BP神经网络在带钢宽展预测中的应用  

The Width Variation Predictive Control Model in Finishing Stand Based on BP Neural Networks

作　　者：刘国平[1] 符强[2] 张华[1] 

机构地区：[1]南昌大学机电工程学院,江西南昌330029 [2]南昌大学自动化工程学院,江西南昌330029

出　　处：《南昌大学学报（工科版）》2003年第3期25-28,共4页Journal of Nanchang University(Engineering & Technology)

摘　　要：传统数学模型对非线性系统进行预测往往造成较大误差,而BP神经网络能取得较好的效果.本文将BP神经网络的思想用于热连轧精轧自然宽展预测,研究并建立了基于BP神经网络的精轧自然宽展预测控制数学模型,经用现场实测数据仿真验证表明该模型大大优于传统的数学预测模型的预测精度.The BP neural network is better than the traditional mathematics models in nonlinear system (prediction.)To meet the demand of improving the predictive accuracy,the author designed the width variation controlment and prediction model based on BP neural networks,which applied the idea of neural net to practical engineering control model.With the good simulation result based upon the actual data from the finishing stand,the BP predictive model presented its better performance than those traditional mathematics models in existence.

关 键 词：BP神经网络 预测 精轧带钢宽 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]