Stone columns for seismic liquefaction mitigation.

Soil liquefaction has been a major source of damage during many past earthquakes. The risk of liquefaction and associated ground deformation can be reduced by various ground-improvement methods including the stone column (gravel drains) technique. Currently, there is a great need for better understanding of stone column liquefaction hazard mitigation mechanisms, particularly in silty soils. In order to address this issue, four dynamic centrifuge model experiments were conducted. Response of saturated silt strata with and without stone columns was analyzed under base dynamic excitation conditions. The underlying mechanism and effectiveness of the stone columns are discussed based on the recorded responses. The test results demonstrated that stone columns can be an effective technique in the remediation of liquefaction induced settlement of non-plastic silty deposits particularly under shallow foundations, or at depths of about 5 m in the free field.

INTRODUCTION

The risk of seismically induced liquefaction and associated ground deformation can be reduced by various ground improvement methods including the stone column technique. A comprehensive literature review by the authors on stone columns (Adalier and Elgamal, 2004) revealed that there is a great need for better understanding of stone column liquefaction hazard mitigation mechanisms, particularly when constructed in silty soils. The possible benefits of stone columns include densification of surrounding non-cohesive soil, dissipation of excess pore water pressure (EPWP) and re-distribution of earthquake-induced or pre-existing stress (due to introduction of the stiffer columns).

When dealing with non-plastic silty soils, only the third benefit can be primarily expected to mitigate liquefaction. Even with the vibro-flotation installation method, densification due to vibrations in silts can be impractical to achieve. In addition, due to very low permeability of the silt, the drainage function of the stone column is practically negligible. Shear stress re-distribution concepts have been previously proposed (Baez, 1995) as means to assess stone columns as a liquefaction countermeasure in such non-plastic silty soils. In order to address this issue, a centrifuge experimental program was conducted.

This paper briefly reports the results of this experimental study which was focused on the possible stiffening benefit, rather than improved drainage and densification due to stone column installation. Saturated silt strata of 8 m and 10 m in thickness (prototype scale) were studied. Using the centrifuge at Rensselaer Polytechnic Institute, Troy, NY-USA, two benchmark model tests were performed first to document the dynamic response characteristics of a silty stratum with and without a surface foundation surcharge. Under the same shaking conditions, the responses of these models, remediated with stone columns, was studied and compared to the benchmark models. Free-field and surface foundation surcharge situations were investigated. Settlement, acceleration, and EPWP in the soil models were recorded. Based on the recorded dynamic responses the underlying mechanism and effectiveness of the stone columns are briefly discussed herein. More detailed discussions on these tests can be found in Adalier et al. (2003).

CENTRIFUGE PHYSICAL MODEL TESTING

The main principle in centrifuge modeling is that a 1/N scale model subjected to a gravitational acceleration of Ng (g is acceleration of gravity) experiences the same stress as the prototype. Thus, stress-strain relationships at similar points in the model and prototype will be equivalent and the behavior of the model will mimic the behavior of the prototype. Consequently, with the help of scaling laws measurements in centrifuge tests can be related directly to an equivalent full-scale prototype. Unless otherwise indicated, all dimensions reported in this paper are in prototype scale, obtained from the actual model units following basic scaling relations. This means that all linear dimensions, including measured deformations, as well as the model time were multiplied by N, and the actual model shaking acceleration and frequency divided by N.

Model 1 test explored the response of a 7.8 m thick, pure silt saturated stratum (relative density, [D.sub.r] = 60%). In Model 2 (Fig. 1), a total of 45, 1.27 m diameter columns were placed (2.5 m center-to-center in a square pattern) vertically in the model laminar container, giving an area replacement ratio ([A.sub.r]) of 20%. Note that the Model 1 configuration is similar to that of Model 2 except it is without stone columns.

[FIGURE 1 OMITTED]

In the Model 3, the response of a 10 m thick saturated silt stratum ([D.sub.r] = 65%) with a rigid footing surcharge (rigid steel rectangular block applying an average vertical contact pressure of 144 kPa) was studied. This surcharge simulated approximately the vertical pressure transmitted by a 10-15 story reinforced concrete building. Model 4 (Fig. 2) investigated the response of the same system employed in the Model 3 but with the inclusion of 36, 1.6-m diameter, stone columns (2.55 m center-to-center spacing). This configuration provided an [A.sub.r] of 30% within the instrumented zone below the footing. Models 1 and 2 were tested at a 50g gravitational acceleration field, whereas Models 3 and 4 were tested at 63g. The soil container used in the Model 1 and 2 tests was a rectangular, flexible-wall laminar box. Models 3 and 4 were constructed in a rigid-wall container. Model response was measured by a large number of miniature transducers, including horizontal accelerometers, pore pressure, and displacement transducers.

