Bearing Capacity Of Driven Model Piles In Dense Sands From Calibration Chamber Tests

Abstract

ABSTRACT Model piles were driven and tested in samples of dense siliceous sands in a large calibration chamber at 3S/lMG, Univ. of Grenoble (France). Tension and compression tests were performed with both open-ended and closed ended pipe piles. The research programme aimed at providing geotechnical engineers with design parameters for piles driven into dense sands, focusing on points of interest for offshore piles, in particular:dense to very dense sands,tension versus compression skin friction capacity,high confining pressures (up to 800kPa) for simulating large penetration depths.comparison with cone penetration tests. New design parameters are proposed for computing bearing capacity of piles driven in dense to very dense siliceous sands. They are compared with those given in the current API RP2A recommendations. INTRODUCTION Bearing capacity of offshore piles driven in dense siliceous sands has been, and continues to be, a controversial topic since API RP2A recommendations were changed in 1984 (1). The two main points of concern, relating to K values for compression and tension pile capacity and limiting values for skin friction and tip resistances, have a direct influence on offshore foundation costs as reflected by steel cost, driving equipment mobilization cost, installation time and cost for possible remedial measures. These changes are still not fully accepted and applied for the North Sea by certifying authorities and some geotechnical consultants preferring to adhere to the more conservative rules of API 82. It is generally agreed that the uncertainty on computed axial capacity of piles driven in sands is large when using current API recommendations, and a statistical pooling of international experts concluded that the API criterion may be conservative by 20 to 40%. In dense to very dense sands (2), This is of particular concern for the North Sea where dense sands are encountered widely and where potential savings are important due to the large number of structures involved. On the other hand, some studies (3,4) have indicated a tendency of the API method to under predict capacity of short piles and over predict that of long piles, due to the so-called "friction fatigue effect". Interesting contributions were made recently by Kraft (5,6) and by Randolph and Dolwin (7) where new design criteria were proposed for computing axial pile capacity in sands, considering that no data provided conclusive evidence of the existence of limiting values and that an optimistic K value of 0,8 leads to the over predicting of the capacity of short piles when using the API criterion (5,6). It is also suggested that the current API design method may be conservative mainly due to underestimation of the end bearing capacity of piles (7). In an effort to improve our knowledge of the behaviour of piles driven in sands, an extensive experimental programme was conducted at 3S/lMG and supported by ELF Aquitaine Production and IFP. Model pile tests were performed in a large calibration chamber