A ground motion construction method considering spectrum characteristics as supplement for conditional mean spectrum

Abstract

In the performance-based earthquake engineering, the ground motion (GM) records are generally selected for performing nonlinear time history analysis to evaluate the performance of the structure. However, due to the limited GM database, it is difficult to select GM sequences with high intensity by amplitude-scaled method for condition mean spectrum (CMS) to reasonably evaluate the performance of the structure. To solve the problem, an efficient method for constructing GM sequences is proposed as a supplement to traditional CMS. To construct GM sequences that meets the specific response spectrum sequence, a target response spectrum (conditional spectrum) sequence with physical meaning is simulated firstly based on the CMS and Latin hypercube sampling. Then, the time domain spectral matching (TDSM) method is introduced to generate GM sequences matching the target response spectrum sequences constructed above by adding adjustment wavelet functions so that available ground acceleration time history with high intensity and long period can be obtained from the limited GM database. However, in the previous studies, the coefficient used to constrain the adjustment wavelet functions usually is taken as an empirical constant value, which makes the efficiency of the iterative calculation unable to be guaranteed by the TDSM method. The particle swarm optimization method is introduced to optimize the TDSM method to find the best coefficient used to constrain the adjustment wavelet functions to improve the calculation efficiency and accuracy. To verify the effectiveness of the proposed method, the GM sequences selected by amplitude-scaled and proposed method are input into single-degree-of-freedom and multi-degree-of-freedom systems. The result shows that the proposed method for constructing GM sequences can be used as a supplement for the traditional CMS.