- 會議論文
利用軋延振動監測鋼帶尾端異常之技術
Tail Pinch Detection Using Rolling Vibrations
摘要
本熱軋軋延尾端異常(Tail pinch, TP)所造成之缺陷為熱軋廠所關注之製程異常事件。傳統上TP 的發生僅能倚靠操作人員目視識別，不僅易錯失監測造成連續軋延損失，低資料完整性亦使得後續的製程診斷難以進行，是以有必要開發一自動監測TP 發生之系統，以利產品品管及製程改善之用。本研究提出利用軋延末端暫態振動訊號監測TP 的發生之技術，透過相對小波能量(relative wavelet energy)的概念，鑑別出異常之軋延訊號。經現場實驗驗證，該法則正確率達90%，已達可自動診斷之目標。
並列摘要
Tail pinches (TP) occurred at the end of the hot rollingprocess and often lead to roll surface damage which could be printed on the subsequent coil and causesurface defects not tolerable for high quality product.During current practice, TP was only detectable by visual inspection, which often miss detects and could not provide systematic quality data for future process improvement. This study was dedicated to developing a technology that could automatically diagnose TP events.It was proposed to utilize the tail out impact vibrationswith the relative wavelet energy to detect TP.Experimental validation showed the proposed approach reached 90% of accuracy and could be applied to the field to replace labor inspection.