Load-factor Stability Analysis of Embankments on Saturated Soil Deposits

A continuum-based finite-element methodology is established for quantifying the stability of earthen embankments built on saturated soil deposits. Within the methodology the soil is treated as a fluid-solid porous medium, in which the soil skeleton’s constitutive behavior is modeled using a smooth elastoplastic cap model that features continuous coupling between deviatoric and volumetric plasticity. In the stability analysis procedure, self-weight of the embankment soils is monotonically increased at rates characteristic of the embankment construction time, until instability mechanisms develop. The transient effects of excess pore pressures and their impact on soil strength are explicitly modeled, allowing for computation of embankment safety factors against instability as a function of construction rate. Details on the proposed method are presented and discussed, including (1) how the construction rate of an embankment can be modeled; (2) how load-based safety factors can differ from resistance-based safety factors; and (3) solved example problems corresponding to a case history of an embankment failure. INTRODUCTION AND MOTIVATION The objective of this work is to develop methods for stability analysis of earthen embankments constructed on saturated soil deposits. Using classical methods and assumptions, stability analysis of such systems typically proceeds using Mohr-Coulomb soil models and various slice-type methods (Nash 1987) and by assuming that the saturated soil has a response behavior that is either fully undrained (short-term response) or fully drained (long-term response). Because the computed factors of safety against instability associated with the fully drained and fully undrained soil assumptions are generally not close in value, with undrained stabilities being significantly less than drained stabilities, the classical methods can leave considerable uncertainty directly attributable to timedependent pore-pressure diffusion effects. Methods of slope stability analysis are thus needed that take into account the rate at which the embankment is constructed together with the spatially/temporally evolving pore-pressure field it generates in the underlying soil deposit. Such stability information is often needed by engineers planning safe yet timely rates of construction for embankment systems. When analyzing geotechnical systems, both load and resistance factors, analogous to those used in structural engineering, can be used to quantify the stability of the system. Load factors of safety against instability are simply the ratio of the load magnitude that first generates instability of the system to the magnitude of the expected load, while the strength or resistance of the system is held constant. Resistance factors of safety against instability are the ratio of actual system strength (or resistance) to reduced system strength at which system instability first occurs, while holding the loading on the system fixed. Although classical slope stability methods (such as slice methods) typically produce resistance-type safety factors, both factors are valid, although not necessarily the same. In Swan and Seo (1999), where continuum/FEM models for computing both resistance and load factors of safety against instability were presented for soil slopes, it was shown that neither Postdoctoral Assoc., Dept. of Civ. and Envir. Engrg., Univ. of Iowa, Iowa City, IA 52242. E-mail: [email protected] Assoc. Prof., Dept. of Civ. and Envir. Engrg., Ctr. for Comp.-Aided Des., Univ. of Iowa, Iowa City, IA 52242 (corresponding author). E-mail: [email protected] Note. Discussion open until October 1, 2001. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on June 30, 1999; revised January 10, 2001. This paper is part of the Journal of Geotechnical and Geoenvironmental Engineering, Vol. 127, No. 5, May, 2001. qASCE, ISSN 1090-0241/01/ 0005-0436–0445/$8.00 1 $.50 per page. Paper No. 21337. AL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGIN method is consistently more or less conservative than the other. In this work, load-based stability analysis is investigated for sand embankments constructed on soft, saturated clay soil deposits. Load-based stability analysis techniques are potentially attractive, because they quite naturally permit stability analysis of embankments as a function of construction time. Geotechnical analysts who utilize load-based safety factors and load-based stability analysis should be well aware of the differences between such methods and the more traditional resistance-based factors of safety used in geotechnical engineering. In load-based stability analysis, the loading on the system is increased until it fails. In some circumstances, if unrealistic soil models are used, load-based stability analysis can yield meaningless results. For example, in the special case of very gentle slope systems (of small steepness), as the unit weight of the soil mass increases, the shear strength of the soil can increase more rapidly than the mobilized shear stresses in the soil mass. If the soil shear strength is modeled with a linear Mohr-Coulomb or linear Drucker-Prager failure envelope, the modeled shear strength continues to increase indefinitely in proportion to increased confining stresses. For cases such as these, the combination of gentle slope systems and unrealistic soil models lead to the result that no amount of increased load will generate failure of the modeled system and the predicted load-based factor of safety will be infinite. To rectify this potential problem with load-based safety factors, analysts who use load-based stability analysis must generally use more realistic soil models. It is well recognized that no soil continues to gain shear strength indefinitely with increased confining stresses. At some point, there is a limit (or saturation) to increase of shear strength from frictional effects. The load-based safety factors and stability analysis can be guaranteed to give meaningful results only when the soil strength models used feature a realistic saturation of friction effects or an ultimate limit on the shear strength of the soil. Consequently, soil models with these characteristics are used in this work. The essence of the load-based stability method to be presented involves constructing representative continuum finiteelement models of embankment systems on saturated foundation soils that have been previously consolidated or overconsolidated. Although the self-weight loading on foundation soils is held fixed, the self-weight loading on the embankment system is monotonically increased, simulating its gradual construction up to the point at which destabilizing mechanisms fully develop. Because the proposed framework computes both time-dependent consolidation and shear failure effects, it is a generalization of preceding works [i.e., Smith and Hobbs (1976)] that have utilized multiphase continuum/