In situ TBM penetration tests and rock mass boreability analysis in hard rock tunnels

Tunnelling and

www.elsevier.com/locate/tust

Tunnelling and Underground Space Technology 22 (2007) 303–316

Underground SpaceTechnologyincorporating Trenchless

Technology Research

In situ TBM penetration tests and rock mass boreability analysis inhard rock tunnels

Q.M. Gong a,*, J. Zhao b, Y.S. Jiang c

a School of Civil and Environmental Engineering, Nanyang Technological University, Singapore 639798, Singaporeb Ecole Polytechnique Federale de Lausanne (EPFL), Rock Mechanics Laboratory, CH-1015 Lausanne, Switzerland

c School of Mechanics and Civil Engineering, China University of Mining and Technology, Beijing 100083, China

Received 2 March 2006; received in revised form 22 June 2006; accepted 4 July 2006Available online 1 September 2006

Abstract

Boreability is popularly adopted to express the ease or difficulty with which a rock mass can be penetrated by a tunnel boring machine.Because the boreability is related to the rock mass properties, TBM specifications and TBM operation parameters, an accurately defin-able quantity has not been obtained so far. In order to analyze and compare rock mass boreability, a series of TBM shield friction testswere conducted in a TBM tunneling site. Two sets of TBM penetration tests were performed in different rock mass conditions duringtunneling in rock. In each step of the penetration test, the rock muck was collected to perform the muck sieve analyses and the shapeof large chips was surveyed in order to analyze the TBM chipping efficiency under different cutter thrusts. The results showed that acritical point exists in the penetration curves. The penetration per revolution increases rapidly with increasing thrust per cutter whenit is higher than the critical value. The muck sieve analysis results verified that with increasing thrust force, the muck size increasesand the rock breakage efficiency also increases. When the thrust is greater than the critical value, the muck becomes well-graded. Themuck shape analysis results also showed with the increase of the thrust, the chip shape changes from flat to elongated and flat. The bore-ability index at the critical point of penetration of 1 mm/rev. defined as the specific rock mass boreability index is proposed to evaluaterock mass boreability.� 2006 Elsevier Ltd. All rights reserved.

Keywords: TBM; Rock mass boreability; Penetration test; Shield friction test; Muck sieve analysis

1. Introduction

Three terms, namely cuttability, boreability and excav-atability, are usually used to describe relative performanceof rock cutting tools in a rock mass. Cuttability is mainlyapplied in coal mining and excavatability in rock slopeexcavation. Boreability is extensively used in rock tunnel-ing. US Commission on Engineering and TechnicalSystems (1984) defined boreability as a value expressingthe boring properties of rock in terms of the penetrationrate with certain numbers/types of cutters and amount ofpressure applied. It expresses the ease or difficulty with

0886-7798/$ - see front matter � 2006 Elsevier Ltd. All rights reserved.

doi:10.1016/j.tust.2006.07.003

* Corresponding author. Tel.: +65 67906895; fax: +65 67921650.E-mail address: [email protected] (Q.M. Gong).

which a rock type can be penetrated by a tunnel boringmachine (TBM). Howarth (1987) defined boreability asthe prediction of the penetration rate of a rock cuttingmachine in a rock mass. From the above definitions, onecan see that the boreability is a comprehensive variablerelated to the rock mass properties, machine specificationsas well as machine operation parameters. It is neither anabsolute nor an accurately definable quantity. When rockmass boreability is compared or calculated, a set of condi-tions must be included or indicated. To be a rock massparameter, rock mass boreability should not includemachine or operational parameters.

Rock mass boreability is a comprehensive parameterreflecting the result of the interaction between rock massesand a TBM. While TBM operation parameters (Thrust,torque and rotation speed) and TBM specifications (cutter

mailto:[email protected]

304 Q.M. Gong et al. / Tunnelling and Underground Space Technology 22 (2007) 303–316

number, cutter diameter, cutter spacing and geometricalarrangement) are specific or normalized, rock mass bore-ability may be described by rock mass conditions (rockmaterial properties, joint properties) and TBM perfor-mance (penetration rate, cutter wear) respectively. In termof rock mass properties, the third version of the NTNUmodel (Blindheim et al., 1983) described the rock massboreability based on three factors: drilling rate index(DRI), bit wear index (BWI), the intensity and orientationof weakness planes in the rock mass. According to thedescription, the rock mass boreability takes penetrationrate and cutter wear into consideration. Sundin and Wan-stedt (1994) proposed a boreability index to measure thefull face boreability of a rock mass based on rock massproperties. It is expressed by the result of rock indentationtest multiplied by a factor for joint and weakness planes.

Based on the machine performance, Wanner andAeberli (1979) proposed the term of specific penetrationto denote the rock mass boreability. The specific penetra-tion is the penetration per revolution, divided by thrustper cutter. Sundin and Wanstedt (1994) defined the valueas the boreability index. Hamilton and Dollinger (1979)used a field penetration index to describe the rock boreabil-ity. The field penetration index, defined as the ratio of theapplied thrust per cutter to the penetration per revolution,actually is the inverse of the specific penetration. In fact,the same parameters are used in these three indexes todescribe the rock mass boreability calculated from themachine performance. In this paper, the latter was adoptedto denote the rock mass boreability. It is noted that theboreability index is a normalized cutter force by revolutionper minute (RPM) and penetration rate. Thus, it facilitatesto compare the performance of different TBMs. But it isnot a rock mass index.

Hamilton and Dollinger (1979) found that the bore-ability index is a function of the thrust. It decreases withincreasing thrust per cutter. This is due to a change inthe efficiency of the cutting action at the cutterhead.Borg (1988) and Bruland (1998) found that a critical

Fig. 1. Simplified geological map of Singa

thrust must be applied to overcome the rocks inherentresistance against breaking. Below this critical thrustvalue almost no penetration rate can be achieved andabove this value the penetration rate increases rapidlywith the increase of thrust force. Therefore, the previ-ously defined boreability index calculated by the TBMperformance data can not accurately represent the rockmass boreability. Only when the thrust force remainssame, the calculated boreability index can demonstratethe different rock mass conditions.

