MECHANIZED EXCAVABILITY RATING FOR HARD-ROCK MINING

By Z. T. Bieniawski, D.Sc. (Eng),1 and Benjamín Celada, Ph.D.2

It should be noted that the RME index does not

When you can measure what you are speaking about, replace the Rock Mass Rating (RMR) or Q-systems as
and express it in numbers, you know something about it;
used in mining and tunneling; indeed, one of the RME
but when you cannot measure it, when you cannot
input parameters, standup time, is determined from the
express it in numbers, your knowledge is of a meager
and unsatisfactory kind.
RMR. However, the approach presented introduces a
—Lord Kelvin (1824–1907) specialized tool relevant to excavating tunnels and drifts.
Possible applications to hard-rock mining are explored.

ABSTRACT INTRODUCTION

After emphasizing the importance of quantitative rock Rock mass classifications, although originally devel-
mass classifications in mining, originally directed to selec- oped for rock tunneling in civil engineering, have been
tion of rock support measures, but subsequently to esti- used in mining for some 35 years, going back to RMR
mates of rock mass properties such as rock mass strength applications in South African hard-rock and coal mining
and rock mass modulus of deformation, current attention [Bieniawski 1972; Laubscher 1976]. In the United States,
calls for a classification specifically for rock mass research investigations by Kendorski et al. [1983] for hard-
excavability by tunnel boring machines (TBMs), which are rock mining, based on RMR, as well as for coal mining
used extensively in tunneling as well as in the mining by Unal [1983] for roof support and Kalamaras and
industry. Bieniawski [1995] for pillar design, also based on RMR,
This paper introduces the Rock Mass Excavability were highly innovative, and their results are used to this
(RME) index for predicting excavability of rock masses by day.
TBMs using a quantification of machine performance and More recently, attention has been paid to rock mass
rock mass conditions. The RME index is based on five classifications aimed at determining rock mass properties,
input parameters aimed at relating rock mass behavior and
i.e., rock mass strength and the rock mass modulus of
machine characteristics: (1) uniaxial compressive strength
deformation. Examples of particularly useful charts for this
of the rock material, (2) drillability/abrasivity, (3) rock
purpose are presented in Appendix B of this paper.
mass jointing at mine drift face, (4) standup time of the
Lately, as machine-bored excavations in tunneling
excavation, and (5) groundwater inflow.
become more common than drill-and-blast tunneling,
Development of the RME index entailed the collection
of extensive data from more than 28 km of tunnels and a need emerged for predicting the performance of tunnel
some 400 case records from projects in Spain involving boring machines (TBMs) based on considerations of
double-shield TBMs. In the process, a number of statistical interaction of rock mass conditions and the TBM opera-
correlations have been established between RME and such tional parameters. If successful, such findings would be of
output parameters as degree of machine utilization, equal interest to mining applications.
advance and penetration rates, thrust and torque of the
cutterhead, and the specific energy of excavation. It was PREVIOUS STUDIES
found that the RME index provides a particularly signifi-
cant correlation for predicting the average rate of advance When one considers the history of underground exca-
(m/day). vation technology, its development, and the major mile-
In essence, the RME index is a classification system stones, the emergence and increasing use of modern TBMs
that features interaction of rock mass conditions with bor- provided both spectacular advantages and achievements, as
ing machine characteristics for use in the early stages of well as complex challenges and problems to designers and
a project. constructors who faced significant shortcomings in our
understanding of the interaction of rock mass conditions
and TBM design and performance.
1 In fact, when Terzaghi introduced his rock load con-
Professor, Superior School of Mines, University of Madrid,
Spain, and President, Bieniawski Design Enterprises, Prescott, cept in 1946, followed by Lauffer’s standup time concept
AZ. in 1958 and Deere’s Rock Quality Designation (RQD) in
2
Professor, Superior School of Mines, University of Madrid, 1964, these design approaches were directed to selection of
Spain, and President, Geocontrol, S.A., Madrid, Spain. rock reinforcement for tunnel construction by drilling and

