CN104376226A - Method and system for calculating band steel thickness reduction amount - Google Patents
Method and system for calculating band steel thickness reduction amount Download PDFInfo
- Publication number
- CN104376226A CN104376226A CN201410708980.0A CN201410708980A CN104376226A CN 104376226 A CN104376226 A CN 104376226A CN 201410708980 A CN201410708980 A CN 201410708980A CN 104376226 A CN104376226 A CN 104376226A
- Authority
- CN
- China
- Prior art keywords
- input
- computation model
- thickness
- parameter
- output
- 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
Landscapes
- Control Of Metal Rolling (AREA)
Abstract
The invention relates to the technical field of metal material processing and discloses a method and system for calculating the band steel thickness reduction amount. The method includes the steps that production data are collected to form input parameters and an output parameter of a teacher training sample, wherein the input parameters include the band steel material, the band steel thickness, the band steel temperature, the annealing temperature, the tension of a heating section and the leveling/finishing elongation, and the output parameter is the band steel thickness reduction amount; a calculation model is built according to the input parameters and the output parameter of the teacher training sample, and the calculation model is trained until the output parameter is obtained through the input parameters; the actual band steel material, the actual band steel thickness, the actual band steel temperature, the actual annealing temperature, the actual tension of the heating section and the actual leveling/finishing elongation are input to the trained calculation model as the input parameters, wherein the output quantity of the calculation model is the actual band steel thickness reduction amount, so that the band steel thickness reduction amount is obtained.
Description
Technical field
The present invention relates to metal material processing technical field, be mainly applicable to the method and system calculating belt steel thickness Reducing thickness.
Background technology
The thickness and precision of steel plate finished product is not only the most basic index of meeting consumers' demand, and is the most important condition ensureing that finished product dispatches from the factory smoothly.By on cold rolling-connect and move back, affect in cold rolling-zinc-plated production process after the thinning various influence factors of product thickness analyze, finally draw the method-mill milling thickness compensation Controlling model improving finished product thickness precision.For connect move back, zinc-plated production, band steel at smooth/polishing machine place because extensibility can cause the thinning of thickness.In addition, can change under being with the heating condition of the yield strength of steel in annealing furnace, band steel be deformed under the condition of tension force, thus causes the change on belt steel thickness direction.And at present, the method that can not detect the change of this belt steel thickness.
Summary of the invention
Technical matters to be solved by this invention is to provide a kind of method and system calculating belt steel thickness Reducing thickness, and it can obtain the Reducing thickness of belt steel thickness.
For solving the problems of the technologies described above, the invention provides a kind of method calculating belt steel thickness Reducing thickness, comprising:
Gather the input and output parameter that production data forms teacher's training sample respectively; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Build computation model by the input and output parameter in described teacher's training sample, and described computation model is trained until described output parameter can be obtained by described input parameter;
Be input in the computation model trained using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
Further, described by the input and output parameter structure computation model in described teacher's training sample, and described computation model is trained until described output parameter can be obtained by described input parameter, comprise: by the input and output parameter in described teacher's training sample by weighted sum, compare with thresholding, the method of nonlinear operation builds described computation model, and adjust by the comparative result described input parameter being input to output numerical value that described computing module obtains and described output parameter the training that weights realize described computation model, until described output numerical value mates with described output parameter.
Further, the described comparative result by described input parameter being input to output numerical value that described computing module obtains and described output parameter adjusts the training that weights realize described computation model, comprising:
If described output numerical value mates with described output parameter, strengthen described weights to realize the training of described computation model;
If described output numerical value does not mate with described output parameter, reduce described weights to realize the training of described computation model.
Further, also comprise:
Gather the input and output parameter that production data forms test samples respectively; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Input parameter in described test samples is input in the computation model trained, the output numerical value obtained is compared with the output parameter in described test samples and mates;
If comparative result is coupling, then illustrates and check successfully to described computation model;
Described the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility are input in the computation model trained as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness, comprise: be input in the successful computation model of inspection using the band steel matter of described reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
Further, also comprise: described teacher's training sample and described test samples are normalized.
