Seite - 1224 - in Book of Full Papers - Symposium Hydro Engineering

period of the structure can greatly affect the structure’s response [9]. In this study, a set of seventy-five pulse-like near-field ground motions collected by Baker [9] is utilized. The pulse in this set has been identified using a wavelet analysis. Sixty non-pulse-like (ordinary) near-field ground motions with no velocity pulse are also used for comparison with the pulse-like record set. Using large number of earthquake records implies an aleatory uncertainty due to record-to-record variability. All ground motions were recorded on firm soil or rock sites. The processed ground motions come from the Next Generation Attenuation project database (http://peer.berkeley.edu/nga/), and are oriented in the fault-normal direction. They are from earthquakes with large magnitudes in the range of 5 < Mw < 8. The acceleration response spectra of the two earthquake sets are shown in Fig. 1. At large periods, the pulse-like ground motions tend to cause larger elastic spectral acceleration than the non-pulse-like ground motions. (a) (b) Fig. 1. Acceleration response spectra of the selected earthquakes: (a) pulse-like, (b) non-pulse-like records 4. DAM MODEL The selected dam model is the tallest non-overflow monolith of Pine Flat gravity dam shown in Fig. 2(a). It is modeled along with its full reservoir using the finite element method; the adopted mesh is illustrated in Fig. 2(b). The water- structure dynamic interaction is taken into account employing Eulerian-Lagrangian formulation. The reservoir length is considered to be five times the dam height and transmitting boundary condition is assigned to its truncated far-end. Zero-pressure condition is assumed for the free surface of the reservoir. The foundation is assumed rigid where the selected earthquakes are uniformly applied. 1224