Academic Open Internet Journal

www.acadjournal.com

Volume 11, 2004

 

THE EFFECT OF DENSITY AND WATER CONTENT UPON THE DYNAMIC PROPERTIES OF RECONSTITUTED MORAINE SAMPLES IN THE SMALL STRAIN RANGE

 

Stanislav Lenart

 

Slovenian National Building and Civil Engineering Institute (ZAG)

Dimičeva 12, SI-1000 Ljubljana, Slovenia

 

E-mail: stanislav.lenart@zag.si

 

 

Abstract

 

The deformation characteristics of reconstituted moraine samples were investigated in a resonant column test at IST, Lisbon. Material from a landslide that had occurred in a seismically very active area in Slovenia was used in the test.

     It was assumed that the behaviour of material with a grain size of 4 mm or less could indicate the general behaviour of the landslide. Particular attention was paid to the behaviour of this material at very small strains, and to the effect of different water contents and densities of the material. Showing the influence of the water content and confining stress an empirical equation defining the small strain modulus in terms of the void ratio and confining stress has been proposed for different water contents.

 

Key Words: resonant column test, shear modulus, damping ratio, small strain

 

 

1       Introduction

 

In November 2000 a very severe landslide occurred in the area known as Stože (1340-1580 m a.s.l.) in the Julian Alps of western Slovenia. Its width was about 300 m, and it was 1.5 km long and up to 50 m thick. The event occurred in glacial moraine and slope debris, covering tectonically highly-fractured dolomite lying on impermeable layers of marly limestone [1]. It was a combination of the effects of two processes – the sliding of soil in the upper part and debris flow in the lower part.

     Material displaced during the landslide from an altitude of 1400 to 1600 m has been mainly deposited at an altitude of 630 m. According to investigations carried out after the event, there is probably still about 3 million m3 of potentially unstable material, representing a danger of landslide recurrence. As much as 50.9 % of the landslide material consists of moraine, silt and clay fines, with grain sizes smaller than 4 mm. The expert group [2], [3], which was established to study the reasons for the occurrence of the landslide, reached the conclusion that this part of the material could lead to the general behaviour of the landslide. Detailed geotechnical characterization of the landslide material was needed. Besides conventional laboratory tests, e.g.; the routine triaxial test and simple shear test, it was decided that the dynamic properties of the material should also be determined since, the landslide is located in a seismically very active area. In order to evaluate the equivalent shear modulus and damping ratio, resonant column tests were performed at IST in Lisbon.

     It is well-known that failure in soils usually takes place at a strain level of a few percent, and settlements of engineering concern are, in most cases, accompanied by a strain of the order of 10-3 or more. In the case of soils in motion, and this is what occurs in the case of earthquakes, inertial force is something that cannot be neglected. Even if the level of strain is infinitesimal, the inertial force could become significantly large as the time interval over which the deformation occurs becomes short. For this reason, it is necessary to know the behaviour of soil at strain level as small as 10-6.

 

2       Tested material and sample preparation

 

Representative materials from the investigated site are glacial moraine, silt and clay fines. To the author’s knowledge, no investigations into the dynamic properties of materials of this kind have yet been performed. In the tests reconstituted samples were used. They were all prepared by means of wet tamping with the objective of achieving the certain moisture and densities similar to those occurring naturally. The grain size distribution curve (Figure 1) is presented.

 

Figure 1: Gradation curve

 

 

Table 1: Sample parameters

No. of Sample

Water content W [%]

Dry density

gd [t/m3]

Void ratio

e

Saturation

Sr

1

7.00

19.00

0.49

0.40

2

9.97

19.01

0.49

0.58

3

12.87

19.02

0.49

0.75

 

 

     The tested specimens were solid cylinders, 7.0 cm in diameter and 10.0 cm in height. They were prepared with materials having a dry density 1.9 t/m3 and different water contents. For all the test samples the procedure used to perform the resonant column test was the multi-stage technique, in which different magnitudes of effective confining stress are applied to the same specimen.

 

Table 2: Some physical indices of the material used in the research

 

Wp [%]

W1 [%]

Ip [%]

Specific gravity [kN/m3]

15.6

19.9

4.3

28.3

 

3       Deformation characteristics at small strains

 

Figure 2:        Damping ratio as a function of shear modulus

 

It has been shown [4], [5], [6] that the soil starts to exhibit hysteretic and non-linear behaviour in the strain range between 10-6 and 10-3, where the secant stiffness decreases with the increasing of the strain level. Various authors have suggested several equations to connect the damping ratio and the shear modulus when both are functions of strain. A simple relationship eqn (1) was derived by Hardin and Drnevich [7]:

                                      ,                                                                         (1)

where G = the secant modulus; and G0 = the initial shear modulus.

