Geotechnical risks, mitigation measures and performance of ...

175© 2019 The Hong Kong Institution of Engineers HKIE Transactions | Volume 26, Number 4, pp.175–189 • https://doi.org110.33430/V26N4THIE-2019-0017

* The first author who was at the age of 35 or below on the closing date of submission for the HKIE Outstanding Paper Award for Young Engineers/Researchers 2019.

Geotechnical risks, mitigation measures and performance of tunnel boring machine (TBM) tunnels for the Tuen Mun – Chek Lap Kok Link Project, Hong Kong

ABSTRACT

The Tuen Mun – Chek Lap Kok Link (TM-CLKL) project implemented latest innovation designs and ground improvement techniques in large slurry tunnel boring machines (TBM) tunnelling for constructing the sub-sea TBM tunnel in Hong Kong. Face stability, ground settlement, blow out, water ingress and ovalisation are always the major geotechnical risks for TBM tunnelling in the industry. This paper summarises the design considerations in terms of ground improvement and special arrangements of TBM tunnelling to mitigate the geotechnical risks and the improvement of the overall performance of TBM tunnelling under the TM-CLKL project. Field monitoring data such as surface settlement, ovalisation of segmental lining and ground improvement technique such as barrettes are presented.

KEYWORDS Geotechnical risks; ovalisation; ground improvement; settlement; sub-sea tunnel; TBM tunnelling; TBM confinement pressure

Lucas Kwan Lok Chan* and Alan Kwok Lun KwongAECOM Asia Company Limited, Hong Kong, People’s Republic of China

CONTACT Lucas Kwan Lok ChanReceived 15 March 2019

[email protected]

1. Introduction

The Tuen Mun – Chek Lap Kok Link (TM-CLKL) project is currently under construction and it comprises a 9 km long dual two-lane carriageway between Tuen Mun and the North Lantau, Hong Kong (Figure 1). The alignment commences at a connection with the toll plaza at Lung Mun Road of Tuen Mun Area 46. It heads southeast on an elevated structure over the Lung Mun Road and the lands on a new reclamation (referred to as the Northern Landfall (NLF)) adjacent to the Tuen Mun River Trade Terminal (RTT). The alignment turns south and heads into an approximately 4.1 km long sub-sea tunnel passing under the Urmston Road. After crossing the Urmston Road, the alignment daylights at the eastern edge of the newly reclaimed land, Hong Kong – Zhuhai – Macao Bridge (HZMB) Hong Kong Port, near the Hong Kong International Airport at Chek Lap Kok (referred to as the Southern Landfall (SLF)) and then head eastwards on a marine viaduct to connect with the North Lantau Highway at Tai Ho of Lantau.

2. Geotechnical risk, mitigation and performance of tunnel boring machine (TBM) tunnelling in TM-CLKL

The selection of tunnelling method, geotechnical risk mitigation and performance of the TM-CLKL tunnel is divided into three areas for the discussion in this paper, namely the NLF, sub-sea section and the SLF.

Figure 1. –General layout of TM-CLKL.

Figure 1. General layout of TM-CLKL.

2.1. Tunnelling method and geotechnical risk mitigation of TBM tunnelling at the NLF

At the northern connection of the TM-CLKL project, the road alignment was designed to connect the land viaduct with the sub-sea tunnel. The traditional method of constructing a launching shaft is to construct a cut-and-cover (C&C) box by diaphragm wall, followed by an excavation and lateral support (ELS) system to form a rectangular shaft. It is also well known that the C&C tunnel box method has a lot of drawbacks in terms of limited

L K L CHAN AND A K L KWONG

176HKIE Transactions | Volume 26, Number 4, pp.175–189

working space, long working programme and the risk of propping and re-propping the struts during excavation and backfilling.

