Page 366 - Engineering Rock Mass Classification_ Tunnelling, Foundations and Landslides

Chapter 26 Geological Strength Index 321

3. Fair
4. Poor
5. Very poor

A 6 Â 5 block in the matrix of Figure 26.1 is picked up first according to actual and
undisturbed rock mass classification and discontinuity surface condition. Then a corre-
sponding GSI is read. According to Hoek (1998) and Marinos and Hoek (2000), a range
of values of GSI (or RMR) should be estimated instead of just a single value. This prac-
tice has a significant impact on the design of slopes and excavations in rocks. Drastic
degradation in GSI, RMR, and Q-values is found to occur in openings after squeezing
and rock bursts. This is also seen in openings, hence the need for evaluating the GSI
of rock mass in the undisturbed condition (D ¼ 0). Back analysis of both a model
(polyaxial strength criterion) and its parameters (from the observed behavior of rock
structures) is an ideal method of the rock mass characterization, and GSI is the first step
in this direction.

Figure 26.1 is used judiciously for crushed/disintegrated and laminated/sheared
rocks. Similarly, hard, thick laminated rocks in the last row of Figure 26.1 may
not be applicable, because they may have a higher strength classification (see
Table 5.2, Class II).

The GSI chart has been subsequently quantified by Cai et al. (2004) by incorpo-
rating the rock block volume (Vb) formed by the joints or discontinuities and the joint
condition factor JC (see Table 4.6). The suggested quantification is also shown in
Figure 26.1. The block volume (Vb), affected by the joint set spacing and
persistence, can broadly be known by the joint spacing given for six different rock
classes in Figure 26.1. The value of joint condition factor, JC, controlled by joint
roughness, weathering, and infilling material, can be obtained by Eq. (26.4) from
Cai et al. (2004).

JC ¼ JW JS ð26:4Þ
JA

where JW ¼ large-scale joint or discontinuity waviness in meters from 1 to 10 m
(Table 26.1), JS ¼ small-scale smoothness in centimeters from 1 to 20 cm (Table 26.2),
and JA ¼ joint alteration factor (Table 26.3).

Cai and Kaiser (2006), based on the proposed quantitative chart (Figure 26.1), and

using surface fitting techniques, suggested the following equation to calculate GSI from

JC and Vb:

GSIðVb, JCÞ ¼ 1 26:5 þ 8:79 ln JC þ 0:9 lnVb ð26:5Þ
þ 0:0151 ln JC À 0:0253 lnVb

where JC is a dimensionless factor defined by Eq. (26.4) and block volume Vb is in cm3
(see the section Calibration of RMi from Known Rock Mass Strength Data in
Chapter 10).

To avoid double-accounting, groundwater condition and in situ stresses are not con-
sidered in GSI because they are accounted for in computer models. GSI assumes that
the rock mass is isotropic; therefore, only cores without weak planes should be tested
in triaxial cells to determine qc and mr as GSI downgrades strength according to schis-
tocity. This classification reduces many uncertainties in rock mass characterization. An
undisturbed rock mass should be inspected for classification; however, heavy blasting
creates new fractures.