Effect of Mesh Size in Dynamic FE Modeling of Resonant Column Tests

This study investigates variations in soil dynamic properties based on strain levels using a finite element model of the resonant column (RC) device. The RC test is recognized for its unique ability to shed light on how soil reacts to dynamic loading. However, several factors can influence the measurement of dynamic properties in RC, therefore, further investigation is needed. These factors include the uniformity of the applied strain field, base-fixity, strain localization, top-platen coupling, sample geometry, and uniformity of the soil. To better understand how these factors impact measured shear wave velocity and damping ratio, finite element analysis has been conducted using Abaqus/Explicit, a commercial continuum mechanics-based finite element package. The modeling employed the specific geometric configuration of the RC setup at the University of Waterloo. Initially, low-strain properties of sand (shear modulus, Poisson’s ratio, damping ratio) have been used and shear strain has been varied as an input loading parameter. Torsional loading was then applied, ranging from shear strains of 10⁻⁵ to 10⁻⁴. The mesh size has been varied to understand its effect on RC device results. The finite element model considered the free vibration of the cylindrical sand specimen following forced vibration, enabling the evaluation of dynamic properties. Modal analysis of the RC setup was conducted to confirm the dominant contribution of the first torsional mode. By comparing the damping ratios and resonant frequencies at different shear strains obtained through finite element modeling with laboratory results, we found a close match, with a difference ranging from 0.50% to 3.5%.