CN117912581A - Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system - Google Patents
Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system Download PDFInfo
- Publication number
- CN117912581A CN117912581A CN202410073787.8A CN202410073787A CN117912581A CN 117912581 A CN117912581 A CN 117912581A CN 202410073787 A CN202410073787 A CN 202410073787A CN 117912581 A CN117912581 A CN 117912581A
- Authority
- CN
- China
- Prior art keywords
- data
- input variable
- prediction model
- intelligent
- sintering
- 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 92
- 238000005245 sintering Methods 0.000 title claims abstract description 65
- 238000012549 training Methods 0.000 claims abstract description 60
- 238000004519 manufacturing process Methods 0.000 claims abstract description 24
- 238000002485 combustion reaction Methods 0.000 claims abstract description 14
- 230000000630 rising effect Effects 0.000 claims abstract description 14
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 claims abstract description 7
- 229910052799 carbon Inorganic materials 0.000 claims abstract description 7
- 239000000463 material Substances 0.000 claims abstract description 7
- 238000009826 distribution Methods 0.000 claims description 35
- 238000004364 calculation method Methods 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000012545 processing Methods 0.000 claims description 5
- 238000005265 energy consumption Methods 0.000 abstract description 10
- 229910000831 Steel Inorganic materials 0.000 abstract description 5
- 239000010959 steel Substances 0.000 abstract description 5
- 238000004134 energy conservation Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 4
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 description 2
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000010304 firing Methods 0.000 description 2
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910002092 carbon dioxide Inorganic materials 0.000 description 1
- 239000001569 carbon dioxide Substances 0.000 description 1
- 229910002091 carbon monoxide Inorganic materials 0.000 description 1
- 238000013434 data augmentation Methods 0.000 description 1
- 238000009851 ferrous metallurgy Methods 0.000 description 1
- 239000005431 greenhouse gas Substances 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/10—Analysis or design of chemical reactions, syntheses or processes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/70—Machine learning, data mining or chemometrics
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Crystallography & Structural Chemistry (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Analytical Chemistry (AREA)
- Chemical Kinetics & Catalysis (AREA)
- General Engineering & Computer Science (AREA)
Abstract
The invention provides a sintering process CO/CO 2 intelligent prediction method and system considering unbalanced output data, which relate to the field of energy conservation and consumption reduction in the steel sintering process, and comprise the steps of adopting a minimum absolute shrinkage and selection operator method to determine input variables affecting a CO/CO 2 intelligent prediction model: the thickness of the material layer, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the temperature of the rising point, the position of the rising point, the sintering end point temperature and the sintering end point position; expanding the input variable by using a data enhancement method to obtain a training data set; building a CO/CO 2 intelligent prediction model based on the mixed kernel correlation vector machine, and training the CO/CO 2 intelligent prediction model by using training data; and predicting actual production data in real time by using the trained CO/CO 2 intelligent prediction model to obtain a CO/CO 2 predicted value, and reflecting the combustion degree of carbon-containing energy sources in the sintering production process. The invention has the following effects: provides an effective method for realizing energy conservation, consumption reduction and green manufacturing in the sintering process.
Description
Technical Field
The invention relates to the field of energy conservation and consumption reduction in a steel sintering process, in particular to an accurate prediction method and system for energy consumption indexes in the sintering process, and particularly relates to an intelligent prediction method and system for CO/CO 2 in the sintering process, which take unbalanced output data into consideration.
Background
Has remarkable scientific and practical significance for the research of energy conservation, consumption reduction and emission reduction in the iron and steel industrial process.
Sintering is the second most energy intensive process in ferrous metallurgy. In order to analyze the energy consumption of the sintering process in real time, a model is usually required to be built to predict some energy consumption indexes, and the ratio of CO to CO 2 (CO/CO 2) is adopted as the sintering energy consumption index. However, existing CO/CO 2 must be manually detected at the sintering site with a hand-held sensor, which is both expensive and dangerous. CO/CO 2 represents the ratio of carbon monoxide to carbon dioxide, directly reflects the combustion degree of carbon-containing energy sources in the sintering production process, has important practical value as an index for measuring sintering energy consumption, and has great significance for reducing the steel energy consumption in China and reducing the greenhouse gas emission in the steel industry. Thus, it is very necessary and significant to achieve accurate prediction of CO/CO 2.
