Title :
Diagonal Recurrent Neural Network Decoupling Control on the Loopers´ Height and Tension System
Author :
Li, Bo-qun ; Zhang, Ke-jun ; Fu, Jian ; Sun, Yi-kang
Author_Institution :
Inf. Eng. Coll., Univ. of Sci. & Technol. Beijing
Abstract :
For improving further automatic gauge control accuracy and quality in hot strip mills, this paper gives appropriate mathematics model and whole transferring functions based on coupling process analysis and practical date in order to complete decoupling control of the loopers´ height and tension system. The control strategy based on DRNN is introduced to decouple the looper system. It can cope with changes in rolling conditions by controlling the rotational speed of the main motor in addition to the existing looper motor current. The simulation results have shown the effectiveness of this algorithm and after decoupled, the loopers´ control performance gets much better
Keywords :
gauges; hot rolling; mechanical variables control; neurocontrollers; recurrent neural nets; rolling mills; transfer functions; automatic gauge control; building model; coupling process analysis; diagonal recurrent neural network decoupling control; hot strip mills; looper height; looper motor current; looper tension system; mathematical model; motor rotational speed; transfer function; Automatic control; Control systems; Couplings; Educational institutions; Milling machines; Recurrent neural networks; Signal analysis; Stress; Strips; Sun; Building model; DRNN; Decoupling; Looper system;
Conference_Titel :
Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on
Conference_Location :
Dalian
Print_ISBN :
1-4244-0332-4
DOI :
10.1109/WCICA.2006.1712872