Similitude and Landslides Behavior for Shaking Table Test

Geotechnical engineers usually encounter large-scale problems with complex material composition. Thus, it is hard to use full-size in-situ test to observe the failure mechanism. By law of similitude and simplification, small-scale model tests were established to simulate 1-g gravitational field. This study was according to the large-scale model slope shaking table tests of the model slope conducted by Lin and Wang in September 2009 at NCREE. The failure time was estimated according to the acceleration responses, particle image velocimetry (PIV) analysis, and marker displacement measurements. The failure plane was estimated from results of the finite differential numerical analysis. The critical acceleration was estimated based on stress path during loading. By using the established numerical mode, the similarity between model and prototype was discussed.
Shallow failure sliding occurred on the slope surface, no significant displacement on the crest and toe, and no obvious difference in acceleration responses were observed. As landslide occurred, the crest and toe started moving toward down-slope. The crack occurred behind the crest, and the acceleration response became shifted. After the crack occurred and the markers at the crest started moving rapidly, the critical displacement was defined by the marker''s horizontal displacement. The failure surface was defined by maximum shear strain distribution when the numerical model''s horizontal displacement was equal to critical displacement. The location of failure surface at the top of the slope is similar to the crack measured from image. According to the law of similitude proposed by Meymand (1998), the prototype seismic slope response could be predicted from model test. Based on results of numerical simulation of prototype seismic behavior, it was found that the normalized deformation and shear strain distribution were quite similar. Therefore, the numerical model established according to the slope model test could be used to simulate in-situ seismic slope response when boundary effects and energy dissipations were insignificant.