In this paper, two sets of penetration tests were carriedout at tunnelling sites in Singapore, in order to analyzerock mass boreability. Before these tests, in situ shield fric-tion tests were conducted to decide the shield friction, inorder to calculate the actual thrust acted on the TBM cut-terhead and the thrust force per cutter. The sieve tests ofmuck were conducted to analyze the chipping efficiency.The muck shape was surveyed to obtain the effect of thrustforce on chipping. By analyzing the penetration test resultsand comparing with other TBM penetration tests at othersites, a specific rock mass boreability index (SRMBI) isproposed to evaluate the rock mass boreability in differentrock mass conditions. The SRMBI is defined as the bore-ability index at the penetration rate of 1 mm perrevolution.

2. Brief description of the sites

The deep tunnel sewerage system (DTSS) is a megainfrastructure project, aiming at long term solution to meetthe needs in wastewater conveyance, treatment and dis-posal in Singapore. The deep tunnel intercepts the flowsin existing gravity sewers, upstream of the pumping sta-tions, and routes the flows entirely by gravity to two newcentralized wastewater treatment plants located at thesouth-eastern and south-western coasts of Singapore. Thetreated effluent from the new treatment plants will be dis-charged through deep sea outfalls into the Singapore sea-water (Fig. 1).

pore and layout of the DTSS project.

Q.M. Gong et al. / Tunnelling and Underground Space Technology 22 (2007) 303–316 305

DTSS Contract T05, the Kranji Tunnel, is 12.6 km inlength, with an internal diameter of 3.6 m. The BukitTimah granite of different weathering grades is found alongthe tunnel alignment, between Seletar and Kranji, as shownin Fig. 1. The properties of granite with different weather-ing grades were studied by Zhao et al. (1994), Zhao et al.(1995) and Zhao (1996). DTSS Contract T06, the Queen-sway Tunnel, is 9.6 km in length, with an internal diameterof 3.3 m. The Bukit Timah granite of different weatheringgrades is found along most of the tunnel alignment. TheJurong Formation of interbedded, highly variable andmostly fractured mudstone, siltstone, sandstone, and con-glomerate rocks is found in a limited section (south end)of the T06 tunnel. The Old Alluvium that consists of indis-tinctly bedded sandy clay or clayey sand with a very stiff tovery dense consistency is found in the Northern end of thistunnel (Fig. 1). The tunnels are constructed by twoHerrenknecht EPB TBMs at each site with a bored diame-ter of 4.90 m and 4.45 m respectively. While the groundconditions expected are variable, these machines weredesigned so that both hard rock and soft ground couldbe excavated. The cutter head was equipped with a combi-nation of both rippers and disc cutters. Precast segmentsare installed as initial ground support. A final concrete lin-ing is added with a corrosion-resistant membrane.

3. Shield friction test

3.1. Forces acting on the cutterhead

In order to assess the rock mass boreability, the forceexerted by the TBM cutterhead on the rock mass need tobe calculated. The forces acting on the front shield of theTBM are shown in Fig. 2. In this figure, Fsteer denotesthe total force of steering cylinders (kN), which is adoptedto calculate the thrust force per cutter based on the specifi-cations of these machines; Frock is equal to the force actedon rock mass by the cutterhead (kN); Ffront represents thefriction force of front shield (kN); Wfront stands for thetotal weight of front shield (kN); Fep is the force from earthpressure (kN); and Fsupport is the force supporting theshield by the tunnel invert. According to equilibrium offorces in the horizontal direction, the equation is shownas follows:

Frock

Fep Wfront

Fsteer

Front Shield

Center

Ffront

Fsupport

Fig. 2. Schematic representation of mec

F steer ¼ F rock þ F ep þ F front ð1ÞIn Eq. (1), the steering force and the earth pressure actedon the cutterhead are available from the TBM records.When the friction force of the front shield is known, theforce acted on the rock mass can be calculated. In situ teststherefore were performed to find out the friction force.

3.2. In situ tests and results

When a ring length of the tunnel is excavated, the TBMis stopped to install ring segments in the rear shield. Afterseveral rings excavated, the cutters in the cutterhead areusually thoroughly checked. At this moment, the shieldfriction tests can be carried out without interfering the nor-mal tunnelling progress. The test procedure is outlinedbelow:

(1) The muck in the cutterhead chamber is emptied first.(2) The cutterhead is then extracted from the tunnel face

to a distance dependent on the capacity of the steer-ing hydraulic cylinders.

(3) The cutterhead is pushed forward to the tunnel facewith cutterhead rotation similar to the normalexcavation.

(4) During the extracting and pushing forward process,the variations of the total steering force, the positionsof the steering cylinders with time are recorded.

During the process, the steering cylinders must keepmovement in the same pace. The uneven movement ofthe steering cylinders results in the breakage of the rockmass in the tunnel wall by cutterhead and discs. Particu-larly the peripheral discs are close to and easily crush intothe tunnel wall, when the cutterhead is pushed forward.The uneven movement among different cylinders alsoresults in the internal confrontation of these cylinders.The position variation of each steering cylinder from thebeginning to the end during the test represents the realmovement of the cutterhead, as shown in Figs. 3, 5 and7. These cylinders may not be located at the same positionat the beginning. Thus, they have different original records.

A series of retracting and pushing tests were carried outin T06 north tunnel. The shield friction force was

Shield Rear Shield

Total Thrust

hanical forces at TBM front shield.

100

105

110

115

120

125

03:05:50 03:06:10 03:06:30 03:06:50

Time (hh:mm:ss)

Cyl

inde

r po

sitio

n (m

m)

A B

C D

Fig. 3. Movement of steering cylinders with time at Ring N1690, T06 inretracting front shield friction test (A–D denote the different set of steeringcylinders).

0

50

100

150

200

250

03:05:50 03:06:10 03:06:30 03:06:50

Time (hh:mm:ss)

Thr

ust (

kN)

0

20

40

60

80

100

120

Tor

que

(kN

m)

Thrust

Torque

RPM=5.0 Min.=36 kN

Fig. 4. Variation of the total steering thrust and torque at Ring N1690,T06 in retracting front friction test.