Disclosure of Invention
In order to solve the problems of high risk and high cost of acquiring CO/CO 2 data in the existing sintering process and simultaneously realize energy saving, consumption reduction and green manufacturing in the sintering process, the invention provides an intelligent prediction method and system for CO/CO 2 in the sintering process, which take unbalanced output data into consideration, and mainly comprise the following steps:
S1: determining key process parameters affecting CO/CO 2 data by adopting a minimum absolute shrinkage and selection operator method, and taking the key process parameters as input variables of a CO/CO 2 intelligent prediction model, wherein the input variables are as follows: the thickness of the material layer, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the temperature of the rising point, the position of the rising point, the sintering end point temperature and the sintering end point position;
S2: expanding the input variable by using a data enhancement method to obtain a training data set;
S3: building a CO/CO 2 intelligent prediction model based on the mixed kernel correlation vector machine, and training the CO/CO 2 intelligent prediction model by using training data;
S4: and predicting actual production data in real time by using the trained CO/CO 2 intelligent prediction model to obtain predicted data of CO/CO 2, wherein the predicted data is used for reflecting the combustion degree of carbon-containing energy sources in the sintering production process.
Further, the process of obtaining the training data set in step S2 is as follows:
(2-1) if the input variable has corresponding CO/CO 2 data, directly adding the input variable and the corresponding CO/CO 2 data thereof into the training data set;
(2-2) if the input variable x does not have the corresponding CO/CO 2 data y m, obtaining k normal CO/CO 2 data related to the input variable x by applying a k-nearest neighbor algorithm based on the mahalanobis distance;
(2-3) then determining y m, which is ultimately related to the input variable x, by the following calculation formula:
where y i is k normal CO/CO 2 data related to the input variable x;
(2-4) finally adding the input variable x and its final associated CO/CO 2 data y m to the training dataset.
Further, in step S3, the process of establishing the intelligent prediction model of CO/CO 2 is as follows:
(3-1) let the training data set be expressed as N represents the number of training data, x (t) represents the t training data, y (t) represents CO/CO 2 data corresponding to the t training data, and the intelligent CO/CO 2 prediction model obtained based on the mixed kernel correlation vector machine is as follows:
Wherein ω (t) represents a weight, K h represents a mixing kernel, and x represents any input variable;
(3-2) constructing a hybrid kernel K h using a polynomial kernel and a gaussian kernel:
Kh=τKG+(1-τ)Kpoly
Where K G represents a Gaussian kernel function, K poly represents a polynomial kernel function, and τ represents a scale factor.
Further, in step S3, the calculation process of the intelligent prediction model of CO/CO 2 is as follows:
(3-3) let g= [ g (1), g (2), …, g (T), …, g (N) ] T, where g (T) = y (x (T) |ω) +ε (T), ω= [ ω (0), ω (1),..;
(3-4) if g (t) is an independent distribution, the likelihood function is:
wherein, The basis functions are represented by the functions,
(3-5) Assuming that the parameter ω (t) obeys a gaussian distribution with zero mean and a variance α -1 (t), there are:
Wherein α= [ α (0), α (1), α (t), α (N) ] T, α (t) represents a super-parameter of the ω prior distribution, N represents the number of training data, let a=diag (α);
the posterior distribution of (3-6) ω is derived from the likelihood distribution and the prior distribution:
wherein, Σ 2 represents the variance of ε (t);
(3-7) performing integral processing on omega to obtain a marginal distribution determined by parameters alpha and sigma 2:
wherein, I represents an identity matrix.
