摘要
在炼钢生产过程中,铁水脱硫过程是发生了一系列物理和化学变化的复杂工业过程.为建立各控制量与脱硫剂使用量间的模型,本文利用减法聚类法实现对ANFIS网络的结构识别,并在此基础上对ANFIS进行参数识别,完成铁水脱硫系统的建模任务.最后,通过将该方法与普通的BP算法建模进行比较,说明ANFIS在收敛速度及建模精度方面的优越性.
At the process of steel making, the process of hot metal desulphurization is a complicated industry process in
which a series of physical and chemical changes will take place. In order to establish the model of control
variables and desulphurization reagent, the subtractive clustering is introduced to identify the structure of ANFIS
and then an adaptive neural-fuzzy inference system (ANFIS) is constructed to identify the parameters in this
paper. Finally, the advances in the convergence velocity and in the accuracy of the modeling are illustrated by
means of the comparison with the property of back-propagation network.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2004年第2期218-221,共4页
Pattern Recognition and Artificial Intelligence
关键词
自适应神经模糊推理系统
铁水脱硫
减法聚类法
反向传播网络
Adaptive Neural-Fuzzy Inference System
Hot Metal Desulphurization
Subtractive Clustering
Back-Propagation Neural Networks
