Observational construction of large-cross-section junction for TBM U-turn

T. Aoyagi, H. Fukuda, A. Uneda Kajima Corporation T. Nishioka, Y. Naka Hanshin Expressway Public Corporation ABSTRACT: Hanshin Expressway Public Corporation plans to organize the integrated Kyoto-highways. Among the highway networks, the highway Route 1 is supposed to be artery to ease the chronic traffic congestions between city center and suburb of Kyoto. Inariyama tunnel in Route1, which has to be twin tunnel of eastbound and westbound, is comprised of the urban and mountain areas. At the urban area with thin overburden, the tunnels plan to be constructed by introducing TBM. The rest of the tunnel in mountain area is schemed to be constructed by NATM. The launched TBM is supposed to be arrived at the shaft in the urban area after U-turned in the NATM zone. The junction for TBM U-turn, which makes up H-shaped large cross sectional cavern turns into the large cross -sectional tunnel with complicated three-dimensional stress state. This paper describes both construction and design for the large cross sectional junction, and verifies rock st ability for the large cavern by the comprehensive monitoring. 1 INTRODUCTION Inariyama Tunnel on the planned Kyoto Highway Route 1 (provisional name) is a twin-bore tunnel consisting of eastbound and westbound two-lane tunnels. In the total tunnel length of 2.5 km, a 1.45km-long section from the Yamashina-side portal is planned as a NATM tunnel, and a 1.05 km section from the Fushimi-side portal is located in an urban area as a combination of a cut-and-cover section and a shield tunnel section. The shielded Tunnel Boring Machine (TBM) launched from the Fushimi-side w e s t b o u n d s h a f t w i l l m a k e a U-t u r n a t a predetermined location and excavate the eastbound tunnel (See Photo.1). The connecting tunnel to enable the TBM to make a U-turn will form a large-cross-section junction between the twin tunnels where a complex threedimensional stress distribution is expected under the excavations.
Fushimi Ward Shafts for shield tunneling
Biwako Sosui Canal

This paper outlines the design of the large-crosssection junction and reports on its construction by the results of predictive analyses based on measurement results. As of this writing (October, 2001), excavation of the upper section of the twin tunnel and the connecting tunnel has been completed.

Photo.1 Overview of the Yamashina portal.

Tunnel section 2500m Yamashina Ward

Approx. 145m

Keihan Railway

JR Nara Line

U-turn junction
Mt. Inari
Fault F1

Direction of excavation 1.69%

Shield tunnel section Approx. 900m Cut-and-cover tunnel section 150m

NATM tunnel section Approx. 1450m

Figure 1. Longitudinal profile of Inariyama Tunnel.

2 OUTLINE OF U-TURN JUNCTION CONSTRUCTION

B A
Transitional

2.1 Overview of the construction The geology of the site mainly consists of shale, hybrid rocks, sandstone and chert, and intrusions of porphyroid are s een at places. Soft and medium-hard rocks are dominant, but there are many interlayers of fractured and faulted rocks of orogenic origin. The largest fault, Fault F1 (Inariyama Fault) located on the west side, consists of clay or of gravel inte rspersed with clay. The U-turn junction was planned to enable the TBM that has been employed for driving the urban area to turn about in the NATM tunnel section so as to launch the TBM for excavation of the eastbound tunnel. The U-turn junction was located on the NATM tunnel side to avoid the Inariyama Fault (F1) in the fractured zone in Figure 1. As shown in Figure 2, the cross -sectional shape of the junction was designed to form the minimum space required for the extraction, relocation and U-turn of the cylindrical-shield TBM about 11 m in diameter and 11 m long. The overburden depth is about 50 m at the center of the connecting tunnel, and it gradually decreases toward the urban area. As shown in Figure 4, the horseshoe-shaped excavated cross section at the junction is about 18.0 m wide and 16 m high, and the large cav ern has a cross sectional area of 240 m2. The excavated cross section is enlarged and lowered from the cross section of the main tunnel at the transitional zone toward the junction.

Transitional

U-turn junction

zone

zone

zone Eastbound tunnel

Shield tunnel section

NATM tunnel section Direction of movement Westbound tunnel

Figure 2. Plan of U-turn junction area.