An intelligent prediction system for a sintering process CO/CO 2 that takes into account unbalanced output data, comprising:
The input variable determining module is used for determining key process parameters affecting the CO/CO 2 data according to a minimum absolute shrinkage and selection operator method, the key process parameters are used as input variables of the CO/CO 2 intelligent prediction model, and the input variables are as follows: the thickness of the material layer, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the temperature of the rising point, the position of the rising point, the sintering end point temperature and the sintering end point position;
The training data set acquisition module is used for expanding the input variables by utilizing a data enhancement method to obtain a training data set;
The model building module is used for building a CO/CO 2 intelligent prediction model based on a mixed kernel correlation vector machine modeling method and training the CO/CO 2 intelligent prediction model by using training data;
the real-time prediction module is used for predicting actual production data in real time by utilizing the trained CO/CO 2 intelligent prediction model to obtain predicted data of CO/CO 2, and the predicted data is used for reflecting the combustion degree of carbon-containing energy sources in the sintering production process.
Further, the process of obtaining the training data set is as follows:
(2-1) if the input variable has corresponding CO/CO 2 data, directly adding the input variable and the corresponding CO/CO 2 data thereof into the training data set;
(2-2) if the input variable x does not have the corresponding CO/CO 2 data y m, obtaining k normal CO/CO 2 data related to the input variable x by applying a k-nearest neighbor algorithm based on the mahalanobis distance;
(2-3) then determining y m, which is ultimately related to the input variable x, by the following calculation formula:
where y i is k normal CO/CO 2 data related to the input variable x;
(2-4) finally adding the input variable x and its final associated CO/CO 2 data y m to the training dataset.
Further, the establishment process of the intelligent CO/CO 2 prediction model comprises the following steps:
(3-1) let the training data set be expressed as N represents the number of training data, x (t) represents the t training data, y (t) represents CO/CO 2 data corresponding to the t training data, and the intelligent CO/CO 2 prediction model obtained based on the mixed kernel correlation vector machine is as follows:
Wherein ω (t) represents a weight, K h represents a mixing kernel, and x represents any input variable;
(3-2) constructing a hybrid kernel K h using a polynomial kernel and a gaussian kernel:
Kh=τKG+(1-τ)Kpoly
Where K G represents a Gaussian kernel function, K poly represents a polynomial kernel function, and τ represents a scale factor.
Further, the calculation process of the intelligent CO/CO 2 prediction model is as follows:
(3-3) assuming that the observed target value is g= [ g (1), g (2), …, g (T), …, g (N) ] T, wherein g (T) = y (x (T) |ω) +ε (T), ω= [ ω (0), ω (1),..;
(3-4) if g (t) is an independent distribution, the likelihood function is:
wherein, The basis functions are represented by the functions,
(3-5) Assuming that the parameter ω (t) obeys a gaussian distribution with zero mean and a variance α -1 (t), there are:
Wherein α= [ α (0), α (1), α (t), α (N) ] T, α (t) represents a super-parameter of the ω prior distribution, N represents the number of training data, let a=diag (α);
the posterior distribution of (3-6) ω is derived from the likelihood distribution and the prior distribution:
wherein, Σ 2 represents the variance of ε (t);
(3-7) performing integral processing on omega to obtain a marginal distribution determined by parameters alpha and sigma 2:
wherein, I represents an identity matrix.
The technical scheme provided by the invention has the beneficial effects that: according to the method, the influence of different production parameters on the CO/CO 2 is considered, the minimum absolute shrinkage and a selection operator method are utilized to determine the input variables affecting the intelligent CO/CO 2 prediction model, and the data enhancement method is utilized to expand the input variables so as to improve the prediction performance of the model. Building a CO/CO 2 intelligent prediction model based on the mixed kernel correlation vector machine, and training the CO/CO 2 intelligent prediction model by using training data; and predicting actual production data in real time by using the trained CO/CO 2 intelligent prediction model to obtain predicted data of CO/CO 2, wherein the predicted data is used for reflecting the combustion degree of carbon-containing energy sources in the sintering production process. The method solves the problems of dangerous and high cost of acquiring CO/CO 2 data in the existing sintering process, and simultaneously realizes energy saving, consumption reduction and green manufacturing in the sintering process.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a flow chart of a method for intelligent prediction of CO/CO 2 during sintering, taking into account unbalanced output data, in accordance with an embodiment of the present invention;
FIG. 2 is a schematic block diagram of modeling of a CO/CO 2 intelligent predictive model in an embodiment of the invention;
FIG. 3 is a graph showing the comparison of predicted and actual values of CO/CO 2 in an example of the present invention;
FIG. 4 is a graph of prediction error for CO/CO 2 in an embodiment of the present invention.