15
blasting. The equipment selected for tunnel excavation was Subsequently, the key objection to QTBM was provided
left to the discretion of the contractor, with little input by by a major study from Norway (where the Q-system was
the designer. Even subsequent modern rock mass classifi- invented) published by Palmström and Broch [2006]. They
cation methods [Bieniawski 1973; Barton 1974] were pre- concluded:
dominantly directed to drill-and-blast tunnels, independent
of TBM characteristics. QTBM is complex and even misleading and shows
Today, this is no longer the case. TBMs have low sensitivity to penetration rate; the correlation
increased in power, size, and type to such an extent that coefficient with recorded data is even worse than
they directly influence tunnel design. Moreover, their conventional Q or RMR or with other basic param-
eters like the uniaxial compressive strength of the
selection is a source of tremendous satisfaction due to
intact rock. It is recommended that the QTBM should
increased safety and higher performance, as well as deep
not to be used.
despair when unexpected ground conditions are encoun-
tered and the TBM may be immobilized for months and This finding is clearly supported by Figure 1.
sometimes has to be rescued by old-fashioned hand mining
or conventional drill-and-blast excavation.
A major problem emerged: how to assess effectively
the interaction between rock mass conditions, as described
by the RMR or Q classification systems, and the design
and performance characteristics of the TBM. Certainly,
some attempts to solve this problem have been made, as
reviewed below, but the state of the art still rests on the
TBM manufacturers and tunnel contractors that must rely
on their experience, ingenuity, and even the will to survive
many adverse conditions.

STATE OF THE ART IN ROCK EXCAVABILITY

Excavability is defined as the ease of excavation and

was investigated as early as Kirsten [1982]. TBM Figure 1.—Advance rates for three TBM tunnels plotted
excavability or performance prediction models were against QTBM [Sapigni et al. 2002].
studied by Barton [2000], Alber [2001], Bieniawski
[2004], Blindheim [2005], and others.
In essence, it is recognized that the choice between a Other attempts were reported by Alber [2001] con-
TBM and drilling and blasting can be quantified based on centrating on contracting practice and probabilistic esti-
rock mass quality and machine characteristics. An example mates of advance rates and project economics. The RMR
of an interdependence function is the QTBM formulation system was used by Grandori et al. [1995] to demonstrate
[Barton 2000]: ranges of effectiveness for TBM performance in different
rock mass quality as a function of machine type: open
QTBM = RQD0/Jn × Jr /Ja × Jw /SRF TBM or double-shield. Bieniawski [2004] reviewed the
concept of rock mass excavability based on the RMR as
× σMASS/F × 20/CLI × q/20 (1)
adjusted for TBMs.
However, there is convincing evidence that complex
where CLI = cutter life index (Norwegian Institute of
equations combining rock mass quality RMR or Q with
Technology), SRF = stress reduction factor, F = average
additional parameters related to TBM characteristics are
cutter load (tnf), q = quartz content (%).
not an effective approach. In other words, it is doubtful
Equation 1 received much attention, but was also
that one formula can include all the factors pertinent to
severely criticized [Blindheim 2005]. In this research, the
rock mass quality, as well as those influencing TBM
above relationship was also tested, but without success
choice and performance.
because of the problem with the definition of rock mass
In fact, expert opinion holds that the RMR and Q-
strength, σMASS, which is based on “inversion of σc to a systems are most effective as they are commonly used,
rock mass strength, with correction for density,” rendering consistent with the purposes for which they were devel-
it unacceptable. Nevertheless, Abrahão and Barton [2003] oped. Thus, adjusting these systems for TBM-sensitive
applied this equation with all 21 parameters (“for which no parameters, such as rock abrasivity and cutter thrust, may
apology is made,” declared the authors), emphasizing that be counterproductive and may only create confusion.
the rock-machine interaction in tunneling is very complex.

16
 THE CONCEPT OF THE ROCK MASS
INPUT DATA FORM f or Rock Ma ss Excava bility
EXCAVABILITY (RME) INDEX
Name of Tunnel .............................................................................................................................
Initial chainage of section:..............................Final chainage of section.........................................
After much overwhelming evidence, such as shown in Length of section:..............................m (should be > 40 m)
Figure 1, we concluded that modifying an existing rock Duration of excavation (days):.......................................................... (number + 1 decimal)

mass quality classification, be it the RMR or Q, for Average Rate of Advance ARA = ...................m/day
determining rock mass excavability was not an effective Lithology:..........................................................................................Average depth:............. .......m
approach for modern engineering practice. Accordingly, ROCK MASS PARAMETERS

