以往對於結構安全診斷，大多採用傅立葉轉換(Fourier Transform, FT)作為分析方法，將地震外力激發出的結構反應訊號由時間域轉為頻率域觀察。然而，傅立葉轉換是以不隨時變的預設基底展開訊號，故只適合處理線性、穩態的訊號，無法獲得訊號的瞬時變動特性。對於地震力等非線性、非穩態的訊號，需以更完善的方法進行分析。 希爾伯特-黃轉換(Hilbert-Huang Transform, HHT)是一個有效的時頻域演算程序，具備後定基底與局部隨適基底兩大特性，因此適合分析非線性、非穩態的訊號。此轉換將訊號展開為時間域和頻率域上的能量分布情形，可進一步以瞬時頻率解讀結構動態訊號的特質，探討結構的安全狀態。 近幾年發展出的HHT SHM分析方法，即是以希爾伯特-黃轉換為核心，整合兩個數值演算程序：時頻域放大函數(T.F.AF)以及模態時間曲線(MTVC)，並定義出模態參數，以統計方法量化結構物的動態特性。 本研究使用有限元素軟體ABAQUS建立不同阻尼比之鋼結構模型，於基底施加地震外力並提取各樓層的加速度歷時資料。接著以HHT SHM方法進行分析，將加速度訊號轉為時頻譜，由時頻譜上擷取模態振動特質。最後比較不同模型的分析結果，探討阻尼比對模態參數的影響。;In the past, Fourier Transform (FT) was usually used to investigate structural health condition. It transforms signals from time domain functions into frequency domain functions. However, Fourier Transform expands the signals by using pre-determined and time-invariant bases. Therefore, it is only suitable for dealing linear and steady signals. Instantaneous properties cannot be obtained by this method. For analyzing nonlinear and unsteady signals such as earthquake waveforms, better method should be applied. Hilbert-Huang Transform (HHT) is an effective algorithm to deal with time-frequency domain signals. It possesses two characteristics, posteriori base and adaptive base. Thus, it is suitable for dealing nonlinear and unsteady signals. Hilbert-Huang Transform expands the signals into energy distribution in both time domain and frequency domain, which makes it possible to interpret the properties of structural dynamic signals by introducing the concept of instantaneous frequency and determine the structural safety as well. A recently developed analytical method called HHT SHM takes Hilbert-Huang Transform as its core, integrating other two numerical steps, time-frequency domain amplification function (T.F.AF) and modal temporal variation curve (MTVC). The method defines modal parameters which quantify the dynamic characteristics with statistical means. This research utilizes a finite element software, ABAQUS, to establish steel structure models with different damping. Apply earthquake forces on the base of the model and obtain the acceleration responses from various floors. HHT SHM method is adopted for analysis to convert acceleration signals into time-frequency spectrum, and the modal vibration characteristics can be extracted from the spectrum. Finally, compare the analysis results from different models and study the influences of damping ratio on the modal parameters.
