摘要
针对目前冷轧生产中因轧辊偏心引起带材厚度波动的问题,提出了一种提高轧辊偏心信号频谱分析精度的方法。该方法综合应用伪点剔除技术、抗混滤波技术和平滑趋势技术对原始轧辊偏心信号进行预处理,再使用改进的快速傅立叶变换算法对预处理后的信号进行分析,其频谱分辨率明显提高。现场测试表明,经该方法分析ABB压头采集到的轧辊偏心信号,其频谱分辨精度明显高于以往用快速傅立叶变换算法的分辨精度,从而可以快速准确地诊断轧辊偏心所造成的影响。
Aiming at strip thickness fluctuation presently caused by roll eccentricity in cold rolling, a method of improving precision in spectrum analysis is proposed. The method uses technique of scalping the false points, filtering and tendency smoothing, to pretreat original roll eccentricity signal, and then to analyze the pretreated signal by modified fast Fourier transform, thus the resolution rate of spectrum can be increased. The method has been on-line tested. The result showed that the resolution rate of spectrum on the roll eccentricity signal analyzed by this method is much better than that analyzed by traditional Fast Fourier Transform, and the influence of roll eccentricity is diagnosed quickly and accurately.
出处
《钢铁》
CAS
CSCD
北大核心
2005年第8期56-59,69,共5页
Iron and Steel
