CN104657602A - Method and system for forecasting thickness of band steel in hot continuous rolling production process - Google Patents
Method and system for forecasting thickness of band steel in hot continuous rolling production process Download PDFInfo
- Publication number
- CN104657602A CN104657602A CN201510052495.7A CN201510052495A CN104657602A CN 104657602 A CN104657602 A CN 104657602A CN 201510052495 A CN201510052495 A CN 201510052495A CN 104657602 A CN104657602 A CN 104657602A
- Authority
- CN
- China
- Prior art keywords
- belt steel
- steel thickness
- data
- cluster
- continuous rolling
- 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
Abstract
The invention provides a method and a system for forecasting the thickness of band steel in a hot continuous rolling production process, and aims to improve the precision of forecasting the thickness of the band steel. The method comprises the following steps: carrying out clustering analysis on acquired data; determining a band steel thickness forecasting model corresponding to each clustering type by an extreme learning machine according to a clustering result; determining the clustering type to which real-time data in the hot continuous rolling production process belongs, and forecasting the thickness of the band steel by calling the corresponding band steel thickness forecasting model according to the clustering type. The system comprises a clustering unit for carrying out clustering analysis on the acquired data, a model determining unit for determining the band steel thickness forecasting model corresponding to each clustering type by the extreme learning machine according to the clustering result, and a forecasting unit for determining the clustering type to which real-time data in the hot continuous rolling production process belongs, and forecasting the thickness of the band steel by calling the corresponding band steel thickness forecasting model according to the clustering type. The method and the system are suitable for the technical field of hot continuous rolling of the band steel.
Description
Technical field
The present invention relates to hot strip rolling technical field, refer to belt steel thickness Forecasting Methodology and system in a kind of hot continuous rolling production run especially.
Background technology
In recent years, along with the high speed development of science and technology, constantly increase being with the demand of the kind of steel, specification and quantity, require also more and more higher to the quality, precision etc. of band steel, especially the thickness problem with steel, it plays a part more and more important in correction technological parameter on the one hand, and it is also the important indicator weighing product quality on the other hand, so be more and more subject to the attention of each steel mill.
Affect belt steel thickness to produce the factor of deviation and be mainly divided into two classes: the first kind is affected by the technological parameter fluctuation of band steel own and causes, comprise the factors such as the irregular and Segregation of Chemical Composition of supplied materials head and tail temperature inequality, watermark, supplied materials thickness; Equations of The Second Kind is changed by mill data and causes, and comprise the factors such as backing roll bias, thermal expansion of rollers, roll wear and Oil Film of Bearing Oil film thickness change, these factors are all unavoidable influence factors.
Traditional belt steel thickness is the calculating undertaken by mathematical model, owing to inherently ignoring and simplifying the On-the-spot factor of much reality in the process of establishing of mathematical model, therefore mathematical model determination belt steel thickness is relied on merely, its error is larger, can not meet the day by day accurate belt steel thickness requirement of user, actual production requirement can not be met simultaneously.
Summary of the invention
The technical problem to be solved in the present invention is to provide belt steel thickness Forecasting Methodology and system in a kind of hot continuous rolling production run, to solve the large problem of simple dependence mathematical model determination belt steel thickness error existing for prior art.
For solving the problems of the technologies described above, the embodiment of the present invention provides belt steel thickness Forecasting Methodology in a kind of hot continuous rolling production run, comprising:
Obtain the data in hot continuous rolling production run, cluster analysis is carried out to the described data obtained;
According to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
Determine the cluster classification belonging to the real time data in hot continuous rolling production run, call corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
Alternatively, the data in described acquisition hot continuous rolling production run, cluster analysis is carried out to the described data obtained and comprises:
Obtain the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
Correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information, extracts the data attribute relevant to belt steel thickness;
Principal component analysis (PCA) is carried out to the described data attribute extracted and chooses major component number, dimensionality reduction is carried out to described data;
Cluster analysis is carried out to the described data after dimensionality reduction.
