TESTING, MODELLING, AND APPLICATIONS OF INTERFACE BEHAVIOR IN DYNAMIC SOIL-STRUCTURE INTERACTION.

Abstract

The behavior of the interface between dry Ottawa sand and concrete has been studied using a new device developed for the cyclic testing of interfaces and joints. The stress conditions existing in the test device are investigated using stress cell measurements and a two-dimensional finite element analysis. A series of cyclic displacement-controlled interface tests are described in which the behavior of the interface is found to be a function of the applied normal stress, the amplitude of the applied displacement, the density of the sand, and the number of applied loading cycles. The (secant) shear stiffness is shown to increase with number of loading cycles, corresponding to an increase in sand density. The results of the laboratory tests are used to determine the parameters for use in a Ramberg-Osgood model to describe the interface shear stress-deformation response. This model is shown to describe the hysteresis behavior of the interface as a function of normal stress, density, and number of loading cycles. The model is used to predict the results of cyclic direct shear tests, and was found to yield satisfactory results. The interface model is implemented in a dynamic one-dimensional finite element procedure in which the soil and interface response are represented by nonlinear springs attached to the nodal points. The finite element procedure is verified by solving some simple problems for which exact or closed-form solutions are available. The response of a stress-controlled sand-concrete interface test is then predicted using the FE procedure with the nonlinear sand-concrete interface model. Although the one-dimensional idealization is a gross approximation to the three-dimensional test condition, reasonable results are obtained. A pile subjected to a harmonic axial load is then analyzed. The computed response is compared to an analytical solution and the observed response of a test pile reported by others. The effects of including interface behavior is demonstrated by solving the pile problem with and without the nonlinear interface effects. The results of this research have provided an improved understanding of the cyclic behavior of dry sand-concrete interfaces. The cyclic behavior has been represented with a simplified model for which the parameters are easily determined from laboratory tests.