双机架铝带连轧机张力系统的模糊神经PID控制  

Fuzzy neural PID control of tension system in aluminum strip cold rolling mill with double-stand

作　　者：李联飞[1] 许展望[1] 连斌忠[1] 郭韡[1] 梁光正[1] 

机构地区：[1]中国重型机械研究院有限公司,陕西西安710032

出　　处：《重型机械》2012年第4期27-32,37,共7页Heavy Machinery

摘　　要：针对1 850 mm双机架铝带冷连轧过程中张力控制系统(ATC)存在参数时变、非线性等问题和传统PID参数不易整定的局限,建立了速度-张力系统数学模型,在误差绝对值积分函数(IAE)和最大灵敏度(Ms)的准则约束下,提出了通过离线极点配置训练模糊神经网络(FNN),在线FNN可以根据系统参数的变化获取恰当PID控制器参数的控制策略。Matlab仿真表明,本控制算法具有较好的动态特性和控制精度,对参数时变性的ATC系统具有良好的控制效果。Aiming at the parameter time - varying, nonlinear problems and the limitations of traditional PID parameter tuning existing in the tension control system during aluminum strip cold tandem rolling, a mathematic model of speed - tension system was built. Under the criteria constraints of minimum integrated absolute error （IAE） and maximum sensitivity （Ms） , a new control strategy is proposed, which train fuzzy neural network （FNN） by off - line dominant pole assignment, and gain appropriate PID controller parameter according to the change of system parameters. The Matlab simulation shows that the control algorithm has good dynamic charac- teristic and the control accuracy, and possesses perfect control effect upon time - varying parameter of tension control system.

关 键 词：铝带冷连轧机 张力控制系统 极点配置 模糊神经PID控制 Matlab仿真 

分 类 号：TG334.9[金属学及工艺—金属压力加工]