摘要
为改善带钢冷连轧机在换辊后由于轧辊的热不稳定状态引起的板厚波动,对轧辊温度场和轧辊径向热膨胀量进行了理论分析和计算。建立了基于人工神经网络的轧辊热膨胀量预报的模型,用机理模型作为教师对神经网络模型进行离线训练,实现了机理模型和非机理模型的有机结合,较好地保证了预报的精度,为进一步提高板厚控制的精度和成品质量奠定了基础。
to improve the precision and stability of rolling process, specially in the epoch after install new rolls, the transformation process of roller temperature field was studied on detail, and the expansion off rolls was calculated on theory. Then a Thermal expansion prediction model based on ANN was put forward, which is integrated with theory model through training off line, The developed predicating model ensure the precision of predication, and improve the reliability of AGC, which is a marked advantage over pure theory schemes.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2006年第5期496-500,共5页
China Mechanical Engineering
基金
国家自然科学基金资助项目(60474044)
河北省自然科学基金资助项目(E2004000221)
关键词
冷连轧机
温度场
热膨胀
人工神经网络
tandem cold rolling mill
temperature field
thermal expansion
ANN