You are on page 1of 4

EXCAVABILITY INDEX

Predicting TBM
excavability 40

+74% +44%
Open TBM

Advance rate (m/day)


30 -7%
-11%
20
-42%
Z T Bieniawski, of Bieniawski Design Enterprises, and Benjamín Double shield
10
Celada and José Miguel Galera, both of Geocontrol SA, Spain,
0
discuss the latest data for the newly introduced Rock Mass Tunnel
length 17% 32% 32% 17% 2%
Class by Less
Excavability (RME) index for the prediction of TBM advance rates RMR I II III IV V than V

80 60 40 20
RMR

ver 500 case histories were Top: Fig 1 - Evinos Tunnel advance rate

O analysed in two years to bring the


Rock Mass Excavability (RME)
index to fruition and demonstrate
versus rock mass quality rating RMR for
two TBM types (Grandori et al, 1995)
2.0
Fair Good Very good Good Fair

m=0.30 m=0.10
Tough

Maen
Pieve

Advance rate (m/h)


its effectiveness on five current projects. It Right: Fig 2 - Advance rates for three TBM
1.5 Varzo
should be emphasised that the RME index tunnels versus QTBM (Sapigni et al, 2002)
does not replace the RMR or Q systems as
1.0
used in tunnel design and construction; to drill and blast tunnels, independent of
indeed one of the RME input parameters, TBM characteristics.
0.5
the stand up time, is determined from the Today, this is no longer the case. TBMs
RMR. However, the approach reported here have increased in power, size and type to
introduces a specialised tool relevant to such an extent that they directly influence 0.1 1.0 10 100 1000 10000 100000
QTBM
TBM tunnel construction, featuring double- tunnel design. Moreover, they can be the
shield and open-type TBMs. source of tremendous satisfaction due to the excavability or performance prediction
This is Part I of two articles on this topic. machine’s increased safety provisions and models were studied by a number of
Part II will be published in a later issue of higher performance rates, as well as deep researchers since 2000: Barton[8], Alber[9],
T&TI and will include applications to single- despair when unexpected ground conditions Blindheim[10], Sapigni at al[11] and
shield TBMs as well as additional case are encountered and the TBM may be Palmström and Broch[12]. Last year, at the
histories leading to further improvements as immobilised for months and sometimes has ITA Congress in Korea, based on case
more experience is gathered with predicting to be rescued by old fashioned hand mining histories from over 400 tunnel sections in
TBM excavability. or drill and blast. In less drastic cases, TBM Spain, the authors introduced the Rock
progress in unfavourable conditions can be Mass Excavability (RME) index[13]. This year,
Previous investigations disrupted and the output decreased new case histories enabled further and more
In the history of tunnelling technology, the severely. specific applications for double-shield TBMs
emergence of modern TBMs stands out as In this situation, a significant problem has which was presented at the RETC in
the major milestone which provided emerged: how to assess effectively the Toronto[14]. Since submission of the RETC
spectacular advantages and achievements, interaction between rock mass conditions, paper, over 100 tunnel sections were
as well as complex challenges to designers as described by the RMR or Q classification analysed from one tunnel in Spain and two
and constructors who faced significant systems, and the design and performance tunnels in Ethiopia, forming the Gibe II
shortcomings in understanding the characteristics of the TBM. Certainly, some hydroelectric project (T&TI April).
interaction of rock mass conditions and TBM attempts to solve this problem have been Most of all, the timeliness of introducing
design and performance. made, but the responsibility still rests on the the RME was confirmed to date by an
In fact, when Terzaghi introduced his rock- TBM manufacturers and tunnel contractors increasing demand to be able to predict
load idea in 1946, followed by Lauffer’s who must rely on their experience, ingenuity TBM advance rates based on both rock
stand up time concept in 1958 and Deere’s and even their will to battle adverse mass quality and machine performance
RQD index in 1964, these design conditions. interaction, because the current indices
approaches were directed to selection of based solely on rock mass conditions have
rock reinforcement for tunnel construction by The meaning of TBM excavability proved insufficient[13].
drill and blast. The equipment for tunnel Excavability is defined as the rate of In this respect, figure 1 shows early
excavation was left to the discretion of the excavation expressed in machine promising research by Grandori[15] in 1995
contractor, with little input by the designer. performance in meters per day. It was involving the RMR for correlation with TBM
Even the subsequent 'modern' rock mass investigated as early as 1982 by Kirsten[4]. rate of advance. This provided some
classification methods, proposed in the Pioneering work by Tarkoy[5], Nelson[6] and interesting trends for the Evinos Tunnel in
1970s, by Wickham at al[1], Bieniawski[2] Ozdemir[7] was performed on rock borebility Greece, comparing the performance of open
and Barton[3] were predominantly directed and disk cutting in the 1990s. TBM TBMs and double-shield TBMs. It showed