[FIGURE 2 OMITTED]

A 100% silt size material "Sil-Co-Sil 120" (Walker and Stewart, 1989) was employed to construct the ground layer in all models. The material representing the stone columns was Nevada 120 sand (http://geoinfo.usc.edu/gees/velacs/). The sand columns had [D.sub.r] of about 65%, although in the field it is possible to achieve [D.sub.r] as high as 90% with crushed stone. Lower [D.sub.r] were desired in this experimental program to have a [G.sub.r] ratio of 5-6 ([G.sub.r] = [G.sub.SC]/[G.sub.S], where [G.sub.SC] is stone column shear modulus, and [G.sub.S] is silt shear modulus). This ratio is a critical parameter for stress concentration or stiffening effects due to introduction of a stone column system (Baez, 1995). The models were saturated with de-aired water under vacuum. Detailed description of model construction and instrumentation is provided by Adalier et al. (2003). In all cases, due to the very low silt permeability, the stone columns did not increase overall drainage or decrease the EPWP build-up rate during the shaking phase in any appreciable way. Therefore, for all tests any change in the behavior of the remediated ground (relative to the unremediated ground) is primarily a result of the stiffening effect of the stone columns.

TEST RESULTS

Model 1 and 2 results

Due to severe space limitation, the test results will be only briefly presented herein. Comprehensive discussions for all model tests are provided by Adalier et al. (2003).

Even though both models attained high EPWP, their dynamic behavior was noticeably different. The decay of accelerations (i.e., loss of strength) in Model 1 was significantly quicker. At corresponding locations, both the softening-induced initial amplification and the subsequent severe attenuation phases are significantly delayed in Model 2 relative to Model 1. This can be attributed to the reinforcing-stiffening effect of the installed stone columns. In general, Model 2 behaved in a stiffer manner. The EPWP traces measured in the top half of both models at corresponding locations, showed fairly close similarity (with Model 2 EPWP build-up being somewhat slower). However, in the bottom half, EPWP build-up in Model 2 was considerably slower than that observed in Model 1. Therefore, the difference in the rate of EPWP build-up between Model 1 and Model 2 soil was more pronounced at depth. Moreover, the entire silt stratum completely liquefied at the end of 12th base input cycle in the Model 1, whereas even at the end of shaking (i.e., 20th cycle), only the top-half of the silt stratum liquefied in Model 2. These EPWP records are consistent with the recorded accelerations, which exhibit much stronger response in the bottom half of Model 2 compared to Model 1. Even at the top half of the silt stratum, it took about two to three times more shaking cycles for the silt to show significant strength degradation in Model 2 compared to Model 1. Therefore, although liquefaction was not prevented by the installed stone columns (in the upper half of the silt stratum) under the strong base input motion applied during these tests, the composite ground had a considerably higher liquefaction resistance than the uniform silt.

Based on reinforcement concepts proposed by Baez (1995), a system of stone columns and silt stratum with parameters such as the one tested in Model 2 ([G.sub.r] = 6, pre-treatment liquefaction Factor of Safety-F[S.sub.pre] = 0.5), an [A.sub.r] of 20%, would have been sufficient to prevent the occurrence of liquefaction (FS [greater than or equal to] 1) in the silt profile. However, the above observations suggest that sufficient vertical stress or confining pressure might be required to "engage" the reinforcing effect of the stone columns as suggested by Baez (1995). The centrifuge tests of Models 1 and 2 indicate that such vertical effective stress might need to exceed about 45 kPa in order for the stone columns to provide significant stress redistribution and mitigate the liquefaction of the loose silt. In practice, this confinement could be obtained with the weight of the structure. Models 3 and 4 test this hypothesis.

Model 3 (model ground with surcharge) results

Models 3 and 4 were subjected to three sinusoidal shaking events. The first shaking event (Shake1) simulated a moderate level of earthquake excitation (10 cycles of 0.08g). Shake2 was stronger excitation (30 cycles of 0.18g) and more clearly revealed the important response characteristics. Shake3 was essentially similar to Shake2. Accordingly, for the sake of brevity Shake2 event will be emphasized herein.

