摘要
板形缺陷模式识别是铜带轧机板形控制中最重要的部分.本文在对多种识别方法对比分析的基础上,提出了一种基于Legendre多项式的板形缺陷识别优化算法.利用其正交特性,结合轧制工艺要求,采用Matlab仿真得到板形基函数多项式的参数设定,代替原Legendre多项式的固定参数,同时根据来料板形特征设置迭代方程.结果表明:该优化算法能够有效地得到板形特征值,既弥补了最小二乘法板形识别的缺陷,又避免了Legendre正交多项式对调节机构的特殊要求;经现场验证,板形平均误差从±8 I减小到±6 I,精度提高了10%左右.
Pattern recognition of flatness defect is the most important part of flatness control in copper strip mill.Based on the comparative analysis of various identification methods,this paper proposes an optimization algorithm for flatness defect identification based on Legendre polynomial.By using its orthogonal characteristics and combining with the rolling process requirements,Matlab simulation is used to obtain the parameter setting of the flatness basis function polynomial instead of the fixed parameters of the original Legendre polynomial.At the same time,the iteration equation is set according to the shape characteristics of incoming material.The results show that the optimization algorithm can obtain the characteristic values of the flatness effectively,which not only makes up for the defect of the flatness recognition by the least square method,but also avoids the special requirement of Legendre orthogonal polynomial on the regulating mechanism.The average flatness error is reduced from±8 I to±6 I,and the accuracy is improved by about 10%.
作者
王海霞
王庆华
Wang Haixia;Wang Qinghua(School of Applied Technology,Soochow University,Suzhou 215325,China;School of Transportation and Logistics,East China Jiaotong University,Nanchang 330013,China;Luoyang Research Institute of Nonferrous Metals Processing,Luoyang 471003,China)
出处
《材料与冶金学报》
CAS
北大核心
2021年第4期275-281,共7页
Journal of Materials and Metallurgy
基金
国家自然科学基金(51805215).
作者简介
王海霞(1975—),女,高级工程师,硕士研究生,E-mail:13776001581@163.com.