Detailed Description
For a clearer understanding of technical features, objects and effects of the present invention, a detailed description of embodiments of the present invention will be made with reference to the accompanying drawings.
The embodiment of the invention provides a method and a system for intelligently predicting CO/CO 2 in a sintering process by considering unbalanced output data, which are realized based on a mixed kernel correlation vector machine and data enhancement. Because CO/CO 2 is an index for measuring sintering energy consumption, the first determination of key sintering process parameters affecting CO/CO 2 (i.e., input variables of the intelligent predictive model of CO/CO 2) using the minimum absolute shrinkage and selection operator method is: the method comprises the steps of material layer thickness, ignition temperature, mixed water, negative pressure of a bellows, vertical combustion speed, rising point temperature, rising point position, sintering end point temperature and sintering end point position, and obtaining a training data set of a model through a data enhancement technology. And finally, establishing a sintering CO/CO 2 intelligent prediction model by using the proposed mixed kernel-based correlation vector machine.
Referring to fig. 1-4, fig. 1 is a flowchart of a method and a system for intelligently predicting CO/CO 2 in a sintering process in consideration of unbalanced output data according to an embodiment of the present invention, fig. 2 is a modeling schematic block diagram of a CO/CO 2 intelligent prediction model according to an embodiment of the present invention, fig. 3 is a graph of a comparison result between a predicted value and an actual value of CO/CO 2 according to an embodiment of the present invention, and fig. 4 is a prediction error graph of CO/CO 2 according to an embodiment of the present invention. The intelligent prediction method of the CO/CO 2 in the sintering process considering unbalanced output data comprises the following specific steps:
(1) Determination of critical sintering process parameters affecting CO/CO 2
The key sintering process parameters affecting the CO/CO 2 are bed thickness, firing temperature, mixed moisture, windbox negative pressure, vertical firing rate, elevation point temperature, elevation point position, sintering end point temperature, and sintering end point position determined using minimum absolute shrinkage and selection operator methods.
(2) Data augmentation technique to expand output training data
To illustrate the effectiveness of the model in the actual production process, 600 sets of production data were obtained. Of which 500 sets of data are used for model training and the remaining 100 sets of data are used for model testing. Meanwhile, a sample set of 200 groups of input data but lacking output data is expanded by utilizing a data enhancement technology, so that the performance of the model is improved.
If the input variable x does not have the corresponding CO/CO 2 data y m, a k nearest neighbor algorithm based on the Markov distance is applied to acquire k normal CO/CO 2 data related to the input variable x; then, y m, which is finally related to the input variable x, is determined by the following calculation formula:
where y i is k normal CO/CO 2 data related to the input variable x;
finally, the input variable x and the CO/CO 2 data y m finally related to the input variable x are added to the training dataset.
(3) CO/CO 2 intelligent prediction model established based on mixed kernel correlation vector machine method
The key sintering process parameters of the material layer thickness, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the rising point temperature, the rising point position, the sintering end point temperature and the sintering end point position are used as input variables of a model, and an intelligent CO/CO 2 prediction model is built based on a mixed kernel correlation vector machine:
(3-1) let the training data set be expressed as N represents the number of training data, x (t) represents the t training data, y (t) represents CO/CO 2 data corresponding to the t training data, and the intelligent CO/CO 2 prediction model obtained based on the mixed kernel correlation vector machine is as follows:
Wherein ω (t) represents a weight, K h represents a mixing kernel, and x represents any input variable;
(3-2) constructing a hybrid kernel K h using a polynomial kernel and a gaussian kernel:
Kh=τKG+(1-τ)Kpoly
Where K G represents a Gaussian kernel function, K poly represents a polynomial kernel function, and τ represents a scale factor.