Alternatively, the described data attribute to pretreated described data and belt steel thickness information carry out correlation analysis, extract the data attribute relevant to belt steel thickness and comprise:
Correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information;
According to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
Alternatively, described according to cluster result, limit of utilization learning machine determines that belt steel thickness forecast model corresponding to each cluster classification comprises:
According to cluster result, each cluster categorical data is divided into training set and test set;
Limit of utilization learning machine carries out the study of belt steel thickness prediction to described training set, and verifies learning outcome with described test set;
Belt steel thickness forecast model corresponding to each cluster classification is exported according to the result.
Alternatively, the described cluster classification determined belonging to the real time data in hot continuous rolling production run, call corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification and comprise:
Obtain the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Call the belt steel thickness forecast model prediction belt steel thickness that this cluster classification is corresponding.
On the other hand, the embodiment of the present invention also provides belt steel thickness prognoses system in a kind of hot continuous rolling production run, comprising:
Cluster cell: for obtaining the data in hot continuous rolling production run, carries out cluster analysis to the described data obtained;
Model determining unit: for according to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
Predicting unit: for determining the cluster classification belonging to the real time data in hot continuous rolling production run, calls corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
Alternatively, described cluster cell comprises:
Pretreatment module: for obtaining the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
Correlating module: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information, extract the data attribute relevant to belt steel thickness;
Dimensionality reduction module: choose major component number for carrying out principal component analysis (PCA) to the described data attribute extracted, dimensionality reduction is carried out to described data;
Cluster Analysis module: for carrying out cluster analysis to the described data after dimensionality reduction.
Alternatively, described correlating module comprises:
Correlation analysis submodule: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information;
Extraction module: for according to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
Alternatively, described model determining unit comprises:
Diversity module: for according to cluster result, each cluster categorical data is divided into training set and test set;
Study authentication module: the study for limit of utilization learning machine, described training set being carried out to belt steel thickness prediction, and with described test set, learning outcome is verified;
Model determination module: for exporting belt steel thickness forecast model corresponding to each cluster classification according to the result.
Alternatively, described predicting unit comprises:
Processing module: for obtaining the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine class Modules: for determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Prediction module: for calling belt steel thickness forecast model prediction belt steel thickness corresponding to this cluster classification.
The beneficial effect of technique scheme of the present invention is as follows:
In such scheme, by carrying out cluster analysis to the data in the hot continuous rolling production run obtained, and the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine, and then the cluster classification belonging to the real time data in hot continuous rolling production run is judged, the cluster classification belonging to real time data calls corresponding belt steel thickness forecast model prediction belt steel thickness.Like this, the belt steel thickness in hot continuous rolling production run is predicted by the belt steel thickness forecast model determined based on cluster analysis, belt steel thickness precision of prediction can be improved, be conducive to revising technological parameter, improve the quality of products, reach production requirement, thus effective solution in hot continuous rolling production run predicts the problem that belt steel thickness error is large.
Accompanying drawing explanation
The method flow diagram of belt steel thickness Forecasting Methodology in the hot continuous rolling production run that Fig. 1 provides for the embodiment of the present invention one;
Fig. 2 is the specific implementation method process flow diagram of S101 in Fig. 1;
Fig. 3 is the specific implementation method process flow diagram of S102 in Fig. 1;
Fig. 4 is the specific implementation method process flow diagram of S103 in Fig. 1;
The structural representation of belt steel thickness prognoses system in the hot continuous rolling production run that Fig. 5 provides for the embodiment of the present invention two;
Fig. 6 is the concrete structure schematic diagram of in Fig. 5 101.
Embodiment
For making the technical problem to be solved in the present invention, technical scheme and advantage clearly, be described in detail below in conjunction with the accompanying drawings and the specific embodiments.
The present invention is directed to the problem that existing simple dependence mathematical model determination belt steel thickness error is large, belt steel thickness Forecasting Methodology and system in a kind of hot continuous rolling production run is provided.
