摘要
板形控制系统的控制对象是一个有惯性、纯滞后、大扰动、多变量、强耦合的非线性系统,难以建立精确的数学模型。以板形机理及其控制方法的基础上进行改进,提出了一种基于支持向量机的板形预报模型,结合BP网络提供参数的智能PID实现板形闭环控制,从而实现板型控制系统智能化。针对板型控制系统的需要对神经网络进行了改进,有效提高其收敛速度、精度及有效性。生产实践表明该预报控制系统对控制精度和鲁棒性都有较大改进,有效提高了轧制前段板形的平直度和板材的成材率。
The strip control system is a nonlinear system with inertia,delay,disturb ,multi-variable and hard to establish its simulation model. Based on traditional strip mechanism and its control system,it adopts a support vector machine（ SVM ） based farecaster and closed-loop controller respectively,which realized the intellectualization,and also proposes a new train algorithm to satisfy the actual requirements of the control system so that the convergence ,precision and validity are promoted. The practices in production prove the proposed intelligent farecast and control system have effectively ameliorated the control precision and robusticity far CVC rolling mill,as a result,the flatness is increased while the rejection rate is decreased in the production procedure.
出处
《机械设计与制造》
北大核心
2008年第11期120-122,共3页
Machinery Design & Manufacture
