GROUND MOTION SUITE SELECTION FOR THE PREDICTION OF THE SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS

Engineering community is increasingly moving towards the use of data from seismic networks in the design and evaluation of structures. While such an approach could yield an increasingly realistic design process, there are also significant risks to the use of strong motion suites for the prediction of seismic response of structures. Inherent in any process are the common mistakes about the selection of assumptions, dependent on the user's constraints as well as knowledge and experience. However, the variability in the ground motion suite introduces an even more significant uncertainty to the process affecting the prediction. The effect of the variability in the ground motions on the predicted performance of urban infrastructure is significant and therefore requires a careful selection of ground motion suite from existing seismic networks as well as a statistically relevant methodology for the refinement of the results of such analyses. For gravity dams, the concern about the selection of the motions is even greater, as the expected variability in the performance of the structure is significantly higher due to the brittle response of the material.

Spectral acceleration, the response of a single degree of freedom system to a ground motion recording, is the most commonly used ground motion index for determining the performance of a structure. However, this index is ineffective for the prediction of damage on gravity dams. In this context, the primary goal of this study is to determine the significant properties of ground motion recordings that are effective in determining the damage occurring on gravity dams during earthquakes. Common ground motion selection algorithms are compared in order to provide the practitioners with guidelines on the use of a ground motion suite for the prediction of the seismic response of gravity dams. A significant number of nonlinear finite element analyses were performed with 70 different ground motion recordings on three different dam geometries in order to identify the effect of the interaction of the frequency and the duration content of the motion and the system properties on the seismic response. A new approach for the selection of the ground motion suite for the evaluation of the performance of gravity dams is suggested based on the statistical analysis of the large pool of combined results.