基于人工神经网络的烧结终点预报系统  

Adaptive Prediction System of Sintering Burning Through Point Based on Artificial Neural Network Model

作　　者：汤程[1] 李希胜[1] 

机构地区：[1]北京科技大学,北京100083

出　　处：《微计算机信息》2007年第17期256-257,245,共3页Control & Automation

基　　金：国家自然科学基金(60472095)

摘　　要：烧结终点提前预报对于稳定终点,进而提高烧结矿强度和产量、降低能耗有重要意义。为解决烧结过程大滞后环节和烧结终点难以测量的困难,文章介绍了烧结终点的软测量方法;提出了一个新的预报参数——风箱废气温度曲线拐点;将BP神经网络应用于烧结终点的预报,实现了准确地预报烧结终点的状态和生产操作指导。　The on- line measure and prediction of burning through point （BTP） in sintering process is significant to optimize operation which aims improving the strength and yield , In order to overcome the long time delay and Sinter burning Through Point of Sinter Process, The present paper describes a soft- measurement method of BTP and proposes a new predictive parameter ,the mathematics inflexion point of waste gas temperature curve in the middle of the strand. Prediction for BTP with back propagation artificial neural network model, The proposed predict can be used to predict future value of BTP and to help operator to adjust sinter process.

关 键 词：烧结终点 预报 人工神经网络 学习算法 

分 类 号：TF046.4[冶金工程—冶金物理化学]