RBF-NN多模型切换的轧机液压活套控制方法研究

Control method for hydraulic looper of rolling mill based on RBF-NN

本文对轧机液压活套系统的工作特性进行了分析。该系统为多变量非线性系统,在位置与张力变化较大时,使用传统的控制方法该系统的稳定性较差;而神经网络在控制领域的广泛应用表明,它用于多变量非线性系统较为适合。实验结果表明,当位置与张力变化较大时,采用RBF神经网络控制方式使系统仍然稳定。

This paper analyzes the characteristics of hydraulic loop system of rolling mill. The system belongs to a multivariate and non-line system, therefore, it is unsuitable to use the traditional control for its tension control. Neural network control has been widely applied in the control field. It is found that the neural network control is more appropriate to the multivariate and non-line system. Experimental results show that RBF-NN control is suitable to make the system more stable even if a fast change of position and tension is being existed.