The tunnel invert level varies from -15 mPD - -40 mPD, with a 5% downward gradient and it changes to 1.994% gradient towards the North Ventilation Shaft (NVS) (Figure 2). The tunnel mainly encountered mixed ground conditions which comprised of reclamation fill, vibro compacted sand, marine deposits, alluvial sand, alluvial clay and granite. In view of this, two TBMs were launched at the tunnel portal, i.e. at the shallowest tunnel section at the NLF.

Figure 2 shows the general arrangement of TBM tunnelling and the geology at the NLF. In order to overcome the constraints on space, programme and struts as just discussed, a peanut-shaped TBM launching shaft was constructed to maximise the working space for TBM assembly and break-in operation. This peanut-shaped shaft takes full advantages of circumferential hoop stresses by transferring ground pressure from circular diaphragm walls

to Y-panel and the connecting cross struts. The overall arrangement was to limit ground movements and maximise the structural bending moment capacity of the Y-panel and the compressive strength of the cross struts. A 17.6 m boring diameter slurry TBM (TBM S880) was used to construct the three-lane northbound tunnel (with climbing lane) at the NLF while a 14.0 m boring diameter slurry TBM (TBM S882) was used to construct the two-lane southbound tunnel between NLF and SLF. TBM tunnelling method allowed early launching of TBMs to speed up the construction programme. However, sufficient ground cover and ground improvement works were required to resist buoyancy and prevent over-stressing of tunnel lining when the tunnel section is shallow and close to the ground surface. In addition to conventional tympanum which is typically installed at the circular diaphragm wall at the break-in position, additional break-in plug was constructed at the exit of the TBM launching shaft. It consisted of two parts to enhance the face stability and watertightness during

Figure 2. General arrangement of TBM tunnelling and geology of the NLF.

Once the TBM dipped into the ground with covering depth of more than about 1.5

times of its diameter, it could be engaged into fully operational mode for controlling the face

stability and water ingress through slurry confinement pressure. During the TBM tunnelling,

full-time slurry pressure monitoring was conducted to control the surface settlement (volume

loss) and avoid blow-out at the surface or surface collapse (sinkhole).


Once the TBM dipped into the ground with covering depth of more than about 1.5

times of its diameter, it could be engaged into fully operational mode for controlling the face

stability and water ingress through slurry confinement pressure. During the TBM tunnelling,

full-time slurry pressure monitoring was conducted to control the surface settlement (volume

loss) and avoid blow-out at the surface or surface collapse (sinkhole).


HKIE Transactions | Volume 26, Number 4, pp.175–189177

the TBM break-in operation. The first part of the break-in plug was designed to cover the TBM shield plus 5 m additional length to allow the TBMs to come progressively into fully operational mode and provide watertightness (permeability of water: 10-6 m/s or less) since no slurry confinement pressure could be applied to the tunnel face at the beginning. Once the TBMs have lining segments fully installed into the ground and away from the diaphragm wall, the latter part of the break-in plug served to allow the TBMs to gradually apply the slurry confinement pressure to the ground to control the water ingress and ground settlement. During the TBM break-in operation, the ground cover was temporarily topped up to +10.0 mPD to resist the buoyancy until the tunnel invert was backfilled. This required a delicate design and performance observations for balancing the load from the additional surcharge and the movement in the segmental lining.

The NVS was constructed at the transition zone between the two-lane and three-lane

northbound tunnel for retrieving the TBM S880 and reassembly of the 14.0 m boring diameter

slurry TBM (TBM S881) before continuing excavating down to the two-lane sub-sea tunnel.

Due to the close tunnel alignment near the NVS, a row of hit and miss diaphragm wall panels

was constructed to reduce the stresses acting on the adjacent tunnel lining. The NVS was then

flooded to +2.5 mPD to balance the slurry confinement pressure for retrieval of TBM S880.

On the other hand, a steel tube with mass concrete fill was constructed to provide confinement

pressure for TBM S882 to cross the NVS and continue to the sub-sea TBM tunnelling (Figure

3).

