See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321625388

Rock Mass Strength derived from Rock Mass Parameters instead of Rock Mass
Classification

Article · December 2017

CITATIONS READS

0 2,578

2 authors, including:

William H. Gibson
Mining One Consultants
15 PUBLICATIONS 87 CITATIONS

SEE PROFILE

All content following this page was uploaded by William H. Gibson on 07 December 2017.

The user has requested enhancement of the downloaded file.

Rock Mass Strength derived from rock mass
parameters instead of rock mass classification

Gibson, W.H., Pastine S.

SRK Consulting, Buenos Aires, Argentina

ABSTRACT: The assessment of strength of rock masses is a key element for the analysis of many rock excavations, both open pit
and underground. There are attempts to assess the rock mass strength using Rock Mass Classification, the most popular one relates
the Hoek-Brown Failure Criterion with the Geological Strength Index (GSI) (Hoek, 2002). To introduce quantification on the GSI
assessment and move away from an estimate based on observation a procedure has been presented (Hoek, 2013) to calculate GSI
using measurable parameters RQD and joint condition. This paper presents a procedure to evaluate the rock mass strength using
those concepts taking RQD, spacing (or fracture frequency) and joint condition as the main parameters.

on measured parameters has been presented by Hoek

1. INTRODUCTION
[1] and this paper explore the use of that concept to
The assessment of rock mass strength is a key assess the rock mass strength. It is proposed to use
element in any rock excavation. For both, open pit spacing, RQD and joint condition as the main
and underground excavations the rock mass strength parameters to calculate the rock mass strength.
defined by a strength envelope is one of the main
inputs to be considered to assess stability.
2. ROCK MASS STRENGTH ASSESSMENT
There are empirical methods that relate rock mass BASED ON ROCK MASS
classifications with design parameters like slope CHARACTERISATION GSI
angles, open-spans and support, Bieniawski [6],
Laubscher [2], Barton [3]. These methods do not The rock mass strength can be evaluated using the
require an explicit definition of strength of the rock Hoek – Brown failure criterion, Hoek [7]. The
mass, but when numerical models are used as a tool strength envelope written in terms of the principal
of analysis it is necessary to have the rock mass stresses is shown in Eq. (1).
strength quantified. The strength envelope could be a
1  3   3 
linear, like Mohr – Coulomb or non-linear like the    mb  s  (1)
one suggested by Hoek – Brown [4].  ci  ci   ci 
The non-linear Hoek – Brown criterion relates the Where ci is the unconfined compressive strength of
strength envelope with the rock mass classification the intact rock (UCS), and the parameters mb, s and a
through the GSI index which is assessed mainly by are related with the rock mass rating through the GSI.
observation of the rock mass, Hoek [1]. This These relationships are shown in Eq (2), (3) and (4).
procedure has a lack of rigor not compatible with
modern engineering. A method to assess GSI based
Fig. 1. Quantification of GSI by Joint Condition and RQD (Hoek [1])

