张潮1，郭强1，王京1
(1.北京科技大学高效轧制国家工程研究中心,北京 100083)
中图分类号：TG335.5文件标识码：A 摘要:在铝带冷轧过程中，板形和板厚两个指标之间存在复杂的耦合关系，对于板形通道单独进行单独控制的方法会对厚度通道造成影响，反之亦然。在某些工况下，甚至会造成整个系统的不稳定。本文在建立单机架冷轧板形板厚耦合机理模型的基础上，利用时滞神经网络结合PID控制器建立板形板厚综合解耦控制系统。根据宁夏某铝带生产现场数据进行仿真研究，仿真结果验证了本文方法的有效性和鲁棒性。 |
关键词: 铝带；板形控制；厚度控制；解耦控制；神经网络
Flatness and Thickness Decoupling Control of Aluminum Cold Rolling Mill Using Delayed Neural Network
ZHANG Chao1, GUO Qiang1,WANG Jing1
(1.NationalEngineeringResearchCenter for Advanced Rolling Technology,University of Science and Technology Beijing, Beijing 100083)
Abstract: In aluminum cold rolling process, flatness and thickness control systems interact each other intensively. Adjusting one of these two factors will bring disturbance to another, and even make the system unstable sometimes. The flatness and thickness dynamic coupling model was built, and the PID controller based on delayed neural network was used to decouple the MIMO system to weaken the coupling effects. Simulation of the field data from one factory in NingxiaProvince validates the effectiveness of the algorithm for the coupling system.
Key Words: aluminous ； flatness control ； gauge control ； decoupling control ； neural network