Dynamic response of bridge piles on soft ground using NUS Geotechnical centrifuge.

Banerjee, S. ; Paul, D.K. ; Lee, F.H. 等

Pile foundations are widely used for both inland and offshore structures, especially under adverse soil conditions. Of particular interest is the case of bridges which are structurally simple, yet highly response-sensitive to soil-structure interaction effects. The performance of these structural systems under seismic excitations has been the subject of considerable attention in recent years, particularly after the numerous failures of pilesupported bridges during recent major earthquakes. Series of dynamic centrifuge tests were performed on the model piles placed in soft kaolin beds in order to study the interference of the vibration of piles on the free-field ground motion. The experimental investigation had been carried out at Geotechnical Centrifuge Laboratory of Centre for Soft Soil Engineering of National University of Singapore under a joint venture of IITRoorkee and National University of Singapore. The investigation reveled the pile-soil interaction under the effect of super-structural load

INTRODUCTION

Pile foundations are widely used for both inland and offshore structures such as bridges, port and harbour structures, tall structures like water tanks, chimney etc. especially under adverse soil conditions. The effect of pile-foundation-soil interaction on the response of such structures under earthquake excitation is still not well understood. The performance of these structural systems under seismic excitations has been the subject of considerable attention in recent years, particularly after the numerous failures of pile-supported bridges during major earthquakes. To develop a better understanding on structure--foundationsoil interaction a series of centrifuge testing on piles in clay under earthquake excitation have been undertaken under a joint collaboration between Indian Institute of Technology, Roorkee, India and National University of Singapore.

Since the early 1980's, many model earthquake tests on the centrifuge have been conducted using stacked ring or laminar containers, which were designed to permit the soil to deform in horizontal shear mode with minimal reflection from the ends. But almost all the researchers concentrated their studies for pile on sand or liquefiable soil.

Nevertheless, few works had also been carried out on pile-clay model but most of them under static condition. In Cambridge Geotechnical Laboratory, number of static analysis had been carried out on integral abutment bridge resting on pile foundations in both clay and sand (Ellis and Springman 2001). Pile research in National University of Singapore has so far focused mainly on the static response of laterally loaded pile foundation in clay to examine the effect of excavation on piles installed behind the retaining walls (Ilyas et al.2004). At University of California, Davis model piles in clay were tested in a flexible shear beam (FSB) container at a centrifugal acceleration of 30g to evaluate a dynamic beam on a nonlinear Winkler foundation (BNWF) analysis method (Boulanger et al. 1999, Christina et al. 2001). These results provide experimental support for the use of dynamic p-y analysis methods as well as dynamic BNWF analysis methods in seismic soil-pile-structure interaction problems involving pile-group systems.

The objective of the present study is to perform dynamic centrifuge test on the pile model placed in a kaolin bed in order to study the dynamic response of pile foundation in clay.

TEST SET-UP

The tests were performed on a centrifuge at National University of Singapore at a centrifugal acceleration of 50g. Soft kaolin clay (bulk unit weight=16kN/m3, W.C=66%, LL=80%, PL=35%) bed models were prepared by pouring clay slurry into the laminar container, followed by episodes of consolidation to re-create the desired strength profile and stress history.

White kaolin powder was mixed with water in a ratio of 1: 1 and the slurry was poured in the rubber packed laminar box in layers and the transducers were placed at the prescribed positions. The laminar box has an inner dimension of 524 mm long, 300 mm wide and 420 mm deep.

The base plate of the box is connected to shaking apparatus consisting of actuator, accumulator and motor pump assembly. Total four accelerometers were placed as shown (Figure 1). Before pouring the slurry to the box a perforated pipe underlying a sand bed of 10 mm thickness had been fixed at the bottom to provide the drainage path and above that a thin layer of geotextile had been placed.

[FIGURE 1 OMITTED]

1g consolidation was carried out on all clay bed models to pre-compress the latter so as to reduce the time needed for centrifuge consolidation. The models were connected to the vacuum chamber by perforated pipe in order to get linear effective stress envelope.

After two weeks of 1g consolidation the laminar box was mounted on the Centrifuge along with all the equipments for the shaking. Two tests were conducted; one with steel pile with no load on top and other with the pile with 1kg load on top. The 18mm-diameter and 25 cm-long model piles actually simulate prototype piles of 900 mm-diameter and 12.50 m depth under 50g model 'gravity'. Now models were consolidated at 50g for 10hrs and then three earthquake time histories were fired to the shaking table.