SEPTEMBER 2007 Tunnels & Tunnelling International


EXCAVABILITY INDEX RECORD of DATA Form for Rock Mass Excavability RME

Name of Tunnel..............................................................................................................................
Initial chainage of section............................... Final chainage of section...................................
Length of section..............................m (should be > 40 m)
Duration of excavation (days)............................................................ (number + 1 decimal)

Average Rate of Advance ARA = ......................m/day


60
Average rate of advance (ARAT) (m/day)

Lithology...................................................................... Average depth....................m


50 Rock Mass Rating RMR: range................................... average....................
ARAT = 0.422RME07 – 11.61
40 R = 0.658
ROCK MASS EXCAVABILITY PARAMETERS

30 1) Uniaxial compressive strength of intact rock ( c


): range...........................average.............MPa

20 2) Drilling Rate Index DRI: range.................................... average.................................

3) Type of homogeneity at excavtionn face..............................Nº of joints per meter.......................


10
Orientation of discontinuities with respect to tunnel axis
0
0 10 20 30 40 50 60 70 80 90 100
(perpendicular, parallel or oblique)................................................................
RME07

Above: Fig 3 - Correlation between RME 4) Stand up time.....................hours 5) Groundwater inflow at tunnel face..............liters/sec
index and the average rate of advance
for double-shield TBMs (Bieniawski et Rock Mass Excavability RME range.....................................average.....................
al., 2006). The colour symbols represent
individual tunnels TBM PERFORMANCE PARAMETERS

Average speed of cutterhead rotation............................rpm Applied Thrust........................kN


that RMR Class III provided a peak in
production for double-shield TBMs, while Specific Penetration....................................mm/rev Torque.........................m. kN
they would not be recommended for neither Rate of Penetration........................................mm/min
Class V (very poor) nor Class I (very good)
Nº cutters changed:...................................... Rate of TBM utilization.............................%
rock masses. However, recent data reported
by Della Valle from Spain[16], showed that OTHER observations
an attempted correlation between the TBM
rate of advance and the RMR resulted in a Ex. in situ stress conditions, squeezing rock, rockbursts, etc......................................................
considerable scatter, although there was a .......................................................................................................................................................
trend similar to that discovered by Grandori.
Figure 2 shows the results from Norway of
an attempted correlation using the QTBM not the correct approach. Much better these aspects are important for comparing
index. It was found that the scatter of the correlations can be obtained with the new different tunnelling case histories.
results was so large and QTBM so complex, RME index. This is demonstrated in figure 3 Table 1 (above) shows the ‘record of data
by involving 21 parameters, that its use was depicting the work on RME[14] presented in form’ developed for two purposes; on
not recommended in the technical the RETC proceedings. existing projects, for correlation between the
literature[10,12]. We have also examined RME index and the TBM parameters; and for
QTBM as an option but could not make it RME input parameters planned projects, to estimate the TBM rate
work for our case histories, for a number of The database analysed includes the of advance, and some machine parameters.
reasons, including the problem of following information collected for each Table 1 includes therefore both the rock
determining the rock mass strength. tunnel section (geo-structural region): mass quality parameters, as well as the TBM
It became clear from these and other Geometrical data: tunnel location, length performance parameters. While all the input
analyses[13] that modifying the RMR or Q and diameter. parameters are customary and
rock mass quality classifications for Rock mass data: RMR and its input straightforward, it should be noted that the
prediction of rock mass excavability - for parameters, DRI, discontinuities data, and RMR is required to estimate the Stand-Up-
which these indices were not intended - is water inflow. Time parameter, while the Drilling Rate Index
TBM data: advance rate, utilisation (DRI) is needed because it is significant in
30
1d 1wk 1mo 1yr
90
10yr coefficient, penetration rate, rotation speed, tunnel boring construction. Here the DRI is
80
20
70 torque, thrust, no. of cutters changed and defined[17] in terms of the brittleness value
Immediate 60
collapse
ing
rat 50
the specific energy of excavation[14]. and the Sievers’ J-value. Both these tests
ss
Unsupported roof span (m)

