本論文以柔度耦合子系統分析法(Receptance coupling subsystem analyses, RCSA)與操作模態分析法(Operational modal analysis, OMA)為基礎,發展兩項工具機關鍵技術;第一項是主軸-刀具系統之刀尖點自然頻率與模態估測技術;第二項是工具機防振腳座異常偵測技術。本研究將主軸、刀柄與刀具簡化成數學模型,應用RCSA將主軸、刀柄與刀具之力學模型合成主軸-刀具系統並取得刀尖點之自然頻率與模態,並使用實驗模態法(Experimental Modal Analysis, EMA)鑑別刀尖點之自然頻率與模態,將其與RCSA之結果進行比較與討論。結果顯示RCSA所鑑別刀尖點之自然頻率,皆高於EMA所鑑別之結果,雖存在誤差但仍在可接受範圍內,亦即當系統處於未加工時,應用此法可快速預測刀尖點之自然頻率,用以作為適應性轉速調變技術之初始轉速估測使用。工具機防振腳座提供機台調整水平、抑制振動傳遞與降低噪音等功能,且防振腳座主要影響著工具機六個低頻剛體模態,本研究第二部份是提出一工具機防振腳座健康診斷技術,透過OMA鑑別工具機床身此六個顯著受防振腳座影響之模態,並以模態保證指標(Modal assurance criterion, MAC)與工具機剛設置完畢確認健康正常下之六個剛體模態比對,藉由工具機前六個自然頻率與模態振形改變判斷防振腳座是否異常。此技術亦可應用於工具機在原廠組立完成後先取得振動模態(此時稱參考模態),運送至客戶端工具機設置完成後取得測試模態,參考模態與測試模態經由MAC量化其間振動模態差異,以作為檢測工具機防振腳座是否設置適當之指標;亦可用於工具機運轉一定時日後,透過馬達激振取得測試模態,將其與剛設置完畢之參考模態比較,可持續監測與追蹤防振腳座狀態。 Two important techniques of machine tool based on receptance coupling subsys-tem analyses(RCSA) and operational modal analysis(OMA) are proposed and realized. The first is estimations of the natural frequency and mode shape of the spindle-tool sys-tem of machine tools based on RCSA. The second is the health diagnosis of machine footing of machine tools using OMA. The spindle-tool subsystem of machine tools includes three main components, i.e. a spindle, a holder and a tool. In this paper, all the components are simplified and mod-eled as beam structures. It is worth of noting that the spindle receptance is determined by experimental modal analysis (EMA), whereas the handle and tool are obtained theo-retically. Then the components are integrated to be a spindle-tool system using RCSA. The natural frequency of the spindle–tool system is determined and compared with that from the EMA. Results show that the natural frequency estimated from the RCSA is 10% as compared with that from EMA.The machine footing of machine tools not only provides the machine level adjust-ment but also reduces the vibration transmission and noise. In the second part of this study, a health diagnosis of machine footing using OMA is introduced. From both nu-merical and experimental results show that the first six natural frequency and the corre-sponding mode shapes of a machine tool, i.e. the rigid body modes are good candidates in diagnose the machine footing. Moreover, the first six mode shapes are more sensitive than the natural frequency to the healthy state of machine footing. The first six mode shapes from the machine tool which has been setup properly in the beginning are designated as reference modes, whereas the mode shapes acquired from the machine tool later on are called the targeted modes. The healthy state of ma-chine footings can be assessed and monitored by quantifying the difference between the targeted and the reference mode shapes using modal assurance criteria (MAC). Results show that the healthy state of machine footing can be monitored successfully with rea-sonable accuracy. Moreover, this proposed machine footing monitoring technique can be used in setting up a machine tool in a new environment.