research devoted to rock mass excavability was initiated in Uniaxial compressive strength of intact rock (? c):........................................MPa
Drilling Rate Index DRI:................. Type of homogeneity at excavation face:..................... .........
2004 with the objective of establishing an index, similar to N° of joints per meter:...............Rock Mass Rating RMR: range.........................average…...........
the RMR, but which was specifically directed to predicting Orientation of discontinuities with respect to tunnel axis
rock mass excavability, rather than rock mass quality. This
(perpendicular, parallel or oblique):..............................................................
work was aimed at selecting the appropriate method of
tunnel excavation, having considered rock mass-machine Stand up time:...................hours Groundwater inflow at tunnel face: .............liters/sec

interaction, using TBMs or conventional mechanized exca- Rock Mass Excavability RME range.....................................average....................

vation. The RME concept proposed first by Bieniawski et TBM PARAMETERS

al. [2006] was based on analyses of 387 sections of three Average speed of cutterhead rotation: .......... ................rpm Applied Thrust:...................m . kN

Spanish tunnels comprising 22.9 km in length. In each Specific Penetration:.....................................mm /rev

case, the tunnels studied included detailed data on rock Rate of Penetration:...................................... mm /min

mass characteristics and TBM parameters, as shown in the N° cutters changed:....................................... Rate of TBM utilization: .................. ...............%

RME input data form in Figure 2.

Figure 2.—Input data form for determining the Rock Mass
Excavability (RME) index.

Table 1.—Input ratings for Rock Mass Excavability (RME) index

UCS OF INTACT ROCK (0–25 points)

σc (MPa).................... <5 5–30 30–90 90–180 >180

Average rating .......... 4 14 25 14 0
DRILLABILITY (0–15 points)
Drilling Rate Index .... >80 80–65 65–50 50–40 <40
Average rating .......... 15 10 7 3 0
DISCONTINUITIES AT TUNNEL FACE (0–30 points)
Homogeneity Number of joints per meter Orientation with respect to tunnel axis
Homo-
Mixed 0–4 4–8 8–15 15–30 >30 Perpendicular Oblique Parallel
geneous
Avg.
10 0 2 7 15 10 0 5 3 0
rating
STANDUP TIME (0–25 points)
Hours ........................ <5 5–24 24–96 96–192 >192
Average rating .......... 0 2 10 15 25
GROUNDWATER INFLOW (0–5 points)
L/sec ......................... >100 70–100 30–70 10–30 <10
1
Average rating .......... 0 1 2 4 5
1
Zero for argillaceous rocks.

17
 SELECTION OF RME INPUT PARAMETERS

The RME index is based on the five input parameters

listed in Table 1, together with the ratings associated with
each. Selecting the five parameters involved a Linear
Discriminant Analysis using the R code developed by the
Institute of Statistics and Probability Theory of the Vienna
University of Technology, Austria. As a result of this
analysis, it was found that the parameters with stronger
influence in the average rate of advance (ARA), expressed
in m/day, are: drillability/abrasivity, discontinuity spacing,
and standup time. In addition, it was decided to include the
two basic rock mechanics parameters: uniaxial compres-
sive strength (UCS) of the rock material and water inflow
Figure 4.—Correlation between the RME index and the
because these two factors are known to strongly influence
average rate of advance (m/day) for single- and double-
the TBM advance. Once the five parameters were selected, shield TBMs.
a weighted distribution was performed. These weights
have been statistically analyzed, minimizing the error in CORRELATION BETWEEN AVERAGE RATE OF
the ARA prediction and resulting in the ratings shown in ADVANCE (ARA) AND RME
Table 1.
In practice, four of the input parameters are deter- The average rate of advance (ARA), expressed in
mined from standard site exploration programs: UCS of m/day, is the most significant parameter to compare
the rock material, rock drillability, rock mass jointing performances from several tunnel or drift construction
(spacing, orientation, and condition of discontinuities at projects. The statistical analyses carried out provided the
the tunnel front), and groundwater inflow. The fifth param- correlation depicted in Figure 4 between the ARA and
eter, standup time, is estimated from the well-known RMR RME for single- and double-shield TBMs.
chart (Figure 3), which depicts standup time versus These findings were derived for tunnels with diam-
unsupported active span as a function of RMR (after eters close to 10 m. In order to take into account the influ-
Bieniawski [1989]; see also Appendices A and B of this ence of other tunnel diameters, D, the coefficient kD is
paper). As the case studies on that chart were derived from used. The values of kD can be calculated from the follow-
drill-and-blast tunnels, a correlation obtained by Alber ing expression:
[1993] is used for TBM tunnels. The following equation is
applicable: kD = –0.007D3 + 0.1637D2 – 1.2859D + 4.5158 (3)

