Quantitative Geotechnical Characterization of Seismic Strengthening Effect on Siliceous Soils

Abstract

This dissertation investigates the phenomenon of “seismic strengthening”, which can be relevant to submarine landslide studies in active continental margins. Given that seismic strengthening is complex in nature and highly variable in different types of mixed sediments in the marine realm, this dissertation focuses on two types of siliceous soils only: natural diatomaceous mud offshore Japan, and a silica sand standard from the USA for generic laboratory testing. These materials underwent multi-methodological geotechnical testing, and the results are presented in three manuscripts in collaboration with partner institutions in Norway, France and Austria. The first manuscript of this study presents a comprehensive analysis of the sedimentology and geotechnical properties of four recently acquired sediment cores (5 meters long) on the continental slopes adjacent to the Japan Trench and Nankai Trough. We observe unexpectedly high undrained shear strength and apparent overconsolidation in the Japan Trench slope cores. We propose that this is due to the presence of diatoms (~ 15% dry weight), which amplifies the strength gains via seismic strengthening due to high particle interlocking and surface roughness of the diatom frustules, especially after they are crushed and compacted by earthquakes. This is supported by comparison to samples from the Nankai Trough slope where diatoms are less abundant, and the shear strength follows the expected trend for active margin sediments. Following the observation of seismic strengthening from Japan, we developed dynamic testing procedures to simulate the effect of seismic shaking in the laboratory. In the second manuscript, we conduct undrained cyclic triaxial tests on Ottawa Sand under a mean effective stress of 100 kPa (~ 10 m depth). After the loading, we drain the excess pore pressure and measure the monotonic undrained shear strength. The cyclic loading and excess pore pressure drainage are used to simulate the effect of small to moderate seismic events on sand deposits. The result shows that the first cyclic loading and pore pressure drainage (first seismic event) can increase the sediment’s undrained shear strength by around 30% without much change in relative density (DR). We also observe that as the intensity of the shaking increases, the undrained shear strength increases. However, the undrained shear strength of a sample subjected to multiple seismic events does not show a clear trend to increase. The third manuscript focuses on the the effect of prior seismic events on the cyclic shear strength of Ottawa Sand. The results show that when cyclic loads only induce partial liquefaction (i.e. no failure) and subsequent drainage allows excess pore pressure to fully dissipate, there is a significant increase in the cyclic strength with negligible change of relative density. One seismic event (15 cycles of cyclic shear stress ratio ~ 0.147 and drainage) can increase the cyclic shear strength of an initially DR ~ 25% specimen to be stronger than an initially DR ~ 50% specimen. The results also show that as the intensity of the cyclic loading increases, the increase in the cyclic shear strength increases. However, in the case when cyclic loads lead to full liquefaction and subsequent drainage is allowed, although overall densification is observed, cyclic shear strength can either increase or decrease depending on the permanent deformation of the preceding undrained cyclic loading phase. The cyclic shear strength increases when the previous permanent (compressive) axial strain is less than 1%, and decreases when the permanent axial strain is 5%.