Numerical analysis of earth embankments in liquefiable soil and ground improvement mitigation

Highlights

• The kinematic hardening elasto-plastic model of cyclic mobility (CM) is used to simulate the liquefaction behavior of soil embankment.

• The reinforcement effect of grouting could effectively reduce co-seismic and post-seismic deformation, even if the top sandy layer on both sides of embankment was being completely liquefied.

• It developed local failure on the embankment slope even with ground reinforcement.

• For reinforcement thickness, the mitigation effect was not obvious when it exceeded 6 m.

Abstract

The security and stability of soil structures is seriously affected by earthquake-induced liquefaction, and many earth embankments located in seismically active areas failed to meet the new seismic design code criteria, introduced following the 2011 off the Pacific coast of Tohoku Earthquake. This paper proposes an appropriate in-situ countermeasure, and the seismic performance of the liquefaction mitigation method by grouting is numerically evaluated. The elastoplastic constitutive model of cyclic mobility (CM) model can present the cyclic shearing and liquefaction properties of cohesionless sand and cohesive clay in embankments, by considering stress-induced anisotropy, over-consolidation, and soil structure. A numerical investigation, based on the water-soil coupled finite element and finite difference (FE-FD) methods is conducted to explain the dynamic behavior of embankment in fully saturated sandy deposits, during an earthquake, and long-term post-consolidation in a consistent manner. Special emphasis is placed on the growth of excess pore water pressure (EPWP), acceleration response, and ground deformation (settlement and lateral spreading) in liquefiable soil during earthquake and post-earthquake consolidation (approximately 50 years). A comparative analysis demonstrates that the implemented countermeasure is effective in reducing co-seismic and post-seismic liquefaction-induced deformation, regardless of whether the top sandy layer on both sides of the embankment is completely liquified during the earthquake loading. However, there is local failure on the embankment slope instead of deep-seated failure in the embankment-soil system when grouting the saturated sandy layer, and appropriate countermeasures should be adopted to protect the slope. Furthermore, the ground strengthening effect is not apparent when the reinforcement thickness exceeds 6 m.

Introduction

Earth embankment safety is crucial since it can endanger human life and property. The seismic performance, failure mechanism, and avoiding or minimizing liquefaction-induced deformations of embankments founded on liquefiable soils are significant research concerns regarding structural operations and earthquake countermeasures. Field investigations on seismic disasters reveal earthquake-induced liquefaction as one of the most significant causes of destruction of soil embankments. Earthquake-induced liquefaction may result in significant loss of soil strength and produce large permanent deformations and ground flow with the accumulation of EPWP. In most cases, there was extensive damage mainly due to the liquefaction of the supporting sandy soils [1,2], resulting in severe subsidence and lateral ground spreading. Typical examples of liquefaction failure include the behavior of embankments during the San Fernando earthquake of 1971, and the Hyogoken-Nanbu earthquake of 1995 [3]. Moreover, as a consequence of the 2011 off the Pacific coast of Tohoku Earthquake, many geo-structures were severely damaged due to earthquake-induced liquefaction. Thus, seismic liquefaction evaluation is necessary for embankment design and safety assessment.

Several studies have evaluated the seismic performance of embankments through laboratory tests and numerical simulations. Experiments on this subject included shaking table tests [4,5] and dynamic centrifuge experiments [[6], [7], [8]]. Consistent with the findings of field analysis, the results of the test showed that earthquake-induced liquefaction would cause destruction and collapse of the soil embankment. Although experiments require considerable manpower and material resources, it is difficult to produce universally applicable results based on a small number of experiments. In recent years, numerical simulation has become an effective tool, and several numerical investigations have been carried out on the liquefaction-related seismic behavior of earth embankments [1,[9], [10], [11], [12]]. Most of the above studies were put in the same significant category that dynamic constitutive model appeared to define only for sandy soil, and their parameters were various and difficult to determine. The selection of a constitutive model to describe the behavior of both liquefied and un-liquefied soil was a key factor in the accuracy of the numerical analysis. In this dynamic analysis, soils were modeled by the cyclic mobility model, which was originally developed to describe the features of clean sands, including its cyclic mobility and liquefaction strength, while, also describing the behavior of clayey soils. The main differences between the mechanical behavior of sand and clay were in the collapse rate of structure, and the loss rate of overconsolidation [14]. For sand, the rate of collapse of the structure was faster than the rate of loss of overconsolidation. In contrast, for clay, the rate of collapse of the soil structure is much slower than the rate of loss of overconsolidation. The characteristics of soil (sand and clay), such as stress-strain relationships, shear dilatancy, strain-hardening and softening, cyclic mobility, liquefaction behavior were difficult to be described accurately in a unified way without using this constitutive model. Meanwhile, all the parameters in the CM model could be obtained by laboratory tests. Since the focus of these studies was the earth embankment sitting on a uniform sandy foundation, the dynamic behavior of the earth embankment founded on liquefiable ground comprising a sandy layer with a low permeability layer of clay remains poorly understood. Cohesive soil (clay) had a very low permeability coefficient compared with sand; thus, more excess pore water pressure (EPWP) remained for a much longer period in the non-homogeneous soil deposits compared with uniform sandy soil deposits. The dissipation of EPWP in clayey soil layer was the dominant factor for settlement after seismic excitation, and played a key role in the extent of settlement and lateral deformations in the earth embankment. Due to the 2011 off the Pacific coast of Tohoku Earthquake, in areas such as Edo River embankment at Nishisekiyado and Eai River embankment at Fuchishiri-jyoryu [15], the most severe damage was observed in this type of soil profile, in which sandy layer existed below the embankment body bounded by the clayey layer. Moreover, there was a lack of research on the long-term consolidation characteristics of liquefied soil foundation, especially the long-term liquefaction-induced deformation after seismic loading.

Ground improvement methods, such as densification, cementation, drainage, and reinforcement, have been developed to mitigate the destruction by liquefaction. Some studies have been conducted to test the effectiveness of foundation reinforcement methods in countering liquefaction [[16], [17], [18]], among which, the anti-liquefaction measure by grouting was confirmed to be an effective approach [19]. Additional numerical studies are necessary to confirm the effects of grouting on liquefaction-related performance and liquefaction-resistant behavior for improving the stability of earth embankments in liquefiable ground.

To evaluate the liquefaction-mitigation performance using the grouting technique, a design study based strictly on numerical analysis was taken up. A kinematic hardening elastoplastic dynamic constitutive model, describing the static-dynamic properties of subsoils with different densities in a unified manner under diverse loading types (cyclic or monotonic) and drainage conditions, was applied to accurately represent the seismic performance of soils under earthquake excitations. In this numerical analysis, large liquefaction-induced settlement and lateral ground spreading that occurred during earthquake loading, and in the process of long-term post-seismic consolidation could be determined uniformly. This was different from conventional liquefaction analysis, where only the sandy layer was considered and the two different processes of seismic liquefaction and subsequent long-term consolidation were treated independently. Therefore, this study might provide a reliable evaluation of liquefaction-induced seismic performance of the earth embankment built on the sandy layer and thick clayey layer. Additionally, it numerically evaluated the mitigation effects of using grouting technology.