基于独立分量回归的加热炉钢温预报模型  

Billet Temperature Model of Reheating Furnace Based on Independent Component Regression

作　　者：杨英华[1] 杨劭伟[1] 刘晓志[1] 陈晓波[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁省沈阳市110004

出　　处：《系统仿真学报》2008年第10期2523-2525,共3页Journal of System Simulation

摘　　要：轧钢加热炉系统具有多变量、非线性、大滞后、交叉耦合等特性,钢坯出炉温度的预报模型一直是个难题。仍然采用统计建模的思路,利用独立分量分析方法不依赖分布假设的优点,建立了钢坯温度变量和过程变量之间的独立分量回归预测模型。基于轧钢厂实际生产数据进行了建模与验证实验,误差比较分析表明,该模型能较好地预测钢坯出炉温度,且预测误差指标优于基于改进PCA的预报模型。Reheating furnace is a multivariable nonlinear system with large inertia, net lag and crossed coupling. So the estimate of main gtlide line of its product quality--billet temperature--is all along a puzzle in the process. Independent component regression （ICR） model between billet temperature variable and process variable was established with multi-statistic projection principle and ICA method. Based on a steel works actual data and reckoning model parameter, analyzing check and error indicate that this model can forecast billet steel outlet temperature, and the forecasting error can satisfy industry aoolication accuracy demands.

关 键 词：独立分量分析(ICA) 独立分量回归(ICR) 加热炉 钢温模型 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]