34460 (on centers)

(Westbound tunnel) (Eastbound tunnel)

7750

0 00 R5

7750

0 00 R5

10900

10900

Figure 3. A-A section of NATM tunnel. (Unit:mm)

with l tion rsec in tunne Inte ma with l tion rsec in tunne Inte ma

34460 (on centers)

15840

600 R8

Connecting tunnel

R8

60

Westbound tunnel

Eastbound tunnel

17800

17800

Figure 4. B-B section of U-turn junction. (Unit: mm)

Shield tunnel section

NATM tunnel section

Photo2. Junction area at the upper section.

Figure 5. Bird's eye view of U-turn junction.

step1

step2

Long steel forepile

step3

Connecting tunnel

Eastbound tunnel Westbound tunnel Excavate upper section of main tunnel and install primary lining step4

Rock anchors

Install rock anchors between main tunnels, and excavate connecting tunnel step5

Complete excavation of upper section of connecting tunnel step6

Eastbound tunnel Westbound tunnel Excavate lower section (bench) of main tunnel and connecting tunnel Figure 6. Construction procedures. Install rock anchors between main tunnels and excavate lower section (bench) (2nd layer) Complete lowering of main and connecting tunnels (up to 3rd layer)
3D numerical analysis (linear ela sticity) Unsupported tunnel model Detailed construction plan

2.2 Construction procedures In connection with the support design for the Uturn junction, the excavation method to be applied was determined, and the method was evaluated o n t h e basis of the construction procedures. Figure 6 illustrates the construction steps. Construction procedures are as follows: 1) Excavate the upper section of the twin tunnels with primary supports, respectively 2) Reinforcement of the surrounding ground between the twin tunnels 3) Excavation of the upper section of the connecting tunnel with auxiliary methods 4) Completion of the upper section for the Hshaped junction 5) Lowering for the H-shaped junction by the bench excavation 6) Completion of the H-shaped junction 3 DESIGN OF U-TURN JUN CTION 3.1 Preliminary design overview Figure 7 shows overall support design flowcharts. The first step in design was to apply threedimensional FEM (Finite Element Method) analysis to investigate three dimensional s t r e s s a n d deformation due to connecting tunnel excavation. Then, a two dimensional FEM analysis using the analytical results thus obtained was conducted to reconfirm the state of stresses in the surrounding ground and the support structures in detail.
Calculate stress increase due to connecting tunnel e xcav ation Determine loosening load (Estimate loosened zone from fractured zone )

Detailed design

2D linear FEM numerical analy sis (supported tunnel analysis) 3D shell analysis 2D nonlinear FEM numer ical analysis

END

Compare with result of detailed 3D FEM analy sis

O.K

END

Figure 7. Design flowcharts.

For the region surrounding the junction in which there are complicated support structures, a threedimensional shell structure model was applied to determine the type of final support structure to be adopted. 3.2 Numerical analyses Figure 8 shows 3D FEM model to investigate three dimensional stress state. In this numerical analysis, excavation steps were roughly expressed in order to grasp the stress and deformation concentration of the surrounding ground. Rock property values were determined according to the results of a back analysis conducted on the basis of convergence measurement results for the

excavated portion of the main tunnels (See Table 1). The analysis indicated that the largest loosened zone would occur near the junction, and that the height of the loosened zone from the crown of the connecting tunnel would be about 7 m. These analytical results were used to design the tunnel supports and the inner lining by 2D FEM. Figure 9 shows typical results of maximum shear strain in 2D twin-tunnel model. Increases in displacement due to the excav ation of the connecting tunnel were extracted from the results of the 3D FEM analysis and were converted to loads. The loads thus obtained were made to act on a 2D twin-tunnel model and the construction steps were modeled to check the structural soundness of the tunnel supports during construction.

ground between twin tunnels i s also reinforced employing the same anchors. At the junction, high strength shotcrete and high stiffness rib were introduced to restrict the internal displacement and mixing for the shotcrete were effectively studied to achieve the initial shotcrete strength. In the support structure chosen in this study, the short-term maximum allowable stress was exceeded locally in the junction zone. To allow for the poss ible need for reinforcement, therefore, it was decided to excavate a connecting tunnel that is 200 mm larger in diameter than necessary and, if deemed necessary according to information gained through feedback, use this extra space to accommodate such reinforcements as an additional support or an extra thickness of shotcrete. Figure 11 shows the support patterns in the junction zone.

approx. 200m

Westbound tunnel

Eastbound tunnel

Approx. 260m Approx. 70m

max (max. shear strain) More than 0.3% 0.1% 0.3% Less than 0.1%

Figure 9. Distribution of maximum shear strain by 2D FEM.