A kind of system calculating belt steel thickness Reducing thickness provided by the invention, comprising:
First data acquisition module, forms the input and output parameter of teacher's training sample respectively for gathering production data; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Model construction module, for building computation model by the input and output parameter in described teacher's training sample, and trains until can obtain described output parameter by described input parameter to described computation model;
Data outputting module, for being input in the computation model trained using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
Further, described model construction module, comprising:
Model construction performance element, for by the input and output parameter in described teacher's training sample by weighted sum, compare with thresholding, the method for nonlinear operation builds described computation model;
Model training unit, the training that weights realize described computation model is adjusted, until described output numerical value mates with described output parameter for the comparative result by described input parameter being input to output numerical value that described computing module obtains and described output parameter.
Further, described model training unit, specifically for:
Mate with described output parameter if described comparative result is described output numerical value, strengthen described weights to realize the training of described computation model;
Do not mate with described output parameter if described comparative result is described output numerical value, reduce described weights to realize the training of described computation model.
Further, also comprise:
Second data acquisition module, forms the input and output parameter of test samples respectively for gathering production data; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Model testing module, for being input to by the input parameter in described test samples in the computation model that trains, comparing the output numerical value obtained with the output parameter in described test samples and mating;
If comparative result is coupling, then illustrates and check successfully to described computation model;
Described data outputting module, specifically for being input in the successful computation model of inspection using the band steel matter of described reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
Further, also comprise:
Data processing module, for being normalized described teacher's training sample and described test samples.
Beneficial effect of the present invention is:
The method and system of calculating belt steel thickness Reducing thickness provided by the invention, build computation model by gathering production sample data, and train constructed computation model until computation model can be used; Again actual production data are input to computation model, by the Reducing thickness of computation model output strip thickness, thus obtain the Reducing thickness of belt steel thickness, then belt steel thickness Reducing thickness is compensated to the exit thickness as milling train on order thickness, just can meet the request for utilization of user to thickness and precision.
Accompanying drawing explanation
The process flow diagram of the method for the calculating belt steel thickness Reducing thickness that Fig. 1 provides for the embodiment of the present invention;
The schematic diagram of the method for the calculating belt steel thickness Reducing thickness that Fig. 2 provides for the embodiment of the present invention;
The structured flowchart of the system of the calculating belt steel thickness Reducing thickness that Fig. 3 provides for the embodiment of the present invention.
Embodiment
For setting forth the present invention further for the technological means reaching predetermined goal of the invention and take and effect, below in conjunction with accompanying drawing and preferred embodiment, the embodiment of method and system of the calculating belt steel thickness Reducing thickness proposed according to the present invention and principle of work are described in detail.
See Fig. 1 and Fig. 2, the method for the calculating belt steel thickness Reducing thickness that the embodiment of the present invention provides, comprising:
Step S110: gather the input and output parameter that production data forms teacher's training sample respectively; Wherein, input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Output parameter is the Reducing thickness of belt steel thickness; Particularly, belt steel thickness and strip width can be measured respectively by thickness-measuring equipment and the wide equipment of survey; Annealing temperature can be measured by the temperature measuring equipment in annealing furnace; Bringing-up section tension force can be measured by tension measuring device; Formula can be passed through
calculate smooth/polishing extensibility ε; In formula, v
entryfor strip speed before smooth/polishing, unit is m/min; v
exitfor strip speed after smooth/polishing, unit is m/min.It should be noted that, extensibility is the ratio of the forward and backward strip length change of smooth/polishing rolling.It has been generally acknowledged that band steel before the rolling, after width be constant, if at one time in the cycle, L
entrybe defined as the length of planisher/polishing machine strip steel at entry, unit is mm; L
exitbe defined as the length of planisher/polishing machine outlet band steel, unit is mm, then smooth/polishing extensibility is:
According to constancy of volume principle, can calculate smooth/polishing extensibility with band steel in the speed that smooth/polishing is forward and backward, its formula is:
In embodiments of the present invention, thickness-measuring equipment can comprise: thicknessmeter, feeler etc.; Survey wide equipment can comprise: width gage, survey wide meter etc.; Temperature measuring equipment can comprise: temperature measurer, temperature sensor etc.; Tension measuring device can comprise: tensiometer, tension pick-up etc.
