神经网络预报冷轧带钢表面粗糙度  

Surface Roughness of Cold Rolled Steel Strip Forecast by BP Neural

作　　者：张维维[1] 郭晓波[1] 张宁[1] 赵成林[1] 王丽娟[1] 

机构地区：[1]鞍钢股份有限公司技术中心,辽宁鞍山114009

出　　处：《鞍钢技术》2013年第1期36-39,共4页Angang Technology

摘　　要：分析了影响鞍钢冷轧厂冷轧带钢表面粗糙度的主要因素,采用BP神经网络算法建立了预报冷轧带钢表面粗糙度的数学模型,利用鞍钢冷轧厂数据对模型离线应用表明,该数学模型的计算误差率在0~±3%范围内的比例为100%。The main factors influencing the surface roughness of cold rolled steel strip produced in Cold Rolled Mill of Angang are analyzed and the mathematical model for forecasting the surface roughness is established by BP neural network method.The offline applications of the model based on data available from the Cold Rolled Mill of Angang show that the percentage of this mathematical model with calculation error rate less than 3% is 100%.

关 键 词：冷轧带钢 表面粗糙度 神经网络 数学模型 

分 类 号：TF748[冶金工程—钢铁冶金]