基于T-S云推理网络的板形智能控制对比研究  

Contrastive on flatness intelligent control via T-S cloud inference network

作　　者：张秀玲[1,2] 赵文保[1] 徐腾[1] 赵亮[1] 

机构地区：[1]燕山大学河北省工业计算机控制工程重点实验室,河北秦皇岛066004 [2]国家冷轧板带装备及工艺工程技术研究中心,河北秦皇岛066004

出　　处：《中南大学学报（自然科学版）》2013年第11期4461-4467,共7页Journal of Central South University:Science and Technology

基　　金：国家自然科学基金资助项目(50675186)

摘　　要：将具有处理数据不确定性的云模型和T-S模糊神经网络相结合,设计T-S云推理网络,基于此网络,建立板形识别模型和轧机板形预测模型。针对900HC可逆冷轧机,设计板形控制系统,研发一种简捷的控制器;基于900HC的实测数据先离线训练确定控制器的初始参数,再在线调整控制器的参数,调整方法使用误差反传算法,并与具有相同结构的T-S模糊控制器进行对比。研究结果表明:此系统具有有效性和较好的鲁棒性。Based on T-S fuzzy neural network and the cloud model, which is able to process data with uncertainty, T-S cloud inference network was designed. Flatness recognition model and flatness predictive model were established based on this network. For the 900HC reversible cold rolling mill, flatness control system was designed and a simple controller was developed. The initial parameters of controller were firstly determined through offiine training based on measured data, then they were adjusted online. The error back propagation algorithm was used and compared with the T-S fuzzy controller. The results show that the flatness control system has effectiveness and a better robustness.

关 键 词：云模型 T—S云推理网络 板形识别模型 轧机预测模型 板形控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]