Earthquake-induced nonlinear dynamic response assessment of structures in terms of discrete wavelet transform

Nonlinear dynamic analysis (NDA) needs a large computation time and volume. In this study, an innovative method is proposed based on utilizing discrete wavelet transform to reduce the computational volume and time of NDA. For this purpose, several far-and near-field ground motion records are selected and decomposed into three levels using Wavelet Daubechies 4. In this way, several single-degree-of-freedom systems are generated and modeled using OpenSees with different beam-to-column stiffness ratios (six states). Then, the nonlinear pseudo-velocity response spectra are plotted for two ductility coefficients to show the ability of the proposed method. The results indicate that the nonlinear response spectral error of the near-field earthquakes for all the natural period intervals is less than those of the far-field earthquakes. Also, the error value for the structures, with a period of more than 0.5 sec, is less than 10%. The results also show that the error value of the wavelet nonlinear response spectrum could be neglected by increasing the ductility coefficient of the structures. Also, it is concluded that the error value is about 10% when utilizing incremental dynamic analysis, which is not signif-icant compared to the 70% reduction in the computational time. Consequently, in analyzing and designing structures with a natural period of more than 0.5 sec, the A3 level of decomposition is recommended. In contrast, the A2 level of decomposition is suggested for the structures with a natural period of less than 0.5 sec.

Automated summary

This study proposes an innovative method using discrete wavelet transform to reduce the computational volume and time of nonlinear dynamic analysis (NDA). The results show that the error of the nonlinear response spectrum for near-field earthquakes is smaller than that of far-field earthquakes. Additionally, by increasing the ductility coefficient, the error of the wavelet nonlinear response spectrum can be neglected.