請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61373
標題: | 希爾伯特黃轉換應用於風扇的破壞檢測分析 Hilbert Huang Transform applied to the fan damage detecting analysis |
作者: | Bo-Yu Yang 楊博宇 |
指導教授: | 王昭男(Chao-Nan Wang) |
關鍵字: | 希爾伯特黃轉換,經驗模態分解法,本質模態函數,風扇破壞檢測, HHT,cooling fan,static unbalance,damage detecting, |
出版年 : | 2013 |
學位: | 碩士 |
摘要: | 本文針對電腦風扇葉片缺損的振動特徵,利用希爾伯特-黃轉換 (Hilbert-Huang Transform, HHT)進行分析與檢測。HHT是一種適用於非平穩與非線性訊號的時頻分析工具,其分析流程首先將待分析訊號用經驗模態分解法(Empirical Mode Decomposition,EMD)分解出數個本質模態函數 (Intrinsic Mode Functions, IMFs)。其次再利用Hilbert轉換(Hilbert Transform, HT)求得該待測訊號之瞬時頻率、瞬時振福,進而建立兼具時-頻-能量三者的分佈圖,稱為希爾伯特時頻譜(Hilbert Spectrum)。文中利用此數據,將風扇旋轉頻率上之平均能量除以振動總能量定義為損傷程度,以此作為破壞檢測指標之一。此外並比較不同破壞程度之希爾伯特時頻譜與偏度之歐式距離當作另一個破壞檢測指標。實驗部分將風扇分成無損壞、輕微損壞及嚴重損壞三種情況,分別量測並分析其希爾伯特時頻譜,再計算其診斷指標,用以判斷其破壞程度,依據分析結果:以平均能量除以振動總能量為診斷指標其準確度為87.5%;以不同破壞程度之希爾伯特時頻譜與偏度之歐式距離為診斷指標,轉速為3100(RPM)之風扇其準確度為85.42%,轉速為2500(RPM)之風扇其準確度為58.33%,轉速為2000(RPM)之風扇其準確度為60.42%。 The current study investigates the property of vibration on the damaged cooling fans via Hilbert-Huang Transform (HHT) method. HHT is a time-frequency analysis tool commonly used to test the nonstationary and nonlinear signals. HHT consists of two procedures when applied to the analysis: (a) The utilization of Empirical Mode Decomposition (EMD) to extract Intrinsic Mode Functions (IMFs) from signals to be processed. (b) The utilization of Hilbert Transform (HT) to obtain Instantaneous frequency and Instantaneous amplitude from signals to be processed. Base on data, the degree of damage is defined as the average energy of fan rotation frequency divided by total energy of vibration. The degree of damage, thus, is used as the indicator for damage detecting. Being classified into undamaged, slightly damaged and seriously damaged, the cooling fans with different classification are compared via Hilbert spectrum to fulfill the damage detecting in the current study. According to the analysis, by considering the result of average energy of fan rotation frequency divided by total energy of vibration as the indicator for damage detecting, the accuracy is 87.5%. With the indicator for damage detecting defined as the distance between Hilbert spectrum of different classification of cooling fans and skewness, the accuracy is 85.42% when the rotation speed of cooling fan is 3100(RPM), 58.33% when the rotation speed of cooling fan is 2500(RPM), and 60.42% when the rotation speed of cooling fan is 2000(RPM). |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/61373 |
全文授權: | 有償授權 |
顯示於系所單位: | 工程科學及海洋工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-102-1.pdf 目前未授權公開取用 | 2.18 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。