摘要
采用Elman动态递归网络方法,以生产实测数据为基础,建立了冷连轧机板形预报模型。进而将基于误差反馈和专家经验的闭环模糊控制引入板形预报中,用于修正预报输出、提高预报精度和鲁棒性。仿真结果表明,该方法有效且预报精度优于BP网络方法。预报结果的相对误差限制在±3%以内,实现了冷连轧机板形的高精度预报。
The authors established the shape model of cold tandem mill by Elman dynamic recursion network method, based on actual measured data. Further more, Considering the closed loop fuzzy control based on error feedback and expertise in shape prediction model, it was used to modify the predicted outputs and improve predicted precision and robustness. The simulation results indicate that this method is highly effective, and the predicted precision is better than that from BP network method. The predicted relative errors are limited less than ± 3%, that achieves high precision of the cold tandem mill.
出处
《中国机械工程》
EI
CAS
CSCD
北大核心
2005年第13期1142-1145,共4页
China Mechanical Engineering
基金
河北省自然科学基金资助项目(E2004000206)
