基于BP网络逆模型的轧辊偏心自适应逆控制

Adaptive Inverse Roll Eccentricity Control Based on BPNN Inverse Model

针对冷轧带钢生产中的轧辊偏心控制问题,将神经网络与自适应逆控制相结合,提出一种基于BP网络逆模型的自适应逆控制方法。将系统的动态特性控制和对象的噪声控制分成两个独立的过程,当动态性能达到最优时,对象扰动的影响也减到最小。将该方法应用到350 mm四辊可逆液压轧机上,仿真结果表明,系统具有良好的动态和稳态特性。

For roll eccentricity in cold rolling mill,neural network was combined with adaptive inverse control theory and an adaptive inverse method based on BP neural network was proposed.The dynamic characteristic control and the noise elimination were distributed to two independent courses.When the dynamic characteristic achieved optimal,the object disturb played down to the least.Apply this method to 350 mm four-roller reversing hydraulic mill,simulation results show that the dynamic and steady characteristics of the system are all quite well.