Granular Materials in Civil Engineering: Recent Advances in the Physics of Their Mechanical Behavior and Applications to Engineering Works - Multiscale Geomechanics: From Soil to Engineering Projects - page 64

This procedure, based on the new physical similarity rule resulting from laws of
fracture mechanics governing grain breakage, allows us to reduce the material mass
of a representative sample by a factor of (1/15) 3 ≈ 1/3,400.
In order to cover the range of usual stresses { σ (G 1 ) } applied on the rockfill
within the embankment, the range of stresses { σ (G 2 ) } applied on the similarly
reduced material in the laboratory must be extended up to the value given by
eqiuation [3.29], i.e. ≈4 times { σ (G 1 ) } if we consider the central trend of the
statistic distribution, and a reduction factor equal to 15.
3.3.2. Incidence of scale effect on rockfill slope stability
3.3.2.1. The question of stability assessment
In recent constructions of large embankments, and especially rockfill dams, the
economic interests have directed the design practices towards:
− coarser grain size distributions in order to minimize the quarrying costs;
− steeper slopes in order to minimize the required volumes of material (or at
least slopes identical to the ones of mid-sized embankments).
The stability of such constructions is mostly estimated by extrapolating the
values of their shear strength from typical values for granular materials, without real
measurements for reasons presented above (see section 3.3.1). Therefore, the impact
of such design practices - coarser grain sizes, steeper slopes and higher
embankments - deserves to be analyzed.
Figure 3.14 summarizes the question arising from these combined effects:
starting from the safety factor of a mid-sized embankment of height H 0 and slope β  ,
built with a rockfill material of grain size distribution G(D 0 ) , can we determine the
safety factor of an embankment having a higher height H and steeper slope β , built
with a coarser granular material G(D) ?
Such a method, if available, would be a useful alternative to the design
extrapolation practices described above. Provided that the materials have
geometrically similar grain size distributions, come from the same mineral stock,
and are prepared at the same density, the scale effects analyzed above bring the
answer, as far as the shear strength is concerned, which is the key element of a slope
stability analysis.