论文部分内容阅读
在实际生产中,轧辊偏心往往会导致带材厚度的波动,降低带材的质量。离线辨识轧辊偏心的控制方法在实际生产中效果不明显甚至会其反作用。为了使轧辊偏心能够在线自适应辨识,提出了一种基于改进粒子群算法优化的RBF神经网络的在线辨识方式,建立在线训练模型,对轧辊偏心信号进行在线辨识研究并与未采用该算法的在线辨识方式进行对比。结果表明,基于改进粒子群算法优化的辨识方式速度更快、精度更高,能迅速地辨识出生产过程中的轧辊偏心信号的变化,达到了期望的结果。
In actual production, roller eccentricity often leads to fluctuations in strip thickness, reducing the quality of the strip. Off-line identification roll eccentric control method in the actual production of the effect is not obvious or even the reaction. In order to make roll eccentricity recognize online adaptively, an online RBF neural network optimization method based on improved particle swarm optimization is proposed. An on-line training model is established. The on-line identification of roll eccentricity is studied and compared with the online Comparison of identification methods. The results show that the identification method based on the improved particle swarm optimization algorithm is faster and more accurate, and can quickly identify the change of the roller eccentricity signal in the production process and achieve the desired result.