عنوان مقاله [English]
In time history analysis of structures, it is essential to select an appropriate set of earthquake ground motions using the meaningful statistical methods. This matter is a challenge in earthquake engineering because the characteristics of future earthquakes are unpredictable. In this paper, earthquake ground motion records were selected based on the geological and geotechnical characteristics of the desired site. The record sets proposed here can be used for analyzing any structure with a desired fundamental period, even for sites that have potential to experience near-fault pulse-like events. The records were selected based on the response spectra simulated according to the statistical characteristics of the target Uniform Hazard Spectrum (UHS). The objective is that the response spectra of the selected records have the mean and standard deviation that are consistent with the corresponding values of the target UHS. In addition to the mean, the standard deviation was also considered in the records selection procedure to capture the ground motion aleatory uncertainty. Two appropriate sets of 7 and 30 records, selected from about 300 earthquake ground motion records, were suggested for each soil types of I, II, and III. Additionally, an efficient method referred to as the least sum of squared errors was developed for scaling the records with respect to the standard No. 2800 of Iran. Results demonstrate that using the different methods for scaling the records leads to various scaling factors for records, and as a consequence different structural responses using the method of time history analysis.