mb  GSI  100 
 exp  (2)
mi  28  14 D  3. ROCK MASS STRENGTH ASSESSMENT
BASED ON RQD, JOINT SPACING AND JOINT
 GSI  100  CONDITION
s  exp  (3)
 9  3D  The rock mass strength envelope derived using Hoek
 e  e 20 / 3 
1 1 GSI / 15 – Brown criterion is based on GSI as a parameter for
a (4) the rock mass as per Eq. (2), (3) and (4).
2 6
where mi is a material constant and D is a Through observation of these equations is possible to
Disturbance Factor. conclude that the natural log of the parameters mb/mi,
s and a are linear with GSI as is shown in Eq (5), (6)
A correlation between the GSI index and the rock and (7).
mass quality can be seen in Fig. 1 (Hoek [1]), visual
observation is used to assess GSI based on the joint
conditions and the fracturing of the rock mass.
  m   GSI  100  Where v1, v2, v3 and v4 are the values of variable v
ln b     (5) (ln(mb/mi), ln(s), or ln(6a-3+exp(-20/3)) at the nodes
 mi   28  14 D  and N1, N2, N3, and N4 the Lagrange Interpolating
 GSI  100  Polynomials evaluated using scales A and B.
ln(s )    (6)
 9  3D  The equations of the Lagrange Interpolating
Polynomials are shown in Appendix A.
 GSI
ln(6a  3  e 20 / 3 )  (7) It is interesting to note that with Eq (8) not only can
15
the parameters v be evaluated, but also GSI can be
The lines shown in Figure 1 have been made parallel calculated as well using the same equation.
and equally spaced (Hoek [1]). This change from the
3.1. Scale A
original chart makes GSI vary linearly with
For scale A Hoek [1] has suggested using joint
interlocking and surface quality (Scales A and B in
condition Jc as defined by Bieniawski [8], it seems
Fig 1).
an appropriate method to characterise the conditions
Because of the linear variation of GSI with the axes of the joints, Eq (9) shows Scale A calculation.
of interlocking and surface quality plus the linear
Scale A=1.5 Jc (9)
variation of the natural log of the parameters mb/mi,
s and 6a-3+exp(-20/3) (called parameters v to refer 3.2. Scale B
any of them) it is possible to calculate those For scale B Hoek [1] has suggested to use RQD/2 but
parameters only by knowing their value at the considering the range for scale B is from 0 to 40 only,
corners of the chart, i.e. knowing their value for GSI this leaves out the case with RQD in the range 80 –
85, 45, 0 and 40 at the corners numbered 1, 2, 3 and 100. As point of reference we can consider that a
4 in Fig 1. rock with an RQD in the range 80 to 90 can have 8
to 5 fractures per meter (Eq(12)), i.e not close to
The calculation can be done by Lagrange
intact rock.
Interpolating Polynomials, being possible to
calculate any Hoek – Brown parameter at any In this paper, it is suggested to use both RQD and
location in the chart shown in Fig 1 knowing their joint spacing Js rating as defined by Bieniawski [8]
values at the corners. to quantify the influence of fractures in the rock mass
strength, it is proposed to use Eq (10) for scale B
using the full range of RQD ranging from 0 – 100
and Js rating ranging from 0 - 20.
/5 (10)
The value of GSI is obtained by the sum of Scale A
and Scale B, which gives de expression presented in
equation (11).
GSI  1.5 Jc  Js  RQD / 5 (11)
GSI calculated using equation (11) and GSI
calculated from field data are compared in Fig 3.

Fig. 2 Shape Function

The interpolation functions for each node takes the

value 1 at the node and 0 on the other nodes.
Any parameter v can be a function of the values at
the nodes 1 to 4 as shown in Eq (8).
v  N1v1  N 2 v2  N 3v3  N 4 v4 (8)
 R
Q
GSI D
+
FF or
Js
Spacing

FF: fracture
Joint Condition Jc frequency

Fig. 3 Comparison between mapped and calculated GSI values. Fig. 4 Shifting focus from GSI rock mass parameters

It is important to note that Js rating is calculated The use of ratings to quantify rock mass properties
using spacing (s) of the fractures or indirectly using masks, up to some degree, the real characteristics of
fracture frequency (ff=1/s) and that RQD is related the rock mass. Ratings are arbitrary, experienced
to ff with equation (12) (Hudson [5]) rock mechanics can make sense of them, in contrast,
.
not experience is required to understand spacing of
100 1 0.1 (12) fracture frequency.

Therefore, RQD and Js rating are not independent Joint condition must be included in the calculation
event though both are required as input. using a rating, there is no other option since the
conditions are difficult to quantity. Spacing or
4. SHIFTING THE FOCUS FROM GSI TO ROCK fracture frequency is different because those values
MASS PARAMETERS are measured directly and it is much easier to get a
For many years, the rock mass strength has been “feeling” for them than using a rating.
calculated using GSI, the system has proven to work
but there are some drawbacks in the method; from The aim is to use spacing or fracture frequency
the author’s point of view the main one is that directly on the rock mass strength assessment and to
different rock conditions may have the same GSI as hide the idea of rating. To do that it is possible to use
it can be seen in Fig 1. the graph provided by Bieniawski [6] shown in Fig 5
and fit a curve to it.
Fig 4 shows the concept of shifting the focus from
GSI to more meaningful parameters. Still the
strength parameters will be calculated using GSI but
the rock mechanics engineer is using Joint
Condition, RQD and Fracture Frequency (or Spacing
if preferred) to assess rock mass strength.