Embodiment one
Shown in Fig. 1, belt steel thickness Forecasting Methodology in a kind of hot continuous rolling production run that the embodiment of the present invention provides, comprising:
S101: obtain the data in hot continuous rolling production run, carries out cluster analysis to the described data obtained;
S102: according to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
S103: determine the cluster classification belonging to the real time data in hot continuous rolling production run, calls corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
Belt steel thickness Forecasting Methodology in hot continuous rolling production run described in the embodiment of the present invention, by carrying out cluster analysis to the data in the hot continuous rolling production run obtained, and the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine, and then the cluster classification belonging to the real time data in hot continuous rolling production run is judged, the cluster classification belonging to real time data calls corresponding belt steel thickness forecast model prediction belt steel thickness.Like this, the belt steel thickness in hot continuous rolling production run is predicted by the belt steel thickness forecast model determined based on cluster analysis, the precision of prediction of belt steel thickness can be improved, be conducive to revising technological parameter, improve the quality of products, reach production requirement, thus effective solution in hot continuous rolling production run predicts the problem that belt steel thickness error is large.
In aforementioned hot rolling production process belt steel thickness Forecasting Methodology embodiment in, alternatively, the data in described acquisition hot continuous rolling production run, the described data obtained are carried out cluster analysis S101 and are comprised:
S1011: obtain the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
S1012: correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information, extracts the data attribute relevant to belt steel thickness;
S1013: principal component analysis (PCA) is carried out to the described data attribute extracted and chooses major component number, dimensionality reduction is carried out to described data;
S1014: cluster analysis is carried out to the described data after dimensionality reduction.
Shown in Fig. 2, in the embodiment of the present invention, first obtain the data in hot continuous rolling production run, integrated, missing values process, denoising and normalized are carried out to the described data in the hot continuous rolling production run obtained; And utilize correlation analysis to carry out correlation analysis to the data attribute of the described data after normalized and belt steel thickness information, thus extract the data attribute relevant to belt steel thickness, principal component analysis (PCA) is carried out to the described data attribute extracted, such as, the mode of variance contribution ratio more than 85% can be used to choose major component number, many data attributes are converted into a few integrated data attribute, thus reach the effect of data attribute dimensionality reduction.
In the embodiment of the present invention, carry out cluster analysis for the data after dimensionality reduction, such as, Fuzzy C-Means Clustering (Fuzzy C-Means Clustering Algorithm can be adopted, FCM) algorithm carries out cluster analysis to the data after dimensionality reduction, and the step of FCM algorithm is as follows:
1), initialization subordinated-degree matrix;
2), random selecting cluster centre;
3), cluster centre and objective function is calculated, if the difference upgrading front and back objective function absolute value is less than default threshold value or iterations reaches maximum times, then algorithm stops obtaining final cluster result, upgrade subordinated-degree matrix, otherwise, return step 2) middle continuation renewal.
In aforementioned hot rolling production process belt steel thickness Forecasting Methodology embodiment in, alternatively, the described data attribute to pretreated described data and belt steel thickness information carry out correlation analysis, extract the data attribute relevant to belt steel thickness and comprise:
Correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information;
According to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
In the embodiment of the present invention, such as, the correlation analysis of partial correlation can be utilized to carry out correlation analysis to the data attribute of pretreated described data and product thickness information, and according to the partial correlation coefficient formulae discovery degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value (such as: predetermined threshold value is 85%) extracts, as the parameter of belt steel thickness prediction.
In aforementioned hot rolling production process belt steel thickness Forecasting Methodology embodiment in, alternatively, described according to cluster result, limit of utilization learning machine determines that belt steel thickness forecast model corresponding to each cluster classification comprises:
According to cluster result, each cluster categorical data is divided into training set and test set;
Limit of utilization learning machine carries out the study of belt steel thickness prediction to described training set, and verifies learning outcome with described test set;
Belt steel thickness forecast model corresponding to each cluster classification is exported according to the result.
Shown in Fig. 3, in the embodiment of the present invention, data after dimensionality reduction are divided into k class by FCM algorithm, wherein, k value is according to test of many times, obtained by the number of times choosing Clustering Effect best, each cluster categorical data is preserved respectively, and each cluster categorical data is divided into two parts, a part is as training set, another part is as test set, belt steel thickness prediction study is carried out to described training set data operating limit learning machine, and utilize described test set data to verify learning outcome, belt steel thickness forecast model corresponding to each cluster classification is exported again according to the result, such as, the corresponding belt steel thickness forecast model 1 of cluster 1, the corresponding belt steel thickness forecast model 2 of cluster 2, the like, cluster k corresponding belt steel thickness forecast model k.
