Liquefaction Potential of Silty Soils D. Erteif, MH Maher

Liquefaction potential of silty soils

D. Erteif, M.H. Maher*> "Lockwood-Singh and Associates, 1944 Cotner Avenue,

^Department of Civil and Environmental Engineering,

Rutgers University, Piscataway NJ 08855-0909, USA

Abstract

Cyclic stress controlled tests are performed to investigate the degree of liquefaction of sand containing cohesionless and low plasticity fines such as those found in tailing materials. The experimental parameters varied consists of sand void ratio, silt content, and plasticity of silt. The failure is defined as

5% and 10% peak to peak axial strain since the pore pressure does not fully developed and stopps building up when it reaches a value equal to 90-95% of the initial confining stress.

1 Introduction

Liquefaction of soils is a complex problem on which a great deal of experimental and numerical research has been done (NRC 8). Much of the research done involved liquefaction of sands. Hence, soils containing some fines are frequently encountered rather than clean sands, especially in alluvial or reclaimed deposits where soil liquefaction becomes extremely important for evaluating the stability of the ground during earthquakes. The phenomenon of liquefaction is considered to be the main cause of slope failures in saturated deposits of fine silty sands. Many natural and artificial soil deposits contain silty layers and the permeability of the soil depends on the amount of silt percent. For example, Mid-Chiba Earthquake shook the area of Tokyo Bay in 1980 with a magnitude of 6.1. Records of time change of pore pressures

164 Soil Dynamics and Earthquake Engineering

indicated that there were differences in the rate of pore pressure dissipation in different layers which contained different amount of fines (Ishihara^). The pore pressure dissipated faster in the layers with 12% fines compared to layers with higher percentage of fines.

This paper describes the results of an experimental study on the effect of fine content on pore pressure generation in sand using laboratory stress controlled undrained cyclic triaxial tests. Detailed description of the experimental procedure and the results of the tests are tabulated at (Erten % ).

2 Testing Program and Procedure

Cyclic stress controlled tests were run by a recently developed automated testing system. Loading is controlled by a closed-loop feedback scheme capable of performing stress controlled tests in the standard triaxial environment. Loading conditions are microprocessor controlled and electronic instrumentation is used for data acquisition and processing (Chan \

Li 4 ). The cyclic tests were run to a double amplitude axial strain of 15%. All the tests were performed at a loading rate of one cycle per minute to ensure equalized pore pressure readings.

The properties of the soils tested are given in Table 1 (a) and (b). The grain size distribution and hydrometer analysis for the tested soils are shown in Fig.l.

The method of undercompaction was used for specimen preparation. The control parameter was chosen as void ratio because different amounts of silt were added to the sand by weight such that the maximum and minimum void ratios of each sample was different for every placement condition. The specimens were prepared in 10 layers using the undercompaction wet tamping method in order to achieve a more uniform density. Details of undercompaction method is given by Ladd 3. The specimens were saturated by "vacuum procedure" (Rad ?) until they were fully saturated. Saturation was completed by the help of backpressure saturation. The pore pressure and cell pressure were increased in increments up to 345 kPa. After the saturation process was completed, the specimens were isotropically

Soil Dynamics and Earthquake Engineering 165

Table 1. The Index Properties Of the Soils Used

(a) Sand and Non-Plastic Silt, (b) Silty Clay

Index Properties Ot;towa Sand(C-190) Sil-Co-Sil 125 Composition Ground Silk;a Ground Silica Specific Gravity 2.65 2.65 &max 0.78 1.35 &min 0.5 0.5 Cu 1 7.5 DIG 0.50 0.003 D50 0.6 0.017 090 0.76 0.068 Liquid Limit 22 22 Plastic Limit NP NP Plasticity Index NP NP Particle Shape Rounded Roundeid Sphericity 0.85 0.85 Unified Soil Classification SP ML

Index Properties Silty Clay Specific Gravity 262 Liquid Limit 35 Plastic Limit 25 Plasticity Index 10 Optimum Moisture, % 26 Unified Soil Classification ML Maximum Density, pcf 104 Pass #4, % 100 Pass #200, % 90

SAND SILT CLAY 100

0.1 0.01 0.001 0.0001 Grain Diameter, mm Figure 1: Grain Size Distribution of the Tested Soils

166 Soil Dynamics and Earthquake Engineering consolidated. Effective consolidation stress of 98 kPa. was used for all tests.

