轧辊偏心自动分层确定阈值滤波法

为计算板带轧机中的轧辊偏心信号，对轧制过程中采集的信号进行滤波是一个关键环节。本文提出一种轧辊偏心自动分层确定阈值滤波法：通过使用第二代小波变换方法，根据由轧辊转速计算得出的轧辊偏心频率来确定小波分解层数，并根据不同层中信号频率范围选取不同的阈值，从而对信号进行滤波。通过与类似工况下小波阈值滤波法比较，轧辊偏心自动分层确定阙值滤波法能更准确的反映轧制过程中的检测信号，为更精确提取轧辊偏心信号、进行有效的轧辊偏心补偿奠定基础。

Roll eccentricity filtering method of automatically decision of decomposition level and threshold value

Filtering the signals collected during the process of rolling is the key step for calculating the roll eccentricity of strip rolling mill. A roll eccentricity filtering method for auto＇decision of decomposition level and a threshold value are proposed in this paper. First, the decomposition level of frequency is determined using the second generation wavelet transform. Second, the threshold value is determined according to the rolling speed and signal frequency range of different level. Third, the signals is filtered. Comparing with the wavelet threshold filtering in the similar wording state, the proposed method and threshold value can accurately reflect collected signals during rolling, and provide the foundation for extracting roll eccentricity signal accurately and effective roll eccentricity compensation.