In aforementioned hot rolling production process belt steel thickness Forecasting Methodology embodiment in, alternatively, the described cluster classification determined belonging to the real time data in hot continuous rolling production run, call corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification and comprise:
Obtain the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Call the belt steel thickness forecast model prediction belt steel thickness that this cluster classification is corresponding.
Shown in Fig. 4, in the embodiment of the present invention, first obtain the real time data in hot continuous rolling production run, pre-service is carried out to the described real time data in the hot continuous rolling production run obtained; And utilize correlation analysis to carry out correlation analysis to the data attribute of pretreated described real time data and belt steel thickness information, thus extract the data attribute relevant to belt steel thickness, principal component analysis (PCA) is carried out to the described data attribute extracted, such as, the mode of variance contribution ratio more than 85% can be used to choose major component number, many data attributes are converted into a few integrated data attribute, thus reach the effect of data attribute dimensionality reduction.Calculate the distance of the real time data after dimensionality reduction to each cluster class center, if the real time data after dimensionality reduction is the shortest to the centre distance of certain cluster classification, then described real time data is attributed to this cluster classification, and described real time data is assigned in affiliated cluster classification, and call belt steel thickness forecast model prediction belt steel thickness corresponding to this cluster classification, such as, real time data after dimensionality reduction is the shortest to the distance at cluster 1 center, then described real time data belongs to cluster 1, and the belt steel thickness forecast model 1 calling cluster 1 correspondence predicts belt steel thickness.
Embodiment two
The present invention also provides the embodiment of belt steel thickness prognoses system in a kind of hot continuous rolling production run, because belt steel thickness prognoses system in hot continuous rolling production run provided by the invention is corresponding with the embodiment of belt steel thickness Forecasting Methodology in aforementioned hot rolling production process, in this hot continuous rolling production run, belt steel thickness prognoses system can realize object of the present invention by the process step performed in said method embodiment, therefore the explanation explanation in above-mentioned hot continuous rolling production run in belt steel thickness Forecasting Methodology embodiment, also the embodiment of belt steel thickness prognoses system in hot continuous rolling production run provided by the invention is applicable to, to repeat no more in embodiment below the present invention.
Shown in Fig. 5, the embodiment of the present invention also provides belt steel thickness prognoses system in a kind of hot continuous rolling production run, comprising:
Cluster cell 101: for obtaining the data in hot continuous rolling production run, carries out cluster analysis to the described data obtained;
Model determining unit 102: for according to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
Predicting unit 103: for determining the cluster classification belonging to the real time data in hot continuous rolling production run, calls corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
Belt steel thickness prognoses system in hot continuous rolling production run described in the embodiment of the present invention, by carrying out cluster analysis to the data in the hot continuous rolling production run obtained, and the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine, and then the cluster classification belonging to the real time data in hot continuous rolling production run is judged, the cluster classification belonging to real time data calls corresponding belt steel thickness forecast model prediction belt steel thickness.Like this, the belt steel thickness in hot continuous rolling production run is predicted by the belt steel thickness forecast model determined based on cluster analysis, the precision of prediction of belt steel thickness can be improved, be conducive to revising technological parameter, improve the quality of products, reach production requirement, thus effective solution in hot continuous rolling production run predicts the problem that belt steel thickness error is large.
Shown in Fig. 6, in aforementioned hot rolling production process belt steel thickness prognoses system embodiment in, alternatively, described cluster cell comprises:
Pretreatment module 1011: for obtaining the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
Correlating module 1012: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information, extract the data attribute relevant to belt steel thickness;
Dimensionality reduction module 1013: choose major component number for carrying out principal component analysis (PCA) to the described data attribute extracted, dimensionality reduction is carried out to described data;
Cluster Analysis module 1014: for carrying out cluster analysis to the described data after dimensionality reduction.
In aforementioned hot rolling production process belt steel thickness prognoses system embodiment in, alternatively, described correlating module comprises:
Correlation analysis submodule: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information;
Extraction module: for according to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
In aforementioned hot rolling production process belt steel thickness prognoses system embodiment in, alternatively, described model determining unit comprises:
Diversity module: for according to cluster result, each cluster categorical data is divided into training set and test set;
Study authentication module: the study for limit of utilization learning machine, described training set being carried out to belt steel thickness prediction, and with described test set, learning outcome is verified;
Model determination module: for exporting belt steel thickness forecast model corresponding to each cluster classification according to the result.