DATA PROCESSING

The output signals from the transducers were stored in .xls format using DAISYLAB version 3.0. The acceleration datasets were detrended using MS EXCEL to remove any dc offset and low frequency drift while retaining the cyclic components. This makes it compatible with an in-house programme which is used to calculate the response spectrum. The programme requires two input files: i) The wave file contained the detrended acceleration datasets. The maximum acceleration points are limited to 5000 ii) The period file consisted of period points at which response spectrums will be calculated and the structural damping. The maximum numbers of period points are limited to 150 and damping was assumed as 5%.

RESULTS

Predominant time periods obtained from response spectra are given in the Table 1 and Table 2.

The time period is found to be increasing towards top layers for single pile with no load (Figure 2, Table 1).But at the pile top time period is again reduced to that of at the base. On the contrary, for pile with load on top the time periods at the mid depth and surface are found to be same as at the pile top. Although, time period at base is almost identical with the former case (Figure 2).

[FIGURE 2 OMITTED]

CONCLUSION

The dynamic centrifuge modeling is, thus, found to be a handful tool for pile dynamics problem. Observations obtained are quite justified with respect to the basic principles of soil dynamics. The single pile with no load actually behaves as a single reinforcing bar in soil attenuating vibration at the pile top. For pile with load, due to strong coupling between soil and pile, whole pile-soil system was behaving as a unit showing unique predominant time period. But time period at base remains almost same in two tests because the accelerometer placed at base is far below from the bottom of pile tip. Hence soil pile interaction can not come into action in either of the tests.

ACKNOWLEDGEMENTS

The authors are grateful to IIT, Roorkee, India and NUS, Singapore for providing all necessary financial and academic support without which this study would have been a distinct dream.

REFERENCES

Brandenberg, Scott J., Singh, P., Boulanger, Ross W., Kutter Bruce L. (2001). "Behavior of piles in laterally spreading ground during earthquakes", CALTRAN Workshop, U.C. Davies.

Boulanger, Ross W, Christina, J. Curras, Kutter Bruce L., Wilson, D. (1999). "Seismic soil-pile-structure interaction experiments and analyses", Journal of Geotechnical and Geoenvironmental Engg. ASCE, Vol-125, No. 9. l-127, No. 7., Vol-130, No. 3.

Christina, J. Curras, Boulanger, Ross W., Kutter Bruce L., Wilson, D. (2001). "Dynamic experiments and analyses of a pile-group-supported structure", Journal of Geotechnical and Geoenvironmental Engg. ASCE, Vo

Ellis, E.A. & Springman, S.M. (2001). "Modeling of soil-structure interaction for a piled bridge abutment in plane strain FEM analysis", Computers and Geotechnics, Vol-28, pp.-79-98.

Ilyas, T., Leung, C.F., Chow, Y.K. and Budi, S.S. (2004). "Centrifuge model study of laterally loaded pile groups in clay", Jn. of Geotechnical and Geoenvironmental Engg. ASCE

S. BANERJEE

Department of Civil Engineering, National University of Singapore, Block E1A, #07-03, No. 1 Engineering Drive 2, Singapore-117576;

D. K. PAUL

Department of Earthquake Engg. IIT-Roorkee, India--247667;

F. H. LEE

Department of Civil Engineering, NUS, Singapore-117576.

Table 1. Predominant time period for pile model with load

 Predominant Time
 Period (sec)

 Mid Pile
Excitations Earthquakes Base depth Surface top

First excitation Small (0.026g) 0.26 1.75 4 0.26
 Medium (0.05g) 0.26 2 4.1 0.26
 Large (0.11g) 0.5 2.6 4.1 0.5

Second excitation Small (0.026g) 0.3 1.75 4.2 0.15
 Medium (0.05g) 0.26 2.8 4 0.26
 Large (0.11g) 0.5 3 4.1 0.5

Third excitation Small (0.026g) 0.26 2 4.2 0.26
 Medium (0.05g) 0.3 3 4.1 0.3
 Large (0.11g) 0.26 2.8 4.1 0.26

Table 2. Predominant time period for pile model with load

 Predominant Time
 Period (sec)

 Mid Pile
Excitations Earthquakes Base depth Surface top

First excitation Small (0.026g) 0.28 1 1 1
 Medium (0.05g) 0.26 1.75 1.75 1.75
 Large (0.11g) 0.13 1.75 1.75 1.75

Second excitation Small (0.026g) 0.26 1.75 1.75 1.75
 Medium (0.05g) 0.3 1.75 2 2
 Large (0.11g) 0.33 2 2 2