Using the Hilbert-Huang Transform to Improve Structural System Identification for Vibration Signal of Bridges

• Main Authors: Te-Yu Liu, 劉德俞
• Other Authors: Wei-Ling Chiang
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/01990282420341402478

博士 === 國立中央大學 === 土木工程研究所 === 99 === Most of the bridges in Taiwan are aging and may require repair or replacement. Ambient vibration tests are utilized to measure the response of old bridges, which are difficult to equip with accelerometers. The Fourier transform method is traditionally used to gain important information about ambient vibration records. However, the development of information technology over the last 20 years has meant that there are more kinds of signal management technologies, including the Hilbert–Huang transform (HHT) methods. In this paper we investigate the characteristics of aging bridges by comparison with their structural properties. We analyze their ambient vibration, and discuss the possible influence of the frequency function. For accurate prediction of the bridge response under external excitationon, information on the dynamic characteristics of the bridge, including the natural frequency and damping ratio is needed. In this paper, we propose a novel and effective tool for the parametric identification of bridges, which uses the output response of the numerical model and ambient vibration of the bridge to identify the natural frequencies and damping ratios. Our approach is based on the empirical mode decomposition (EEMD) method, the random decrement technique (RDT), the Ibrahim Time Domain (ITD) method and frequency response function (FRF) method. Only one acceleration sensor is required. First, the noise of the bridge acceleration is measured using the EEMD method to determine the response of each mode. Second, the free vibration modal response is obtained by RDT. Finally, the natural frequencies and damping ratios for the 3DOF model and the bridge can be identified using the ITD method and the frequency response function approach for each free vibration modal response. According to the simulation results, the accuracy of the proposed method in identifying natural frequencies and damping ratios is remarkable. In the last part of this thesis, the bridge- vehicle interaction (VBI) problem is studied by numerical methods (ODE method and ANSYS numerical software). These data together with the ODE method were used to calibrate the numerical simulation model of ANSYS software. Change the value of element parameters to simulate different damages of bridge. The output displacement of the damage bridge is decomposed by EEMD method. Through the second derivative of the displacement component (IMF) with respect to the axial variable along the beam, the damage position can be identified. By simulative damage detection to the ANSYS model, it proves the feasibility of the method of damage detection based on the flexibility difference curvature.