Figure 3. General arrangement of NVS for TBM S880 break-out and TBM S882 crossing.

2.1.1. Performance of TBM tunnelling at the NLF

TBM tunnel design depends on the interaction between the excavation and tunnel

support systems and the ground. The excavation and support sequence influences the

pressures acting on the tunnel, ovalisation of tunnel segmental lining and volume loss. In

TBM tunnelling, settlement will occur ahead of the tunnel face due to stress relief, therefore,

the pressures acting on the lining will be less than the full overburden pressure assumption.

Settlement continues at a constant rate until the closed tunnel segmental lining takes full

effect and the settlement will then cease. To control the volume loss, it is required to closely

monitor the temporary support pressure and timely installation of the permanent tunnel

segmental lining. After the tunnel rings are installed, the tunnel rings tend to deform

Figure 3. General arrangement of the NVS for TBM S880 break-out and TBM S882 crossing.

Once the TBM dipped into the ground with covering depth of more than about 1.5 times of its diameter, it could be engaged into fully operational mode for controlling the face stability and water ingress through slurry confinement pressure. During the TBM tunnelling, full-time slurry pressure monitoring was conducted to control the surface settlement (volume loss) and avoid blow-out at the surface or surface collapse (sinkhole).

The NVS was constructed at the transition zone between the two-lane and three-lane northbound tunnel for retrieving the TBM S880 and reassembly of the 14.0 m boring diameter slurry TBM (TBM S881) before continuing excavating down to the two-lane sub-sea tunnel. Due to the close tunnel alignment near the NVS, a row of hit and miss diaphragm wall panels was constructed to reduce the stresses acting on the adjacent tunnel lining. The NVS was then flooded to +2.5 mPD to balance the slurry confinement pressure for retrieval of TBM S880. On the other hand, a steel tube with mass concrete fill was constructed to provide confinement pressure for TBM S882 to cross the NVS and continue to the sub-sea TBM tunnelling (Figure 3).

2.1.1. Performance of TBM tunnelling at the NLF

TBM tunnel design depends on the interaction between the excavation and tunnel support systems and the ground. The excavation and support sequence influences the pressures acting on the tunnel, ovalisation of tunnel segmental lining and volume loss. In TBM tunnelling, settlement will occur ahead of the tunnel face due to stress relief, therefore, the pressures acting on the lining will be less than the full overburden pressure assumption. Settlement continues at a constant rate until the closed tunnel segmental lining takes full effect and the settlement will then cease. To control the volume loss, it is required to closely monitor the temporary support pressure and timely installation of the permanent tunnel segmental lining. After the tunnel rings are installed, the tunnel rings tend to deform (sometimes referred to as ovalisation) when it reacted to the load. Ovalisation monitoring is therefore required to check the structural integrity of the tunnel.

Convergence confinement curve method (Panet, 2001; Almog et al., 2015) was used to simulate the interaction between the ground and tunnel support system so as to estimate the required support pressure and the corresponding ovalisation in tunnel design. The critical sections of TBM tunnelling were identified and simulated by finite element programme, i.e. Plaxis 2D, to ensure upper and lower bound effects are considered in the design with appropriate relaxation factor.

Regular ovalisation monitoring was conducted after the tunnel segmental lining are installed. Figure 4 shows that the break-in plug and jet grout columns installed on the sides of the tunnels enhanced the short-term stability during the excavation (TBM S880 Ring No. R0015 - R0085, TBM S882 Ring No. R0012 - R0042). The ovalisation magnitude



was slightly increased in the soft ground sections which consists of reclamation fill, marine deposit and alluvium layers (TBM S880 Ring No. R0085 - R0160, TBM S882 Ring No. R0042 - R0130). There was a downward trend when the geology changed from soft ground condition to mixed ground condition or full-face rock condition.