10 a
m 40 It is important to note that a tunnel section have been standardised in the tunnelling
8 ck
6 Ro studied for RME is defined as a geological industry and are performed by commercial
30
5 structural region of the tunnel, that is, the or university laboratories.
4 RMR
20
3 60
70 80 same section for which the RMR is Figure 4 shows the RMR chart for
2 2 50 determined, and within which uniform estimation of the Stand-Up Time parameter.
40
characteristics exist, such as discontinuity Since this chart was originally developed for
No support
1 30 spacing and conditions, the same rock type drill-and-blast tunnels, a correlation is
required
20 lithology and the RMR not varying by more available between the RMRD&B and
0 than half-class (10 points). A section is not RMRTBM based on work by Alber[9] for
10-1 1 10 100 103 104 105 106
Stand-up time (hrs)
arbitrarily delineated by the number of full TBM tunnels. The following equation is
days tunnelled or by having only the same applicable:
Above: Fig 4 - Stand up time as a rock material. In fact, the section selected
function of RMR and unsupported tunnel should be preferably longer than 40m, the RMRTBM = 0.8 x RMRD&B + 20
span (Bieniawski, 1989). Red squares time of excavation should be given in days
represent tunnelling cases, green to one decimal fraction and the TBM Construction by TBM generally results in
squares are mining data utilisation should not be less than 30%. All higher RMR values than for the same tunnel

Tunnels & Tunnelling International SEPTEMBER 2007


EXCAVABILITY INDEX

Uniaxial compressive strength of intact rock [0-25 points] Table 3:


c (MPa) <5 5-30 30-90 90-180 >180 Effectiveness of the crew Crew factor (FE )
Rating 4 14 25 14 0 handling TBM and terrain

Drillability [0-15 points] Less than efficient 0.88


Efficient 1.00
DRI <80 80-65 65-50 50-40 <40
Very efficient 1.15
Rating 15 10 7 3 0

Discontinuities in front of the tunnnel face [0-30 points] Table 4:


Homogeneity Number of joints per meter Orientation with repect to tunnel axis Tunnel length excavated Adaptation
(km) factor (FA )
Homogeneous Mixed 0-4 4-8 8-15 15-30 >30 Perpendicular Oblique Parallel
0.5 0.68
Rating 10 0 2 7 15 10 0 5 3 0
1.0 0.80
Stand up time [0-25 points] 2.0 0.90
Hours 4.0 1.00
<5 5-24 24-96 96-192 >192
6.0 1.08
Rating 0 2 10 15 25
8.0 1.12
Groundwater inflow [0-5points] 10.0 1.16
Liter/sec >100 70-100 30-70 10-30 <10 12.0 1.20

Rating 0 1 2 4 5
60

Average rate of advance (ARAT) (m/day)


section excavated by drilling and blasting histories but also further considerations of ARAT = 0.661RME07 – 20.45
because of the favourable circular tunnel possible adjustment factors that might 50
R = 0.867
shape and lesser damage to the surrounding influence rock mass excavability and better
40
rock mass by the process of machine prediction of the TBM rates of advance. In
boring. the meantime, some considerations based 30
on TBM experience in Spain, are elaborated
Determination of the RME Index below. 20

The RME index is obtained from summation 1) Influence of the TBM crew:
10
of the five input parameters in Table 2 In tunnel construction, it seems evident
(above) which tabulates the ratings that the qualifications and experience of the 0
0 10 20 30 40 50 60 70 80 90 100
appropriate for the ranges listed. Note that TBM crew, who handle the tunnelling RME07
the values given are the average ratings, for machine every day, have an important Abdalajis Gilgel Gibe II - Inlet Guadarrama
a more precise determination of these input influence on the performances achieved.
ratings, convenient graphs can be found In order to include this effect, the Above: Fig 5 - RME data for three tunnels
elsewhere[12]. Once the RME is determined, experience gained during the construction of excavated with double-shield TBMs in
a TBM average rate of advance (ARA) may the Guadarrama tunnels was used and is rock with strength σc < 45MPa
be estimated from figure 3 or the latest ones defined as shown in Table 3.
that follow. In addition, other correlations 2) Influence of the excavated length: 60
have also been obtained by the authors[13]
Average rate of advance (ARAT) (m/day)