In aforementioned hot rolling production process belt steel thickness prognoses system embodiment in, alternatively, described predicting unit comprises:
Processing module: for obtaining the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine class Modules: for determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Prediction module: for calling belt steel thickness forecast model prediction belt steel thickness corresponding to this cluster classification.
The above is the preferred embodiment of the present invention; it should be pointed out that for those skilled in the art, under the prerequisite not departing from principle of the present invention; can also make some improvements and modifications, these improvements and modifications also should be considered as protection scope of the present invention.
Claims (10)
1. a belt steel thickness Forecasting Methodology in hot continuous rolling production run, is characterized in that, comprising:
Obtain the data in hot continuous rolling production run, cluster analysis is carried out to the described data obtained;
According to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
Determine the cluster classification belonging to the real time data in hot continuous rolling production run, call corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
2. method according to claim 1, is characterized in that, the data in described acquisition hot continuous rolling production run, carries out cluster analysis comprise the described data obtained:
Obtain the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
Correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information, extracts the data attribute relevant to belt steel thickness;
Principal component analysis (PCA) is carried out to the described data attribute extracted and chooses major component number, dimensionality reduction is carried out to described data;
Cluster analysis is carried out to the described data after dimensionality reduction.
3. method according to claim 2, is characterized in that, the described data attribute to pretreated described data and belt steel thickness information carry out correlation analysis, extracts the data attribute relevant to belt steel thickness and comprises:
Correlation analysis is carried out to the data attribute of pretreated described data and belt steel thickness information;
According to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
4. method according to claim 1, is characterized in that, described according to cluster result, and limit of utilization learning machine determines that belt steel thickness forecast model corresponding to each cluster classification comprises:
According to cluster result, each cluster categorical data is divided into training set and test set;
Limit of utilization learning machine carries out the study of belt steel thickness prediction to described training set, and verifies learning outcome with described test set;
Belt steel thickness forecast model corresponding to each cluster classification is exported according to the result.
5. method according to claim 1, is characterized in that, the described cluster classification determined belonging to the real time data in hot continuous rolling production run, calls corresponding belt steel thickness forecast model prediction belt steel thickness comprise according to described cluster classification:
Obtain the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Call the belt steel thickness forecast model prediction belt steel thickness that this cluster classification is corresponding.
6. a belt steel thickness prognoses system in hot continuous rolling production run, is characterized in that, comprising:
Cluster cell: for obtaining the data in hot continuous rolling production run, carries out cluster analysis to the described data obtained;
Model determining unit: for according to cluster result, the belt steel thickness forecast model that each cluster classification is corresponding determined by limit of utilization learning machine;
Predicting unit: for determining the cluster classification belonging to the real time data in hot continuous rolling production run, calls corresponding belt steel thickness forecast model prediction belt steel thickness according to described cluster classification.
7. system according to claim 6, is characterized in that, described cluster cell comprises:
Pretreatment module: for obtaining the data in hot continuous rolling production run, carry out pre-service to the described data obtained, described pre-service comprises: missing values process, denoising and normalized;
Correlating module: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information, extract the data attribute relevant to belt steel thickness;
Dimensionality reduction module: choose major component number for carrying out principal component analysis (PCA) to the described data attribute extracted, dimensionality reduction is carried out to described data;
Cluster Analysis module: for carrying out cluster analysis to the described data after dimensionality reduction.
8. system according to claim 7, is characterized in that, described correlating module comprises:
Correlation analysis submodule: for carrying out correlation analysis to the data attribute of pretreated described data and belt steel thickness information;
Extraction module: for according to the formula of correlation coefficient determination degree of correlation, the data attribute degree of correlation being exceeded predetermined threshold value extracts, as the parameter of belt steel thickness prediction.
9. system according to claim 6, is characterized in that, described model determining unit comprises:
Diversity module: for according to cluster result, each cluster categorical data is divided into training set and test set;
Study authentication module: the study for limit of utilization learning machine, described training set being carried out to belt steel thickness prediction, and with described test set, learning outcome is verified;
Model determination module: for exporting belt steel thickness forecast model corresponding to each cluster classification according to the result.
