摘要
铝热连轧机是一个极其复杂的系统,为了精确地建模预报末机架出口处的板凸度,提出了利用延时反馈BP网络对板凸度时间序列的预报方法,来实现对铝带板凸度的精确建模和预测。通过某"1+4"铝热连轧机现场板凸度的预报试验,采集了大量试验数据。结果表明,板材相对凸度的预报精度可达到0.01%。验证了该方法的有效性和优越性。
Aluminum hot tandem rolling mill is an extremely complicated system. In order to predict the crown of aluminum strip at the finisher delivery more accurately, this article has put forward a method using the time-lapse feedback BP network to forecast the time-se- quencing of the crowns, and the reconfiguration of BP network was trained with LM algo- rithm. A large number of test data were gathered from the on-the-spot experiment of a ＂1 ＋ 4＂ aluminium hot tandem rolling mill, to simulate and verify this method. The simulation re- sults show that the predictive precision of relative crown（ % ） of the strip can reach 0.01%, indicating an available and advanced method.
作者
龚寄
骆拓
GONG Ji LUO Tuo(Mechanical Engineering College, Hunan International Economics University, Changsha 410205, China Changsha Research Institute of Mining and Metallurgy, Changsha 410012, China)
出处
《轻合金加工技术》
CAS
北大核心
2016年第11期19-22,共4页
Light Alloy Fabrication Technology
关键词
铝热连轧机
板凸度
延时反馈BP网络
LM算法
时间序列预报
hot tandem rolling mill ford aluminum strip
crown of strip
time-lapse feed- back BP network
LM algorithm
prediction of time-sequcing
