Identification of Hydraulic Structural Parameters under Limited Input Earthquake Records

Abstract

For large hydraulic structures such as concrete gravity dams, it is hard to collect the input signal from the bedrock. On behalf of identifying the modal frequency and damping ratio of a concrete gravity dam, the Random Decrement Technique (RDT) with the Hilbert-Huang Transform (HHT) were combined to derive the relationship between the nonlinear nonstationary signal and the modal parameter. This was achieved by the following process: (1) the original signal was filtered using a band pass filter to reflect structural intrinsic vibration characteristics; (2) the multiple intrinsic modal functions were then obtained using Empirical Mode Decomposition (EMD); (3) RDT was used for extracting free attenuation functions from the intrinsic modal functions afterward; and (4) finally, Hilbert Transform (HT) was used to the free attenuation functions, combined with the least squares fitting method, to gain the structure modal frequency and damping ratio. The feasibility of this method was verified using three earthquake acceleration records (2007 and 2008) from the Shuikou concrete gravity dam. For analyzing the accuracy, merits and demerits of this means, our consequences were compared with those obtained using the Peak Point Picking Method (PPP) in the frequency domain, the Autoregressive model with Exogenous (ARX) in the time domain, the Natural Excitation Technique (NExT) united with Hilbert-Huang Transform in the time-frequency domain, and Finite Element Method (FEM). Findings from this study indicate that the proposed approach can be utilized for nonlinear nonstationary signals well.