双机架铝带连轧机张力系统的模糊神经PID控制

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。

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

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 characteristic and the control accuracy,and possesses perfect control effect upon time-varying parameter of tension control system.