Step S120: be normalized teacher's training sample, is namely converted into the variable in (0,1) scope.
Step S130: build computation model by the input and output parameter in teacher's training sample, and under the condition of setting accuracy, computation model is trained until output parameter can be obtained by input parameter;
Be specifically described this step, this step specifically comprises:
By the input and output parameter in teacher's training sample by weighted sum, compare with thresholding, the method for nonlinear operation builds computation model, and adjust by comparative result input parameter being input to output numerical value that computing module obtains and output parameter the training that weights realize computation model, mate until export numerical value with output parameter.This step is further detailed, mates with output parameter if export numerical value, strengthen weights to realize the training of computation model; If export numerical value not mate with output parameter, reduce weights to realize the training of computation model.
Step S140: be input in the computation model trained as input parameter using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility, the output parameter of computation model is the Reducing thickness of actual belt steel thickness.
The method that the embodiment of the present invention provides is further detailed, also comprises:
Gather the input and output parameter that production data forms test samples respectively; Wherein, input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Output parameter is the Reducing thickness of belt steel thickness; Particularly, belt steel thickness and strip width can be measured respectively by thickness-measuring equipment and the wide equipment of survey; Annealing temperature can be measured by the temperature measuring equipment in annealing furnace; Bringing-up section tension force can be measured by tension measuring device; Formula can be passed through
calculate smooth/polishing extensibility ε; In formula, v
entryfor strip speed before smooth/polishing, unit is m/min; v
exitfor strip speed after smooth/polishing, unit is m/min.It should be noted that, extensibility is the ratio of the forward and backward strip length change of smooth/polishing rolling.It has been generally acknowledged that band steel before the rolling, after width be constant, if at one time in the cycle, L
entrybe defined as the length of planisher/polishing machine strip steel at entry, unit is mm; L
exitbe defined as the length of planisher/polishing machine outlet band steel, unit is mm, then smooth/polishing extensibility is:
According to constancy of volume principle, can calculate smooth/polishing extensibility with band steel in the speed that smooth/polishing is forward and backward, its formula is:
In embodiments of the present invention, thickness-measuring equipment can comprise: thicknessmeter, feeler etc.; Survey wide equipment can comprise: width gage, survey wide meter etc.; Temperature measuring equipment can comprise: temperature measurer, temperature sensor etc.; Tension measuring device can comprise: tensiometer, tension pick-up etc.
Test samples is normalized, is namely converted into the variable in (0,1) scope.
Input parameter in test samples is input in the computation model trained, the output numerical value obtained is compared with the output parameter in test samples and mates;
If comparative result is coupling, then illustrates and check successfully to computation model;
If comparative result is not for mate, then illustrate computation model inspection unsuccessful, need the parameter of optimization teacher training sample and/or the weights of Adjustable calculation model until comparative result coupling.
Then step S140 specifically comprises: be input in the successful computation model of inspection using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of computation model is the Reducing thickness of actual belt steel thickness.
