Mohammad Davoodi - Assistant Professor, International Institute of Earthquake Engineering and Seismology (IIEES)Tehran, Iran
Mohammad Kazem Jafari - Professor, (IIEES) Tehran, Iran
Masoud Amel Sakhi - PhD Student of Earthquake Engineering, (IIEES) Tehran, Iran

Seismic behavior study of dams can be performed by four different methods: observations made from dam response during earthquakes, experiments on prototype dams to determine dynamic properties such as blasting and ambient tests, experiments on reduced scale models such as shaking tables or centrifuge testing and finally, analytical studies. Earthquake records are used to evaluate dynamic response of dams, but in the lack of available earthquake records, blasting tests can be used to determine dam’s dynamic response characteristics for validating numerical models. Blasting tests have one drawback with regard to earthquake studies; they are concerned with vibrations of much smaller amplitude than those normally encountered in earthquake events. It should be said that, blasting tests can be only used to calibrate dam behavior in linear elastic ranges. It is, however, a valuable contribution, at least in the initial phase of loading, to the real non-linear numerical modeling of dam behavior under earthquake loadings. Since the amplitude and the frequency contents of blasting records vary efficiently with time, the classical signal processing techniques have limits to show exact characteristics of the signal. To overcome this limitation, time – frequency domain (TFD) analysis are presented. In this paper, a summary of processed results of blasting records that have conducted on Masjed Soleiman rockfill dam, the highest rockfill dam in Iran, is presented. Blasting test records are processed in time – frequency domain by wavelet method. Using this technique, frequency contents of blasting spectra are conducted. Classical method and obtained results are compared with each other and differences between two methods are discussed and described.