RMRTBM = 0.8 × RMRD&B + 20 (2) CORRELATIONS OF RME WITH OTHER

PARAMETERS
A number of significant correlations were obtained in
this study in addition to those discussed above.

Specific Energy of Excavation

The concept of specific energy of excavation (Es) for

mechanized tunneling and mining is “borrowed” from the
petroleum and gas drilling industry, where it has been used
for many years [Teale 1965]. Most recently, this concept
was applied to assess the ease of mechanical excavation
involving this expression:

Es = F/A + 2π N T/ A × ARA (4)

where Es = specific energy of excavation (kJ/m3);

F = total cutterhead thrust (kN);
A = excavated face area (m2);
Figure 3.—Standup time as a function of RMR and N = cutterhead rotation speed (rps);
unsupported span [Bieniawski 1989]. T = applied torque (kN·m);
and ARA = average rate of advance (m/s).

18
 The above equation consists of two terms. The first Cutterhead Thrust (FC) and Torque (T)
represents the specific energy of the cutterhead thrust from
static loading, while the second is the specific energy of Figures 6–7 show the correlation of RME with both FC
rotation incurred by the rotating cutterhead. In this study, and T values, providing acceptable coefficients of R=0.64
the specific energy of rotation (Er) was related to the RME and R=0.71, respectively, for single- and double-shield
in Figure 5. TBMs.

LATEST FINDINGS

The construction of the famous Guadarrama tunnels

involving two tubes, each 9.5 m in diameter and 28 km
long, using four double-shield TBMs, led to the intro-
duction of an adjustment to the predicted ARA obtained
from a given RME, incorporating the effect of the length
of the tunnel excavated and the influence of the crew skills
when dealing with the TBM and the terrain. This can be
represented as

ARA R (5)
ARA T =
FL × FC
Figure 5.—Correlation between the RME index and the
specific energy of excavation. where ARAT = predicted true value of ARA from the
correlation with RME;
ARAR = recorded average rate of advance,
m/day, achieved in a tunnel section;
FL = factor of experience as a function of
tunnel length excavated;
and FC = factor of effectiveness by the crew
handling the TBM and the terrain.

Based on the results obtained during construction of

the Guadarrama and Abdalajís Tunnels, Tables 2–3 show
the values appropriate for the coefficients FL and FC.

Table 2.—RME adjustment factor (FL)

Figure 6.—Correlation between the RME index and
TBM torque.
Tunnel length Adjustment
excavated (km) factor (FL)
0.5 ............................................ 0.50
1.0 ............................................ 0.86
2.0 ............................................ 0.97
4.0 ............................................ 1.00
6.0 ............................................ 1.07
8.0 ............................................ 1.12
10.0 .......................................... 1.15
12.0 .......................................... 1.20

Figure 7.—Correlation between the RME index and

cutterhead thrust.

19
 Table 3.—RME adjustment factor (FC) 3. The effect of blasting damage, because in hard-rock
mining drilling and blasting, unless smooth blasting
is used, may have an adverse effect on stability com-
Effectiveness of the crew Adjustment
handling TBM and terrain factor (FC)
pared to machine boring.