See Fig. 3, the system of the calculating belt steel thickness Reducing thickness that the embodiment of the present invention provides, comprising:
First data acquisition module 100, forms the input and output parameter of teacher's training sample respectively for gathering production data; Wherein, input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Output parameter is the Reducing thickness of belt steel thickness;
Particularly, the first data acquisition module 100, comprising:
Belt steel thickness computing unit, for measuring belt steel thickness based on thickness-measuring equipment; In embodiments of the present invention, thickness-measuring equipment can comprise: thicknessmeter, feeler etc.;
Strip width computing unit, for based on survey wide device measuring strip width; In embodiments of the present invention, survey wide equipment can comprise: width gage, survey wide meter etc.;
Annealing temperature computing unit, for measuring annealing temperature based on the temperature measuring equipment in annealing furnace; In embodiments of the present invention, temperature measuring equipment can comprise: temperature measurer, temperature sensor etc.;
Tension force computing unit, for measuring bringing-up section tension force based on tension measuring device; In embodiments of the present invention, tension measuring device can comprise: tensiometer, tension pick-up etc.
Smooth/polishing extensibility computing unit, for passing through formula
calculate smooth/polishing extensibility ε; In formula, v
entryfor strip speed before smooth/polishing, unit is m/min; v
exitfor strip speed after smooth/polishing, unit is m/min.
Data processing module 200, for being normalized teacher's training sample, is namely converted into the variable in (0,1) scope.
Model construction module 300, for building computation model by the input and output parameter in teacher's training sample, and trains computation model until can obtain output parameter by input parameter under the condition of setting accuracy;
Particularly, model construction module 300, comprising:
Model construction performance element, for by the input and output parameter in teacher's training sample by weighted sum, compare with thresholding, the method for nonlinear operation builds computation model;
Model training unit, adjusting for the comparative result by input parameter being input to output numerical value that computing module obtains and output parameter the training that weights realize computation model, mating until export numerical value with output parameter.
Wherein, model training unit, specifically for:
If comparative result mates with output parameter for exporting numerical value, strengthen weights to realize the training of computation model;
If comparative result does not mate with output parameter for exporting numerical value, reduce weights to realize the training of computation model.
Data outputting module 400, for being input in the computation model after training using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of computation model is the Reducing thickness of actual belt steel thickness.
The structure of the system that the embodiment of the present invention provides is further detailed, also comprises:
Second data acquisition module, forms the input and output parameter of test samples respectively for gathering production data; Wherein, input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Output parameter is the Reducing thickness of belt steel thickness;
Particularly, the second data acquisition module, comprising:
Belt steel thickness computing module, for measuring belt steel thickness based on thickness-measuring equipment; In embodiments of the present invention, thickness-measuring equipment can comprise: thicknessmeter, feeler etc.;
Strip width computing module, for based on survey wide device measuring strip width; In embodiments of the present invention, survey wide equipment can comprise: width gage, survey wide meter etc.;
Annealing temperature computing module, for measuring annealing temperature based on the temperature measuring equipment in annealing furnace; In embodiments of the present invention, temperature measuring equipment can comprise: temperature measurer, temperature sensor etc.;
Tension force computing module, for measuring bringing-up section tension force based on tension measuring device; In embodiments of the present invention, tension measuring device can comprise: tensiometer, tension pick-up etc.
Smooth/polishing extensibility computing module, for passing through formula
calculate smooth/polishing extensibility ε; In formula, v
entryfor strip speed before smooth/polishing, unit is m/min; v
exitfor strip speed after smooth/polishing, unit is m/min.
Data processing module 200, also for being normalized test samples, is namely converted into the variable in (0,1) scope.
Model testing module, for being input to by the input parameter in test samples in the computation model that trains, comparing the output numerical value obtained with the output parameter in test samples and mating;
If comparative result is coupling, then illustrates and check successfully to computation model;
If comparative result is not for mate, then illustrate computation model inspection unsuccessful, need the parameter of optimization teacher training sample and/or the weights of Adjustable calculation model until comparative result coupling.
In this case, data outputting module 400, specifically for being input in the successful computation model of inspection using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of computation model is the Reducing thickness of actual belt steel thickness.
