應用平穩小波包於結構物之微動反應以識別其動態特性

Abstract

本研究利用平穩小波包配合ARV（Auto-Regressive Vector）模式建立一系統識別方法。先以頻譜圖峰值決定欲選取之平穩小波包頻帶，再配合ARV 法，利用此系統識別方法分析結構物受到振動後量測所得之反應，能夠求得該結構物之動態特性，包括自然振動頻率、阻尼比與模態振形。

本研究先利用數值模擬資料進行系統識別，驗證本文所提程序之正確性與可行性，並對可能的影響參數進行分析。識別結果並與文獻中使用其他方法之識別結果作比較，顯現本識別方法有其優點。最後利用本識別方法分析實際振動試驗反應訊號，實際振動試驗包括八層樓鋼結構振動台試驗、國道3 號高速公路高屏溪橋微動試驗。

This study applies stationary wavelet packet and the ARV (Auto-Regressive Vector) to build a systematic method of identification. First, the frequency band of stationary wavelet packet is decided by FFT (Fast Fourier Transform,) then, by adopting ARV, this systematic identifying method is able to acquire the dynamic characteristics of structure, including natural frequencies, damping ratios, and modes by analyzing the ambient vibration responses.

This study identifies numerical simulation data to validate and verify the process of the systematic method and analyzes other possible influential parameters. To show the benefits of this systemic method of identification, the results are compared to ones by different methods from other references. This study also includes practical experiments using the systematic method: an eight-level steel structure shaking table test, and a vibration test on Kaoping River Cable-Stayed Bridge on National Freeway 3.