基于EEMD的轧辊偏心补偿控制

针对傅里叶变换、小波算法等传统信号处理方法在非线性信号的提取与重构中存在的缺陷，提出了基于聚合经验模态分解的轧辊偏心信号提取新方法。另外，针对传统自动厚度控制系统(AGC)在偏心补偿控制中的不足，设计了有偏心补偿环节的 AGC系统。新方法将轧制力信号分解为多个不同特征模态函数，从中提取表征偏心信号的特征模态函数，并用此重构偏心信号，最后将新方法重构的偏心信号投入到此系统中控制轧件厚度。仿真及实验结果表明，利用聚合经验模态分解方法重构得到的轧辊偏心模型可以很大程度减小厚度波动，补偿效果优于小波算法。

Control of roll eccentricity compensation based on EEMD

Traditional signal processing methods such as Fourier transform and wavelet algorithm showed some limits at extracting and reconstructing the nonlinear signal.A new roll eccentricity signal extraction method,which was based on the ensemble empirical mode decomposition,was presented.Furthermore,in view of the limits of the traditional automatic gauge control (AGC)system in the control of the eccentricity compensation,an AGC system with the unit of eccentricity compensation was designed.This new method divided the rolling force signal into a plurality of different intrinsic mode functions,and extracted part of intrinsic mode functions that characterized ec-centricity signal to reconstruct eccentricity signal.Finally the remodelling eccentricity signal reconstructed by the new method was input into this system to control the thickness of the rolled piece.The simulation and experiment results showed that the reconstructed eccentricity model with the ensemble empirical mode decomposition method can greatly reduce the thickness fluctuation,and the compensation effect was better than wavelet algorithm.