The Reducing thickness of belt steel thickness is tried to achieve by the embodiment of the present invention, first band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility production data is collected as input amendment, collect thickness Reducing thickness and form teacher's training sample set and test samples collection respectively as output sample production data, and teacher's training sample and test samples are normalized, namely the variable in (0,1) scope is converted into; Computation model (as BP neural network model) is built by the input and output parameter in teacher's training sample, and train until output parameter can be obtained by input parameter this computation model under the condition of setting accuracy, and by the threshold value of the training network weight of gained and each neural unit stored in weights file; Test with the BP neural network model that test samples set pair trains again.Check successfully at BP neural network model, be input in the successful BP neural network model of inspection using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input quantity, the output quantity of BP neural network model is the thickness Reducing thickness of actual band steel.Wherein, the successful BP neural network model of inspection is utilized to compare (material is IF steel) as shown in table 1 the result of calculation and actual result of even moving back product thickness Reducing thickness.
Table 1
As seen from the above table, closely, therefore the embodiment of the present invention has the high feature of precision for the Reducing thickness of the belt steel thickness obtained by the successful BP neural network model of the inspection of the embodiment of the present invention and actual measurement thickness Reducing thickness.
The method and system of the calculating belt steel thickness Reducing thickness that the embodiment of the present invention provides, build computation model by gathering production sample data, and train constructed computation model; Again the computation model trained is tested; Again actual production data are input to the successful computation model of inspection, by the Reducing thickness of computation model output strip thickness, thus obtain the Reducing thickness of belt steel thickness, again belt steel thickness Reducing thickness is compensated to the exit thickness as milling train on order thickness, just can meet the request for utilization of user to thickness and precision.By the method and system that the embodiment of the present invention provides, the thickness Reducing thickness model of 3 layers of BP neural network containing 6 input nodes and 1 output node can be built.Input layer variable is wherein band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility, output quantity is belt steel thickness Reducing thickness, thus obtain the Reducing thickness of belt steel thickness, for mill milling thickness compensation Controlling model provides accurate Data safeguard, meet the request for utilization of user to thickness and precision further.Being calculated in the process of belt steel thickness Reducing thickness by BP neural network model, also data being normalized, thus improve the frequency of training of BP neural network model.Because the embodiment of the present invention will be with steel matter as one of input parameter of BP neural network model, thus when producing new steel grade product, also can estimate out the Reducing thickness of its belt steel thickness, guidance is provided to the production of new steel grade product, thus improve the applicability of the embodiment of the present invention.
It should be noted last that, above embodiment is only in order to illustrate technical scheme of the present invention and unrestricted, although with reference to example to invention has been detailed description, those of ordinary skill in the art is to be understood that, can modify to technical scheme of the present invention or equivalent replacement, and not departing from the spirit and scope of technical solution of the present invention, it all should be encompassed in the middle of right of the present invention.
Claims (10)
1. calculate a method for belt steel thickness Reducing thickness, it is characterized in that, comprising:
Gather the input and output parameter that production data forms teacher's training sample respectively; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Build computation model by the input and output parameter in described teacher's training sample, and described computation model is trained until described output parameter can be obtained by described input parameter;
Be input in the computation model trained using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
2. the method for claim 1, it is characterized in that, described by the input and output parameter structure computation model in described teacher's training sample, and described computation model is trained until described output parameter can be obtained by described input parameter, comprise: by the input and output parameter in described teacher's training sample by weighted sum, compare with thresholding, the method of nonlinear operation builds described computation model, and adjust by the comparative result described input parameter being input to output numerical value that described computing module obtains and described output parameter the training that weights realize described computation model, until described output numerical value mates with described output parameter.
3. method as claimed in claim 2, is characterized in that, the described comparative result by described input parameter being input to output numerical value that described computing module obtains and described output parameter adjusts the training that weights realize described computation model, comprising:
If described output numerical value mates with described output parameter, strengthen described weights to realize the training of described computation model;
If described output numerical value does not mate with described output parameter, reduce described weights to realize the training of described computation model.