Figure 8. Three-dimensional symmetric FEM model. Table 1. Physical properties of ground for 3D FEM analysis. Unit weight Modulus of elasticity Poisson's ratio 21 kN/m 3 800 N/mm 0.35
2

Stress smaller than short-term allowable stress Stress equal to or larger than short-term allowable stress

The 3D shell analysis method was used to model even complicated support structural details. In the analysis, loosening loads determined through the 3D numerical analysis were applied to confirm the structural soundness of the tunnel supports in the junction zone of the completed structure. The analysis revealed that as shown in Figure 10, in the completed structure, large local stresses occur at the arch between the main tunnel and the connecting tunnel. Table 2 shows support structure specifications reflecting the results of the analysis. 3.3 Supporting pattern for the junction According to the concentration of ground stress and deformation during excavation, high -strength rock anchor are introduced to prevent the extension of the loosening zone. Furthermore, the loosed

Figure 10. Distribution of shotcrete stress by 3 D sh e ll model. Table 2. Support specifications. Junction zone Ground classification Steel arch Shotcrete Rock bolt Rock anchor D1 H-250 (t =400 N/mm 2) t=300 mm (28=36 N/mm 2) l=6m@1.2m (TD24) 26, 32@1.2m (l =8~13 m) Transitional zone D1 H-200 (t =400 N/mm 2) t=250 mm (28=24 N/mm 2) l=6m@1.2m (TD24) 26, 32@1.2m (l =8 m)

Support pattern (plan)

A
15600 17400 14000

Section B-B (cross section of connecting tunnel)

Steel arch support NH-250x250x9x14 Shotcrete (ck=36N/mm 2) t=300 Rock anchor threaded steel bar (lower section) 32 L=8.0m ctc2000 @2.4m

15600

Eastbound tunnel
Rock anchor B Threaded steel bar 32 L=13.0m ctc2000 @2.4m
300
3

00
S.L

13310

00 72 r1=

12632

15600

Westbound tunnel
17400 14000

Grouted-in-place self-drilling rock bolt @1.2m L=6.0m

Rock anchor threaded steel bar (lower section) 26 L=8.0m ctc1000 @2.4m

C
Silica-resin-grouted long steel pipe forepiling 114.3 x61@0.6m L=9m Grouted-in-place self-drilling rock bolt @1.2m L=6.0m

Section A-A (cross section of janction)

34463 Distance between tunnels

Section C-C (cross section of transitional zone)

Silica-resin-grouted forepiling L=5.0m(120 ) @0.6m Grouted-in-place self-drilling rock bolt L=6.0m@1.2m
=7

300

250

Rock anchor threaded steel bar (upper section) 32 L=12m ctc1000 @2.4m
30 0

25 0 R5

6000

r1 =

15310

86

00

S.L

r1

6 =8

00

30

Shotcrete (ck=36N/mm2) t=300

S.L

12430

S.L

Shotcrete (ck=24N/mm2) t=250 Steel arch support NH-200x200x8x12

10 0

Rock anchor threaded steel bar (lower section) 26 L=8m ctc1000 @2.4m

LEVEL

Steel arch support NH-250x250x9x14

11152

8065

8065 16130

Figure 11. Support patterns in junction zone

4 COMPUTER-BASED MONITORING AND FEEDBACK 4.1 Measurement results Due to scarce information about rock properties in advance, updated results for monitoring were reflected to the predictive support designs. Especially, rock soundness of the foreground was evaluated by using drilling energy for rock bolts at the each cutting face in connection with comprehensive monitoring. Figure 12 shows the locations of the measurement cross-sections, and Table 3 shows the measurements conducted in different cross sections. The cross section at the end of the connecting tunnel (Section A3) in which the highest degree of stress concentration was expected to occur during the excavation of the connecting tunnel was defined as the principal measurement cross section. Figure 13 shows the measurement items. The measurement results for that cross section were used as the basis of the computer-based monitoring and feedback method of construction. Figure 14 to 16 show measurement results in the junction zone of the Sections A3 in the eastbound tunnel. Comparison of the measurement results for the westbound and the eastbound tunnel reveals that both displacements and support member stresses in the eastbound tunnel are larger than those in the

