The one-dimensional consolidation of saturated soil was numerically analyzed, with the consideration of the acceleration, by finitedeformation finite-element-method. Effects of the infinitely small deformation assumption and the acceleration were numerically examined after investigating those fundamental conditions for dynamic analyses, e.g. the number of elements, time increment and loading duration, under which stable solutions could be obtained. Conclusions say that the assumption of the small deformation tends to overestimate the consolidation settlement and consolidation time and that the account of the acceleration will not strongly affect the consolidation behavior.
Consolidation theories, currently used in practical purposes, such as the Terzaghi's and Mikasa's ones do not consider the acceleration of soil and pore-water particles although the consolidation of soils essentially is a dynamic phenomenon. Additionally, they assume the infinite smallness of deformation although finite deformation usually occurs due to consolidation. Making no account of the acceleration and assuming the smallness of deformation seem to mainly due to the mathematical limitation by which dynamic and finite deformation theories could not yield analytical solutions. Zienkiewicz and Shiomi (1984) formulated dynamic equations that governs two-dimensional problems, however they did not consider the finiteness of deformation. This study aims at understanding consolidation phenomenon through numerical analyses with full account of the finiteness of deformation and the acceleration of soil and water particles. Governing equations were developed based on the concept of the mixture (Atkin and Craine, 1976), and finite element technique was applied to them to be algebraic system of equations (ex, Zienkiwicz and Taylor, 2000). This paper will demonstrate as some of conclusions that the assumption of small deformation should not be made if circumstance permits and that the consideration of acceleration would not affect so seriously that the difference in numerical solutions from the quasi-static approach would be practically negligible.