10. system according to claim 6, is characterized in that, described predicting unit comprises:
Processing module: for obtaining the real time data in hot continuous rolling production run, and pre-service, correlation analysis and dimension-reduction treatment are carried out to described real time data;
Determine class Modules: for determine with dimensionality reduction after the shortest cluster classification of real time data distance, and described real time data to be assigned in this cluster classification;
Prediction module: for calling belt steel thickness forecast model prediction belt steel thickness corresponding to this cluster classification.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510052495.7A CN104657602A (en) | 2015-02-02 | 2015-02-02 | Method and system for forecasting thickness of band steel in hot continuous rolling production process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510052495.7A CN104657602A (en) | 2015-02-02 | 2015-02-02 | Method and system for forecasting thickness of band steel in hot continuous rolling production process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104657602A true CN104657602A (en) | 2015-05-27 |
Family
ID=53248716
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510052495.7A Pending CN104657602A (en) | 2015-02-02 | 2015-02-02 | Method and system for forecasting thickness of band steel in hot continuous rolling production process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104657602A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160437A (en) * | 2015-09-25 | 2015-12-16 | 国网浙江省电力公司 | Load model prediction method based on extreme learning machine |
| CN105929812A (en) * | 2016-04-18 | 2016-09-07 | 北京科技大学 | Strip steel hot continuous rolling quality fault diagnosis method and device |
| CN107274016A (en) * | 2017-06-13 | 2017-10-20 | 辽宁大学 | The strip exit thickness Forecasting Methodology of the random symmetrical extreme learning machine of algorithm optimization that leapfrogs |
| CN108256260A (en) * | 2018-02-05 | 2018-07-06 | 北京科技大学 | A kind of continuous casting billet quality Forecasting Methodology based on extreme learning machine |
| CN108994090A (en) * | 2018-07-06 | 2018-12-14 | 南京钢铁股份有限公司 | A kind of diagnostic method of down-pressing system of rolling mill failure |
| CN110110447A (en) * | 2019-05-09 | 2019-08-09 | 辽宁大学 | It is a kind of to mix the feedback limit learning machine steel strip thickness prediction technique that leapfrogs |
| CN110275868A (en) * | 2019-06-21 | 2019-09-24 | 厦门嵘拓物联科技有限公司 | A kind of multi-modal pretreated method of manufaturing data in intelligent plant |
| WO2021004198A1 (en) * | 2019-07-10 | 2021-01-14 | 江苏金恒信息科技股份有限公司 | Plate performance prediction method and apparatus |
-
2015
- 2015-02-02 CN CN201510052495.7A patent/CN104657602A/en active Pending
Non-Patent Citations (9)
| Title |
|---|
| HIROYUKI MORI等: "Deterministic Annealing Clustering for ANN-Based Short-Term Load Forecasting", 《IEEE TRANSACTIONS ON POWER SYSTEMS》 * |
| 刘莹莹等: "基于粗糙集和神经网络的薄带钢厚度预测", 《济南大学学报（自然科学版）》 * |
| 彭开香等: "基于信息融合的带钢厚度预测控制", 《北京科技大学学报》 * |
| 彭开香等: "基于贝叶斯神经网络的带钢厚度预测与控制", 《北京科技大学学报》 * |
| 李彩华等: "混合式短期边际电价预测模型", 《电力系统自动化》 * |
| 林树宽等: "基于聚类和主成分分析的神经网络预测模型", 《小型微型计算机系统》 * |
| 梁辉等: "基于减法聚类的带钢厚度数据驱动建模", 《北京科技大学学报》 * |
| 梁辉等: "基于数据驱动的热连轧厚度建模与控制方法", 《北京工业大学学报》 * |
| 陈德花等: "福建汛期降水主分量逐步回归预测模型研究", 《气象》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105160437A (en) * | 2015-09-25 | 2015-12-16 | 国网浙江省电力公司 | Load model prediction method based on extreme learning machine |
| CN105929812A (en) * | 2016-04-18 | 2016-09-07 | 北京科技大学 | Strip steel hot continuous rolling quality fault diagnosis method and device |
| CN107274016A (en) * | 2017-06-13 | 2017-10-20 | 辽宁大学 | The strip exit thickness Forecasting Methodology of the random symmetrical extreme learning machine of algorithm optimization that leapfrogs |
| CN108256260A (en) * | 2018-02-05 | 2018-07-06 | 北京科技大学 | A kind of continuous casting billet quality Forecasting Methodology based on extreme learning machine |