In general, TBM S880 and TBM S882 performed satisfactorily from the TBM launching shaft to the NVS through the newly reclaimed land and alluvium/ weak CDG with the highest production rate of five nos. of ring per day. Only minor volume loss was recorded on the ground

surface across this zone and the ovalisation was found well below the predicted values. Table 1 summarises the overall performance of TBM S880 and TBM S882 at the NLF.

2.2. Tunnelling method and geotechnical risk mitigation of TBM tunnelling at sub-sea section

The TM-CLKL project was the first-time in adopting TBM soft ground tunnelling method for a sub-sea road tunnel construction in Hong Kong. TBM sub-sea tunnelling was proposed after the following considerations:

580 (R0377 / R0302)

structure with diaphragm wall, development load of 20 kPa

Table 1. Summary of TBM tunnel design at the NLF.

Regular ovalisation monitoring was conducted after the tunnel segmental lining are

installed. Figure 4 shows that the break-in plug and jet grout columns installed on the sides of

the tunnels enhanced the short-term stability during the excavation (TBM S880 Ring No.

R0015 to R0085, TBM S882 Ring No. R0012 to R0042). The ovalisation magnitude was

slightly increased in the soft ground sections which consists of reclamation fill, marine

deposit and alluvium layers (TBM S880 Ring No. R0085 to R0160, TBM S882 Ring No.

R0042 to R0130). There was a downward trend when the geology changed from soft ground

condition to mixed ground condition or full-face rock condition.

Figure 4. Ovalisation monitoring of TBM S880 and TBM S882 at the NLF.

In general, TBM S880 and TBM S882 performed satisfactorily from the TBM

launching shaft to the NVS through the newly reclaimed land and alluvium / weak CDG with

the highest production rate of five nos. of ring per day. Only minor volume loss was recorded

on the ground surface across this zone and the ovalisation was found well below the predicted

values. Table 2 summarises the overall performance of TBM S880 and TBM S882 at the NLF.

Item TBM S880 (Northbound tunnel at the NLF)

TBM S882 (Southbound tunnel at the NLF)

Tunnel length 631.15 m 631.51 m

No. of rings installed 380 rings 295 rings

Duration of TBM drive 190 calendar days 134 calendar days

Overall progress of TBM drive (Total length / total duration)

Best TBM drive

(Max excavation length per day)

3.32 m/day

10.07 m/day

4.71 m/day

12.87 m/day

Table 2. Overall performance of TBM tunnelling at the NLF.

Figure 4. Ovalisation monitoring of TBM S880 and TBM S882 at the NLF.


Table 1. Overall performance of TBM tunnelling at the NLF.

Item TBM S880 (Northbound tunnel at the NLF)

TBM S882 (Southbound tunnel at the NLF)

Tunnel length 631.15 m 631.51 mNo. of rings installed 380 rings 295 ringsDuration of TBM drive 190 calendar days 134 calendar daysOverall progress of TBM drive(Total length / total duration) 3.32 m/day 4.71 m/day

Best TBM drive(Max excavation length per day) 10.07 m/day 12.87 m/day

The tunnel invert level varies from -40 mPD - -50 mPD in the sub-sea section. The

gradient changes from 1.994% - 3.00% near the cross passage CP45 location (Figure 5) and

maintains at 0.67 % grade until reaching the SLF near the CP9 location (Figure 8). The tunnel

mainly encountered mixed ground conditions which comprised of alluvial sand, alluvial clay

and granite. Figure 5 shows the sub-sea tunnel alignment and the simplified geology of sub-

sea tunnel section.

Figure 5. Simplified geological profile of sub-sea tunnel section.



● No submarine utilities diversion is needed (existing CLP Power 132 kV cable and telecommunication cables connecting Tuen Mun and the Hong Kong International Airport).

● Less impact on the navigation in Urmston Road (major shipping channel in Hong Kong).

● Limited marine operations for the immersed tunnel construction at the restricted area of the airport.

