Ahmed-W. Elgamal

Observed and predicted earthquake response of La Villita Dam

Abstract La Villita earth and rockfill dam in Mexico has been subjected to five major earthquakes during the period 1975–1985. The recorded time histories of crest acceleration and associated permanent displacements imply the occurrence of seismically induced localized stick-slip slope deformations. In this paper, the upstream-downstream seismic response of La Villita dam is investigated in detail. The dam vibration characteristics are represented by a one-dimensional shear-wedge elasto-plastic numerical model. Available bedrock acceleration histories are employed as input excitation and seismic response is calculated using a step-by-step time integration procedure. Crest acceleration histories are computed and further employed to predict the observed permanent displacements. For that purpose, an elastic-perfectly plastic single-degree-of-freedom (SDOF) model is used in a Newmark-type sliding mass analysis. The predicted SDOF permanent displacements are found to be similar to their recorded counterparts. Moreover, this yielding SDOF model is found to capture the peculiar asymmetric response characteristics associated with the evolution of stick-slip displacements, and observed in the recorded crest accelerations of La Villita Dam. Finally, the employed numerical procedures are found to demonstrate the possible development of additional permanent crest displacements during future seismic events.

Interlaboratory Studies to Evaluate the Repeatability of Dynamic Centrifuge Model Tests

As part of the VELACS (VErification of Liquefaction Analysis Using Centrifuge Studies) Project, a series of dynamic centrifuge model tests were duplicated at different Universities (U.C. Davis, UC Berkeley, Caltech, Cambridge, Colorado, MIT, Princeton, and RPI) to assess the repeatability of the test results and to provide a data base for a Class A (before the event) type prediction of the test results. This paper will present the results obtained on different centrifuges, the evaluation of the repeatability of the results, and discussions on the use fulness of the results for the validation of constitutive models and numerical procedures. The experience and lessons learnt from this practice show that standardization of dynamic centrifuge modeling is very important in order to achieve repeatability of test results. The results resented here are based on the materials published in a Proceedings of the VELACS Symposium.