Less than efficient....................... 0.88 As a matter of fact, all of the above effects were incor-
Efficient ....................................... 1.00 porated into the Mining Basic RMR (MBR) classification
proposed by Kendorski et al. [1983].
Very efficient ............................... 1.15
There are various types of excavating machines used
in mining. In modern hard-rock mines, machine excavation
This produces a refined RME07 correlation depicted in is used to construct access drifts and chambers, while in
Figure 8 devoted specifically to double-shield TBMs. coal mines, continuous miners and shearers are common.
In each case, to access mineral deposit production, mines
SPECIAL CONSIDERATIONS FOR APPLICATIONS employ roadheaders and/or open-type TBMs.
IN MINING The RME index can be applied directly to evaluate
excavability of mine drifts and chambers. However, at the
Applications of rock mass classifications in mining time of writing, work on correlations between the RME
require some special considerations compared to civil and ARA is still in progress for roadheaders and open
engineering for a number of reasons. The three most TBMs. In fact, investigations to determine a correlation
important are— between the RME and ARA for open-type TBMs began
last year, with results expected to be presented by June
1. The effect of in situ stresses, since mines are usually 2007.
deeper than tunnels; As far as applications to roadheaders and similar
2. The effect of the induced stresses, because in mining machines are concerned, we are still in the process of data
the stress field changes as mining advances and also collection and would welcome any case histories of
due to adjacent excavations; and RME applications in this respect by interested parties. In

60

50
ARAT = 0,813RME07 - 32,56
σci≤45 MPa

40
ARAT (m/día)

30 σci>45 MPa

20
ARAT = 0,597RME07 - 24,88

10

0
50 55 60 65 70 75 80 85 90 95 100
RME07

Figure 8.—Correlation between the RME07 and the average rate of advance for double-shield TBMs. For RME<50, TBMs in
double-shield mode are not recommended.

20
addition, applications in room-and-pillar and longwall Bieniawski ZT [1973]. Geomechanics classification of
mining will require modifications to the actual structure of jointed rock masses. J S Afr Inst Civ Eng Dec:382–398.
the RME index due to the specific nature of such mining Bieniawski ZT [1989]. Engineering rock mass classi-
operations. For example, the ratings for the standup time fications: a complete manual. New York: John Wiley &
parameter may require an adjustment factor due to the Sons.
degree of fracturing in the roof strata and due to the effect Bieniawski ZT [2004]. Aspectos clave en la elección
of the induced stress in order to better assess the stability del método constructivo de túneles (in Spanish). In: Pro-
of the rock mass in these types of mining operations. ceedings of Jornada Técnica (Madrid, Spain), pp. 1–37.
Also in: Ingeopress, No. 126, pp. 50–68.
CONCLUSIONS Bieniawski ZT, Celada B, Galera JM, Álvares M
[2006]. Rock mass excavability (RME) index: a new way
After 3 years of studies and analyses of more than 400 to select the optimum tunnel construction method. In: Pro-
case histories, RME seems to provide a tool that enables ceedings of the ITA World Tunnelling Congress (Seoul,
tunnel designers and constructors to estimate the perform- South Korea).
ance of TBMs. Future work will focus on extending the Blindheim OT [2005]. A critique of QTBM. Tunnels &
RME to all types of TBMs and improving the existing Tunnelling Int 6:32–35.
correlations with the significant operational output param- Grandori R, Jager M, Vigl, L [1995]. The Evinos
eters. Extending this work for more applications to mining TBM tunnel. In: Proceedings of the Rapid Excavation and
provides challenging opportunities. Tunneling Conference. American Society of Civil
Engineers, pp. 132–146.
ACKNOWLEDGMENTS Kalamaras ES, Bieniawski ZT [1995]. A rock mass
strength concept incorporating the effect of time. In: Pro-
This paper was made possible by the dedicated work ceedings of the ISRM Congress (Tokyo, Japan). Balkema,
of the professional staff at Geocontrol, S.A., Madrid, pp. 295–302.
Spain, including the first two recipients of the Bieniawski Kendorski FS, Cummings RA, Bieniawski ZT, Skin-
Scholarship for tunneling research at the Superior School ner EH [1983]. Rock mass classification for block caving
of Mines, Universidad Politécnica de Madrid: Doña María mine drift support. In: Proceedings of the 15th ISRM
Álvarez Hernández and currently Don José Carballo Congress (Melbourne, Victoria, Australia). Balkema, pp.
Rodrígez. These Bieniawski Scholarships were generously B101–B113.
funded by Geocontrol, S.A. Kirsten HAD [1982]. A classification for excavation.
Civ Eng S Afr 7:293.
REFERENCES Laubscher DH, Taylor HW [1976]. The importance of
geomechanics of jointed rock masses in mining operations.
Abrahão RA, Barton N [2003]. Employing TBM In: Proceedings of the Symposium on Exploration for
prognosis model. Tunnels & Tunnelling Int 35(12):20–23. Rock Engineering (Johannesburg, South Africa), pp. 119–
Alber M [1993]. Classifying TBM contracts. Tunnels 128.
& Tunnelling Int Dec:41–43. Palmström A, Broch E [2006]. Use and misuse of rock
Alber M [2001]. Advance rates for hard-rock TBMs mass classification systems with particular reference to the
and their effects on project economics. Tunnelling Undergr Q-system. Tunnelling Undergr Space Technol 21(6):575–
Space Technol 15(1):55–60. 593.
Barton N [1974]. Engineering classification by the Sapigni M, Berti M, Bethaz E, Bustillo A, Cardone G
Q-system. Rock Mech 6(4):183–236. [2002]. TBM performance estimation using rock mass
Barton N [2000]. TBM tunneling in jointed and classifications. Int J Rock Mech Min Sci 39:771–788.
faulted rock. Rotterdam, Netherlands: Balkema. Teale R [1965]. The concept of specific energy in rock
Bieniawski ZT [1972]. Engineering classification for drilling. Int J Rock Mech Min Sci 2:57–73.
rock masses. Rep S Afr Coun Sci Ind Res, ME 327/Rock Unal E [1983]. Design guidelines and roof control
Mech, Pretoria, South Africa. standards for coal mine roofs [Dissertation]. University
Park, PA: The Pennsylvania State University.

