橋樑受風的相關研究主要為風洞試驗，而試驗的內容包括橋樑斷面模型和全橋模型風洞試驗，利用模擬風場特性來量得橋樑縮尺後的反應，並求得風力係數和顫振導數，再經由轉換預測橋樑真實的結構反應。但我們並不完全了解這些實驗所計算出來的顫振導數，是否真實的模擬對應到全橋上。因此，本文主要探討的方向，係藉由全橋模型風洞試驗與實場量測的反應，來分析顫振導數，再與斷面模型實驗做比較，探討兩者的差異。 在系統識別方面，本文的識別流程係把振態分割觀念加入隨機遞減法(RD法)，並識別出各振態不同風速下的自然頻率與阻尼比，再代入橋體運動方程式，計算出無耦合下的顫振導數(H1*、A2*、A3*、P1*)，最後把全橋模型風洞試驗以及實場量測分析出來的顫振導數結果，與斷面模型試驗兩相比較。 由研究結果可知，斷面實驗所得到的顫振導數(H1*、A2*、A3*)，以及由近似公式算出的顫振導數(P1*)，其氣動力效應相對上比從全橋模型風洞試驗和實場量測得到的氣動力效應來得大。 Regarding the wind effects on bridges, wind tunnel tests have been most important tools. Specifically, wind tunnel tests include section model and full bridge model tests, in which the characteristics of the wind field are simulated as well. Then, the response of the bridge scale is measured and the coefficients and flutter derivatives of the wind are obtained. Nevertheless, we do not know exactly whether the flutter derivatives derived from section modal tests correspond to the real aerodynamic behavior of the whole bridge. Correspondingly, this study adopts the full-bridge model wind tunnel tests and the results of field measurements in order to analyze the flutter derivatives. In turn, the results are compared with those obtained form the section model tests and the differences between the two sets of results are discussed. As far as the system identification is concerned, this study applies the modal splitting concept to the random decrement method (RD method) in order to identify the natural frequencies and damping ratios of different modes at various wind speeds. Then, these data are substituted into the equation of bridge motion to obtain uncouple flutter derivatives (H1*、A2*、A3*、P1*). Eventually, the flutter derivatives derived from both the full-bridge model tests and the field measurements are compared with the results from the bridge section model test. Based on the research findings, the aerodynamic effects of both the flutter derivatives obtained from a section model tests experiment (H1*、A2*、A3*) and the flutter derivatives calculated from an approximate formula (P1*) are shown relatively larger than those derived from the full-bridge model tests and field measurements.