(3-3) Let g= [ g (1), g (2), …, g (T), …, g (N) ] T, where g (T) = y (x (T) |ω) +ε (T), ω= [ ω (0), ω (1),..;
(3-4) if g (t) is an independent distribution, the likelihood function is:
wherein, The basis functions are represented by the functions,
(3-5) Assuming that the parameter ω (t) obeys a gaussian distribution with zero mean and a variance α -1 (t), there are:
Wherein α= [ α (0), α (1), α (t), α (N) ] T, α (t) represents a super-parameter of the ω prior distribution, N represents the number of training data, let a=diag (α);
the posterior distribution of (3-6) ω is derived from the likelihood distribution and the prior distribution:
wherein, Σ 2 represents the variance of ε (t);
(3-7) performing integral processing on omega to obtain a marginal distribution determined by parameters alpha and sigma 2:
wherein, I represents an identity matrix.
(4) Sintering process CO/CO 2 prediction
The modeling principle block diagram of the intelligent prediction model of the CO/CO 2 in the sintering process considering unbalanced output data is shown in fig. 2, and the structure of the modeling principle block diagram can better reflect the dynamic performance in the actual sintering production process. The graph of the comparison result of the predicted value and the actual value of CO/CO 2 is shown in FIG. 3, and the graph of the prediction error of CO/CO 2 is shown in FIG. 4. As can be seen from fig. 3 and 4, the prediction error of the predicted result of the CO/CO 2 during sintering is within [ -5%,4% ], which can be interpreted that the predicted value of CO/CO 2 can effectively follow the actual value. Therefore, the modeling method can realize effective prediction of CO/CO 2, can meet the requirement of accurate prediction of energy consumption in the actual production process, and has important engineering practice significance for promoting intelligent manufacture and green manufacture of the sintering process.
The foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and scope of the invention are intended to be included within the scope of the invention.
Claims (8)
1. An intelligent prediction method for a sintering process CO/CO 2 by considering unbalanced output data is characterized in that: the method comprises the following steps:
S1: determining key process parameters affecting CO/CO 2 data by adopting a minimum absolute shrinkage and selection operator method, and taking the key process parameters as input variables of a CO/CO 2 intelligent prediction model, wherein the input variables are as follows: the thickness of the material layer, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the temperature of the rising point, the position of the rising point, the sintering end point temperature and the sintering end point position;
S2: expanding the input variable by using a data enhancement method to obtain a training data set;
S3: building a CO/CO 2 intelligent prediction model based on the mixed kernel correlation vector machine, and training the CO/CO 2 intelligent prediction model by using training data;
s4: and predicting actual production data in real time by using the trained CO/CO 2 intelligent prediction model to obtain a predicted value of CO/CO 2, wherein the predicted value is used for reflecting the combustion degree of carbon-containing energy sources in the sintering production process.
2. The intelligent prediction method for the CO/CO 2 in the sintering process taking unbalanced output data into consideration as claimed in claim 1, wherein: the process of obtaining the training data set in step S2 is as follows:
(2-1) if the input variable has corresponding CO/CO 2 data, directly adding the input variable and the corresponding CO/CO 2 data thereof into the training data set;
(2-2) if the input variable x does not have the corresponding CO/CO 2 data y m, obtaining k normal CO/CO 2 data related to the input variable x by applying a k-nearest neighbor algorithm based on the mahalanobis distance;
(2-3) then determining y m, which is ultimately related to the input variable x, by the following calculation formula:
where y i is k normal CO/CO 2 data related to the input variable x;
(2-4) finally adding the input variable x and its final associated CO/CO 2 data y m to the training dataset.
3. An intelligent prediction method for sintering process CO/CO 2 taking into account unbalanced output data as claimed in claim 1: in step S3, the establishment process of the intelligent CO/CO 2 prediction model is as follows:
(3-1) let the training data set be expressed as N represents the number of training data, x (t) represents the t training data, y (t) represents CO/CO 2 data corresponding to the t training data, and the intelligent CO/CO 2 prediction model obtained based on the mixed kernel correlation vector machine is as follows:
Wherein ω (t) represents a weight, K h represents a mixing kernel, and x represents any input variable;
(3-2) constructing a hybrid kernel K h using a polynomial kernel and a gaussian kernel:
Kh=τKG+(1-τ)Kpoly
Where K G represents a Gaussian kernel function, K poly represents a polynomial kernel function, and τ represents a scale factor.
