經驗模態分解法（Empirical Mode Decomposition）配合希爾伯特轉換（Hilbert transform）具有可適性（adaptive）的特點，能有效分析非穩態性及非線性時間序列行為。若將經驗模態分解法所得之內建模態函數（Intrinsic mode function）進一步推導即可求得隨時間、頻率變化的阻尼損失因子，稱之為希爾伯特阻尼頻譜（Hilbert Damping Spectrum）。 本文應用希爾伯特阻尼頻譜並配合從實測地震資料識別不含土壤效應之摟房結構系統技巧，分析大樓地震歷時記錄以作高樓損傷評估之探討。研究結果顯示，判別希爾伯特阻尼頻譜曲線之變異性，確實可指出結構損傷情形。 It is effective to analysis nonlinear and nonstationary properties of time series by the feature of adaptive from the Empirical Mode Decomposition with Hilbert Transform. The Empirical Mode Decomposition method can be further developed to estimate damping loss factor from Intrinsic Mode Functions, called Hilbert Damping Spectrum. Its damping value is dependent upon both time and frequency. In this paper, Hilbert Damping Spectrum with the skill to separate building without soil-structure interaction from buildings seismic response data provides a excellent damage assessment for structural system identification. Comparing the variability of Hilbert Damping Spectrum curves has been found to detect damage, clearly.