During Shake2, asymmetry in accelerations indicative of lateral deformations (Aydingun and Adalier, 2003) was observed at a6 and a3 (both under the footing edge at 3 and 5 m depth respectively) followed by a significant attenuation phase. At a8, 3 m beneath the foundation, notable gradual attenuation of accelerations was also observed. Likewise, the footing accelerations were gradually attenuated as the foundation soils became softer due to EPWP development. Compared to Shake1, a stronger (both in magnitude and spatial extent) negative EPWP build-up tendency was observed at the central foundation zones (as the magnitude and the spatial extent of horizontal normal strains in the foundation grew with stronger base input excitation). However, away from these expansive zones, significant positive EPWP was attained (i.e., P1, P2, P3, and P6).

The footing was observed to undergo large settlements of 0.47 m (Fig. 3) during Shake2. It is noted that in every shaking event, both in Model 3 and Model 4, over 90% of the foundation settlements occurred during shaking. These large settlements were partially a result of migration of underlying foundation soil towards the free field, where the ground surface was observed to have negligible net vertical deformations (compaction settlements were largely masked by the heave).

[FIGURE 3 OMITTED]

Model 4 (model ground with 36 stone columns and with surcharge)

Based on the recorded response, it may be concluded that the stone column application in the foundation layer has led to the following effects:

i) The acceleration spikes that appeared at a1, a2, and a3 in the Model 3 were not observed in Model 4. Absence of this cyclic mobility effect indicates that shear strains were smaller (Elgamal et al., 1996; Aydingun and Adalier, 2003) than those in Model 3. Moreover, accelerations in the silt were slightly stronger than those measured in Model 3. Overall, the relatively high recorded accelerations (including those of the footing), clearly indicate that the stone columns largely preserved overall foundation stiffness.

ii) Overall increased foundation stiffness during shaking also reduced the outward migration of soils beneath the footing, in turn reducing the negative EPWP tendency that was strongly observed at P7, P5, and P4 of Model 3. In general, EPWP in Model 4 was somewhat slower, reached lower ultimate values, and dissipated faster than in Model 3.

iii) The seismic shaking was effectively transferred (actually with some amplification) directly from the base of the deposit to the footing (see a9 record) by the stiff composite ground (i.e., silt-stone column). Thus the composite soil block under the foundation sustained enough of its initial stiffness to transmit and amplify the base accelerations to the footing. The overall foundation shear strength also provided resistance to the heavy foundation penetration during shaking and much reduced the vertical settlements.

iv) And most importantly, foundation settlements were reduced by about 50% (Fig. 3).

CONCLUSIONS

Centrifuge model test results indicated an overall stiffer foundation material response during shaking in the models remediated by stone columns. Stone columns somewhat retarded the EPWP build-up (in the soil between columns), increased the foundation soil overall stiffness, and significantly reduced the surcharge-footing settlements. In the free-field situation, the stiffening effect provided by the stone columns was only primarily effective in reducing pore-pressures at depths below 5 m (45 kPa) approximately. In practice, this confinement level could be obtained with the weight of the structure. Near ground surface, the installed columns were only of marginal effect in reducing pore pressures. However, this issue does not substantially affect the important situations of remediation below shallow foundations, where the deformation mechanism is totally different, and the stone columns were found to reduce settlements by about 50%. Moreover, improved shear and bending stiffness (i.e., higher [G.sub.r]) of columnar inclusions may further restrain the lateral outflow of EPWP-softened soils, which was observed to significantly contribute to the vertical foundation settlement.

REFERENCES

Adalier, K., Elgamal, A.., Meneses, J. and Baez, I.J. (2003). "Stone columns as liquefaction countermeasure in non-plastic silty soils", J. Soil Dynamics and Earthquake Engineering, Vol. 23(7), 571-584.

Adalier, K. and Elgamal, A. (2004). "Mitigation of liquefaction and associated ground deformations by stone columns", J. Engineering Geology, Vol. 72(3-4), 275-291.

Aydingun, O. and Adalier, K. (2003). "Numerical analysis of seismically-induced liquefaction in earth embankment foundations. Part I. Benchmark model", Canadian Geotechnical J., 40(4), 753-765.

Baez, J.I. (1995). "A design model for the reduction of soil liquefaction by vibro-stone columns", Ph.D. Thesis, University of Southern California.

Elgamal, A., Zeghal, M., Taboada, V., Dobry, R. (1996). "Analysis of site liquefaction and lateral spreading using centrifuge testing records", Soils and Foundations, 36(2), 111-121.

Walker, A.J. and Stewart, H.E. (1989). "Cyclic undrained behavior of nonplastic and low plastic silts", Technical Report, NCEER-89-0035, 220pp.

K. ADALIER

Department of Civil and Environmental Engineering, Florida State University, Panama City, FL, USA

A. ELGAMAL

Department of Structural Engineering, University of California, San Diego, CA, USA