Soil Plug Behavior of Piles in Sand

Abstract

Abstract It has been proposed by Randolph (2) that granular soil plugs in open ended pipe piles can develop much greater resistances than allowed in present design (3). A simple mathematical model extending Randolph's concept to include load deformation behavior was developed and checked against more sophisticated numerical solutions. A series of small scale laboratory tests and a large scale field test were conducted to verify the model predictions. A "lock up" phenomenon wasexperimentally confirmed and the simplified model was validated. Additional work is needed to explore remaining questions, particularly rate effects and cyclic loading. Introduction In an open-ended pipe pile loaded to failure, either the full end bearing at the pile tip will be mobilized or the soil plug will slip. Whichever is the weakest component will determine the mechanism. Conventional design practice (3) is to estimate the internal plug resistance using the same unit shear components used for external shaft resistance. This is a reasonable approach inmany cases, but in relatively free draining and/or strongly dilative soils or for very slow loading it may be overly conservative. Randolph (2) has shown that a soil plug can have much greater capacity than conventional prediction methods suggest. In the present study Randolph's basic idea was investigated both analytically and experimentally. The theory was extended to include load-deformation predictions using a simple one dimensional model. Analytical verification of the model was carried out using finite element analysis. In addition, a series of laboratory tests was conducted using three inch diameter pipes with sand plugs. Finally, large scale tests were carried out using an eighteen- inch diameter prototype pile with asand plug. The results of these studies are detailed below. Ultimate Plug Capacity Consider the schematic diagram in Fig. 1 showing a differential element of the soil plug. Several simplifications are made. Most importantly, the vertical stress, ?, is assumed uniform across the cross section and the failure (or slip) shear stress at the pile wall is proportional to the vertical stress, i. e, r-?. The form of the constant of proportionality, ? is discussed later. Theassumed stress state is not rigorously correct; for example there are no complementary shear stresses on the horizontal plane, but as will be shown it is a good approximation. Summing forces on the differential element leads to the simple differential equation(Mathematical Equation)(Available In Full Paper) In all cases the simple theory under predicts (is more conservative than) the more rigorous FEM. The finite element results showed clearly that near the plug base the mean stress tended to build up near the pile wall, thus deviating from the one-dimensional assumption. This effect was more pronounced in the higher capacity plugs. The greatesterrors occurred for situations where the wall friction was the greatest. On the other hand these comparisons should be judged in light of the high sensitivity of the solution (exponential behavior) to minor variation in parameters.