wes tbound tunnel. This is thought to be due to stiffness in geological conditions between the eastbound and the westbound tunnel. The control limit for the displacement is shown in Table 4. As shown, the tunnel displacement of the eastbound tunnel is between Level 1 and Level 2. According to the results of ground displacement measurement in Figure 17, elastic displacement of ground occurs at depths of more than 10 m from the tunnel wall surface, and there seems to be a point of inflexion at a depth of 4 to 5 m from the tunnel wall surface. This result indicates that the depth from the tunnel wall surface of the loosened zone in the surrounding ground is about 5 m. This agrees closely with the loosened zone (about 4m) estimated through the 3D FEM numerical analysis. As for stress measurement, t h e short-term allowable stress was exceeded in the eastbound tunnel. The fact, however, that displacements and support member stresses increased gradually as construction progressed caused concern for further stress increases in the subsequent stages of construction. For this reason, back and predictive analyses were conducted to investigate the need t o s t r e n g t h e n t h e s upport structure.

Eastbound tunnel

Junction
Displacement (mm)
Eastbound tunnel Shield tunnel section Westbound tunnel NATM tunnel section

05/19/01 0.0 -4.0 -8.0 -12.0 -16.0 -20.0

06/08/01

06/28/01

07/18/01

08/07/01

08/27/01

09/16/01

10/06/01

10/26/01 90

70

Horizontal convergence
Connecting tunnel Bench excavation

60 50

Second bench

40

Third bench

First bench

30 20 10 0

Section A3

Section A1

Section A2

-24.0 -28.0

Section C

Section B

Figure 14. Results of A displacement measurement. (Section A3)

Eastbound tunnel
36 90 80

Figure 12. Measurement cross sections. Table 3. Measurements to be conducted in different crosssections. Measuring section Measuring items A1 A2 A3 B C Convergence Steel arch support stress
30

Crown Shoulder(Left) Shoulder(Right)

70 60 50

24 18 12 6 0 -6
05/19/01 06/08/01 06/28/01 07/18/01

Connecting tunnel

Bench excavation

40

Second bench

First bench

Third bench

30 20 10 0

Shotcrete stress Rock bolt axial stress Ground displacement Figure 15. Eastbound and westbound tunnels A3 ) Measured shotcrete stress. (Section
Westbound tunnel only

08/07/01

08/27/01

09/16/01

10/06/01

10/26/01

Eastbound tunnel
0 -40 90

Crown Shoulder(Left)

80 70 60 50

-80 -120

Shoulder(Right)
Connecting tunnel Bench excavation

40

Third bench

First bench

-160 -200 -240

05/19/01 06/08/01 06/28/01

Second bench

30 20 10 0
10/06/01 10/26/01

07/18/01

08/07/01

08/27/01

09/16/01

Figure 16. Measured steel support stress. (Section A3)

+5.5mm 10m

5m

Measuring items Horizontal convergence Subsidence crown Ground displacement Rock bolt axial stress Steel arch support stress Shotcrete stress C1 ~ C4 S

Remarks
E1 M E2 M E3 M

L=12m L= 6m

Figure 13. Measuring items (Section A3) Figure 17. Measured results of ground displ acement measur ement . (Section A3 in westbound tunnel)

Progress of the face(m)

Stress(N/mm2)

Progress of the face(m)

Stress N/mm

progress of the face (m)

Subsidence crown

80

Table 4. Displacement for the control limit. (Unit: mm)

Control level Level I Level II Level III

Displacement Horizontal convergence Subsidence crown Horizontal convergence Subsidence crown Horizontal convergence Subsidence crown

Upper section 26 13 39 20 52 26

4.2 Back analysis and predictive analysis In the back analysis, convergence (subs idence of crown, horizontal convergence) and support stress at the time of completion of the main tunnel were traced. For the purpose of tracing, back calculations were made, parameterizing the modulus of deformation of ground, the coefficient of initial lateral pressure and the modulus of elasticity of shotcrete. The present paper reports on the study results for the eastbound tunnel, which is at a more advanced stage of excavation than the westbound tunnel and which has to be driven under poor ground conditions. The analytical results show that the modulus of deformation (E=863 N/mm2) of ground is closer to a value (E=800 N/mm2) for "DI" ground (according to Japan's ground classification system for road tunn e l s ) t h a t w a s a s s u m e d f r o m t h e o utset, indicating that the ground in the U-turn junction zone for the eastbound tunnel is indeed a DI ground which expresses medium-hard-rock. The modulus of elasticity of shotcrete obtained from the analysis agrees with the 28-day modulus of elasticity for ck=36 (N/mm2).
Table 5. Results of back analysis. Input Data (Measured value) Subsidence of crown Horizontal convergence Support stress Shotcrete H-beam 9.0 mm 8.8 mm 165 N/mm
2 2

