and x is change to be solved
Amount, xiIt is supporting vector, α and b are to solve for the solution that linear equation obtains, and m is sample number.
7. according to the method described in claim 1, it is characterized in that:
In the step S3, during establishing the AWLS-SVR models, further include:Utilize K-fold cross-validation methods and net
Lattice search optimizes the AWLS-SVR models.
8. according to the method described in claim 7, it is characterized in that:
It is described to optimize the AWLS-SVR models using K-fold cross-validation methods and grid data service, specially:
A. network coordinate is established, enables C=[2-3,27], δ=[2-3,27], step-length 1;
B. the training sample data are divided into K subset;
C. for each group (C, δ) in grid, using any one subset as test set, remaining K-1 subset as training set,
Test set is predicted after training pattern, the mean square error of statistical test result:
Wherein,For predicted value;YiFor sample actual value;
D. the test set is changed to another subset, then remaining K-1 subset is taken to count mean square error again as training set
Difference until after K subset is all once predicted, takes the average value of K group mean square errorsPrediction as the group (C, δ) misses
Difference;
E. parameter combination (C, δ) is replaced, repeats step b to d, training pattern under each parameter combination sequentially in calculating grid
And compare one by one, the parameter combination of mean error minimum is the optimal parameter combination in grid section.
9. according to the method described in claim 1, it is characterized in that:
The step S6, specially:The relative error M of the predicted value is calculated, utilizes adaptive training sample method of adjustment tune
The whole sample data.
The relative error M of the predicted value is calculated, if M≤δ, deletes the data group at earliest moment in the sample data with more
The new sample data;
If M>Nearest deleted data group is added to the sample data, increases sample data length by δ, wherein, δ is
Error setting value.
10. a kind of boiler slag phosphorus content measuring apparatus based on adaptive weighted AWLS-SVR, which is characterized in that including:Extremely
A few processor, at least one processor and the computer program instructions being stored in the memory, when the calculating
Machine program instruction realizes method as claimed in any one of claims 1-9 wherein when being performed by the processor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810030797.8A CN108241789A (en) | 2018-01-12 | 2018-01-12 | The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810030797.8A CN108241789A (en) | 2018-01-12 | 2018-01-12 | The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108241789A true CN108241789A (en) | 2018-07-03 |
Family
ID=62699561
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810030797.8A Withdrawn CN108241789A (en) | 2018-01-12 | 2018-01-12 | The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108241789A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109541143A (en) * | 2018-11-28 | 2019-03-29 | 西安建筑科技大学 | A kind of prediction technique that the constituent element clinker actual constituent transitivity containing volatilization changes over time |
CN113588597A (en) * | 2021-06-25 | 2021-11-02 | 河钢股份有限公司 | Method for improving analysis precision of furnace slag |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120232772A1 (en) * | 2011-03-07 | 2012-09-13 | Toyota Motor Engineering & Manufacturing North America, Inc. | Adaptive air charge estimation based on support vector regression |
CN102778538A (en) * | 2012-07-06 | 2012-11-14 | 广东电网公司电力科学研究院 | Soft measuring method based on improved SVM (Support Vector Machine) for measuring boiler unburned carbon content in fly ash |
CN102831274A (en) * | 2012-08-30 | 2012-12-19 | 中南林业科技大学 | Correction method of distribution of quenching stress field of aluminum alloy thick plate |
CN105045951A (en) * | 2015-05-27 | 2015-11-11 | 华南理工大学 | Soft-measurement method for volatile fatty acid in effluent of anaerobic wastewater treatment system |
CN107133460A (en) * | 2017-04-26 | 2017-09-05 | 中国能源建设集团广东省电力设计研究院有限公司 | A kind of online dynamic prediction method of boiler flyash carbon content |
-
2018
- 2018-01-12 CN CN201810030797.8A patent/CN108241789A/en not_active Withdrawn