● Less environmental impacts as compared with the immersed tunnel construction (e.g. dredging, backfilling and tunnel foundation works).

● Less ecological impact and disturbance to the natural habitat of the Chinese white dolphins and other marine life.

● Marine borrow area, mud disposal ground and contaminated mud pits located nearby the project site.

However, the tunnel profile was required to be embedded in the firm soil strata with sufficient ground cover to provide tunnel stability and minimise the impact on the existing CLP Power submarine cables. Furthermore, the tunnel profile was designed to avoid impact on the tunnel due to sunken ship load and possibility of future dredging, particularly in relation to the future extension of the Urmston Road shipping channel.

The tunnel invert level varies from -40 mPD - -50 mPD in the sub-sea section. The gradient changes from 1.994% - 3.00% near the cross passage CP45 location (Figure 5) and maintains at 0.67% grade until reaching the SLF near the CP9 location (Figure 8). The tunnel mainly encountered mixed ground conditions which comprised of alluvial sand, alluvial clay and granite. Figure 5 shows the sub-sea tunnel alignment and the simplified geology of sub-sea tunnel section.

2.2.1. Performance of TBM tunnelling at sub-sea section

Figure 6 presents the settlement monitoring of the seawall structure at the NLF. Minor settlement was recorded ahead of the TBM S882 cutterhead due to the stress relief.

The settlement became constant when the TBM S882 moved 50 m (3.6 times of TBM diameter approximately) away from the NLF seawall structure.

When TBM S881 was passing underneath the NLF seawall, a sudden drop in settlement curve was recorded and reached the “Action Level” of settlement monitoring. Close monitor of slurry confinement pressure and timely tunnel segmental lining installation were carried out to minimise the ground settlement. Meanwhile, the settlement rate was closely monitored and became constant when TBM S881 moved away from the NLF seawall structure.

This incident reminds that immediate investigation of the cause of settlement and implementation of the mitigation measures are of the essence when sudden drop in settlement is observed. Slurry confinement pressure shall be closely monitored and the TBM annulus gap shall be timely grouted. The settlement rate shall become constant when the TBM moves away in the distance of about two times of TBM diameter to four times of TBM diameter, depending on the geology.

In general, TBM S881 and TBM S882 performed satisfactorily in the sub-sea tunnel section. The tunnel stability was found within the predicted value in the weak CDG layer. A slightly higher ovalisation was found in the TBM S881 Ring No. R1513 - R1573 which were in alluvial clay or alluvial sand layer (Figure 7).

The overall performance of TBM S881 and TBM S882 at the sub-sea tunnel section are summarised in Table 2. TBM S881 and TBM S882 had a better performance in the sub-sea tunnel section after the learning period at the NLF tunnel section. The highest production rate of 13 nos. of ring per day was achieved.

2.3. Tunnelling method and geotechnical risk mitigation of TBM tunnelling at the SLF

At the southern connection, the road alignment is designed to connect the sub-sea tunnel with the marine viaduct at the HZMB Hong Kong Port. The tunnel invert level varies from -50 mPD - -40 mPD with the gradient

Table 2. Overall performance of TBM tunnelling at sub-sea tunnel section.

Item TBM S881(Northbound tunnel at sub-sea tunnel)

TBM S882(Southbound tunnel at sub-sea tunnel)

Tunnel length 4,058.43 m 4,039.89 mNo. of rings installed 1,837 rings 1,837 ringsDuration of TBM drive 406 calendar days 402 calendar daysOverall progress of TBM drive(Total length / total duration) 10.00 m/day 10.05 m/day



changes from 1.990% - 5.84% upwards. The tunnel mainly encountered mixed ground conditions which comprised of alluvial sand, alluvial clay and meta-sediment. Figure 8 shows the general arrangement of the SLF and the generalised geology of TBM tunnel at the SLF section.

