流體膜軸承的不穩定(如油顫)所引發的自激振動會使機台造成零件鬆脫或毀損，更嚴重的是其會引發次現象─乾顫，會使機台造成破壞。這些不穩定的特徵都藏在轉速從零上升到額定轉速的訊號之中。在本論文HHT被應用來說明不穩定的特徵頻率，在Hilbert譜上清楚的呈現不穩定。但是乾顫並不能在Hilbert譜上面被觀察到，所以發展了全Hilbert譜，可以檢測出乾顫的徵兆。經由以上的分析後，更加了解訊號中的不穩定的特徵，因此希望可以有可以抑制不穩定的方法。本論文發現軸中心線是一個不錯的控制指標，每個機械故障都會影響而改變軸中心線的位置，所以軸中心線的位置加上可接受的範圍即可以很好的解釋軸承所發生的現象，在依各現象做出適當的控制，讓機台的運轉效率更佳。 Oil whip induces self-excited vibration in fluid-handling machines, and what is worse, it can cause self-excited reverse precessional full annular rub, known as “dry whip” which is a secondary phenomenon resulting from a primary cause and may lead to a catastrophic failure of machines; that is, “coincidence of oil whip and dry whip” that occurs repeatedly with constant frequency and amplitude in small clearance cases of fluid-handling machines. The clues of these instabilities are hidden in startup vibration signals of this kind of machine. Hilbert-Huang Transform is applied for instability interpretation. The instability can be clearly indicated by using Hilbert spectrum, but the malfunction like dry whip cannot be observed by it. Therefore, in this paper, first, the Full Hilbert Spectrum (FHS) is developed to illustrate the symptoms of dry whip. Next, the position of shaft centerline within an acceptance region (AR) is introduced to predict and prevent instability at an early stage. Thus, the transition from stability to instability in rotor systems can be completely understood. This is a good notion to demonstrate how a rotor system changes from a stable to an unstable state. These findings can help engineers to predict and prevent the instability in turbo-generator.
