2.4 NEGATIVE SKIN FRICTION ON PILE GROUPS
2.4.3 Centrifuge Model Tests on Pile Groups Subject to NSF
More recently, some researchers started to utilize the powerful and versatile Geotechnical Centrifuge model tests to study the problem of NSF on piles, especially pile groups. Tomas (1998) carried out downdrag tests on 4-pile, 9-pile and 16-pile
40
groups in centrifuge by constructing an embankment in-flight around the piles in two lifts. It was noted that the model pile group were jacked into the soil in-flight but the centrifuge was stopped subsequently to mount the sand hoppers to produce surcharge loading in that study. The centrifuge was then spun back to 100g to re-consolidate the sample and later the surcharge loading was applied. For the 4-pile group with center- to-center spacing of 3.5 times the pile diameter, no group effects was observed and the magnitude of downdrag forces in each individual piles within the pile group was found to be comparable to that obtained from single isolated piles. For the 9-pile and 16-pile groups, the inner pile was observed to develop unexpectedly large downdrag force comparable to that of a single pile and thus group effects was again not observed.
Tomas attributed this zero group effect to the possibility of development of full slip along the pile length, but acknowledged that the probability of a bending effect on the inner pile could not be discounted. He suggested that at a pile spacing of 3.5 times the pile diameter, pile group effect may not exist under a large soil settlement. It should be noted that quarter-bridge strain gauge configurations had been used in the instrumentation of the model piles used in Tomas’s centrifuge model study, and the great susceptibility of such configuration to thermal effects was not taken into account when analyzing the test data.
Lee (2001) also attempted to study the pile group effect of NSF using centrifuge modeling technology. However, great difficulty was encountered in the instrumentation of the model piles with full-bridge strain gauge circuit. Instead, only 2 active gauges were instrumented at each level inside the model piles while 2 passive gauges were installed in a dummy box outside the model piles. As such, it was found that there was a strong thermal effect in the pile instrumentation system which posed
41
great difficulties in the interpretation of the test results. No conclusive results on pile group effects on NSF were effectively obtained from the centrifuge model tests.
Lee et al. (2002) studied the negative skin friction mobilized on single piles and pile groups in centrifuge due to settlement induced by the self-weight consolidation and lowering of groundwater. 2-pile, 4-pile and 5-pile groups with center-to-center spacing varying from 2 to 6 times the pile diameter were conducted. It was found that the mobilization of NSF decreases with decrease of pile spacing and numbers of piles.
The inner pile developed a smaller NSF than the outer piles. However, it would develop the same amount of NSF as an isolated pile if pile spacing was larger than 6 times of pile diameter. It was found that evaluation of NSF on grouped piles using the concept of effective pile number proposed by Shibata et al. (1982) matched the test results well.
Chan et al. (2003) conducted centrifuge model tests to study the behavior of single piles “floating” in clay and subjected to NSF caused by self-weight consolidation.
Significant downdrag settlement of about 300 mm was observed for the model pile which posed severe serviceability problem. The model test was subsequently extended to “floating” pile groups consisted of 3×3 piles with 2.5d and 3.0d spacing (d is the pile diameter) to study the shielding effects of perimeter piles on the inner pile (Ng et al., 2005). It was observed that the measured maximum dragload of the center pile in the group at 2.5d and 3.0d spacing was only 53% and 75% of the measured maximum dragload of an unprotected single pile, respectively, and the measured downdrag settlement of the center pile was reduced to about 57% and 80% of the single pile. The test data was further substantiated by 3D back-analysis using ABACUS with the clay modeled by modified Cam-clay model with elasto-plastic slip considered at the pile- soil interface.
42
In summary of the above literature review on pile groups subject to NSF, it can be seen that the understanding of group effects of NSF is far from conclusive. For example, the limit pile spacing beyond which no group effects exists was reported to vary from 2.5d (Koerner and Mukhopadhyay, 1972) to 3.5d (Tomas, 1998) to 6d (Lee, 2002). Some test results indicated that the inner pile experienced negligible NSF (Okabe, 1977) while some test results showed that the downdrag force developed on the inner pile can be larger than that developed on the corner pile (Little, 1994), or is comparable to a single pile (Tomas, 1998). On the other hand, difficulties in the instrumentation of model piles have hampered fruitful test results to be obtained when using the centrifuge modeling technique to study NSF on pile groups (Tomas, 1998;
Lee, 2001). Although some sensible test data had been obtained by some researchers in their centrifuge model tests (Lee et al. 2002; Chan, et al. 2003; Ng et al. 2005), it was noted that the model piles were invariably installed at 1-g condition and was subsequently spun up to high-g for these model tests, posing a question of the correct modeling of stress condition around the piles. Obviously, the effects of NSF on pile groups involved complex pile-soil-pile interactions and further research work is highly desirable in this area.