    Park and Stewart [8] have proposed an equation for sandy soils eqn (2) and separate equation for clayey soils eqn (3):

                                                                                                         (2)

                                                                                                                  (3)

 

    Figure 2 shows the correlations between damping ratio and normalized shear modulus. In the case of the investigated material the correlation is more similar to that expected for clayey soils than that for sandy soils.

 

 

4       Resonant column tests and their results

 

4.1 The resonant column testing equipment

 

The apparatus which was used in the case of the presented tests was of the fixed-free type. It is fixed at its base and excited in torsion at the top. The top end is vibrated in torsional mode by means of an electromagnetic drive system (SEIKEN model DTC-158). The procedure used to perform the resonant column tests was the multistage technique, with effective confining stress magnitudes of 20, 50, 70, 100, 200, 400, 500 and 600 kPa being applied.

 

4.2 The influence of water content and density on the shear modulus and damping ratio in the non-linear range

 

The decreasing of normalized shear modulus shows the non-linear behavior. The strain above which this process starts is called the elastic threshold shear strain, gl. Vucetic [9] proposed taking for its value the strain at which G/G0=0.99. In the cases that this paper presents it starts at 4´10-6. Normal values of the elastic threshold shear strain in the case of sands are near to 5´10-6, and to 5´10-5 in the case of cohesive soils.

 

Figure 3:        Variation in the normalized shear modulus reduction curves with water content and effective confining stress

 

 

Figure 4: Variation in the normalized shear modulus reduction curves with density and effective confining stress

 

    The variations in G/G0 were investigated at various densities, water contents and confining pressures. The normalized shear modulus reduction curve did not depend a lot on the water content or density of the tested soil, but it was clearly affected by the effective confining stress, as shown in Figure 3 and Figure 4. The G/G0 versus log g curves of the tested material were classified for 2 different effective confining stress levels by means of normal consolidation, 3 different water content values and 3 different densities of the material. The curves move to the right as the effective confining stress increases. The effect of different water contents and densities is not significant.

     The damping ratio decreases with increasing confining pressure. The effect of different water contents is small, but it is possible to perceive (Figure 5) a decrease in the damping ratio with decreasing water content. The density of the material has no perceivable effect (Figure 6).

 

Figure 5:        Variation in the damping ratio with water content and effective confining stress

 

 

Figure 6:  Variation in the damping ratio with water content and effective confining stress

 

 

4.3 Effects of void ratio and effective confining stress upon the initial shear modulus

 

The initial shear modulus, G0 was measured under different effective confining stresses s0´ for various states of packing, represented by different void ratios, e. The general form of the equation eqn (4) was used [10], [11] to describe this relationship.

                                          ,                                                                            (4)

where A, n = non-dimensional constants; s0´ = the effective confining stress; and F(e) = a function that describes the effect of the void ratio defined [12] as:

                                         ,                                                                            (5)

where e = void ratio.

    The initial shear modulus, G0 at the end of primary consolidation has been divided by the function F(e) eqn (5) and plotted against the effective confining stress employed in the test. The data points plotted on the log-log scale are related linearly and confined in a narrow range. They can be fitted together using the least-squares regression method, and values of A and n coefficients can be obtained.

     The empirical equations for the tested material on small strain modulus depending on void ratio and effective confining stress in relation to different water contents can thus be defined as:

                        , for w = 13 %                                                           (6)

                        , for w = 10 %                                                           (7)

                         , for w = 7 %                                                            (8)

where G0 and s0´ are in kPa.

     Figure 8 and Figure 9 show the initial shear modulus normalized by the void ratio function eqn (5). It is possible to observe that the density and water content affect the initial shear modulus. The effect of different water contents is more evident, and remains similar when the effective confining stress is changed.

With increased dry density the value of the normalized initial shear modulus also increases. The normalized initial shear modulus versus dry density can be plotted as a straight line. The slope of this line appears to increase with increasing effective confining stress.