Again based on the data from the


ARAT = 0.839RME07 – 40.83
such as those between the RME and the Guadarrama tunnels, it is known that 50
R = 0.763
penetration rate (PR) for double-shield increased performance is obtained as the
40
TBMs, as well as correlations with the tunnel excavation increases. This factor may
specific energy of excavation (MJ/m3), thrust be defined as shown in Table 4. 30
and torque[14]. They will be discussed in 3) Influence of the tunnel diameter:
Part II of this series. It should be noted that the correlation in 20
figure 3 was derived for tunnels with
10
Recent analyses and results diameters close to 10m. In order to take into
After the introduction of the RME index last account the influence of different tunnel 0
0 10 20 30 40 50 60 70 80 90 100
year, comments were requested from the diameters, D, a coefficient FD was proposed
RME07
leaders in the field and their contributions, such that:
together with the new case histories that Above: Fig 6 - RME data for San Pedro
became available, led to ‘fine tuning’ of the FD = - 0.007D3 + 0.1637D2 – 1.2859D + 4.5158 Tunnel excavated with open-type TBM in
index (adjusting the ratings of some input rock with strength σc > 45MPa
parameters) and to re-runs of the correlation Therefore, for D = 10m, FD = 1, while for
analyses. Particularly valuable suggestions D = 8m, FD = 1.12 but for D = 12m, FD = 0.5,
were gratefully received from Dr Remo that is, one-half of the coefficient for D = ARAR
Grandori, President of SELI, Italy, Dr Evert 10m. ARAT =
Hoek, of Canada, Dr Sigismund Combining the effects as seen above, FE•FA•FD
Babenderede, of Germany, Dr Felipe the ‘real’ average rate of advance (ARAR)
Mendaña, of Spain, Dr Harvey Parker, of that is determined from tunnels whose
USA, and Ing. F Antonini of SELI. excavation diameter is different from 10m, Influence of the type of TBM
Most of all, the extensive use of the RME can be used to obtain the ‘theoretical’ To ensure full application of the RME, studies
index on its new tunnels by SELI, in Ethiopia, average rate of advance (ARAT), which is have been carried out to establish specific
provided not only new valuable case the one correlated with the RME: correlations between the ARA and the RME

SEPTEMBER 2007 Tunnels & Tunnelling International


EXCAVABILITY INDEX

60 Left: Fig 7 - Comparison of TBM 45MPa.


Average rate of advance (ARAT) (m/day)

Double shield working with performance versus excavability index In this case, in better exacavability
50 telescopic shield
RME for open and double-shield TBMs conditions, RME > 75, the use of open
40 boring in rock with strength σc < 45MPa TBMs clearly gives better performance than
by double shield TBMs.
30 figure 5 depicts the ARAT values, in m/day, However, in terrains whose excavability is
for three tunnels. The correlation co-efficient between 65 < RME < 75 both types of
20
of R=0.867 is significantly high. The data tunnelling machines provide similar results.
Open TBM
10 from one tunnel in Ethiopia has been kindly Finally, in terrains whose excavability is
provided by Dr Remo Grandori of SELI. between 50 < RME < 65, double-shields
0
0 10 20 30 40 50 60 70 80 90 100 For open-type TBM excavation of 49 allow better performance than open TBMs.
RME07 tunnel sections, figure 6 shows the ARAT
values for the San Pedro Tunnel in Spain, Future lines of research
excavated in rock having a σc > 45MPa. The To complete this research it is proposed to
60
Average rate of advance (ARAT) (m/day)