21
 APPENDIX A.—EXAMPLE OF RME CALCULATIONS

The figure below presents an example of the actual procedure for calculating RME07 for one of the case histories
plotted in Figure 8.

22
 APPENDIX B.—GENERAL GUIDELINES ON ROCK MASS CLASSIFICATIONS
(BASED ON THE RMR SYSTEM: 35 YEARS LATER)

The Rock Mass Excavability classification features Most of all—users, please beware! It is not recom-
one parameter—standup time—depicted in Figure 3, mended to apply any rock mass classification system on its
which is determined from the Rock Mass Rating (RMR), own, be it NATM, RMR, or Q. Instead, both RMR and Q
as shown in Appendix A. Since RMR was first introduced should always be used to cross-check the results and com-
in 1972 and published internationally in 1973, it is appro- pare recommendations, even if known correlations exist
priate to briefly summarize some of the lessons acquired between these two systems, which sometimes turn out to
about rock mass classifications after 35 years of use be oversimplifications.
throughout the world.
The most important aspect to remember is that the
main application of RMR is not just recommendations for
rock support (because they change as technology changes),
but estimation of rock mass properties for design and
numerical purposes, i.e., the modulus of deformation of the
rock mass, rock mass strength, and standup time. Figures
A–1 and A–2 depict these strength and deformation
relationships. In fact, the RMR case histories for these
purposes still remain the prime data for analyses and
correlations and are published in full [Bieniawski 1989].
There are three general guidelines to be observed for
good engineering practice:

1. Rock mass classifications, either quantitative sys-

tems, such as RMR and Q, or descriptive methods
(New Austrian Tunneling Method (NATM) or Geo-
logical Strength Index (GSI)), are most effective if
not used on their own, but incorporated within the
overall engineering design process.
2. Rock mass classifications on their own should only Figure A–1.—Correlation between the ratio of rock
be used for preliminary planning purposes and not as mass strength, σM, and UCS of rock material, σc, as a
final rock reinforcement. For preliminary design and function of RMR [Kalamaras and Bieniawski 1995].
planning purposes, the two quantitative RME and
Q-systems are excellently suited. They quantify rock
mass conditions, enable estimates of rock mass prop-
erties, and provide the reference bases for expected
rock mass conditions.
3. The two predominant quantitative rock mass classifi-
cations, RMR and Q, are particularly essential for
monitoring rock conditions during construction or
mining to enable effective comparison of predicted
conditions from site investigation with those encoun-
tered. For this purpose, descriptive classifications
(those not based on quantitative input data) are defi-
cient. They do not provide a continuous quantifi-
cation of the encountered conditions, even if based
on deformation measurements during construction,
because contractual specifications in many countries
prevent enough measurements to be taken since they
interfere with the mining or tunneling schedule.

Figure A–2.—Correlation between the modulus of

deformation of the rock mass, EM, and RMR.

23