At the tip of the SLF, the TBM tunnels passed underneath the HZMB Hong Kong Port project seawall structure which is supported by stone columns. The stone columns and the alluvial sand layer would possibly form a drainage path and hence induce a high risk of blow-

out or surface collapse in this area. A clear separation was designed and close monitor of slurry confinement pressure was conducted during the TBM tunnelling.

After passing through the HZMB Hong Kong Port seawall structure, the tunnel was constructed underneath the South Ventilation Building (SVB) (Figure 8) and arrived the TBM parking plug (Figure 8) which served as a temporary parking in case the cofferdam formed by the C&C tunnel was delayed. This was a significant risk mitigation measure as it was conceived that it bought a high risk to the TBM

Figure 6. Ground settlement monitoring of seawall structure at the NLF.

Figure 6. Ground settlement monitoring of seawall structure at the NLF. Figure 6. Ground settlement monitoring of seawall structure at the NLF.



face stability at great depth if it is stopped for a very long period of time without advancing. In this area, the tunnel stability was sensitive to the loading and unloading sequence due to the SVB basement construction and the tunnel lining was designed to resist different combinations of loading in terms of the short time overburden loading from the ground, unloading from the SVB excavation, dewatering and the long-term creeping of the fill, marine deposit and the alluvium.

The geology changes from firm to stiff alluvium layer to alluvial clay and marine deposits layer when the tunnel profile goes upwards. To reduce the excess pore pressure possibly generated by the reclamation process and to improve the consolidation behaviour of the alluvium layers, deep band drain was installed in areas at the southern part of the South Ventilation Shaft (SVS) (Figure 8) in the alluvium and terminated at the tunnel invert level. The TBM tunnel profile was designed to end at the location where it would encounter the marine deposits layer to

(R1767 / R2083) Alluvial sand the SLF Table 3. Summary of TBM tunnel design at sub-sea tunnel section.

Figure 7. Ovalisation monitoring of TBM S881 and TBM S882 at sub-sea tunnel section.

The overall performance of TBM S881 and TBM S882 at the sub-sea tunnel section

are summarised in Table 4. TBM S881 and TBM S882 had a better performance in the sub-

(R1767 / R2083) Alluvial sand the SLF Table 3. Summary of TBM tunnel design at sub-sea tunnel section.


The overall performance of TBM S881 and TBM S882 at the sub-sea tunnel section

are summarised in Table 4. TBM S881 and TBM S882 had a better performance in the sub-



reduce the risk of any potential lateral movement in a newly reclaimed ground. The TBM then break out at the connecting cofferdam and would then be dismantled and retrieved through the cofferdam.

Numerical analysis was adopted to simulate the TBM tunnelling at the newly reclaimed ground where excess pore pressure might still have not been dissipated completely and the potential long-term creeping of the alluvial clay. The results show that tunnel long-term stability in terms of ovalisation could be significantly affected by the long-term consolidation process and the potential excess pore pressure. In order to mitigate this potential risk identified from the numerical findings, a mitigation method was carried out to reduce the potential long-term ovalisation by constructing some barrette panels on the tunnel sides.

Figure 9 shows the barrette panel layout along the TBM tunnel at the SLF and a typical section of TBM tunnel with barrette panels.

2.3.1. Performance of TBM tunnelling at the SLF

Detailed ground investigation and instrumentation monitoring were carried out to study the state of consolidation and the strength of the material for the TBM tunnel design. Ovalisation becomes the controlling parameter of the TBM tunnel design in the soft ground and under the consolidation process. Detailed numerical analysis were conducted to study the TBM tunnel behaviour in the consolidating ground and impact on the tunnel stability due to the loading and unloading construction activities.

Figure 8. General arrangement of TBM tunnelling and geology of the SLF.

At the tip of the SLF, the TBM tunnels passed underneath the HZMB Hong Kong Port

project seawall structure which is supported by stone columns. The stone columns and the

alluvial sand layer would possibly form a drainage path and hence induce a high risk of blow-

out or surface collapse in this area. A clear separation was designed and close monitor of

slurry confinement pressure was conducted during the TBM tunnelling.