130

150

170

190

210

230

16:06:16 16:06:36 16:06:56 16:07:16 16:07:36

Time (hh:mm:ss)

Cyl

inde

r po

sitio

n (m

m)

A B

C D

Fig. 5. Movement of steering cylinders with time at Ring N2014, T06 inretracting front shield friction test (A–D denotes the different set ofsteering cylinders).


calculated respectively from these retracting tests and push-ing tests. Some test results were ignored due to the unevenmovement of the cylinders. The calculated front shield fric-tion forces from the different ring tests are listed in Tables 1and 2 and some samples of the data analysis are shown in

Figs. 3–8. Some points shown in these figures are abnormaldue to the uneven movement of the steering cylinders. Inthe calculation of the average value of the friction force,these points are not taken into account.

It can be seen from Tables 1 and 2 that the frictionforces obtained from the retracting tests are much less thanthat from the pushing tests. The torque for retracting thecutterhead is also much less than that for pushing. Theresults from the pushing cutterhead friction tests indicatethe process of pushing the cutterhead forward is more com-plicated than the retracting process. In the process of push-ing the cutterhead, more factors besides the friction forceinfluence the cutterhead movement. The fringe of the shieldand the gauge cutters may cause local instability of the tun-nel wall. Some rock chips fallen from the cutterhead mayblock the clearance between the shield and the tunnel wall.This increases greatly the force needed for pushing the cut-terhead. Although the TBM is in the condition of pushingthe cutterhead during excavation, the above mentioned fac-tors influencing the pushing force are almost non-existentin the continuously tunnelling process. Because the cutter-head closely contacts with the tunnel face, it prevents thelocal instability of the tunnel wall, and the excavated rockblocks and chips from the tunnel face are collected bybuckets in time. In addition, the cutterhead vibration ben-efits its movement during boring.

Theoretically, the friction force should be the same forboth the pushing and retracting processes in these tests.The friction force in the retracting friction tests is thereforeassumed to represent the true friction force between thetunnel sidewall and the front shield.

Based on the calculated results, the minimum valueranges from 36.0 kN to 87.32 kN and their average is about50 kN. Several factors may contribute to the variation ofthe friction force, such as the granite weathering grade,groundwater condition and obstacles in front of the cutter-head. For the actual estimation, a friction force of 50 KN isa reasonable value to be used in the calculation of the bore-ability index.

4. TBM penetration tests

4.1. Stepped TBM penetration test in T05

A stepped TBM penetration test was conducted in thenorth drive of the T05 site. Before the penetration test,the TBM cutters were checked. The cutter wear conditionswere recorded and the cutters which satisfied the replace-ment criterion were replaced. The cutter condition is listedin Table 3. The cutterhead chamber was emptied by thescrew conveyor.

After the cutterhead chamber was emptied, the tunnelface mapping was carried out through the four buckets inthe cutterhead designed to collect the muck. The four buck-ets are evenly distributed in the cutterhead. The tunnel facecan be well observed from the four openings. The tunnelface mapping at the test ring is shown in Fig. 9. Photos

Table 1Results for retracting friction tests in T06

Ring No. Rock mass Friction force (kN) Torque (kN m)

Minimum Average Maximum Minimun Average Maximum

N1690 Fresh to slightly weathered granite 36.0 43.15 72.94 37.66 52.32 64.34N1988 Fresh to slightly weathered granite 87.32 104.53 111.98 – – –N2014 Fresh to slightly weathered granite 38.0 73.52 200.33 40.45 55.84 74.45N2129 Fresh to slightly weathered granite 36.0 98.28 266.06 27.33 57.56 97.26

Table 2Results for pushing friction tests in T06

Ring No. Rock mass Friction force (kN) Torque (kNm)

Minimum Average Maximum Minimum Average Maximum

N1745 Moderately to highly weathered granite (UCS, 20.57 MPa) 962.58 1224.25 1494.73 68.21 79.56 103.5N2074 Fresh to slightly weathered granite 919.44 1079.70 1355.01 61.75 108.7 154.06N2124 Fresh to slightly weathered granite 982.1 1053.57 1257.49 160.73 191.29 253.47

0

500

1000

1500

2000

2500

16:06:16 16:06:36 16:06:56 16:07:16 16:07:36

Time (hh:mm:ss)

Thr

ust (

kN)

0

10

20

30

40

50

60

70

80

Tor

que

(kN

m)

Thrust

Torque

RPM=7.3 Min.=38 kN

Fig. 6. Variation of the total steering thrust and torque at Ring N2014,T06 in retracting front friction test.

140

150

160

170

180

190

200

210

17:47:52 17:48:12 17:48:32 17:48:52 17:49:12

Time (hh:mm:ss)

Cyl

inde

r po

sitio

n (m

m)

A B

C D

Fig. 7. Movement of steering cylinders with time at Ring N2124, T06 inpushing front shield friction test (A–D denotes the different set of steeringcylinders).

0

200

400

600

800

1000

1200

1400

1600

1800

17:47:52 17:48:12 17:48:32 17:48:52 17:49:12

Time (hh:mm:ss)

Thr

ust (

kN)

0

50

100

150

200

250

300

350

400

450

Tor

que

(kN

m)

Thrust

TorqueRPM=1.6 Min.=982 kN

Fig. 8. Variation of the total steering thrust and torque at Ring N2124,T06 in pushing front friction test.


taken from the cutterhead chamber showing the rock massconditions of the upper right and left openings are pre-sented in Figs. 10 and 11. The rock mass is composed of

pink light grey, fresh granite. One joint set is observed inthe rock mass. The joint plane is smooth and undulating,and the joint wall is slightly altered with non-softeningmineral coating. The joint spacing is about 400 mm. Theangle between the tunnel axis and the joint plane is about30–40�. The tunnel face was dry during excavation. Coresamples were taken from the tunnel face. Uniaxial com-pressive strength tests, point load tests and Brazilian ten-sile strength tests were conducted in the laboratory.The rock uniaxial compressive strength corrected to thediameter 50 mm is 172.9 MPa, the point load strength(Is50) is 7.84 MPa, and the Brazilian tensile strength is10.55 MPa.

