Swelling Rock - Rock Mechanics Based on an Anisotropic Jointed Rock Model - page 196

of swelling pressure (Fig. 8.10, lower left). On the other hand, confi ned swelling leads
to a closure of fl ow paths, which reduces the permeability of the rock mass and thus re-
duces and even interrupts further water access. As a result, a further increase of swelling
pressure is inhibited.
As described in Section 7.3.2, the permeability of a rock mass parallel to a discontinuity
set D is controlled by the normal displacement

δ n,D on the discontinuities of this set. In
the coupled model, in addition to the elastic and viscoplastic strain components

ε n,D el

ε n,D vp , respectively, the swelling strain ε n,D q also contributes to

and δ n,D :

(8.17)

ε n,D el is small compared with the normal
strain components due to viscoplastic and swelling behavior, it is neglected. Thus, nor-
mal displacements

Since the elastic normal strain component

δ n,D on the discontinuities, and thus permeability changes in the
coupled model, are calculated on the basis of ε n,D vp and ε n,D q alone:

(8.18)

For unleached Gypsum Keuper s D stands for the mean spacing of bedding-parallel
discontinuities s B or joints s J .

8.6 Characteristic Parameters of the Gypsum Keuper

The coupled model contains a large number of rock mechanical parameters. However,
many of these parameters could be reliably estimated from the results of rock mechan-
ical investigations on unleached Gypsum Keuper in the area of Stuttgart.

In dilatometer tests (Section 15.2) a geometric mean of the deformation modulus of
4900 MPa was determined for initial loading (Fig. 8.12). The corresponding mean value
of 4100 MPa obtained from plate loading tests (Section 15.3) reveals good agree-
ment (Fig. 8.13). A defi nite anisotropy could not be derived from the test results.
Thus, an isotropic deformability of the rock mass can be assumed as mentioned in
the previous section.
Based on the results of mapping (Section 13.9) the friction angles of the bedding-par-
allel discontinuities B and joints J were estimated from observed surface characteristics
as φ J = 35°. The corresponding values for cohesion were estimated from

persistence to range from 0 to 0.2 MPa for the bedding-parallel discontinuities and from
0.1 to 0.6 MPa for the joints.

φ B = 30° and

The order of magnitude of these shear parameters has been confi rmed by an in-situ
shear test which was conducted in the Wagenburg tunnel (BMV 1975) and by back
analyses carried out for the temporarily unsupported underground turning loop of
the urban railway in Stuttgart (Wittke et al. 2006). The evaluated shear parameters
are compiled in Fig. 8.14.