基于改进BP神经网络的轧机扭振智能控制  

Intelligent Control of Torsional Vibration of Rolling Mills Based on Improved BP Neural Network

作　　者：时培明[1] 李冰洋[1] 

机构地区：[1]燕山大学电气工程学院,河北秦皇岛066004

出　　处：《噪声与振动控制》2015年第5期134-138,共5页Noise and Vibration Control

基　　金：河北省自然科学基金(E2012203194);河北省高等学校创新团队领军人才培育计划项目(连铸;轧钢自动化技术及应用)(LJRC013)

摘　　要：针对轧机传动系统扭振控制问题,建立考虑负载转矩的轧机传动系统动力学模型。考虑到扭振模型比较复杂和参数不易测量的特点,提出基于神经网络的状态观测器,并对标准BP网络进行优化处理。设计基于改进BP神经网络状态观测器的智能控制系统,并利用SIMULINK对轧机实例进行仿真。结果表明设计的智能控制系统对轧机传动系统的扭振有良好的控制效果。The dynamic model of the drive system of rolling mills was established to study the torsional vibration control problem in the loading condition. Due to the complicated characteristics and the difficulty in parameters measurement of the torsional vibration model, the state observer method was proposed based on neural network, and the typical BP network was optimized. The intelligent control system based on the improved BP neural network observers was designed and the real mill examples were simulated using SIMULINK. The results show that the designed smart control system of the rolling mill drive system has a good effect for torsional vibration control.

关 键 词：振动与波 轧机传动系统 扭振 改进BP神经网络 状态观测器 

分 类 号：TM341[电气工程—电机] TH113.1[机械工程—机械设计及理论]