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120232772A1 (en) * | 2011-03-07 | 2012-09-13 | Toyota Motor Engineering & Manufacturing North America, Inc. | Adaptive air charge estimation based on support vector regression |
CN102778538A (en) * | 2012-07-06 | 2012-11-14 | 广东电网公司电力科学研究院 | Soft measuring method based on improved SVM (Support Vector Machine) for measuring boiler unburned carbon content in fly ash |
CN102831274A (en) * | 2012-08-30 | 2012-12-19 | 中南林业科技大学 | Correction method of distribution of quenching stress field of aluminum alloy thick plate |
CN105045951A (en) * | 2015-05-27 | 2015-11-11 | 华南理工大学 | Soft-measurement method for volatile fatty acid in effluent of anaerobic wastewater treatment system |
CN107133460A (en) * | 2017-04-26 | 2017-09-05 | 中国能源建设集团广东省电力设计研究院有限公司 | A kind of online dynamic prediction method of boiler flyash carbon content |
Non-Patent Citations (2)
Title |
---|
俞金寿: "《工业过程先进控制技术》", 31 August 2008, 华东理工大学出版社 * |
阳春华等: "氧化铝蒸发浓度的自适应加权 LSSVR 预测", 《控制工程》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109541143A (en) * | 2018-11-28 | 2019-03-29 | 西安建筑科技大学 | A kind of prediction technique that the constituent element clinker actual constituent transitivity containing volatilization changes over time |
CN109541143B (en) * | 2018-11-28 | 2021-07-06 | 西安建筑科技大学 | Prediction method for actual components and physical property of slag containing volatile components along with time change |
CN113588597A (en) * | 2021-06-25 | 2021-11-02 | 河钢股份有限公司 | Method for improving analysis precision of furnace slag |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108446711A (en) | A kind of Software Defects Predict Methods based on transfer learning | |
CN103631681A (en) | Method for online restoring abnormal data of wind power plant | |
CN110018670A (en) | A kind of industrial process unusual service condition prediction technique excavated based on dynamic association rules | |
CN111680870B (en) | Comprehensive evaluation method for quality of target motion trail | |
CN104748807A (en) | Online power station main steam flow calculation method based on flow correction | |
CN110400231B (en) | Failure rate estimation method for electric energy metering equipment based on weighted nonlinear Bayes | |
CN109388774A (en) | A kind of thermal power plant NOx prediction model characteristics of variables extracting method based on method of comparison | |
Muehleisen et al. | Bayesian calibration-what, why and how | |
CN108241789A (en) | The carbon containing measuring method of boiler slag and equipment based on adaptive weighted AWLS-SVR | |
CN106649919A (en) | Method and system for predicting carbon content of fly ash in coal-fired power plant boiler | |
CN111352408A (en) | Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor | |
CN109407649A (en) | A kind of fault type matching process based on fault signature variables choice | |
CN109409425A (en) | A kind of fault type recognition method based on neighbour's constituent analysis | |
CN116805439A (en) | Drought prediction method and system based on artificial intelligence and atmospheric circulation mechanism | |
CN113392497B (en) | Method for measuring field performance degradation of photovoltaic module according to geographical region | |
CN114528764A (en) | Soft measurement modeling method and device based on integral optimization and instant learning | |
CN101446828A (en) | Nonlinear process quality prediction method | |
CN111090932B (en) | Method for marking wind farm in transportation suitable for medium complex terrain | |
CN117612651A (en) | Method for predicting manganese content of converter endpoint | |
CN112541296A (en) | SO2 prediction method based on PSO-LSSVM | |
Ruliandi | Geothermal power plant system performance prediction using artificial neural networks | |
CN110927478B (en) | Method and system for determining state of transformer equipment of power system | |
Middleton et al. | Comparison of RADM and OSCAR precipitation chemistry data | |
CN107870611A (en) | Workshop analogue means and workshop analogy method | |
CN107300535A (en) | The method of near-infrared quick detection organic fertilizer active constituent content |
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 | ||
WW01 | Invention patent application withdrawn after publication | ||
WW01 | Invention patent application withdrawn after publication |
Application publication date: 20180703 |