CENTRIFUGE MODEL TESTS OF NAILED SOIL SLOPES

抄録

This paper presents the results of 24 centrifuge model tests of nailed soil slopes and vertical walls, constructed out of dry Leighton buzzard sand. The walls were initially supported by fluid pressure from flexible rubber bags against the face, and excavation was modelled by gradually draining the fluid from the bags. The finished model walls were 200 mm high and were initially tested at 30 g acceleration to correspond to a prototype structure 6.0 m high. If failure was not obtained, the acceleration was increased progressively to a maximum of 80 g. No surcharges were applied. The main parameters varied in the tests were the wall slope, nail length, nail surface roughness, nail inclination, facing stiffness and facing roughness. Observations were made of the mechanism of failure when it occurred, of soil pressures on the facing, and of pre-failure deformations. Failure was always by pull-out rather than breakage of the nails, and a series of pull-out tests of the model nails was conducted to aid interpretation of the results. Failure surfaces were seen to have the shape of logarithmic spirals, and limit equilibrium analyses based on these surfaces agreed well with experimental observations. Prior to failure, earth pressures on the facing compared reasonably well with those calculated by Coulomb's method, except at the base of the wall. One test set out to model a full-scale trial wall, and although the construction process could not be exactly modelled in the small-scale centrifuge tests, comparisons were sufficiently good to give confidence in the model test results.