Page 213 - Engineering Rock Mass Classification_ Tunnelling, Foundations and Landslides

Chapter 13 Strength Enhancement of Rock Mass in Tunnels 175

attempted carefully as mentioned earlier. At present, a non-linear back analysis may be
difficult, and it does not give unique (or most probable) parameters.

The proposed strength criterion is different from Mohr-Coulomb’s strength theory
(Eq. 26.12), which works well for soils and isotropic materials. There is a basic difference
in the structure of soil and rock masses. Soils generally have no pre-existing planes of
weaknesses so planar failure can occur on a typical plane with dip direction toward s3.
However, rocks have pre-existing planes of weaknesses like joints and bedding planes,
and as such, failure occurs mostly along these planes of weaknesses. In the triaxial tests
on rock masses, planar failure takes place along the weakest joint plane. In a polyaxial
stress field, a wedge type of failure may be the dominant mode of failure if s2 >> s3.
Therefore, Mohr-Coulomb’s theory needs to be modified for anisotropic and jointed rock
masses.

The new strength criterion is proved by extensive polyaxial tests on anisotropic tuff
(Wang & Kemeny, 1995) and six other rocks. It is interesting to note that the constant A is
the same for biaxial, triaxial, and polyaxial tests (Singh et al., 1998). Further, the effec-
tive in situ stresses (upper bound) on ground level in mountainous areas appear to follow
Eq. (13.14) (qcmass ¼ 3 MPa, A ¼ 2.5), which indicates a state of failure of earth crust
near the water-charged ground due to the tectonic stresses.

The output of the computer program SQUEEZE shows that the predicted support
pressures are of the order of those observed in 10 tunnels in the squeezing ground
condition in the Himalayas in India. There is a rather good cross-check between the the-
ory of squeezing and the observations except in a few cases. Thus, Eqs. (13.14) and
(13.15) assumed in the theory of squeezing are again justified partially (Singh &
Goel, 2002).

In the NJPC project tunnel excavated under high rock cover of 1400 m through
massive to competent gneiss and schist gneiss, the theory predicted rock burst condition
(Jr/Ja ¼ 0.75, i.e., > 0.5). According to site geologists Pundhir, Acharya, and Chadha
(2000), initially a cracking noise was heard followed by the spalling of 5–25 cm thick
rock columns/slabs and rock falls. This is mild rock burst condition. Another cause of
rock burst is the Class II behavior of gneiss according to the tests at IIT. According
to Mohr-Coulomb’s theory most severe rock bursts or squeezing conditions were
predicted under rock cover more than 300 m (qc ¼ 27 MPa and qcmass ¼ 15.7 MPa). Mild
rock burst conditions were actually met where overburden was more than 1000 m.
However, polyaxial theory (Eq. 13.14) suggested mild rock burst conditions above
overburden of 800 m. Thus, polyaxial theory of strength is validated further by the
SQUEEZE program (Singh & Goel, 2002). Recently, Rao, Tiwari, and Singh (2003)
developed the polyaxial testing system. Their results were replotted and parameter
A was found to increase slightly from 3.8 to 4.2 for dips of joints from 0 to 60, although
qcmass changed drastically. Thus, the suggested hypothesis appears to be applicable
approximately for the rock masses with three or more joint sets.

Poor Rock Masses

Squeezing is found to occur in tunnels in the nearly dry weak rocks where overburden
H is more than 350 Q1/3m. The tangential stress at failure may be about 2gH assuming
hydrostatic in situ stresses. Thus, mobilized compressive strength is 2 g 350Q1/3 ¼ 700
gQ1/3 T/m2 (Eq. 13.9).