4. the method for claim 1, is characterized in that, also comprises:
Gather the input and output parameter that production data forms test samples respectively; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Input parameter in described test samples is input in the computation model trained, the output numerical value obtained is compared with the output parameter in described test samples and mates;
If comparative result is coupling, then illustrates and check successfully to described computation model;
Described the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility are input in the computation model trained as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness, comprise: be input in the successful computation model of inspection using the band steel matter of described reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
5. method as claimed in claim 4, is characterized in that, also comprise: be normalized described teacher's training sample and described test samples.
6. calculate a system for belt steel thickness Reducing thickness, it is characterized in that, comprising:
First data acquisition module, forms the input and output parameter of teacher's training sample respectively for gathering production data; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Model construction module, for building computation model by the input and output parameter in described teacher's training sample, and trains until can obtain described output parameter by described input parameter to described computation model;
Data outputting module, for being input in the computation model trained using the band steel matter of reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
7. system as claimed in claim 6, it is characterized in that, described model construction module, comprising:
Model construction performance element, for by the input and output parameter in described teacher's training sample by weighted sum, compare with thresholding, the method for nonlinear operation builds described computation model;
Model training unit, the training that weights realize described computation model is adjusted, until described output numerical value mates with described output parameter for the comparative result by described input parameter being input to output numerical value that described computing module obtains and described output parameter.
8. system as claimed in claim 7, is characterized in that, described model training unit, specifically for:
Mate with described output parameter if described comparative result is described output numerical value, strengthen described weights to realize the training of described computation model;
Do not mate with described output parameter if described comparative result is described output numerical value, reduce described weights to realize the training of described computation model.
9. system as claimed in claim 6, is characterized in that, also comprise:
Second data acquisition module, forms the input and output parameter of test samples respectively for gathering production data; Wherein, described input parameter comprises: band steel matter, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility; Described output parameter is the Reducing thickness of belt steel thickness;
Model testing module, for being input to by the input parameter in described test samples in the computation model that trains, comparing the output numerical value obtained with the output parameter in described test samples and mating;
If comparative result is coupling, then illustrates and check successfully to described computation model;
Described data outputting module, specifically for being input in the successful computation model of inspection using the band steel matter of described reality, belt steel thickness, strip width, annealing temperature, bringing-up section tension force and smooth/polishing extensibility as input parameter, the output quantity of described computation model is the Reducing thickness of actual belt steel thickness.
10. system as claimed in claim 9, is characterized in that, also comprise:
Data processing module, for being normalized described teacher's training sample and described test samples.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410708980.0A CN104376226A (en) | 2014-11-28 | 2014-11-28 | Method and system for calculating band steel thickness reduction amount |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410708980.0A CN104376226A (en) | 2014-11-28 | 2014-11-28 | Method and system for calculating band steel thickness reduction amount |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104376226A true CN104376226A (en) | 2015-02-25 |
Family
ID=52555130
