舞钢4200mm轧机AGC的模糊神经网络滑模控制

Fuzzy Neural Network Sliding Mode Control in 4200 mm Mills of Wugang

针对舞钢4 200mm轧机在工作过程中的高阶、非线性、强耦合等特性及客户对板材的质量要求越来越高,而现有的控制策略难以满足精确轧制要求的实际情况,提出一种将目前控制领域较为先进的模糊神经网络控制策略与滑模变结构控制策略相结合的新型控制算法,并将其应用到轧机AGC控制系统中,仿真与试验结果表明,该控制器能有效克服轧机的上述不良特性且有较好的鲁棒性与跟踪性能,能满足精确轧制的要求。

Aimming at the high-order,nonlinear and strong coupling characteristics in rolling process at the 4 200 mm mills of Wugang,higher quality requirements and the existing control strategy to meet requirements of precision rolling actual situation,the paper proposes a kind of advanced control strategy in control field,fuzzy neural network control strategies and sliding model variable structure of combining new control algorithm.The simulation and experimental results in mill AGC control system show that the controller can effectively avoid afore-mentioned bad characteristics of rolling mill and has good robustness and tracking performance,and is satisfied in rolling precision.