ABSTRACT
In geotechnical physical modelling, scaling laws should be respected to represent correctly the behaviour of prototype. In the 1g reduced models, the scaling law with respect to the stress level “σ* = 1” is not fulfilled. Therefore, a gradient in soil properties with depth cannot be simulated in such a model and it was decided to prepare a laboratory soil with homogeneous properties. A model with a geometry scale of 1/10 simulates a real soil with constant mechanical properties (soil stiffness and undrained shear resistance) ten times higher. Also, the frequency in the dynamic tests should be
10 times the prototype frequency. Hence, assuming a prototype input with a frequency of around 1 Hz, the model should be tested at 3.16 Hz, which is close to 1 Hz and far below the fundamental frequency of the clay model layer. Even though the
1 INTRODUCTION
Groups of Vertical Rigid Inclusions beneath a granular layer (IR) are widespread in France as a means of ground reinforcement for soft soil. Less common, but also highly efficient as a ground improvement system, Mixed Module Columns (CMM) present significant advantages for construction in seismic areas. A short Stone Column (length about 2 m) is linked to a rigid inclusion in CMM (Bustamante et al. 2006). Both CMM and IR systems lead to an increase in bearing capacity and a reduction in settlement. However, CMM have the advantage that the upper part of the system is more flexible and less seismic energy is transmitted upward and downward, to and from the superstructure.
