基于局域均值分解的轧制力响应频率提取

Signal extraction for the rolling force's response frequency based on local mean decomposition

在轧制生产过程中,轧制力信号是用于轧件厚度控制的基本信号,该信号包含有大量其他信息,这些信息主要通过频率和幅值表示,其中频率信息也反映了轧机在轧制力传感器安装点处的综合响应。采用局域均值分解(LMD)时频分析的方法,先将频率信息的实测信号进行LMD分解,再利用希尔伯特变换求取各个乘积函数(PF)分量的瞬时频率,最后通过对多组轧制力传感器实测数据的求解得到该综合响应中各频率的大小。利用这一方法,对某厂1700mm五机架冷连轧四辊轧机第一机架的多卷轧制力传感器数据的频率信息进行提取,得到其响应的基频为8.5Hz,三倍频为25.5Hz。

During the rolling process ,rolling force signal is the essential signal for the gauge control . The signal contains a lot of other information which is mainly represented by its frequency and amplitude and the frequency information reflects the global response of the mill at the installation position of the sensor .The method of local mean decomposition （LMD） time-frequency analysis is used in this research .First ,the measured signals are decomposed by LMD and then Hilbert transform was used to get the instantaneous frequency of each component of production function （PF） .At last ,the global frequency response was solved by the measured data of several sets of rolling force sensors .Using this method ,the fundamental response frequency of the rolling force from the first stand in a 1700 mm five-stand tandem strip mill has been extracted .The basic fundamental frequency is found to be 8 .5 Hz and the triple frequency ,24 .5 Hz .