基于机器学习的拉矫延伸率预测模型及数值分析

doi:10.12068/j.issn.1005-3026.2022.02.012

东北大学学报（自然科学版） ›› 2022, Vol. 43 ›› Issue (2): 236-242.DOI: 10.12068/j.issn.1005-3026.2022.02.012

基于机器学习的拉矫延伸率预测模型及数值分析

陈兵，韩烬阳，唐晓垒，夏搏然

(北京科技大学机械工程学院，北京100083)

修回日期:2021-05-31 接受日期:2021-05-31 发布日期:2022-02-28
通讯作者: 陈兵
作者简介:陈兵(1976-)，男，湖北荆州人，北京科技大学副教授.
基金资助:
中央高校基本科研业务费专项资金资助项目(FRF-GF-19-009B).

Prediction Model for Elongation of Tension Leveling Based on Machine Learning Algorithm and Numerical Analysis

CHEN Bing， HAN Jin-yang， TANG Xiao-lei， XIA Bo-ran

School of Mechanical Engineering， University of Science and Technology Beijing， Beijing 100083， China.

Revised:2021-05-31 Accepted:2021-05-31 Published:2022-02-28
Contact: CHEN Bing
About author:-
Supported by:
-

摘要/Abstract

摘要： 在冷轧弯曲矫直过程中，针对拉矫机工艺参数设置问题，利用经验公式、有限元仿真建立的延伸率模型预测精度不高.为提高预测精度，基于传统解析模型与机器学习算法进行研究，比较了两种方法预测模型的精度，得到机器学习算法的延伸率预测模型要比数值解析模型的拟合优度高.比较BP神经网络算法和支持向量机(SVM)算法，得到两种机器学习算法的预测模型精度基本一致.为进一步提高预测精度，采用Adam算法对BP神经网络进行优化，采用遗传算法对SVM预测模型的参数进行优化，最终得到最优预测模型的均值绝对百分比误差MAPE以及拟合优度R^○2分别为13.4%和0.953，可以为实际生产提供技术指导.

关键词: 支持向量机；BP神经网络；延伸率；预测模型优化；冷轧薄板

Abstract: In the process of cold rolling bending straightening， aiming at the setting of process parameters of tension straightener， the prediction accuracy of elongation models established by empirical formula and finite element simulation is not high. To improve the accuracy， the traditional analytical models and machine learning algorithms are studied. The accuracies of the two methods are compared. It is found that the elongation prediction model of machine learning algorithm has higher goodness of fit (R²) than the numerical analytical model. Comparing BP neural network algorithm with SVM (support vector machine) algorithm， the prediction model accuracies of the two machine learning algorithms are basically the same. In order to further improve the prediction accuracy， the BP neural network is optimized by Adam algorithm， and the parameters of SVM prediction model are optimized by genetic algorithm. Finally， the MAPE(mean absolute percentage error) and R^○2 of the optimal prediction model are 13.4% and 0.953 respectively， which can provide technical guidance for actual production.

Key words: SVM; BP neural network; elongation; prediction model optimization; cold rolled sheet

中图分类号:

TG333.2

陈兵，韩烬阳，唐晓垒，夏搏然. 基于机器学习的拉矫延伸率预测模型及数值分析[J]. 东北大学学报（自然科学版）, 2022, 43(2): 236-242.

CHEN Bing， HAN Jin-yang， TANG Xiao-lei， XIA Bo-ran. Prediction Model for Elongation of Tension Leveling Based on Machine Learning Algorithm and Numerical Analysis[J]. Journal of Northeastern University(Natural Science), 2022, 43(2): 236-242.