 

Figure 7:        Normalized initial shear modulus versus effective confining stress (for different water contents)

 

 

Figure 8:        Normalized initial shear modulus as a function of water content and effective confining stress

 

Figure 9:        Normalized initial shear modulus as a function of the specimen dry density and effective confining stress

 

5       Conclusions

 

The shear modulus and damping ratio of reconstituted moraine samples were investigated at small strains (10-6~10-4) using resonant column tests. The non-linear behaviour of the tested material is presented as a decreasing of the normalized shear modulus, G/G0. In the case of the reconstituted moraine samples, it starts at an elastic threshold shear strain of 4´10-6 with no significant effect of different dry densities or water contents. The initial shear modulus, G0 increases and the damping ratio, ξ decreases as the effective confining stress increases. The normalized modulus reduction curves depend on the confining pressure, but they are very little affected by density or different water content for a given tested soil. The initial shear modulus itself increases with a decrease in the water content. The density of the material has no perceivable effect upon the damping ratio. The damping ratio decreases as the water content decreases.

    An equation predicting values of G0 on reconstituted moraine samples has been suggested. It describes the initial shear modulus as a function of the void ratio and effective confining stress only. The equation was established for three different water contents. The initial shear modulus of the reconstituted moraine samples normalized by the void ratio function increases with the increasing level of the effective confining stress. This increase becomes smaller when the water content is lower. The normalized initial shear modulus versus dry density has been plotted as a straight line, whose slope increases with the increase in effective confining stress.

 

6       Acknowledgment

 

The author wishes to thank Prof. A. Gomes Correia and Prof. J. A. Santos, of the Technical University of Lisbon (IST), who kindly made their testing equipment available for the research. Their valuable advice is much appreciated. The financial support of the Ministry of Education, Science and Sport Republic of Slovenia is also gratefully acknowledged.

 

7       References

 

[1] Petkovšek, B. Geological Characteristics of the Stože Landslide. Ujma, Ministrstvo za obrambo, Uprava RS za zaščito in reševanje, No. 14-15, pg. 98-101, 2001. (in slovenian)

[2] Majes, B. and Petkovšek, A. How to reduce new possible debris flows, which jeopardize Log pod Mangartom. Proc. 23rd Slov. Civil Eng. Assembly, Bled, Lopatič, J. and Saje, F., Editors, Slovensko društvo gradbenih konstruktorjev, Ljubljana, pp. 9-20, 2001. (in slovenian)

[3] Petkovšek, A Geological-geotechnical investigations of the Stože Landslide, Ujma, Ministrstvo za obrambo, Uprava RS za zaščito in reševanje, No. 14-15, pp. 109-117, 2001. (in slovenian)

[4] Correia, A.G., Santos, J. in Barros, J.M.C., Niyama, S. An approach to predict shear modulus of soils in the range of 10-6 to 10-2 strain levels, Proc. Fourth Internat. Conf. on Recent Advances in Geotech. Earthquake Eng. and Soil Dynamics, San Diego, California, Paper No. 1.22, 2001.

[5] Lenart, S. Resonant column test and cyclic shear test on a reconstituted moraine samples, Research Report, Technical University of Lisbon, Portugal, 2002.

[6] Santos, J.A. in Correia, G.A. Shear modulus of soils under cyclic loading at small and medium strain level, Proc. 12th World Conference on Earthquake Engineering, Paper ID 0530, Auckland, New Zealand, 2000.

[7] Hardin, B.O. in Drnevich, V.P. Shear modulus and damping in soils: design equations and curves, Journal of the Soils Mechanics and Foundation Division, ASCE, Vol. 98, No. SM7, July, pp. 667-692, 1972.

[8] Park, D. in Stewart, H.E. Suggestion of Empirical Equations for Damping Ratio of Plastic and Nonplastic Soils based on the Previous Studies, Proc. Fourth Internat. Conf. on Recent Advances in Geotech. Earthquake Eng. and Soil Dynamics, San Diego, California, Paper No. 1.21, 2001.

[9] Vucetic, M. Cyclic Characterization for Seismic Regions Based on PI. Proc. of the 13th International Conference on Soil Mechanics and Foundation Engineering, Vol. 1, pp. 329-332, 1994. [10] Hardin, B.O. The nature of stree-strain behavior of soils. Proc. Earthquake Engineering and Soil Dynamics Conference, Pasadena, ASCE, V.I, pp. 3-90, 1978.

[10] Hardin, B.O. in Richart, F.E. Elastic wave velocities in granular soils, Journal of SMF Div., Proc. of ASCE, Vol. 89, pp. 33-65, 1963.

[11] Ishihara, K. Soil Behaviour in Earthquake Geotechnics. Tokyo, Oxford University Press, pp. 85-152, 1996.

 

 

Technical College - Bourgas,

All rights reserved, © March, 2000