台灣光子源儲存環館之微動量測與模態識別

Abstract

為使同步加速器Χ光源精準運作，故對環境振動量要求非常嚴格。本研究配合儲存環館進行不同興建階段微動量測與分析。從歷次量測反應之自身功率頻譜密度函數得知該區土層顯著振動頻率為2~4Hz；儲存環館工地現場附近有28~32Hz、57~60Hz穩定振動源。由歷次量測結果發現筏式基礎板能有效減低5Hz以上之微振量；相對於素地反應，基礎板明顯降低了4~100Hz位移均方根值，水平向可從80nm降至30nm，垂直向可由85nm降至40nm。經由ARV(Auto-Regressive Vector)方法分析量測資料識別儲存環館模態參數(內外環館共分別識別7和9個模態)，並與設計階段之有限元素模型所得比較。量測所得自然振動頻率明顯大於有限元素模型者，內外環館第一模態頻率差值分別可達45%與70%；另外只有外環館第一模態與第四模態、內環館第一模態與第三模態之振形與有限元素模型者較為一致(MAC值大於0.6)，其餘模態MAC值均在0.3以下。顯示設計之有限元素模型與興建完成之結構有明顯差異，現地結構之勁度較設計者高。

Environment induced vibrations should be controlled under very small level to precisely operate a synchrotron Χ light source. This work measures the ambient vibrations of the storage ring building in its various construction stages in National Synchrotron Radiation Research Center. The measurement shows a couple of important facts. The predominant frequency of the site is around 3 Hz, and there are external stable vibration sources with frequencies of 28~32 Hz and 57~60 Hz. The mat foundation of the building considerably reduces the vibrations with frequencies larger than 5 Hz. The root mean square values of displacement in the range of 4 to 100 Hz decrease from 80 nm to 30 nm in horizontal directions and from 85 nm to 40 nm in vertical direction. The auto-regressive vector time series model is applied to identify the modal parameters of the building from the measured ambient vibration responses. The building is divided into two parts by expansion joints, namely internal ring building and external ring building. Seven and nine modes are identified for the internal and external buildings, respectively. These identified natural frequencies are significantly larger than those obtained from the finite element model for design. The identified first fundamental frequencies of the internal ring and external ring buildings are larger than the designed ones by 45% and 70%, respectively. The identified 1st and 4th mode shapes of the external ring building and the 1st and 3rd mode shapes of the internal ring building show reasonable agreement with those obtained from the finite element model with MAC larger than 0.6. The inconsistency between the identified results and the analytical results of finite element model indicates the significant differences between the constructed building and the design one. The real stiffness of the building is larger than the designed.