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410708980.0A Pending CN104376226A (en) | 2014-11-28 | 2014-11-28 | Method and system for calculating band steel thickness reduction amount |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104376226A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107686958A (en) * | 2017-09-30 | 2018-02-13 | 马鞍山钢铁股份有限公司 | A kind of accurate control method of continuous hot-dipping galvanizing steel plate finished product thickness |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101097581A (en) * | 2006-06-27 | 2008-01-02 | 宝山钢铁股份有限公司 | Processing parameter setting method of tension straightening machine set correct roller |
| CN103390114A (en) * | 2013-08-02 | 2013-11-13 | 首钢总公司 | Detection judging system and method thereof for thickness precision of band steel finished products |
-
2014
- 2014-11-28 CN CN201410708980.0A patent/CN104376226A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101097581A (en) * | 2006-06-27 | 2008-01-02 | 宝山钢铁股份有限公司 | Processing parameter setting method of tension straightening machine set correct roller |
| CN103390114A (en) * | 2013-08-02 | 2013-11-13 | 首钢总公司 | Detection judging system and method thereof for thickness precision of band steel finished products |
Non-Patent Citations (5)
| Title |
|---|
| 吴海淼等: "冷轧带钢边部减薄及其影响因素分析", 《钢铁研究》 * |
| 岳晓奎等: "《飞行器组合导航鲁棒滤波理论及其应用》", 31 December 2013 * |
| 李文婷: "基于改进型粒子群算法的热轧带钢宽度神经网络预报模型的研究", 《中国优秀硕士学位论文全文数据库 工程科技I辑》 * |
| 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》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107686958A (en) * | 2017-09-30 | 2018-02-13 | 马鞍山钢铁股份有限公司 | A kind of accurate control method of continuous hot-dipping galvanizing steel plate finished product thickness |
| CN107686958B (en) * | 2017-09-30 | 2019-08-27 | 马鞍山钢铁股份有限公司 | A kind of accurate control method of continuous hot-dipping galvanizing steel plate finished product thickness |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103286142B (en) | Dynamical roll gap compensation method during a kind of cold continuous rolling dynamic variable specification | |
| CN104741388B (en) | A kind of Rolling Thickness control method | |
| CN108126988B (en) | Bar minus deviation On-line Measuring Method and system | |
| CN106825069B (en) | A kind of cold-strip steel high precision plates shape surface roughness on-line intelligence control method | |
| CN102189117B (en) | Cold rolled steel strip straightness feedforward control method based on transverse performance detection | |
| CN102658298B (en) | Plate-shape quality online judgment method applicable to hot-rolled thin strip steel | |
| CN103341503B (en) | Self-adaptation convexity change hot rolled plate shape control model | |
| CN106540968A (en) | The compensation method of cold rolled sheet shape measured value and device | |
| CN103376085B (en) | Cold-rolled strip steel shape intelligent comprehensive evaluation method | |
| CN102172639B (en) | Dynamic substitution regulating method for excessive bending of working roll of cold rolling mill | |
| CN101507977A (en) | System error comprehensive compensation technique of strip-mill strip-shape detection device | |
| CN107913911B (en) | It is a kind of to monitor thickness control to the feedforward compensation method and system of entrance tension | |
| CN101869914A (en) | Thickness control method of finish roller strip steel and device | |
| CN107008758A (en) | Cold-strip steel high accuracy plate shape surface roughness On-Line Control Method and system | |
| CN109261748A (en) | A kind of straightener rolling schedule regulating system and method based on rolling schedule parameter | |
| CN102397886B (en) | System for correcting plate detection error due to transverse temperature difference of cold-rolled strip | |
| CN102286751B (en) | Method for judging pickling speed in cold rolling pickling technology | |
| CN107111800A (en) | Method for manufacturing parts based on the analysis to weighted statistical index | |
| CN102886384A (en) | Flatness defect identification method of 20-roller Sendzimir rolling mill based on support vector machine | |
| CN107520259B (en) | A kind of cold rolling new steel grade resistance of deformation coefficient rapid correction method | |
| CN104376226A (en) | Method and system for calculating band steel thickness reduction amount | |
| CN103377318B (en) | Cold-strip steel online plate shape statistical method | |
| CN108817100A (en) | A method of judging that certain breast roller causes strip exit thickness to fluctuate | |
| CN104707869A (en) | Hot rolling width model fast dynamic self-adapting method | |
| CN108655182A (en) | A kind of hot-strip production method of overall length thickness consecutive variations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 100041 Shijingshan Road, Beijing, No. 68, No. Applicant after: Shougang Group Co. Ltd. Address before: 100041 Shijingshan Road, Beijing, No. 68, No. Applicant before: Capital Iron & Steel General Company |
|
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150225 |