Total rate of advance (benchThird difference: two layer First Second Displacement layers) due bench to hardnessbenchground conditions in bench between the eastbound and 5 westbound tunnels. 11 10 52 Compared with the results of measurements 5 5 3 26 conducted at the time of completion of the upper 16 15 8 78 section of the 7main tunnel and at the 39 time of 8 4 completion of 21 upper section of the connecting the 21 10 104 tunnel, the predicted values5show good agreement 11 10 52 with the measurement results. Comparison of the resultant axial stress (axial strength) that occurs when the stress level reaches the strength of the steel arch support and shotcrete with the predicted maximum resultant axial stress shows that the factor of safety is 1.00, indicatin g that the strength has no margin of safety. It was decided, therefore, to apply an extra layer of shotcrete.
310

Lower section

Steel stress (N/mm 2 )

290 270 250 230 210 190 170 150

ma in tu nn up el pe rs va ec ti ti on on of Lo co we nn ri ec ng ti ng tu nn up el pe r (E se as ct t) io n 1s t la ye r lo of we r (E se as ct t) io (E 2n n as d t) la ye 3r lo r d (W we of la es r ye t) se r 1s ct of t io lo n la we ye r r se of ct lo io we n (w r es se Ex t) ct ca (W io 2n va es n d ti t) la on ye 3r of r d lo la of co we ye nn r r ec se of ti ct ng io lo n we tu r nn se el ct lo io we n r se ct io n Ex ca

239 230 206 193 165

Eastbound Measured value

Figure 18 Comparison between predictive and measured value (Steel arch support stress).

)
36 Eastbound 30 24
17.9 18.2

Shotcrete stress (N/mm

Ex

ca

va

ti

on

of

Measured value

14.0 N/mm

18
14.0

16.2

16.9

Output Data (B ack analysis result) Modulus of deformation of ground Coefficient of initial lat eral pressure Modulus of elasticity of shotcrete 863 N/mm 1.88 24 N/mm 2
2

12
nn el up pe rs ec of tio co n nn Lo ec we tin rin g g tu nn el up pe rs (E ec as tio t) 1s n tl ay er lo of w er (E se as ct (E t) io as 2n n t) d 3r lay d (W er la lo ye es of we t) ro rs 1s fl ec tl ow tio ay er n er se of ct lo io w n Ex er ca (w se va ct es (W tio io t) n n 2n es of t) d 3r lay co d nn er la ec lo of tin yer we of g rs tu ec nn low tio er el n se lo ct w io er n se ct io n n tu

ca

va t

io n Ex

of m

ai

Figures 18 and 19 show the results of a predictive analysis conducted on the basis of the physical property values obtained from the back analysis. In the predictive analysis, analytical steps were determined so as to express the difference in the

Figure 19 Comparison between predictive and measured value (S hotcrete stress).

Ex ca va tio n

5 CONCLUSION The construction work reported in the present paper involved the construction of a tunnel junction with an unprecedentedly large cross section, and very few detailed geotechnical data were available. For these reasons, it was decided to allow for a certain amount of variation (allowance for reinforcement: 200mm) in support design and take the computer-based monitoring and feedback approach. As a result, an extra layer of shotcrete was applied, and so far the tunnel construction has been executed safely and under predicted conditions. We will proceed with the construction by the monitoring and feedback method to complete the work safely. In closing, we would like to thank the members of the Committee on Tunnel Technology for Kyoto Highway, particularly Dr. Toshihisa Adachi, Chairman, and Dr. Yuzo Onishi, Secretary, for their generous guidance and assistance provided in connection with this study.

6 REFERENCE
Aoyagi, T. & Uneda, A.2001. Design of the large-cross sectional junction for TBM Uturn(Part 2). Proceedings of the 56 th JSCE annual c o n f e r e n c e , O c t o b e r 2 0 0 1 ( i n Japanese) Nishioka, T & Ishihara, H. 2001. Design of the large-cross sectional junction for TBM Uturn (Part 1). Proceedings of the 56 th JSCE annual c o n f e r e n c e , O c t o b e r 2 0 0 1 ( i n Japanese)