Verification of the Newark Analysis of Permanent Displacements of Earth Embankments Subject to Earthquake Loads on a Geotechnical Centrifuge

摘要

The objective was verification of the Newmark (1965) sliding block analysis to predict permanent displacements of earth dams subjected to earthquake loading. The tests were conducted on a new Bumpy Road; shaking table aboard the Cambridge University Geotechnical Centrifuge. Clay embankment models of a triangular cross section with 1:2 slopes that represent prototypes of 6.9 m height were each submitted to between three and six earthquakes. The sliding block calculations consistently under estimated displacements; however, relative errors in displacements were smaller for large than small earthquakes. The yield of clay at stresses below the shear strength, dynamic magnification of base acceleration and strain softening during shaking are cited as factors contributing to unconservative predictions of displacements. Another, unexpected mode of embankment failure was observed: strain softening and pore pressure generation during one test caused a slope failure 3.5 s (two hours in prototype terms) after the model earthquake. The displacement during this slump was about 2.8 m in prototype terms - on the order of ten times the displacements which occurred during any actual earthquake. (Author)