Axial strain, volume changes, and pore pressure changes were recorded.

3 Results and Discussion

Fig.2 shows the cyclic stress ratio vs. number of cycles on loosely constituted sand specimen. As the cyclic stress ratio decreases, the number of cycles required to induce 2.5, 5.0, 10.0 % strain and initial liquefaction almost collapse on each other. This indicates that the development of initial liquefaction occurs following sudden increases of cyclic strains even at the same cycles. In medium condition, as shown in Fig.3 the trends seems more or less the same but with higher stress ratios. The initial liquefaction immediately follows 10% strain faster than in loose condition. For specimens prepared by addition of 10% non-plastic silt in both loose and medium conditions, the cyclic strength curve is rather steep. For medium condition, the curves showed the same trend as loose sand and initial liquefaction occurred after a few cycles following 10% strain. For lower stress ratios, initiation of 10% cyclic strain was immediately followed by initial liquefaction (Figs. 4 & 5). The results of cyclic tests on sands with 10% low-plasticity fines are plotted in Fig. 6 in terms of the cyclic stress ratio versus the number of cycles. It may be observed that addition of both non-plastic and low- plasticity fines accelerated the pore pressure generation (Figs. 7 and 8). This effect is more pronounced with the addition of non-plastic fines. The cyclic resistance curves showing cyclic stress ratio,

(5.0% double amplitude criterion in 20 stress cycles) versus void ratio are displayed in Fig. 9. As silt content increases, the resistance curves shift towards the lower void ratios. The cyclic resistance at a given void ratio decreases with increases in silt content. It can be said that as the amount of silt increases, the soil becomes more susceptible to liquefaction and the effect of changes in cyclic stress ratio is more pronounced. The increase in cyclic resistance for the same decrease in void ratio is not same for all the specimens. As the silt content increases up to 30%, the specimen is more effected from void ratio changes.

Soil Dynamics and Earthquake Engineering 167

0.45 Loose Condition • Initial Liquefaction Ottowa Sand 20-30 X 2.5% strain 0.4 0=98 kPa. O 5.0% strain A 10% strain 035

0.3

0.25 X29

0.2 10 "100 Number of Cycles, N

Figure 2: Cyclic Stress Ratio vs. Number of Cycles

Not Failed at the End of 1000th Cycle at S=0.2

Medium Condition • Initial Liquefaction - Ottowa Sand 20-30 X 2.5% strain 0=98 kPa. O 5.0% strain 4- 10% strain 0.35

0.3

0.25

1 10 100 Number of Cycles, N

Figure 3: Cyclic Stress Ratio vs. Number of Cycles

0.45 : 'e= 0.64-6.60 ' '; Content of fines: 10% 0.4 L Type of fines: Non-plastic J

! # 2.5% strain ! 0.35 A 5.0% strain - : o 10.0% strain ; 0.3

0.25 Cycli c Stres s Ratio , a /2 o 0.2 ^ e ^ 10 iwr 1000 Number of Cycles, N

Figure 4: Cyclic Stress Ratio vs. Number of Cycles

168 Soil Dynamics and Earthquake Engineering

0.4 I e= 0.54-0.58 '. Content offine s : 1 0% ; 0.35 1 Type of fines: Non-plastic

0.3

0.25 '- 0 -_

i c Stres s Ratio , r /2 a >> 0.2 - • 2.5% strain o - U A 5.0% strain : O 10.0% strain 0.15 10 100 Number of Cycles, N

Figure 5: Cyclic Stress Ratio vs. Number of Cycles

0.45 e=0.52-0.58 Content of fines: 10% § 0.4 Type of fines : Low-plasticity

•g 0.35 oJ g 0.3

^ 0.25

10 100 1000 Number of Cycles to Liquefaction, N

Figure 6: Cyclic Stress Ratio vs. Number of Cycles

0.4 : ' e=o! 54-0.58 ' I Content of fines :20% 0.35 1 Type of fines:Non-plasti c - < • 2.5% strain ' A 5.0% strain - o 0.3 g>- O o 10.0% strain - 04 0.25 oo

