ABSTRACT
In previous papers, (e.g. Lignola et al. 2008, Flora et al. 2007, Croce et al.; 2006; Croce et al. 2008) the authors have suggested to simulate geometrical and mechanical defects of columns by means of probabilistic models to be calibrated with
1 INTRODUCTION
The use of jet-grouting with earth supporting functions has greatly increased in recent years. At the design stage, however, there is still a relevant degree of uncertainty, mainly due to the lack of reliable methods for predicting the diameter and the position of the columns, as well as the mechanical properties of the cemented soil (soilcrete), all varying along columns axis due to a number of reasons. As a matter of fact, these defects make jet grouted columns far from being perfectly cylindrical, homogeneous bodies. It follows that, most times, jet grouting is designed on the basis of subjective rules of thumb, subsequently refined on site by means of trial and error tests (trial fields). In order to improve the reliability of design analyses, two complementary research efforts are needed: on one hand, field trials must be carefully analysed to find the relevance of columns defects in the largest possible variety of subsoil conditions and jet grouting technology; on the other hand, more satisfactory design methods taking into account the intrinsic variability of jet columns must be conceived. This paper focuses on this second aspect, with reference to the particular case of provisional support in tunnelling (open structure, Fig. 1.a) or in shafts (closed structure, Fig. 1.b), realized in cohesionless soils (ideal for jet grouting effectiveness) or at
statistical analyses of the results of available field trials. These models have been then introduced into mechanical analyses, carried with different approaches, to take into account defects into the design of jet grouted structures. The first part of this paper is devoted to a brief summary of this activity with reference to the above mentioned supporting structures (Fig. 1); in the second, the limits of the proposed predictive methods are highlighted, and possible improvements suggested. These latter have been commented by reporting the results of preliminary analyses.