4. A method for intelligent prediction of sintering process CO/CO 2 taking into account unbalanced output data as claimed in claim 3: in step S3, the calculation process of the intelligent prediction model of CO/CO 2 is as follows:
(3-3) let g= [ g (1), g (2), …, g (T), …, g (N) ] T, where g (T) = y (x (T) |ω) +ε (T), ω= [ ω (0), ω (1),..;
(3-4) if g (t) is an independent distribution, the likelihood function is:
wherein, The basis functions are represented by the functions,
(3-5) Assuming that the parameter ω (t) obeys a gaussian distribution with zero mean and a variance α -1 (t), there are:
Wherein α= [ α (0), α (1), α (t), α (N) ] T, α (t) represents a super-parameter of the ω prior distribution, N represents the number of training data, let a=diag (α);
the posterior distribution of (3-6) ω is derived from the likelihood distribution and the prior distribution:
wherein, Σ 2 represents the variance of ε (t);
(3-7) performing integral processing on omega to obtain a marginal distribution determined by parameters alpha and sigma 2:
wherein, I represents an identity matrix.
5. An intelligent prediction system for a sintering process CO/CO 2 taking unbalanced output data into consideration, which is characterized in that: comprising the following steps:
The input variable determining module is used for determining key process parameters affecting the CO/CO 2 data according to a minimum absolute shrinkage and selection operator method, the key process parameters are used as input variables of the CO/CO 2 intelligent prediction model, and the input variables are as follows: the thickness of the material layer, the ignition temperature, the mixed moisture, the negative pressure of the bellows, the vertical combustion speed, the temperature of the rising point, the position of the rising point, the sintering end point temperature and the sintering end point position;
The training data set acquisition module is used for expanding the input variables by utilizing a data enhancement method to obtain a training data set;
The model building module is used for building a CO/CO 2 intelligent prediction model based on a mixed kernel correlation vector machine modeling method and training the CO/CO 2 intelligent prediction model by using training data;
The real-time prediction module is used for predicting actual production data in real time by utilizing the trained CO/CO 2 intelligent prediction model to obtain a predicted value of CO/CO 2, and the predicted value is used for reflecting the combustion degree of carbon-containing energy sources in the sintering production process.
6. An intelligent prediction system for sintering process CO/CO 2 taking into account unbalanced output data as defined in claim 5 wherein: the process of obtaining the training dataset is as follows:
(2-1) if the input variable has corresponding CO/CO 2 data, directly adding the input variable and the corresponding CO/CO 2 data thereof into the training data set;
(2-2) if the input variable x does not have the corresponding CO/CO 2 data y m, obtaining k normal CO/CO 2 data related to the input variable x by applying a k-nearest neighbor algorithm based on the mahalanobis distance;
(2-3) then determining y m, which is ultimately related to the input variable x, by the following calculation formula:
where y i is k normal CO/CO 2 data related to the input variable x;
(2-4) finally adding the input variable x and its final associated CO/CO 2 data y m to the training dataset.
7. An intelligent prediction system for sintering process CO/CO 2 taking into account unbalanced output data as defined in claim 5 wherein: the establishment process of the intelligent CO/CO 2 prediction model comprises the following steps:
(3-1) let the training data set be expressed as N represents the number of training data, x (t) represents the t training data, y (t) represents CO/CO 2 data corresponding to the t training data, and the intelligent CO/CO 2 prediction model obtained based on the mixed kernel correlation vector machine is as follows:
Wherein ω (t) represents a weight, K h represents a mixing kernel, and x represents any input variable;
(3-2) constructing a hybrid kernel K h using a polynomial kernel and a gaussian kernel:
Kh=τKG+(1-τ)Kpoly
Where K G represents a Gaussian kernel function, K poly represents a polynomial kernel function, and τ represents a scale factor.
