基于径向基神经网络的薄板平整轧制力预报模型

上海交通大学学报（自然版） ›› 2011, Vol. 45 ›› Issue (06): 924-928.

基于径向基神经网络的薄板平整轧制力预报模型

段雪厚，王石刚，徐威，唐成龙

(上海交通大学机械与动力工程学院，上海 200240)

收稿日期:2010-07-08 出版日期:2011-06-29 发布日期:2011-06-29
基金资助:
国家自然科学基金资助项目(50975171)

A Model to Predict Temper Rolling Force of Thin Gauge Strip with RBF Neural Networks

DUAN Xue-Hou, WANG Shi-Gang, XU Wei, TANG Cheng-Long

(School of Mechanical Engineering, Shanghai Jiaotong University, Shanghai 200240, China)

Received:2010-07-08 Online:2011-06-29 Published:2011-06-29

摘要/Abstract

摘要： 冷轧薄板在平整轧制时具有轧件厚度薄、压下率小的特点，其平整轧制力往往计算困难，精度难以保证.针对上述情况，提出一种基于参数修正的轧制力数学模型来计算其平整轧制力.同时，为进一步提高计算精度，运用RBF(Radial Basis Function)神经网络来预测该平整轧制力数学模型的计算误差，并将该误差与数学模型的计算值相结合，完成对其的修正.离线仿真结果表明，薄板平整轧制力数学模型在经过自身修正参数及RBF神经网络的2次修正后，计算精度可达到6%以内，具有较高的工程应用价值.

关键词: 平整轧制力预报, 薄板, 径向基神经网络, 参数修正

Abstract: The temper rolling process of thin gauge strip has the features of thin thickness and small reduction, thus it is difficult to calculate its temper rolling force and get a highprecision value. Aiming to solve this problem, this paper established a mathematical model to calculate its temper rolling force based on parameters modification. Meanwhile, the paper predicted the calculation error of the mathematical model above by using RBF neural networks in order to improve its precision further, and completed the modification for temper rolling force model by combining the calculation error and calculated values of original mathematical model. The offline simulation indicates that the precision of new model of temper rolling force for thin gauge strip can reaches 6% after twice modification by its modified parameters and RBF neural networks, which makes it have high value of application in real project.

Key words: prediction of temper rolling force, thin gauge strip, radial basis function(RBF) neural networks, parameters modification

中图分类号:

TG335.5

段雪厚, 王石刚, 徐威, 唐成龙. 基于径向基神经网络的薄板平整轧制力预报模型[J]. 上海交通大学学报（自然版）, 2011, 45(06): 924-928.

DUAN Xue-Hou, WANG Shi-Gang, XU Wei, TANG Cheng-Long. A Model to Predict Temper Rolling Force of Thin Gauge Strip with RBF Neural Networks[J]. Journal of Shanghai Jiaotong University, 2011, 45(06): 924-928.

参考文献

［1］白振华, 康晓鹏, 龙瑞兵, 等. 工程实用平整轧制压力模型及其自学习技术研究［J］. 钢铁, 2008, 43(10): 5154.

BAI Zhenhua, KANG Xiaopeng, LONG Ruibing, et al. Practical temper rolling force model and its self study technology［J］. Iron and Steel, 2008, 43(10):5154.

［2］Le H R, Sutcliffe M P F. A robust model for rolling of thin strip and foil［J］. International Journal of Mechanical Sciences, 2001, 43(6): 14051419.

［3］Fleck N A, Johnson K L. Towards a new theory of cold rolling thin foil［J］.International Journal of Mechanical Sciences, 1987, 29(7): 507524.

［4］唐成龙, 王石刚, 梁庆华, 等. 适用于双机架薄板平整机稳态轧制负荷的数学模型［J］. 上海交通大学学报, 2009, 43(1): 1924.

TANG Chenglong, WANG Shigang, LIAN Qinhua, et al. Mathematical models of stable roll load of double stands thin gauge strip temper mill［J］. Journal of Shanghai Jiaotong University, 2009, 43(1): 1924.

［5］王秀梅, 吕征, 王国栋, 等. 轧制力预报中的神经网络和数学模型［J］. 东北大学学报：自然科学版, 1999, 20(3): 319321.

WANG Xiumei, L Zheng, WANG Guodong, et al. Neural network model and mathematical model for prediction of rolling force［J］. Journal of Northeastern University(Natural Science), 1999, 20(3): 319321.

［6］张俊明, 刘军, 俞小峰, 等. 一种RBF神经网络在某冷连轧机组轧制力计算中的组合应用［J］. 塑性工程学报, 2008, 15(1): 133137.

ZHANG Junming, LIU Jun, YU Xiaofeng, et al. Application of the combination of a sort of RBF neural network in roll force calculation of certain tandem cold mill［J］. Journal of Plasticity Engineering, 2008, 15(1): 133137.

［7］董敏, 刘才, 李灵峰. RBF网络优化设计及在轧机轧制力预报中的应用［J］. 钢铁, 2005, 40(1): 3436.

DONG Min, LIU Cai, LI Lingfeng. Optimized design of RBF network and its application in rolling force prediction［J］. Iron and Steel, 2005, 40(1): 3436.

［8］ZHAO Zhongqiu, HUANG Deshuang. A mended hybrid learning algorithm for radial basis function neural networks to improve generalization capability［J］. Applied Mathematical Modelling, 2007, 31(7): 12711281.

［9］YING Yusheng, WANG Jingcheng, CHEN Chunzhao. A model to predict rolling force of finishing stands with RBF neural networks［J］. Journal of Shanghai Jiaotong University (Science), 2005, 10(3): 256259.

[1]	丁明, 孟帅, 王书恒, 夏玺. 六自由度波浪补偿平台的神经网络自适应反馈线性化控制[J]. 上海交通大学学报, 2022, 56(2): 165-172.
[2]	吴韬, 莫时旭, 向勇斌, 邹泽群, 郑艳. 刚性基底弹性转动约束矩形板各受载条件下屈曲分析[J]. 上海交通大学学报, 2022, 56(1): 114-126.
[3]	师明华1，王旭2，谢磊1，赵路军3. 执行阀黏滞过程的控制性能评估[J]. 上海交通大学学报（自然版）, 2014, 48(07): 1009-1014.
[4]	程妍雪1，庞永杰1，杨卓懿1，冯亮2. 基于径向基神经网络模型的耐压壳6σ设计[J]. 上海交通大学学报（自然版）, 2014, 48(04): 493-497.
[5]	周杰1,戴洪1,陈勇锋1,陆俊江2,姚小兵2. 拉深翻边复合成形工艺参数多目标优化设计[J]. 上海交通大学学报（自然版）, 2014, 48(03): 412-416.
[6]	史冬岩，王青山，石先杰，庄重. 任意边界条件下正交各向异性薄板自由振动特性分析[J]. 上海交通大学学报（自然版）, 2014, 48(03): 434-438.
[7]	张骥超1, 刘罡2, 林忠钦1, 高伟3, 李淑慧1. 侧围外板冲压工艺稳健性优化设计[J]. 上海交通大学学报（自然版）, 2012, 46(07): 1005-1010.
[8]	葛美玲，姚莉秀，李鹃，杨杰. 基于优化CANDIDE模型的人脸动画 [J]. 上海交通大学学报（自然版）, 2010, 44(11): 1491-1500.