Fig. 5 Input chart for RMR parameter discontinuity spacing

(Bieniawski 2011)
 20 5. CONCLUSION
18
Spacing Rating Js
16 The procedure presented to assess rock mass strength
14 uses the main two features present in a rock mass i.e.
12 joint condition and fracturing. The method keeps the
10 y = 1.4614x0.3431 focus on those parameters while moving rating and
8 R² = 0.9997 GSI to an intermediate level. The rock mechanic
6
4
engineer can assess the rock mass directly from
2 geotechnical parameters using the Hoek – Brown
0 failure criterion for the rock. The procedure is shown
0 200 400 600 800 1000 1200 1400 1600 1800 2000 in Fig 8.
Spacing [mm]

Fig. 6 Best fit rating Js for Spacing

Fig 6 shows the curve fitted to the relation spacing –

rating with a very good correlation. Fig 7 is the fit for
fracture frequency and rating used in Scale B, with a
very good correlation as well.

20
18
16
y = 15.6297x‐0.3431
Spacing rating Js

14
12
R² = 1.0000
10
Fig. 8 Diagram showing interrelation between input data,
8 calculation parameters and strength parameters
6
4
2 It is worth noting that the procedure presented
0 reproduces exactly the strength parameters obtained
0 10 20 30 40 50 60 using the conventional method based on GSI and
FF fracture frequency [1/m] programs like RocLab. This can be seen in Fig 9,
comparing the red and green lines from the output of
Fig. 7 Best fit rating Js using Fracture Frequency the program RM_Strength. This program can be
obtained for free contacting the main author at
Equations 13 and 14 show the relation between [email protected] or
spacing (s) or fracture frequency (ff) with Js. [email protected].
.
1.46 (13)
The authors would like to thank Professor Carlos
.
15.63 (14) Carranza-Torres for his thorough review and
comments about the paper.
Using equation (14) on equation (10) it is possible to
calculate Scale B from rock mass parameters directly
without using, at least not explicitly, the values of
GSI or rating. Equations (15) and (16) show the
formulae to calculate Scale B.
.
1.46 /5 (15)
.
15.63 /5 (16)
Fig. 9 Comparison results using conventional strength parameters assessment and proposed method. (Red and Green lines)

6. Bieniawski, ZT. 2011 Misconceptions in the Application

of Rock Mass Classifications and Their Corrections.
REFERENCES ADIF Seminar on Advanced Geotechnical
Characterization for Tunnel Design, Madrid
1. Hoek, E, Carter, T.G. Diederich, M.S. 2013.
Quantification of the Geological Strength Index Chart. 7. Hoek, E. C. Carranza-Torres. B. Corkum. 2002. Hoek –
47th US Rock Mechanics Geomecanics Symposium. Brown failure criterion 2002 Edition. Roclab program
Rocscience.
2. Laubsher,D.H. 1990, A geomechanics classification 8. Bieniawski, ZT, 1989. Engineering rock mass
system for the rating of rock mass in mine design, J1 S Classifications: a complete manual John Wiley and
Afr Inst Min and Metall, Vol 90, No10 , pp257,273 Sons, New York.

3. Barton, N.R., R. Lien, and J. Lunde. 1974 Engineering

classification of rock masses for the desing of tunnel
support, Rock Mech. 6(4), 189-239

4. Hoek, E. P.K. Kaiser, W.F. Bawden 1995. Support of

Underground Excavations in Hard Rock. Balkema

5. Hudson J.A, J. Harrison, Engineering Rock Mechanics,

and Introduction to the Principles. Pergamon, 1997
 APPENDIX A
Lagrange Interpolating Polynomials evaluation

Where:
Jc is joint condition evaluated using Bieniawski [8]
Jci is join condition evaluated at node i
SB is Scale B that correspond to Js + RQD/5
SBi = Js + RQD/5 evaluated at node i.
Table A1 shows the values of Scale A and B
evaluated at nodes 1 to 4.

Table A1. Values of Scale A (joint condition) and Scale B

(RQD/5 +Js) at the nodes.
Scale A Scale B
Nodes GSI
SA=1.5*Jc SB= Js + RQD/5
1 45 40 85
2 45 0 45
3 0 0 0
4 0 40 40

View publication stats