Properties of soils using laboratory seismic methods and numerical modelling
Date of Issue2016-05-31
School of Civil and Environmental Engineering
Due to increasing complexity of the infrastructure in urban living environment, the determination of geotechnical parameters such as soil stiffnesses (shear and Young’s moduli) and damping ratio at different strains are becoming increasingly important in geotechnical engineering. However, derivation of these parameters requires additional resources and time to conduct advanced geotechnical testing. One such advanced geotechnical test is the wave propagation test which can be incorporated into various tests to determine the parameters simultaneously. It is the intention of this study to further develop the capabilities of the wave propagation test in order to realise its full potential of determining both the soil stiffness and damping ratio. The versatility of the wave propagation test will be demonstrated through its incorporation into a cyclic triaxial test to determine both the dynamic and mechanical properties of geo-materials. Despite recent popularity, the wave propagation test has yet to be standardised. Furthermore damping ratio was seldom determined using the wave propagation test. Hence, concerted effort was expended to develop proper procedures for the determination of damping ratio. This study was separated into four test series supplemented by numerical modelling. Test Series I investigates three different configurations of transducers used for wave propagation test: ultrasonic, bender/extender element and the hybrid bender element – ultrasonic system. Despite the non-invasive nature of the ultrasonic system, it was unable to provide accurate and reliable S-wave velocity, hence shear stiffness, due to poor transducer design. Development of the hybrid bender element – ultrasonic system provided better understanding of the ultrasonic system and its limitations. The bender element test system was eventually selected due to its stronger signal. Prior to subsequent tests, it was further improved with better waterproofing. Test Series II can be separated into two parts. The first part evaluates the different techniques used to determine the material damping ratio from the P- and S-wave signals. The Hilbert transform method (HTM) was shown to be the most reliable and robust method to determine damping ratio. As the HTM has never been applied in the field of geotechnical test, it is verified experimentally by comparing the results derived for Ottawa 20-30 sand with those reported in the literature. Numerical simulations were further conducted using a simplified viscoelastic material model to verify the HTM. Meanwhile, the study of the numerical model revealed the discrepancy in the analytical solution which is used to obtain Young’s modulus E from the flexure natural frequency. A correction factor χm was introduced to account for the discrepancy caused by the Timoshenko beam effect and to reconcile the difference in actual and calculated E. The study allowed better understanding of the input parameters of the viscoelastic model. The establishment of the testing framework allows the investigation of the effects of changing net confining pressures and degrees of saturation (Sr) on P- and S-wave velocities and damping ratios (Vp, Vs, ξp and ξs, respectively). The relationship of Vp with Skempton pore pressure parameter B, matric suction and Sr were examined. The limitation of using B-value to determine full saturation was highlighted and Vp was proposed as a better proxy for Sr as it was found to increase sharply when Sr is greater than about 80 - 90% to about 1500 m/s at full saturation. The experiments also showed that the effect of matric suction on Vs is not isotropic compared with isotropic effective confining stress. Trends of ξp and ξs with Sr were shown to exhibit significant variability. However, values of ξp and ξs were generally close to each another. The damping ratios were also shown to decrease with increasing confining pressure and to increase with increasing Sr. Test Series III demonstrated the robustness of the testing framework through tests on soil-rubber mixtures – an engineered material where changes in stiffness and damping characteristics can be controlled and varied significantly. Results of Vs and ξ obtained were compared with those obtained from the resonant column test reported in the literature. The rubber can also be treated as a third phase to observe the response of wave propagating through a material with different phases. In Test Series IV, undisturbed residual soil specimens were tested in a cyclic triaxial set-up equipped with bender elements, local displacement transducers (LDT) and proximity sensors. Both dynamic and mechanical properties were determined. Dynamic parameters (E and Poisson’s ratio ν) determined using the LDT was found to be inaccurate due to slippage between the rubber membrane and the soil specimens. Results of the dynamic soil properties (G/Gmax and ξ) with shear strain were shown to agree with the literature. The residual soil specimens were sheared under consolidated undrained condition and constant water content condition for the saturated and unsaturated specimens, respectively. The effective shear strength parameters (ϕ’ and c’) were observed to fall within the range shown in the literature. The effective cohesion intercept was shown to increase with depth as the deeper specimens have a higher density. The total cohesion was also shown to increase with both the matric suction and depth.