8. An intelligent prediction system for sintering process CO/CO 2 taking into account unbalanced output data as defined in claim 7 wherein: the calculation process of the intelligent CO/CO 2 prediction model is as follows:
(3-3) let g= [ g (1), g (2), …, g (T), …, g (N) ] T, where g (T) = y (x (T) |ω) +ε (T), ω= [ ω (0), ω (1),..;
(3-4) if g (t) is an independent distribution, the likelihood function is:
wherein, The basis functions are represented by the functions,
(3-5) Assuming that the parameter ω (t) obeys a gaussian distribution with zero mean and a variance α -1 (t), there are:
Wherein α= [ α (0), α (1), α (t), α (N) ] T, α (t) represents a super-parameter of the ω prior distribution, N represents the number of training data, let a=diag (α);
the posterior distribution of (3-6) ω is derived from the likelihood distribution and the prior distribution:
wherein, Σ 2 represents the variance of ε (t);
(3-7) performing integral processing on omega to obtain a marginal distribution determined by parameters alpha and sigma 2:
wherein, I represents an identity matrix.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410073787.8A CN117912581A (en) | 2024-01-18 | 2024-01-18 | Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410073787.8A CN117912581A (en) | 2024-01-18 | 2024-01-18 | Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN117912581A true CN117912581A (en) | 2024-04-19 |
Family
ID=90696306
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410073787.8A Pending CN117912581A (en) | 2024-01-18 | 2024-01-18 | Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117912581A (en) |
-
2024
- 2024-01-18 CN CN202410073787.8A patent/CN117912581A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101211383A (en) | Blast furnace molten iron silicon content feature analysis and prediction method | |
CN104807039A (en) | Variable dimensionality reduction modeling method for boiler combustion optimization | |
CN103729571B (en) | Modeling method for utilization rate of carbon monoxide in iron-making process of blast furnace | |
CN117289652A (en) | Numerical control machine tool spindle thermal error modeling method based on multi-universe optimization | |
CN112149714A (en) | Method for determining energy efficiency characteristic index reference value of coal-electric unit based on data mining | |
CN112765902A (en) | RBF neural network soft measurement modeling method based on TentFWA-GD and application thereof | |
CN115077640A (en) | Real-time prediction method for main steam flow in urban solid waste incineration process | |
CN117912581A (en) | Sintering process CO/CO taking unbalanced output data into account2Intelligent prediction method and system | |
CN113158355B (en) | All-condition optimization design method for low-temperature liquid expander | |
CN111798023B (en) | Comprehensive coke ratio prediction method in steelmaking sintering production | |
CN110262393B (en) | Gray theory sectional weighted thermal error modeling method with lag data processing | |
CN116127345B (en) | Converter steelmaking process mode design method based on deep clustering generation countermeasure network | |
CN109376446B (en) | Polarization curve fitting method based on particle swarm optimization | |
CN116029433A (en) | Energy efficiency reference value judging method, system, equipment and medium based on grey prediction | |
CN115512455A (en) | Heating furnace steel billet temperature prediction method integrating mechanism and model migration | |
CN115290218A (en) | Soft measurement method and system for wall temperature of boiler water wall of thermal generator set | |
CN111639435B (en) | Banbury mixer rotor energy consumption prediction method for bolt lifting action in rubber mixing process | |
CN112100819B (en) | Casting riser design method based on integrated calculation and data driving | |
CN114818462A (en) | Flow field modeling method based on discrete invariance grid convolution operator | |
CN114282680A (en) | Vehicle exhaust emission prediction method and system based on machine learning algorithm | |
Wang et al. | A deep learning model for the accurate prediction of the microstructure performance of hot rolled steel | |
Rui et al. | Reliability Evaluation of Piston Pump Based on Comprehensive Evaluation Index | |
CN112380779B (en) | Robust soft measurement method and system for sintering end point | |
CN110608660B (en) | Eddy current sensor displacement and temperature prediction method | |
Zhang et al. | Soft Sensor Modelling Method Using Improved LWPLS for Fermentation Monitoring of Pichia Pastoris. |
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 |