The penetration test was divided into seven steps, asshown in Table 4. Before Step 1, the TBM was operatedfor 5 min (step 0) with full thrust to verify that the TBMis in normal operation condition. It is also useful for thecomparison of the test results. The TBM was then operatedat different total thrust levels, from 40% to 100% of thedesigned maximum cutter load. According to the sugges-tions by Bruland (1998), the penetration test duration

Table 3TBM cutter wear conditions in T05 penetration test

Tool type Tool position Tool wear (mm)

Double cutters 1/3 10/62/4 8/65/7 5/46/8 10/12

Face single cutters 9 110 1411 812 913 1214 1515 1316 1417 1518 719 720 121 9

Transition single cutters 22 023 1524 025 1026 127 028 529 830 10

Gauge cutters 31 332a 432b 7

Remarks: the maximum allowable cutter wear is 20 mm. Thus, the overallcutter wear is normal.

RQ

D=

100No joints

1

No jointsNo joints

Spacing=400 mm

Pink light grey,fresh granite

Fig. 9. Tunnel face map before TBM penetration test in T05.

Fig. 10. Tunnel face at upper right opening in the tested ring of T05tunnel.

Fig. 11. Tunnel face at upper left opening in the tested ring of T05 tunnel.


should be a time corresponding to approximately 30 revo-lutions of the cutterhead at each thrust level. In order toobtain a high quality test result, the test duration at eachthrust level was designed to be 10 min that include the timeused to stabilize the thrust force at the beginning of each

step, and the speed, in terms of revolution per minute(RPM), was fixed to 10, corresponding to 100 revolutionsper step.

Due to the extremely low advance rate in the first threesteps, the test duration was changed to 5 min. The thrustand torque variations with time are shown in Figs. 12and 13. From the figures, it can be seen that the thrustwas not stable during the first two steps. Thus, these dataare not used to analyze the penetration rate.

The penetration test results are listed in Table 5. In eachstep, the start time, end time, RPM, torque and the steeringforce were recorded automatically in the TBM data acqui-sition system. The time duration for each step, penetrationper revolution, force per cutter and rock mass boreabilityindex are calculated based on the above data. It shall benoted that the total number of cutters installed in the cutter-head is 33, including double cutters, face cutters, transitioncutters and gauge cutters. The uneven distribution of theforce in different type of cutters is not considered. The fric-tion force between the front shield and the tunnel wall isdeducted from the computation of the force per cutterand rock mass boreability index.

The correlations between force per cutter, torque andTBM penetration are shown in Figs. 14 and 15. The pene-tration increases with increasing torque and force per

Table 4Design of the penetration test in T05 tunnel

TBM penetration test Thrust level (%) Thrust force inmain cylinder (kN)

Average thrust force insteering cylinder (bar)

Test time (min) Muck samples (>30 kg)

Step 0 100 9200 165 5 Yes (No. 0)Step 1 40 3896 95 10 Yes (–)Step 2 50 4870 100 10 Yes (–)Step 3 60 5633 103 10 Yes (No. 3)Step 4 70 6453 108 10 Yes (No. 4)Step 5 80 7450 121 10 Yes (No. 5)Step 6 90 8484 142 10 Yes (No. 6)Step 7 100 9200 165 10 Yes (No. 7)

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10:38:56 10:48:56 10:58:56 11:08:56 11:18:56 11:28:56 11:38:56 11:48:56

Time (hh:mm:ss)

Tot

al s

teer

ing

thru

st(k

N)

Fig. 12. Average steering thrust variation with time in T05 penetration test.

0

500

1000

1500

2000

2500

3000

3500

4000

10:38:56 10:48:56 10:58:56 11:08:56 11:18:56 11:28:56 11:38:56 11:48:56

Time (hh:mm:ss)

Tor

que

(kN

.m)

Fig. 13. Torque variations with time in T05 penetration test.

Table 5Penetration test results in T05 tunnel

Teststep

Start time(hh:mm:ss)

End time(hh:mm:ss)

Duration(min)

Advance(mm)

Penetration(mm/rev.)

Torque(kNm)

Average steeringforce (bar)

Force per cutter(kN/cutter)

BI ((kN/cutter)/(mm/rev.))

0 10:39:36 10:44:06 4.5 70 1.62 1696.82 163.05 237.12 146.493 10:56:46 11:02:26 5.67 10 0.18 667.51 101.94 147.68 817.254 11:03:46 11:13:06 9.33 26 0.29 673.93 107.81 156.28 543.425 11:13:16 11:22:56 9.67 44 0.47 810.29 119.69 173.67 368.326 11:24:06 11:30:46 6.67 59 0.92 1199.27 140.46 204.07 222.407 11:31:26 11:56:36 25.17 481 1.98 2432.27 163.30 237.50 120.05


0.0

0.5

1.0

1.5

2.0

2.5

100.00 130.00 160.00 190.00 220.00 250.00

Force per cutter (kN)

Pene

trat

ion

(mm

/rev

)

Fig. 14. Variation of penetration with the force per cutter in T05penetration test.

0.0

0.5

1.0

1.5

2.0

2.5

500 1000 1500 2000 2500Torque (kNm)

Pene

trat

ion

(mm

/rev

)

Fig. 15. Variation of penetration with the torque in T05 penetration test.

Table 6Penetration test results in T06 tunnel

Time(hh:mm:ss)

Penetration(mm/rev.)

Torque(kN m)

Thrust percutter (kN/cutter)

BI (kN/cutter/mm/rev.)

