ABSTRACT

Cyclic lateral load tests on model pipe piles driven in inundated, saturated sand were conducted aboard the Cambridge University geotechnical centrifuge. The tests provided qualitative and quantitative data on the interaction between pile and sand. The consistency of the test data was confirmed by the fact that the recorded data for various model piles followed the scaling laws of centrifugal modelling.

Experimental p-y curves were compared with curves recommended by API. The test data indicate that the initial stiffnesses of p-y curves vary as the square root of depth rather than linearly and that the ultimate resistance is underestimated near the pile head and overestimated at depth by the API curves. Despite these differences, computed response of piles under test conditions, using a slightly modified version of the API curves, showed good agreement with measured response. Particularly interesting is the good agreement between computed and measured response of a 5 ft diameter pile. This size is representative of the large piles in use whereas the data base for the API curves consists almost entirely of data from 2 ft diameter piles.

The possibility of computing p-y curves directly using representative constitutive equations for sand was explored with encouraging results.

INTRODUCTION

The response of an offshore pile foundation in sand to wave and earthquake loading is usually determined using a simple model of soil-pile interaction. The reaction between pile and soil, at any depth z, is represented by a non-linear Winkler spring called a p- curve in which p is the net soil reaction per unit length of pile and y is the pile displacement at depth z. The p-y curves most commonly used for sands are those recommended by the American Petroleum Institute1 in RP2A which presents appropriate p-y curves for both static and cyclic loading.

The API p-y curves are based on those developed by Reese et al from the results of lateral load tests on 2 ft diameter steel pipe piles driven in dense sand at Mustang Island 3. Although proprietary information from other tests not reported in the open literature has, no doubt, had some bearing on the API recommendations for constructing p-y curves, the data base is still very limited considering the dramatic increase in the diameters of the piles now being used offshore and the wide variety of sands in which such piles are driven. The lack of field data is primarily due to the difficulty and expense of conducting experimental parametric studies on full size piles. These difficulties an be overcome by resorting to centrifugal testing.

Barton4 conducted lateral loading tests on single piles and pile groups in the Cambridge University geotechnical centrifuge for a wide variety of test conditions. The cyclic loading tests on single piles in inundated sand will be described in this paper. Data from the tests provide information on the shapes of p-y curves, variations in their initial stiffness and ultimate resistance with depth and the applicability of the p-y curves used in practice for analysis of cyclic loading to much larger pile sizes than those tested at Mustang Island.

This content is only available via PDF.
You can access this article if you purchase or spend a download.