1 0 -_ 0.15 10 100 1000 Number of Cycles, N

Figure 7: Cyclic Stress Ratio vs. Number of Cycles

Soil Dynamics and Earthquake Engineering 169

e= 0.52-0.58 Content of fines:30% 0" 0.35 Type of fines:Non-plasti c

A O

•£0

0.15 10 100 Number of Cycles, N

Figure 8: Cyclic Stress Ratio vs. Number of Cycles

35 : Sand ; 1.3 : • • - • 10% Silt • 25 ; # "

3.2 • 20% silt "

po o 15 30%silt : fo r Liquefactio n i 2 0 Loa d Cycle s ) 1 ' ' • 1 • 0.45 0.5 0.55 0.6 0.65 0.7 0.75 Void Ratio, e Figure 9: Relation Between Cyclic Strength, Void Ratio and Silt content

Hence, this proved to be wrong for higher silt contents. Changes in void

ratio in specimens prepared by adding more than 60% fines have been

observed to have less effect on the results compared to specimens prepared by lower percentages of fines. It can be said that loose silty sands show similar

cyclic strengths regardless of their silt content. The liquefaction resistance of silty sands for a given number of cycles can be plotted with the values of

cyclic stress ratio causing 5% double amplitude strain, versus silt content is

shown in Fig. 10. From this figure it is possible to predict the potential for 5% double amplitude strain of a soil under earthquake loading by comparing

the value of stress ratios and the number of cycles as a function of silt content.

170 Soil Dynamics and Earthquake Engineering

< )..._ --•-•5 cycles ; 0.35 - *_^m — XO -1 30 0cycle cycles s- ~ x

0.3 (>^_ -•~ x \» : XD * : 0.25 - \ "^ _.» : \ x x ; 0.2 : \ -

n is

Figure 10: Silt percent vs. Cyclic Stress Ratio

4 Conclusions

1) The presence of fines (passing #200 sieve), has substantial effect on cyclic undrained behavior of sands. The difference in behavior between pure sand specimens and specimens prepared by adding 10% non-plastic silt is not significant. The effect is more pronounced when the silt content exceeds 10%. The liquefaction resistance of sands decreases with addition of low plasticity fines (PI-10). This effect is more pronounced for fines which have no plasticity.

2) The state of compactness of several sand and silty sand mixtures is not easy to control. It is almost impossible to prepare these mixtures to the same void ratio. Small differences in void ratio for lower percentages of fines lead to wider discrepancies in state of compactness. The cyclic strength is decreasing as the void ratio of the specimen increases. When specimens are compacted to the same void ratio, the cyclic strength of silty sands is found to be lower than that of the sands.

3) The test results also indicate that as the silt content increases (-60%), the results are less effected by void ratio discrepancies.

Soil Dynamics and Earthquake Engineering 171

ACKNOWLEDGEMENT

These tests were conducted at the Soil Dynamics Laboratory of the Department

of Civil & Environmental Engineering at Rutgers University. The support of the department is gratefully acknowledged.

References

LChan C.K., An Electropneumatic Cyclic Loading System, Geotechnical Testing Journal GTJODJ, 1981, Vol. 4, No. 4, 183-187.

2.Erten D., Effect of Fines Content on Liquefaction Potential of Sands Ph.D.

Thesis, Rutgers University, 1994 S.Ladd, R.S., Preparing Test Specimens Using Undercompaction,

Geotechnical Testing Journal, GTJODJ, 1978, Vol.1, 16-23.

4.Li X.S., An Automated Triaxial Testing System, Advanced Triaxial Testing of &,// aW #oc&, AS7M S7f 977, R.T. Donaghe, R.C. Chaney and M. L.

Silver, Eds., ASTM, Philadelphia, pp. 95-106, 1988 5. Ishihara, K., Pore Pressure Rises During Earthquakes, Proceedings of the

,y, St. Louis MI, 1988, Vol 3, pp. 1201-1204. 6. "Liquefaction of Soils During Earthquakes, National Research Council

(NRC), Committee on Earthquake Engineering, Report No. CETS-EE-001, 1985, Washington, D.C.

7.Rad S.N., Clough W. G., New Procedure for Saturating Sand Specimens", yowma/qfG^^cWcaZE^mgen^ ASCE, Vol. 110, No.9, 1984, 1205- 1217.