15:28:54 1.998 155.783 77.059 38.56815:29:04 1.995 129.102 78.972 39.58515:29:14 1.739 113.825 74.863 43.04915:29:24 2.366 135.342 79.149 33.45315:29:34 1.988 140.075 74.473 37.46115:29:44 1.993 163.314 79.362 39.82015:29:54 2.606 168.478 79.504 30.50815:30:04 2.245 145.455 74.260 33.07815:30:14 2.494 116.837 78.405 31.43815:30:24 2.860 134.696 82.656 28.90115:30:34 2.735 161.808 83.931 30.68815:30:44 2.987 139.215 87.261 29.21415:30:54 3.367 168.263 83.790 24.88615:31:04 2.999 151.049 85.879 28.63615:31:14 3.250 174.933 90.130 27.73215:31:24 3.738 176.009 90.272 24.15015:31:34 3.742 223.992 91.264 24.38915:31:44 4.228 210.221 94.311 22.30615:31:54 4.117 215.170 90.096 21.88415:32:04 4.598 254.976 91.689 19.94115:32:14 4.610 261.646 97.145 21.07315:32:24 4.743 242.926 96.330 20.31015:32:34 4.983 245.724 92.965 18.65615:32:44 4.880 228.080 97.464 19.97215:32:54 4.876 247.015 103.237 21.17315:33:04 5.380 270.253 97.287 18.08315:33:14 5.223 271.974 99.447 19.04015:33:24 5.969 283.809 101.573 17.01715:33:34 6.474 280.796 98.455 15.20815:33:44 6.218 276.923 101.714 16.35815:33:54 5.860 265.089 98.881 16.87415:34:04 6.116 288.327 98.030 16.02915:34:14 5.974 249.166 108.870 18.22415:34:24 5.614 261.861 105.257 18.74915:34:34 5.991 326.843 108.091 18.04215:34:44 6.972 360.839 110.287 15.81915:34:54 7.483 343.626 109.968 14.69615:35:04 7.758 359.979 113.192 14.59015:35:14 8.737 378.698 110.358 12.63115:35:24 8.119 347.069 111.845 13.77615:35:34 6.720 362.561 110.712 16.47515:35:44 7.843 358.903 111.811 14.25615:35:54 8.220 386.660 115.495 14.05015:36:04 7.119 371.598 112.590 15.81515:36:14 7.358 355.030 114.750 15.59515:36:24 8.743 422.163 121.764 13.92715:36:34 8.339 481.119 117.832 14.13015:36:44 7.878 429.909 115.353 14.64215:36:54 9.112 470.576 112.767 12.37615:37:04 8.876 406.455 124.138 13.98615:37:14 7.768 381.926 116.309 14.97315:37:24 8.862 416.784 120.702 13.62015:37:34 8.402 405.810 118.612 14.11715:37:44 8.465 460.033 121.977 14.41015:37:54 8.876 461.539 116.097 13.08015:38:04 8.317 421.732 125.449 15.083


cutter. From Fig. 14, it can be seen that there exists a crit-ical point in this curve. When the force per cutter is higherthan the critical value, the penetration increases rapidly.For the tested rock mass, fresh granite with few joints,the critical point where the granite is fragmented effectivelyis at the penetration rate between 0.5 mm/rev. and 1.0 mm/rev. With the increase of the torque, the increase of the pen-etration shows a different tendency, as shown in Fig. 15.Initially, the penetration almost linearly increases withincreasing torque. Then, when the penetration is more thanthe critical value and the TBM operates efficiently, the tor-que increases slowly compared with the previous stage.Because the thrust force is more than the critical value,the increase of the thrust force makes the cutter indentationprocess more effective and the interaction between two cutsis strengthened. The induced lateral cracks under the actionof the neighbouring cutters can reach each other and coa-lesce to form rock chips more frequently (Gong et al.,2006b). With the increase of the penetration, the rock chip-ping force acted on the two sides of the cutters has toincrease, which also leads the increase of chipping fre-quency. Thus, the increment rate of the rolling force isslower than that of the penetration rate. Correspondingly,the increment rate required to increase the torque is gettingslower with increasing penetration rate.

4.2. Continuous TBM penetration test in T06

A simple penetration test slightly different from the onein T05 was performed in the north drive of the T06 tunnel.In this test, the force per cutter was increased continuously,

0.0

2.0

4.0

6.0

8.0

10.0

12.0

0 30 60 90 120 150

Thrustper cutter (kN)

Pene

trat

ion

(mm

/rev

)Fig. 18. Relation between thrust per cutter and penetration in T06penetration test.


and correspondingly the torque was also increased contin-uously. Then, the TBM penetration per revolutionincreased continuously. The results are listed in Table 6.The rock mass is composed of pink grey granite with rockmaterial uniaxial compressive strength close to 120 MPa.The rock mass is slightly to moderately weathered. TheRQD is close to 80 in the tunnel face and the joint spacingclose to 200 mm, as shown in Fig. 16.

The test results are shown in Figs. 17 and 18. Withincreasing torque and thrust per cutter, the penetrationper revolution increases. The correlation between the thrustper cutter and the penetration also shows that there exists acritical value in the correlation curve. Because the rockmass is relatively easier to excavate, the critical thrust forceis low. The correlation between torque and penetrationshows the same tendency with the stepped penetration testresult in T05.

4.3. Discussions on the boreability index (BI)

The correlation between the boreability index (BI) andthe penetration in the stepped and continuous penetrationtests are respectively plotted in Figs. 19 and 20. As can beseen, the boreability index for a rock mass is not a con-

Fig. 16. Tunnel face in the tested ring of T06 tunnel.

