基于人工神经网络的CSP精轧机组轧制压力预报  

Prediction of rolling load of finishing mill group in CSP based on artificial neural network

作　　者：谭成楠[1] 程晓茹[1] 任勇[1] 熊涛[1] 

机构地区：[1]武汉科技大学材料与冶金学院,湖北武汉430081

出　　处：《武汉科技大学学报》2008年第2期143-146,共4页Journal of Wuhan University of Science and Technology

摘　　要：采用BP人工神经网络的方法,以CSP热连轧六机架精轧机组生产实测数据为基础,建立了高精度的热连轧精轧机组轧制压力预报模型。结果表明,该模型训练平均误差分别为4.63×10-4和8.35×10-4,预报平均误差分别为±2.2%和±1.6%。On the basis of measured data of finishing mill group in CSP, the high accuracy prediction model for rolling load of finishing mill group in CSP is established by means of BP artificial neural network. It is found that： the errors of the model training average 4. 63×10^-4 and 8.35×10^-4 respectively; the prediction errors of the model average ±2.2% and ±1.6% respectively.

关 键 词：轧制压力 人工神经网络 CSP热连轧 模型预报 

分 类 号：TG335[金属学及工艺—金属压力加工]