correlation co-efficient of R=0.763 is also analyse correlation between RME and ARAT
50 high. for single-shield TBMs, mainly in the tunnel
Open TBM
Analyses are not yet completed for sections excavated in low strength rocks, σc
40 tunnels excavated with single-shield TBMs. < 45MPa; where this kind of tunnelling
This will be featured in Part II of this series. machines can provide better results.
30
For this task, data will be used from
20 Double-shield and open TBMs sections excavated by single-shield TBMs
Double shield It was found that improved correlations are as well as those excavated working with
working with
10 telescopic obtained when one differentiates between double-shield TBM working in single-shield
shield
case histories featuring rocks with the mode.
0
0 10 20 30 40 50 60 70 80 90 100 uniaxial compressive strength of the intact These aspects and further case histories
RME07
rock σc > 45MPa and those σc < 45MPa. will be included in Part II of this series.
Above: Fig 8 - Comparison of TBM This is due to an increasing and decreasing
performance versus excavability index trend in the ratings of the σc parameter, Conclusion
RME for open TBMs and double-shield reaching the favourable average value for In closing, the case histories database for
TBMs in rock with strength σc > 45MPa TBM excavability at 45MPa. the Rock Mass Excavability (RME) index has
Figure 7 shows the regression lines which been increased significantly since its
for the typical hard rock TBMs, namely; open were obtained for the machines excavating introduction a year ago. The results obtained
TBMs; single shield TBMs; and double in ground with σc < 45MPa. to date are promising and we welcome
shield TBMs. It is clear that in this case double shield comments and suggestions to:
We started with the analyses of the tunnel TBMs always give better results than open prof-ztb@mindspring.com T&T
sections excavated by double shield TBMs TBMs, if the intact rock strength is less than
because most of the case histories in the 45MPa. ACKNOWLEDGMENTS
database involved this type of machine. Figure 8 shows the regression lines for the
Thus, for double-shield TBM construction, machines operating in the ground with σc > The authors wish to express their gratitude
to the Railway Infrastructure Corp in Spain
REFERENCES (ADIF) for allowing the use of the information
from the Guadarrama, Abdalajís and San
1. GE Wickham, HR Tiederman, EH Skinner, 1972. “Support determination based on geologic Pedro tunnels and to Dr Remo Grandori of
prediction” RETC, New York, p43 SELI, Italy for providing the data on tunnels
2. ZT Bieniawski, 1989. “Engineering Rock Mass Classifications - a Complete Manual” Wiley of hydroelectric complex Gilgel Gibe II, in
3. N Barton, 1988. “Rock mass classification using the Q-system” ASTM Tech Pub 984 Ethiopia.
4. HAD Kirsten, 1982. “A classification for excavation in natural materials”. Civil Eng in S Africa. They also wish to thank the Spanish
5. PJ Tarkoy, 1991. “Determining total hardness for TBM boreability” Canadian Symp Rock Mech construction companies ACS, FCC,
6. PP Nelson, 1993. “TBM performance analysis wrt rock properties” Comp Rock Eng, p261 Ferrovial-Agroman and Vias y
7. L Ozdemir, B. Nilsen, 1993. "Hard rock tunnel boring prediction". Proc. RETC, Boston, p.833. Construcciones; SACYR and San Pedro
8. N Barton, 2000. “TBM Tunneling in Jointed and Faulted Rock”. AA Balkema. UTE for their collaboration in the acquisition
9. M Alber, 2000. “Advance rates for hard rock TBMs”. TUST 15 (1). of data from Guadarrama, Abdalajís West
10. OT Blindheim, 2005. “A critique of QTBM”. T&TI, June ‘05. and San Pedro tunnels, respectively.
11. M Sapigni, M Berti, F Bethaz, A Bustillo, G Cardone, 2002. “TBM performance estimation using Finally, we wish to recognise the
rock mass classifications”. Int J Rock Mech Min Sci, 39, p771 dedicated contributions of the first two
12. A Palmström, E Broch, 2006. “Use and misuse of QTBM”. TUST 21, p575 recipients of the Bieniawski Scholarship for
13. ZT Bieniawski, B Celada, JM Galera, M. Álvares, 2006. “Rock Mass Excavability (RME) index” tunnelling research at the Superior School of
ITA World Tunnelling Congress, Korea. Mines, Universidad Politécnica de Madrid:
14. ZT Bieniawski, B Celada, JM Galera, 2007. “TBM Excavability: Prediction and Machine-Rock Doña María Álvarez Hernández - who was
Interaction”, RETC, Toronto, p1118 involved in the first year of this research,
15. R Grandori, M Jäger, F Antonini, L Vigl, 1995. “Evinos-Mormos Tunnel - Greece” RETC, San and subsequently Don José Carballo
Francisco, p747 Rodrígez - who developed the statistical
16. N Dela Valle, 2006. “Barcelona’s new backbone tunnel”. TBM Tunnelling Symposium, Madrid analysis of the data from the latest case
17. www/drillability.com/13A-98eng.pdf “Rock tunnel boring: Drillability test methods” histories.

Tunnels & Tunnelling International SEPTEMBER 2007

You might also like