0.0

2.0

4.0

6.0

8.0

10.0

0 100 200 300 400 500

Torque (kNm)

Pene

trat

ion

(mm

/rev

)

Fig. 17. Relation between torque and penetration in T06 penetration test.

stant. The rock mass boreability index decreases withincreasing penetration rate. This phenomenon was notedby Hamilton and Dollinger (1979). Although the BI wasadopted to evaluate the rock mass boreability (Wannerand Aeberli, 1979; Sundin and Wanstedt, 1994), it was usu-ally used to analyze the variation of rock mass conditionsunder the same machine operation parameters. In the

y = 208x-0.7856

R2 = 0.9992

0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

800.00

900.00

0.0 0.5 1.0 1.5 2.0 2.5 3.0Penetration (mm/rev)

BI

((kN

/cut

ter)

/(m

m/r

ev))

Fig. 19. Variation of boreability index with the penetration in T05penetration test.

y = 61.205x-0.7009

R2 = 0.9891

0

10

20

30

40

50

0.0 2.0 4.0 6.0 8.0 10.0Penetration (mm/rev)

BI

((th

rust

/cut

ter)

/(m

m/r

ev))

Fig. 20. Variation of boreability index with penetration in T06 penetra-tion test.


process of the TBM penetration, a critical thrust must beexerted in order to effectively fragment the rock (Borg,1988; Bruland, 1998). Otherwise, almost no penetrationrate can be achieved. Gong et al. (2005) and Gong et al.(2006a) simulated the rock chipping process under TBMcutters using UDEC modelling. Only when the cutter loadis more than the critical load, the chipping between twocutters can occur. With the increase of the cutter load,the chipping efficiency increases (Bruland, 1998; Gonget al., 2006b). Thus, the BI is relevant to the chipping effi-ciency under different cutter loads. It can not indepen-dently stand for the rock mass boreability.

From the above test results shown in Figs. 19 and 20,these data points can be well fitted into a power function.The correlation coefficient in the stepped penetration testis as high as 0.9996 and that in the continuous penetrationtest is up to 0.9945. From these functions, it is easy toobtain the boreability index for the penetration rate of1 mm per revolution. It is close to the critical point anddoes not change with the force acted on the tunnel face.The boreability index at penetration 1 mm/rev. is not rele-vant to the machine operation conditions (thrust force,

Fig. 21. Chip shape (a) the longest axis, (b) the middle axis, (c) theshortest axis.

0

10

20

30

40

50

60

70

80

90

100

0.01 0.1

Perc

enta

ge p

assi

ng (

by w

eigh

t)

No.3 No.4

No.5 No.6

No.7 No.0

SandSilt

Grain s

Fig. 22. Rock chip size distribution for differ

RPM and torque) and eliminates the influence of the oper-ation uncertainties on the rock mass boreability. In fact,the index is only dependent on the rock mass conditionsand cutterhead design, especially the cutter diameter, cutterspacing and cutter tip width. Consequently, the boreabilityindex at penetration 1 mm/rev., defined as specific rockmass boreability index (SRMBI) can be used to evaluatethe rock mass boreability. The SRMBI in the steppedand continuous penetration tests is respectively 208 and61.21 kN/cutter/mm/rev.

Compared with the stepped penetration test popularlyused to evaluate the machine performance in a given geol-ogy (Bruland, 1998; Buchi, 2004), the continuous penetra-tion test has never been mentioned before. It was thoughtthat time duration is needed to stabilize the machine per-formance. The result from the continuous penetration testshowed good agreement with that from the stepped pene-tration test. It verified that the cutter force can be reflectedby the instantaneous penetration rate.

5. Muck sieve and shape analysis

In order to study the chipping efficiency at differentthrust levels, the muck samples were collected for sieve testsand muck analysis at each step in the T05 penetration test,as shown in Table 4. To ensure the muck samples to be rep-resentative, the samples were directly taken from the con-veyor belt 3 min after the beginning of each step. In thistest, due to extremely low advance for the first two steps,the corresponding muck samples were not collected.

During the penetration test, muck samples of more than30 kg were taken from the conveyor belt at each test step.These samples were then transported to the laboratoryfor sieve analysis. The sieve analysis was conducted follow-ing the BS code 1377 (1990). Based on the chip shape, asdefined in Fig. 21, the middle axis (b axis) size distributioncan be obtained from the standard sieve analysis. The dis-tribution curves for samples collected from the different test

1 10 100

Gravel Stone

ize (mm)

ent thrust forces in T05 penetration test.

0

20

40

60

80

100

120

140

160

180

0 5 10 15 20 25

Chip number

The

long

est a

xis

leng

th (

mm

)

No.3 No.4No.5 No.6No.7 No.0

Fig. 23. The longest axis length distribution of the largest 25 chips in T05penetration test.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Ratio of the middle axis to the longest axis

Rat

io o

f th

e sh

orte

st a

xis

to th

e m

iddl

e ax

is

No.3 No.4No.5 No.6No.7 No.0

elongated cubic

elongated and flat flat

Fig. 24. Rock chip shape variation for different thrust forces (centre point:mean value; cross: standard deviation) in T05 penetration test.

y = 120.67x-0.6017

R2 = 0.9885

y =y = 170.23x-0.7367

R2 = 0.9996

0

100

200

300

400

500

600

700

800

900

1000

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Penetrati

BI

((kN

/cut

ter)

/(m

m/r

ev))

Fig. 25. Rock mass boreability index va


steps are shown in Fig. 22. With increasing thrust force andconsequently increasing penetration per revolution, thechip size increases. Only in Steps 0 and 7, the muck is wellgraded. That is to say, any increase of the thrust force willresult in the improvement of the TBM efficiency on rockfragmentation. The TBM can efficiently fragment the rockmass only if the force per cutter reaches a critical value.The sieve analysis results also verified that there exists acritical point for a rock mass in the TBM penetrationcurve. When the thrust is more than the critical value,the rock mass can be efficiently fragmented. Otherwise,the rock is crushed into rock powder and small pieces.

After the sieve analysis, the large chips with longest axismore than 37.5 mm were taken for further shape analysis ateach test step. The dimensions of every chip, as defined inFig. 21, were measured. The longest axis (a axis) distribu-tion of the largest 25 rock chips in every step is drawn inFig. 23. The length of the large chips increases generallywith increasing penetration per revolution. Due to themulti-pass cutting, the large chips were also formed inSteps 5 and 6. The chip shape was analyzed based on theratio of the measured three axes. The chip shape can bedescribed in four types (Buchi, 2004), namely (1) flat, (2)elongated and flat, (3) elongated, and (4) cubic accordingto the ratio of the middle axis to the longest axis and theratio of the shortest axis to the middle axis, as shown inFig. 24. In order to find out the shape variation in differentsteps, the largest 25 chips in the longest and middle axis areselected to obtain the statistical mean value and standarddeviation. The results are also shown in Fig. 24. Withincreasing penetration per revolution, the chip shapechanges slowly from flat to elongated and flat. This alsoshows the rock breakage efficiency increases with increas-ing penetration per revolution.

y = 208x-0.7856

R2 = 0.9992

y = 61.205x-0.7009

R2 = 0.9891

168.32x-0.7798

R2 = 0.9964

5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0

on (mm/rev)

granite

granite-gneiss

sandstone with quatzite

Bukit timah granite (T05)

Bukit timah granite (T06)

riations with different penetrations.