After passing through the HZMB Hong Kong Port seawall structure, the tunnel was

constructed underneath the South Ventilation Building (SVB) (Figure 8) and arrived the TBM

Figure 8. General arrangement of TBM tunnelling and geology of the SLF.



Figure 10 presents the short-term ovalisation monitoring data at the SLF. The magnitude of the ovalisation was slightly larger than the NLF tunnel section and the sub-sea tunnel sections since a number of construction activities were carrying out such as excavation of SVB basement and excavation of SVS and the background settlement due to consolidation. The ovalisation magnitude was reduced at the barrette panel area.

Table 3 summarises the overall performance of the TBM S881 and TBM S882 at the SLF. In general, TBM S881 and TBM S882 performed satisfactorily in the soft ground condition and the barrette panel area at the SLF. The overall excavation rate was slightly reduced in the barrette

panel area which may due to the hardness of the barrette panels. Both TBMs were successfully broken through the cofferdam at the C&C tunnel section (Figure 11).

3. TBM confinement pressure design and control of the TM-CLKL tunnel

TBM confinement pressure control is essential to prevent face collapse, blow-out due to excessive slurry pressure, excessive ground settlement or excessive ground heave. The design methodology and site control of TBM confinement pressure will be described in this section.

Figure 9 shows the barrette panel layout along the TBM tunnel at the SLF and a typical

section of TBM tunnel with barrette panels.

Figure 9. Barrette panel layout and typical section of TBM tunnel with barrette panels.

2.3.1. Performance of TBM tunnelling at the SLF

Detailed ground investigation and instrumentation monitoring were carried out to

study the state of consolidation and the strength of the material for the TBM tunnel design.

Ovalisation becomes the controlling parameter of the TBM tunnel design in the soft ground

and under the consolidation process. Detailed numerical analysis were conducted to study the

TBM tunnel behaviour in the consolidating ground and impact on the tunnel stability due to

the loading and unloading construction activities.

Figure 10 presents the short-term ovalisation monitoring data at the SLF. The

magnitude of the ovalisation was slightly larger than the NLF tunnel section and the sub-sea

tunnel sections since a number of construction activities was carrying out such as excavation

Figure 9. Barrette panel layout and typical section of TBM tunnel with barrette panels.


of SVB basement and excavation of SVS and the background settlement due to consolidation.

The ovalisation magnitude was reduced at the barrette panel area.

Figure 10. Ovalisation monitoring of TBM S881 and TBM S882 at the SLF.

of SVB basement and excavation of SVS and the background settlement due to consolidation.

The ovalisation magnitude was reduced at the barrette panel area.

Figure 10. Ovalisation monitoring of TBM S881 and TBM S882 at the SLF. Figure 10. Ovalisation monitoring of TBM S881 and TBM S882 at the SLF.

Table 3. Overall performance of TBM tunnelling at tunnel section between SVS and C&C tunnel.

Item TBM S881(Northbound tunnel between SVS

and C&C tunnel)

TBM S882(Southbound tunnel between SVS

and C&C tunnel)Tunnel length 239.19 m 291.97 mNo. of rings installed 107 rings 121 ringsDuration of TBM drive 25 calendar days 40 calendar daysOverall progress of TBM drive(Total length / total duration) 9.08 m/day 6.99 m/day




3.1. Design methodology of TBM confinement pressure

3.1.1. Design methodology of TBM confinement pressure by limit equilibrium method

The minimum confinement pressure was calculated according to the recommendation in Geotechnical Engineering Office, the HKSAR Government (GEO) Report No. 249 (Geotechnical Engineering Office, 2009) to prevent face instability. The method developed by Kimura and Mair (1981) was adopted in cohesive soil and the method developed by Anagnostou and Kovari (1996) was adopted in frictional soils.