| CN108994090A (en) * | 2018-07-06 | 2018-12-14 | 南京钢铁股份有限公司 | A kind of diagnostic method of down-pressing system of rolling mill failure |
| CN108994090B (en) * | 2018-07-06 | 2020-03-27 | 南京钢铁股份有限公司 | Method for diagnosing faults of rolling mill pressing system |
| CN110110447A (en) * | 2019-05-09 | 2019-08-09 | 辽宁大学 | It is a kind of to mix the feedback limit learning machine steel strip thickness prediction technique that leapfrogs |
| CN110110447B (en) * | 2019-05-09 | 2023-04-18 | 辽宁大学 | Method for predicting thickness of strip steel of mixed frog leaping feedback extreme learning machine |
| CN110275868A (en) * | 2019-06-21 | 2019-09-24 | 厦门嵘拓物联科技有限公司 | A kind of multi-modal pretreated method of manufaturing data in intelligent plant |
| WO2021004198A1 (en) * | 2019-07-10 | 2021-01-14 | 江苏金恒信息科技股份有限公司 | Plate performance prediction method and apparatus |
| GB2600213A (en) * | 2019-07-10 | 2022-04-27 | Jiangsu Jinheng Information Tech Co Ltd | Plate performance prediction method and apparatus |
| GB2600213B (en) * | 2019-07-10 | 2023-06-14 | Jiangsu Jinheng Information Tech Co Ltd | Method and apparatus for predicitng plate performance |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104657602A (en) | Method and system for forecasting thickness of band steel in hot continuous rolling production process | |
| Ding et al. | Effect of input variables on cooling load prediction accuracy of an office building | |
| CN107377634B (en) | A kind of hot-strip outlet Crown Prediction of Media method | |
| Zhang et al. | Carbon emissions reductions and technology gaps in the world's factory, 1990–2012 | |
| Lall et al. | China's competitive performance: a threat to East Asian manufactured exports? | |
| CN104517162A (en) | On-line hardness forecasting method of continuous annealing product by means of integrated learning | |
| CN102658297B (en) | Self-learning method for improving quality of first band steel plate shape with changed specification | |
| CN108311544B (en) | A kind of roll-force Parameter Self-learning method and device | |
| CN105678404A (en) | Micro-grid load prediction system and method based on electricity purchased on-line and dynamic correlation factor | |
| CN102982466B (en) | A kind of score in predicting method based on user's liveness | |
| CN105204333B (en) | A kind of energy consumption Forecasting Methodology for improving iron and steel enterprise's energy utilization rate | |
| CN103942375A (en) | High-speed press sliding block dimension robust design method based on interval | |
| CN103123483B (en) | Rolling deformation resistance prediction system of new steel based on clustering analysis | |
| CN103093285A (en) | Short-term load forecast method based on artificial neural network | |
| Li et al. | Dynamic time features expanding and extracting method for prediction model of sintering process quality index | |
| Thiede | Digital technologies, methods and tools towards sustainable manufacturing: Does Industry 4.0 support to reach environmental targets? | |
| CN102886384B (en) | Flatness defect identification method of 20-roller Sendzimir rolling mill based on support vector machine | |
| Yin et al. | Dynamic real–time abnormal energy consumption detection and energy efficiency optimization analysis considering uncertainty | |
| Bu et al. | A novel approach to improve the computing accuracy of rolling force and forward slip | |
| Xu | Economic Modeling of Profit and Cost &Quantity in Forge Process of Screw | |
| Shi et al. | Tool life prediction of dicing saw based on PSO-BP neural network | |
| CN102541032B (en) | Bottleneck device forecasting method for re-entrant manufacturing system | |
| CN106559414A (en) | Network attack consequence dynamic quantitative appraisal procedure based on region situation information | |
| CN106407550A (en) | Soft sensor modeling method in industrial process | |
| CN105956318A (en) | Improved splitting H-K clustering method-based wind power plant fleet division method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150527 |
|
| RJ01 | Rejection of invention patent application after publication |