6. Evaluation of rock mass boreability

Because the TBM performance databases are usuallytreated as proprietary, used for consultation, the test dataare regarded as confidential. Almost no test data can befound in the published papers. Only three penetration testresults conducted in granite-gneiss, granite and sandstonewith quartzitic matrix outside Singapore were collected(Buchi, 2004). The correlation between rock mass boreabil-ity index and penetration is fitted very well to the powerfunction, as shown in Fig. 25. The critical point is alsolocated at the penetration between 0.5 mm/rev. and1.0 mm/rev. The rock mass boreability index increases rap-idly with decreasing penetration when the penetration is

Table 7Deviation analysis while using the specific rock mass boreability index with a

Rock type Thrust percutter (kN)

Penetration perrevolution (mm)

Bore(kN/

Granite-gneiss 175.00 2.93 5987.50 0.47 186

105.00 0.86 122122.50 0.90 136140.00 1.29 108157.50 1.76 89175.00 2.43 72192.50 3.17 60210.00 4.17 50

Granite(Emolweni tunnel)

75.00 0.04 1875100.00 0.14 714125.00 0.36 347150.00 0.60 250175.00 1.12 156200.00 1.84 108220.00 2.72 80240.00 3.33 72

Sandstone withquartzitic matrix

100.00 0.14 714120.00 0.22 545140.00 0.34 411160.00 0.56 285180.00 1.15 156200.00 2.69 74220.00 3.84 57100.00 0.14 714

Granite(T05, DTSS)

237.12 1.62 146147.68 0.18 817156.28 0.29 543173.67 0.47 368204.07 0.92 222237.50 1.98 120

Granite(T06, DTSS)

74.86 1.74 4379.50 2.61 3083.79 3.37 2491.69 4.60 1997.14 4.61 2197.29 5.38 1898.46 6.47 15

109.97 7.48 14110.36 8.74 12112.77 9.11 12

less than the critical point. On the contrary, the rock massboreability index decreases slowly with increasing penetra-tion when the penetration is more than the critical value. Itis similar to the results of the penetration tests conducted inDTSS T05 and T06, as shown in Fig. 25. Using the powerfunction, one can easily obtain the rock mass boreabilityindex at the penetration per revolution of 1.0 mm/rev.,namely the specific rock mass boreability index. Since therock mass boreability index is the force per cutter normal-ized by the penetration per revolution and the specific rockmass boreability index is the specific BI at penetration perrevolution of 1 mm/rev., it is a normalized value and elim-inates the influence of the operation uncertainties (thrustforce, RPM and torque), especially the variation of thrust

power of 0.75 to calculate the boreabiltiy index

ability indexcutter/mm/rev.)

Predicted boreabiltiyindex (kN/cutter/mm/rev.)

Deviation (%)

.73 53.88 9.79

.17 212.58 �14.19

.09 135.12 �10.67

.11 130.59 4.05

.53 99.69 8.14

.49 78.97 11.75

.02 62.00 13.91

.73 50.79 16.36

.36 41.35 17.89

.00 1903.23 �1.51

.29 743.77 �4.13

.22 366.28 �5.49

.00 249.70 0.12

.25 156.36 �0.07

.70 107.75 0.87

.88 80.37 0.63

.07 69.06 4.18

.29 735.43 -2.96

.45 523.98 3.94

.76 378.03 8.19

.71 260.01 9.00

.52 151.57 3.16

.35 80.13 7.78

.29 61.36 �7.10

.29 735.43 �2.96

.49 144.94 1.06

.25 750.48 8.17

.42 529.65 2.53

.32 365.54 0.75

.40 221.86 0.24

.05 124.70 �3.87

.05 40.42 6.12

.51 29.84 2.19

.89 24.62 1.05

.94 19.49 2.25

.07 19.45 7.68

.08 17.33 4.19

.21 15.08 0.84

.70 13.53 7.95

.63 12.04 4.65

.38 11.67 5.70


force on the boreability index. In addition, due to the spe-cific rock mass boreabiltiy index being located close to thecritical point, it can be used to denote the rock mass bore-ability at different operational conditions. These test resultsshowed that the specific rock mass boreability index variesin a wide range from 61.2 kN/cutter/mm/rev. to 208 kN/cutter/mm/rev. in different rock masses. It is not onlyrelated to the rock material strength, but also to the jointsystem of the rock mass. For example, the rock strengthof granite in T05 penetration test is 172.9 MPa and its spe-cific rock mass boreabiltiy index is 208 kN/cutter/mm/rev.,while the rock strength of granite-gneiss in Buchi’s test isclose to 250 MPa and its specific rock mass boreabilityindex is only 120.67 kN/cutter/mm/rev. (Buchi, 2004).The specific rock mass boreability index remains a constantin the same rock mass condition if the same TBM is used,and does not change at different operating thrust forces.

