Despite of recent advances in design and construction techniques of dam engineering, there remain many points unclarified with regard to dynamic stability of dams during earthquakes, especially for a dam the foundation of which consists of disadvantageous geology such as weak rock and uncemented deposits. Among the points to be urgently clarified, earthquake characteristics at dam foundations will be most essential. This is mainly because earthquake response of a dam is largely dependent on the characteristics of input ground motions, and partly because dam-foundation interactions have a very significant effect on this kind of dynamic stability.
In 1978, Japanese National Committee on Large Dams published "Earthquake Records on Rock Foundations"(JNCOLD 1978), which Contains 31 digitized accelerograms at 9 dam sites as well as each power spectrum and acceleration response spectra. These data are very valuable not only for evaluating dynamic response of dams but also for characterizing earthquake motions at dam foundations. For earthquake response analysis of embankment dams, damping of dam vibration is usually assumed to be made up of two elements, one is internal damping which derives from non-elastic dynamic properties of dam materials, and the other is radiation damping which derives from dissipation of dam vibration energy into the underlying foundation. As for the former, various kinds of research have been conducted for the evaluation, while as for the latter, empirical values which is sometimes larger than the former have been intuitively employed in the analysis due to lack of the related knowledge. Needless to say, the radiation damping should be evaluated at least as accurate as the internal damping in order to make such dynamic analysis more reliable. This will become of great importance for large embankment dams founded on weak rock or uncemented deposits. From this point of view, earthquake characteristics at dam foundations will be studied in this paper.
Table 1 summarizes dam foundations and earthquakes cited in the "Earthquake Records on Rock Foundations", and the locations of those dams and epicenters are shown in Fig. 1. Among the accelerograms, those with index no. 1–3 and 29–31 are from the same earthquake.
(Figure in full paper)
(Table in full paper)
Maximum accelerations of these records are plotted against epicentral distances in Fig. 2, in which a number attached to the point is an index no. of an accelerogram and that in a parenthesis is the earthquake magnitude. Three broken lines show empirical relationship between maximum acceleration a and epicentral distance Δ expressed as(Tamura et al 1979)
(Equation in full paper)
Although data plotted in Fig. 2 are limited in number, it could be pointed out that the attenuation relationship in Eq.(l)appears adapable to Kisenyama. Dam and Susobana Dam, but not adaptable to Nagawado Dam. Not only Eq.(l),but also anyone of other attenuation equations proposed before could not represent these data to a satisfactory level of accuracy. Maximum acceleration on a rock foundation seems to be dependent upon local geological conditions, as well as earthquake magnitude and epicentral distance.