参考文献

[1]Magura D，Kyslan K，Padmanaban S，et al.Distribution of the strip tensions with slip control in strip processing lines［J/OL］.Energies，2019［2021-04-20］.https://www.mdpi.com/1996-1073/12/15/3010.DOI:10.3390/en12153010.
[2]Shi J L，Ding H，Xu W，et al.Finite element analysis on mechanical properties of 120 billet tension leveler［J/OL］.Materials Science and Engineering，2018 ［2021-04-20］.https://iopscience.iop.org/article/10.1088/1757-899X/397/1/012048.DOI:10.1088/1757-899X/397/1/012048.
[3]肖阳，李启贤.数据挖掘在中国钢铁行业中的应用［C］//中国计量协会冶金分会2012年会暨能源计量与节能降耗经验交流会论文集.南昌，2012:1-4.(Xiao Yang，Li Qi-xian.Application of data mining in China′s iron and steel industry［C］// Proceedings of 2012 Meeting of Metallurgical Branch of China Metrology Association and Experience Exchange Meeting on Energy Measurement，Energy Conservation and Consumption Reduction.Nanchang，2012:1-4.)
[4]Huh H，Lee H W，Sang P R，et al.The parametric process design of tension leveling with an elasto-plastic finite element method［J］.Journal of Materials Processing Technology，2001，113(1/2/3):714-719.
[5]Zhou C L，Jiang Z Y，Wang G D，et al.Deformation work in strip tension leveling［J］.Advanced Materials Research，2011，145:482-487.
[6]马伟然，程军周，任玉成，等.基于神经网络的拉矫机控制模型建立［J］.重型机械，2016(1):6-9.(Ma Wei-ran，Cheng Jun-zhou，Ren Yu-cheng，et al.Establishment of a tension leveler control model based on neural network［J］.Heavy Machinery，2016(1):6-9.)
[7]张清东，刘天浩，王文广，等.宽带钢拉弯矫直工艺参数关系模型研究［J］.钢铁，2007(2):44-46.(Zhang Qing-dong，Liu Tian-hao，Wang Wen-guang，et al.Study on the relationship model of the process parameters of wide-band steel stretch-bending and straightening［J］.Iron and Steel，2007(2):44-46.)
[8]Mathieu N，Potier-Ferry M，Zahrouni H.Reduction of flatness defects in thin metal sheets by a pure tension leveler［J］.International Journal of Mechanical Sciences，2017，122:267-276.
[9]Gui H L，Li Y，Song H，et al.Research on stress neutral layer offset in the straightening process［J/OL］.Advances in Materials Science and Engineering，2015［2021-04-20］.https://www.researchgate.net/publication/286538688_Research_on_Stress_Neutral_Layer_Offset_in_the_Straightening_Process.DOI:10.1155/2015/863610.
[10]丁军，张清东，常铁柱，等.基于神经网络的连续热镀锌线拉矫机工艺参数设定模型［J］.冶金设备，2007(1):33-36.(Ding Jun，Zhang Qing-dong，Chang Tie-zhu，et al.A neural network-based model for setting process parameters of a continuous hot-dip galvanizing line tension leveler［J］.Metallurgical Equipment，2007(1):33-36.)
[11]Misaka Y，Masui T.Shape correction of steel strip by tension leveller［J］.Transactions of the Iron and Steel Institute of Japan，1978，18(8):475-484.
[12]Steinwender L，Kainz A J，Krimpelsttter K，et al.Computational analysis of the curvature distribution and power losses of metal strip in tension levelers［C/OL］//IOP Conference Series Materials Science and Engineering.2010 ［2021-04-15］.https://www.researchgate.net/publication/249232311_Computational_Analysis_of_the_Curvature_Distribution_and_Power_Losses_of_Metal_Strip_in_Tension_Levellers.DOI:10.1088/1757-899X/10/1/012135.
[13]Yoshida F，Urabe M.Computer-aided process design for tension leveling of clad sheet metal［J］.Key Engineering Materials，2000，177:503-508.
[14]Li S Z，Yin Y D，Xu J，et al.Numerical simulation of continuous tension leveling process of thin strip steel and its application［J］.Journal of Iron and Steel Research International，2007，14(6):8-13.
[15]Zhao P J，Zhou C L，Li W N，et al.Analysis of the strip curvature in tension leveling with FEM［J］.Advanced Materials Research，2012，572:308-311.
[16]Morris J W，Hardy S J，Thomas J T.Effects of tension leveling process parameters on cold rolled strip characteristics using a designed factorial analysis approach［J］.Ironmaking & Steelmaking，2013，32(5):443-448.
[17]张婧.拉弯矫直工艺模型与带钢矫后力学性能研究［D］.太原:太原科技大学，2017.(Zhang Jing.Research on tension leveling process and effects on mechanical properties of steel strip［D］.Taiyuan:Taiyuan University of Science and Technology，2017.)
[18]Wang L，Zhang D H，Sun J，et al.Research on the elongation control system of tension leveler for continuous pickling line［J］. Advanced Materials Research，2013，753/754/755:1442-1447.(上接第181页)
[10]Jiang Z Y，Chen S，Molisch A F，et al.Exploiting wireless channel state information structures beyond linear correlations:a deep learning approach［J］.IEEE Communications Magazine，2019，57(3):284-293.
[11]Gmd S B，Smola A J，Williamson R C，et al.New support vector algorithms［J］.Neurocomputing，2000，12(5):1207-1245.
[12]Shao Y H，Zhang C H，Yang Z M，et al.An ε-twin support vector machine for regression［J］.Neural Computing and Applications，2013，23:175-185.
[13]Melki G，Cano A，Ventura S.MIRSVM:multi-instance support vector machine with bag representatives ［J］.Pattern Recognition，2018，79:228-241.
[14]Wu J X，Fan P Z.A survey on high mobility wireless communications:challenges，opportunities and solutions［J］.IEEE Access，2016，4:450-476.
[15]Clerckx B，Kim J.On the beneficial roles of fading and transmit diversity in wireless power transfer with nonlinear energy harvesting［J］.IEEE Transactions on Wireless Communications，2018，17(11):7731-7743.
[16]Xu Y T，Li X Y，Pan X L，et al.Asymmetric ν-twin support vector regression［J］.Neural Computing and Applications， 2018，30(12):3799-3814.
[17]Yang Y W，Gao F F，Ma X L，et al.Deep learning-based channel estimation for doubly selective fading channels ［J］.IEEE Access，2019，7:36579-36589.

基于机器学习的拉矫延伸率预测模型及数值分析

Prediction Model for Elongation of Tension Leveling Based on Machine Learning Algorithm and Numerical Analysis

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 0

编辑推荐

Metrics

本文评价