عنوان مقاله [English]
The successful time history analysis mostly depends on the accessibility of the adequate ground acceleration records given for a specific site soil condition. But, except for a few regions in the world, in most cases, limited numbers of seismic records are available. Designers overcome this limitation by using artificial records that may be constructed by combining sine waves with different phase and amplitudes, or matching spectra with a specific design spectrum. Although, these procedures are acceptable from a mathematical overview, they are not satisfactory procedures for the designer to obtain the best or even probable acceleration records for the considered region. So, these procedures, by generating a lot of artificial records, solve the designers problem. But, using these acceleration records to evaluate the earthquake response of a structure in order to design, rehabilitate or strengthen may lead the designers or clients to make a wrong decision.
In this paper, a new innovative method, based on the wavelet transform to generate artificial acceleration records for a specific area, is presented. First, each seismic record is decomposed to its basis signals. Each basis
signal, besides the specific limited frequency domain, includes the unique physical phenomenon of the area. In other words, each basis signal inherently defines specific characteristics of the region. This is based on the number, type, and mechanical, geotechnical and seismic properties of the soil layers in the region, and also probable scenarios of the passing seismic waves. Hence, as a second step, by randomly combining these basis signals, we obtain other possible seismic records of that area. Therefore, we can generate many artificial seismic records using only limited site specific recorded ground motion. In order to explain the method, the acceleration earthquakes of the Smart1 earthquake network station of Taiwan, recorded during a 5.8 magnitude
earthquake on the Richter scale, in 1985, are used. Similar and different features of the artificial records are compared with that of the real seismic records. Finally, the performance of a 10-story steel frame against artificial
and real acceleration records is compared together. Artificial records may or may not be stronger than real accelerograms, but they, specifically, are the probable acceleration records of that particular area.