過往對於結構物的非破壞檢測大多採用傅立葉轉換(Fourier Transform﹐FT)作為數值分析方法,傅立葉轉換需預設基底函數展開訊號,然而該基底函數無法順應時間的變動,只適合解析線性、穩態的訊號,因此對於非線性、非穩態的訊號如地震等,所得到的結果無法反應結構物的真實情形。不同於傅立葉轉換,希爾伯特-黃轉換(Hilbert-Huang Transform﹐HHT)能夠由訊號本身分解出基底函數,具備後定基底與局部隨適基底的特性,因此能夠解析非線性、非穩態的訊號。透過希爾伯特-黃轉換能夠將訊號展開為能量在頻率域和時間域的分布情形,並可進一步計算出各時刻下的瞬時頻率和瞬時能量,求得結構物的動態特性,並了解結構物的狀態。 近幾年發展的HHT SHM 分析方法便是以希爾伯特-黃轉換作為核心,並整合時頻域放大函數(Time-Frequency domain Amplification Function﹐T.F.AF)與模態時間曲線(Modal Temporal Variation Curve﹐MTVC)兩個數值分析方法,以及定義模態參數,量化結構物受地震力影響後產生的動態特性,以評估結構物健康情形。 本研究以有限元素軟體ABAQUS 建立不同斷面厚度的鋼結構模型,找出各斷面的自然振動頻率,同時對基底處輸入不同地震加速度大小,並收集樓頂層的加速度歷時資料。由收集到的資料以HHT SHM 分析方法,將資料轉換為結構物的模態參數,來探討改變地震力和斷面,對模態參數造成的影響。;In the past, Fourier Transform (FT) was mostly used as a numerical analysis method in non-destructive detection of structures. The Fourier Transform requires a predetermined basis to expand the signal. However, this basis which is only suitable for analyzing linear and steady signal cannot adapt to changes with time. For nonlinear, nonstationary signals such as earthquakes and structure responses, the results obtained from Fourier Transform cannot reflect the real situation of the structure. Unlike Fourier Transform, Hilbert-Huang Transform (HHT) can decompose the original signal into basis with posterior and locally adaptive characteristics, which can analyze nonlinear and nonstationary signals. By Hilbert-Huang Transform, the signal can be expanded into the distribution of energy in time-frequency domain, which is convenient to obtain the dynamic characteristics of the structure and understand its condition as well. In recent years, HHT SHM analysis method is developed based on Hilbert-Huang Transform, integrating two numerical analysis methods of time-frequency domain amplification function (Time-Frequency domain Amplification Function﹐T.F.AF) and modal time variation curve (Modal Temporal Variation Curve﹐MTVC). The method defines modal parameters to quantify the dynamic characteristics of a structure affected by seismic forces and assess the status of the structure. A finite element software, ABAQUS, is introduced into this study to construct steel structure models with different section thicknesses. We then figure out natural frequencies from these models separately, entering seismic forces with different magnitudes on the base, and collect response data on the top floor. HHT SHM analysis method is applied to convert the collected data into the modal parameters of the structure. Finally, we can discuss the influences of different scale seismic forces and different thickness sections on the modal parameters.