Evaluation of Linearity of First Mode of Vibration In Resonant Column Testing

Abstract

The shear strain along the height of a specimen in resonant column testing (ASTM D4015-92) is assumed to vary linearly when evaluating its dynamic shear modulus and damping ratio. To examine this assumption, specimens of dry sand, mine tailings, and cemented sand are tested at different confinement and shear strain levels. The coupling between the cemented-sand specimen and the top and bottom platens is enhanced using three different agents: gypsum cement, Portland cement, and epoxy resin. The mode shapes are measured using a miniature accelerometer that is attached to the specimen at predefined elevations and a reference accelerometer placed at the top of the specimen (driving plate). The dynamic properties of the specimens are computed from the responses of both accelerometers. The measured mode shapes for dry sands and mine tailings are linear but the cemented-sand specimen displayed a nonlinear behavior in terms of the mode shape and change in dynamic properties with the strain level. The accuracy of the measured mode shape and dynamic properties is improved as the coupling between the specimen and the platens is enhanced. The epoxy resin produced best coupling, followed by Portland and gypsum cements, with the latter two only effective at low confinement and strain level.