All of the listed penetration test results show that thepower of the fitted penetration test curves varies in a nar-row range from 0.60 to 0.79 in different rock mass condi-tions. These curves are approximately offset curves withdifferent specific rock mass boreability index values, asshown in Fig. 25. If a power is assumed to be 0.75, the devi-ations at each thrust level in every test are listed in Table 7,based on the specific rock mass boreability index and thepower function. The maximum deviation of the predictedrock mass boreability index is close to 18% for the gran-ite-gneiss rock mass. The deviations in other rock massesare less than 10%. Only if the granite rock masses are takeninto consideration, the maximum deviation of the predictedrock mass boreability index is less than 8.5%. TBMs areusually operated with a limited thrust force under the hardrock conditions. In most cases, TBMs can efficiently frag-ment rock mass and the penetration is more than 1 mmper revolution. Generally, the estimation of the rock massboreability index is slightly lower than the actual valuewhen the power is taken to be 0.75. For the penetrationprediction, the estimation is conservative. Thus, the specificrock mass boreability index can be estimated using thepower of 0.75. Together with the cutter thrust force andthe penetration rate recorded by TBM data acquisition sys-tem, the specific rock mass boreability index of graniticrock masses can be evaluated in tunnels excavated by sim-ilar TBMs.

7. Conclusion

In order to evaluate the rock mass boreability, a series ofTBM front shield friction tests were conducted in DTSST06 tunnel site. It was found that the friction force betweenthe front shield and the tunnel wall is about 50 kN. Then,two types of TBM penetration tests were conducted inDSTT T05 and T06 tunnel sites, respectively. With theincrease of the thrust force per cutter, the penetration perrevolution increases. However, a critical point exists in thesecorrelation curves. The penetration per revolution increasesrapidly with increasing thrust per cutter when the thrust per

cutter is higher than the critical value. The critical point islocated at the penetration per revolution between 0.5 mm/rev. and 1.0 mm/rev. The muck sieve test results verifiedthat with increasing thrust force, the muck size increasesand the rock breakage efficiency also increases. Only whenthe thrust force is higher than the critical value, the muckis well-graded. The muck shape analysis results also showedthat with the increase of the thrust, the chip shape changesfrom flat to elongated and flat.

Besides the penetration tests conducted at DTSS sites inSingapore, some penetration test data were collected. Thecorrelation between the rock mass boreability index andthe penetration per revolution is fitted well to a power func-tion. The correlation coefficient is close to 1. Because theboreability index is a function of the thrust, the boreabilityindex at 1.0 mm/rev. is selected and defined as the specificrock mass boreability index, to evaluate the rock mass bore-ability without the influence of TBM operational parame-ters. The powers of the fitted power functions fall in anarrow range. A power exponent of 0.75 is suggested to esti-mate the specific rock mass boreability index of graniticrock masses in tunnels excavated by similar TBMs.

Acknowledgements

The authors wish to thank Prof. Amund Bruland of theNorwegian University of Science and Technology for mak-ing some useful suggestions and corrections, Dr. ErnstBuchi of Geo’96 (Switzerland) for providing three penetra-tion test results and valuable discussions about the penetra-tion test. The authors would like to thank Sembcorp,Zublin AG (Singapore), Public Utility Board of Singaporeand Herrenknecht AG for allowing these tests conducted inthe tunneling sites.

References

Blindheim, O.T., Johansen, E.D., Johannessen, O., 1983. Criteria for theselection of full face tunnel boring or conventional tunneling.Norwegian Tunnelling Technology, Norwegian Soil and Rock Engi-neering Association 2, 33–38.

Borg, A., 1988. Hard rock tunnel boring in Norway NorwegianTunnelling Today Norwegian Soil and Rock Engineering Association.Norwegian Tunnelling Society, NFF 5, 109–112.

Bruland, A., 1998. Hard rock tunnel boring. Doctoral thesis, NorwegianUniversity of Science and Technology, Trondheim.

BS code 1377, 1990. British Standard Methods of test for soils for civilengineering purposes. Part 2 Classification Tests. British StandardsInstitution.

Buchi, E., 2004. Paper on reuse of TBM muck with sieve curves and 3pages of penetration test results. Private communications.

Gong, Q.M., Zhao, J., Jiao, Y.Y., 2005. Numerical modelling of the effectsof joint orientation on rock fragmentation by TBM cutters. Tunnellingand Underground Space Technology 20 (2), 183–191.

Gong, Q.M., Jiao, Y.Y., Zhao, J., 2006a. Numerical modeling of theeffects of joint spacing on rock fragmentation by TBM cutters.Tunnelling and Underground Spacing Technology 21 (1), 46–55.

Gong, Q.M., Zhao, J., Hefny, A.M., 2006b. Numerical simulation of rockfragmentation process induced by two TBM cutters and cutter spacingoptimization. Tunnelling and Underground Space Technology 21 (3–4),263.


Hamilton, W.H., Dollinger, G.L., 1979. Optimizing tunnel boringmachine and cutter design for greater boreability. RETC Proceedings,Atlanta 1, 280–296.

Howarth, D.F., 1987. Mechanical rock excavation—assessment of cutta-bility and boreability. RETC proceedings 1, 145–164.

Sundin, N.O., Wanstedt, S., 1994. A boreability model for TBM’s. In:Nelson, P., Laubach, S.E. (Eds.), Rock mechanics models andmeasurements challenges from industry. Proceedings of the 1st NorthAmerican Rock Mechanics Symposium, The University of Texas atAustin, Balkema, Rotterdam, pp. 311–318.

Commission on Engineering and Technical Systems of USA, 1984.Geotechnical site investigation for underground projects, vol. 1 andvol. 2, p. 182.

Wanner, H., Aeberli, U., 1979. Tunnelling machine performance in jointedrock. In: Proceedings of 4th Congress of the International Society forRock Mechanics, Montreux, vol. 1, pp.573–580.

Zhao, J., 1996. Construction and utilization of rock caverns in Singapore,Part A: the Bukit Timah granite bedrock resource. Tunnelling andUnderground Space Technology 11 (1), 65–72.

Zhao, J., Broms, B.B., Zhou, Y., Choa, V., 1994. A study of theweathering of the Bukit Timah granite, Part B: Field and laboratoryinvestigation. Bulletin of the International Association of EngineeringGeology 50, 105–111.

Zhao, J., Zhou, Y., Sun, J., Low, B.K., Choa, V., 1995. Engineeringgeology of the Bukit Timah granite for cavern construction in Singa-pore. Quarterly Journal of Engineering Geology 28, 153–162.

Documents

In situ TBM penetration tests and rock mass boreability analysis in hard rock tunnels