Technical Note
Behavior of EPS geofoam in true triaxial compression tests

https://doi.org/10.1016/j.geotexmem.2007.10.005Get rights and content

Abstract

This paper investigates the behavior of EPS geofoam in a true triaxial apparatus using 70 mm×70 mm×140 mm prismatic brick-shaped specimens. The specimens are subjected to different stress paths in the deviator (π) plane by means of stress-controlled loading, in which the axial stress is imposed at a rate of 75 kPa/min in the major principal direction. Stress–strain characteristics and volume change behavior have been recorded, and the yield surface has been deduced from the experimental data. The following observations have also been made for the geofoam: (a) it is an elastoplastic hardening material with plastic contractive volume change under compressive loading, (b) it softens stiffness-wise under confining stress, (c) the onset of contractive volume change corresponds quite well to the proportional limit, and (d) yielding is a slightly decreasing function of the intermediate principal stress. The study found that yielding can be represented reasonably well by a Drucker–Prager yield surface.

Introduction

Geotechnical applications of EPS geofoam abound. For instance, due to its extreme lightweight and high strength/stiffness to weight ratio, EPS geofoam is an excellent material for building road embankments, in flexible pavements and in fill constructed on top of very soft compressible soils (e.g., Refsdal, 1985; Duskov, 1991; Skuggedal and Aaboe, 1991; Frydenlund, 1991; McDonald and Brown, 1993). The lightweight property of EPS has been further exploited for optimal repair of failed slopes (Jutkofsky et al., 2000), while its high compressibility makes it a compressible buffer material of choice (Horvath, 1997; Bathurst et al., 2006; Zaman and Bathurst, 2007).

The use of EPS geofoam in geotechnical applications such as those described above can be better optimized and improved through rigorous field and laboratory testing. Moreover, our understanding of the material behavior can be enhanced when experimental investigation is carried out in tandem with computer modeling, the latter being particularly useful in filling the knowledge gaps where expensive experimentation alone cannot do it. Yielding and hardening behavior, and mechanical properties such as stiffness and deformability of EPS geofoam, have already been studied using conventional laboratory uniaxial and triaxial compression tests by previous investigators (e.g., Zou and Leo, 1998; Atmatzidis et al., 2001; Chun et al., 2004). Given that two of the three principal stresses are identical, conventional triaxial tests can only access two limiting lines of the yield surface in the general stress space, when the Lode angle equals 0° or 180°. Such conditions are rarely met in practice except at points of symmetry and/or under very particular loading conditions. Generally, true triaxial conditions prevail (near edges, unsymmetrical structures or loading, etc.). As such, only the true triaxial tests can establish the yield surface in such general stress conditions. For instance, when Wong and Leo (2006) proposed an elastoplastic hardening model based on conventional triaxial test data, they also pointed out that under triaxial conditions, the Drucker–Prager and Mohr–Coulomb yield criteria will give similar results. Only experimental results through more comprehensive true triaxial tests can best decide which yield criterion is more appropriate.

In view of the foregoing, a series of true triaxial tests with three distinct principal stresses have been performed to investigate the response of EPS geofoam under various stress paths and to deduce its yield locus. In recent years, various investigators have designed and undertaken true triaxial tests for the purpose of studying the mechanical behavior of geomaterials such as rocks (Hamison and Chang, 2000), sand (Wang and Lade, 2001) and unsaturated soil (Matsuoka et al., 2002). These tests have been useful in enabling the investigators to develop constitutive models for these materials. In the true triaxial tests performed in this paper, the prismatic brick-shaped specimen is loaded independently in three orthogonal directions corresponding to directions of the principal stresses. This makes it possible to investigate the material response under a number of stress paths without rotation of the principal stress/strain axes. Of particular interest in this study is the determination of an appropriate yield function for EPS geofoam.

Section snippets

EPS geofoam

EPS geofoam is manufactured in different grades with the density typically ranging from 10 to 30 kg/m3. The lower the density, the more compressible is the material and the lower is its yielding stress. In this paper, the material investigated is the 16 kg/m3 EPS geofoam, but the qualitative results obtained will be applicable to the geofoam material of different densities.

True triaxial test apparatus

Fig. 1a shows the true triaxial testing device used in this study. The device was originally developed by the Geotechnical

Experimental results

Plots of major deviator stress versus major principal strain ((σIσIII)−εI), volume strain versus major principal strain (εvε1), and yield points on π-plane have been obtained from the true triaxial test after corrections have been allowed for bedding errors. It is also worth noting that the response of EPS geofoam to stress loading is quite unique in the sense that the material does not have a well-defined peak failure point which is commonly seen in geomaterials. In fact, the slopes of the

Conclusions

A program of true triaxial experiments has been undertaken to investigate the mechanical behavior of EPS geofoam. The results confirm that EPS is an elastoplastic hardening material which also softens stiffness-wise under increasing confining pressure. The onset of contractive volume change corresponds quite well to the proportional limit in the stress–strain curve. The study also found that the EPS geofoam can be modeled as a Drucker–Prager material.

Acknowledgments

Financial support from the Hong Kong Polytechnic University and the Research Grants Committee of the Hong Kong Special Administrative Region Government of China for some parts of this work is gratefully acknowledged. Mr. Y.P. Leong provided assistance to all laboratory testing works, which is also gratefully acknowledged.

References (21)

There are more references available in the full text version of this article.

Cited by (0)

1

Tel.: +852 2766 6081; fax: +852 2334 6389.

2

Tel.: +33 4 72047268; fax: +33 4 72047156.

3

Tel.: +852 27666065; fax: +852 23346389.

View full text