退火爐風扇失效嚴重程度研究

A Study for the Failure Severity of Annealing Recirculation Fans

何秋誼(Chiou-Yi Hor) ； 吳崇勇(Chung-Yung Wu) ； 黃俊奎(Chun-Kuei Huang) ； 謝松憲(Sung-Hsien Hsieh)

《應用聲學與振動學刊》 12卷1&2期 (2020/12) Pp. 25-31

https://doi.org/10.6282/JASV.202012_12(1_2).0003

摘要

近幾年來，有許多學者將研究重心投注在殘壽（或故障嚴重程度）預估領域。解決方法主要有物理模型與資料驅動方法，其中後者因可應用在複雜設備與運轉狀態，故逐漸為人所注意。隨著電腦計算能力逐年提升，人工智慧技術在資料驅動方法也占有一席之地，且已經廣泛應用在各種工業領域。本研究將探討其中的迴歸技術，說明此種技術如何應用在估算退火爐風扇設備故障嚴重程度。藉由這種技術的導入，可以提高檢測自動化與準確度，提早讓現場維修人員在設備狀況改變前採取因應措施，避免可觀的意外停機損失。

關鍵字

設備監診 ； 殘壽預估 ； 故障嚴重程度 ； 人工智慧 ； 迴歸

並列摘要

In recent years, many scholars have focused their effort on the field of equipment remaining useful or failure severity estimation. Two main solutions to address this issue are physical models and data-driven methods. With the metric of being able to apply on complex facilities and operating conditions, the latter one is gradually received attention. Since the computing power has been tremendously raised recently, and artificial intelligence (AI) technology plays a place in data-driven methods, and has been widely adopted in various industrial fields. This research will explore one of AI technique, regression, and explain how this technique is applied to the severity of equipment failures. With the introduction of this technology, the degree of automation and accuracy can be improved, which allows field maintenance staff to take countermeasures before equipment conditions change and thus avoid considerable unexpected downtime losses.

並列關鍵字

Facility Diagnosis ； Remaining Useful Life Estimation ； Severity ； Artificial Intelligence ； Regression