The minimum confinement pressure in frictional soils is derived from the “silo model” by Horn (1961) where a 3D sliding mechanism in limit equilibrium is assumed to occur along a slip plane AB-EF-LM as shown in Figure 12. The sliding surface is assumed to be bounded within a wedge (volume inside ABDCEF) plus the silo (volume inside DCEF-KNLM) above.

Item TBM S881

(Northbound tunnel between SVS

and C&C tunnel)

TBM S882

(Southbound tunnel between SVS

and C&C tunnel) Tunnel length 239.19 m 291.97 m

No. of rings installed 107 rings 121 rings

Duration of TBM drive 25 calendar days 40 calendar days

Overall progress of TBM drive (Total length / total duration)

Best TBM drive

(Max excavation length per day)

9.08 m/day

17.60 m/day

6.99 m/day

17.84 m/day

Table 5. Overall performance of TBM tunnelling at tunnel section between SVS and C&C

tunnel.

Figure 11. TBM S881 and TBM S882 breakthrough at the SLF. Figure 11. TBM S881 and TBM S882 breakthrough at the SLF.

An iterative process is applied to the limit equilibrium method to derive a balance between the driving force from the wedge and the silo, and the resisting force from the sliding planes ABEF (back), ADE (side) and BCF (side). The minimum confinement pressure is determined by varying the inclination angle ω until it provides the maximum driving shear force.

The maximum confinement pressure was calculated according to the recommendation in the GEO Report No. 249 (Geotechnical Engineering Office, 2009) to prevent excessive pressure leading to blow-out. The calculation was based on the dead weight of the soil above the tunnel as a conservative approach. Both Ultimate Limit State (ULS) and Serviceability Limit State (SLS) were checked to assess which state controls the design of the confinement pressure.

Figure 12. Silo model by Horn (1961).

An iterative process is applied to the limit equilibrium method to derive a balance

between the driving force from the wedge and the silo, and the resisting force from the sliding

Figure 12. Silo model by Horn (1961).

3.1.2. Design methodology of TBM confinement pressure by finite element analysis

The interaction between the tunnels and surrounding environment is a complex and non-linear 3D problem. Convergence-Confinement Method (Panet, 2001; Almog et al., 2015) was adopted to predict the stress state and estimate the ground settlements in a 2D finite element analysis. The confinement pressures applied at the tunnel boundary in the mesh is an overall average of the slurry pressure at the face, shield pressure and grout pressure around the lining (Figure 13). The overall average pressure is based on the work by Aristaghes and Autuori (2003) who suggested that the equivalent over-pressure,

planes ABEF (back), ADE (side) and BCF (side). The minimum confinement pressure is

determined by varying the inclination angle ω until it provides the maximum driving shear

force.

The maximum confinement pressure was calculated according to the recommendation

in the GEO Report No. 249 (Geotechnical Engineering Office, 2009) to prevent excessive

pressure leading to blow-out. The calculation was based on the dead weight of the soil above

the tunnel as a conservative approach. Both Ultimate Limit State (ULS) and Serviceability

Limit State (SLS) were checked to assess which state controls the design of the confinement

pressure.

3.1.2. Design methodology of TBM confinement pressure by finite element analysis

The interaction between the tunnels and surrounding environment is a complex and

non-linear 3D problem. Convergence-Confinement Method (AFTES, 2001; Almog et al.,

2015) was adopted to predict the stress state and estimate the ground settlements in a 2D finite

element analysis. The confinement pressures applied at the tunnel boundary in the mesh is an

overall average of the slurry pressure at the face, shield pressure and grout pressure around

the lining (Figure 13). The overall average pressure is based on the work by Aristaghes and

Autuori (2003) who suggested that the equivalent over-pressure, Pf' = 0.5Pgrout' + 0.5Pslurry' ,

applied to the tunnel crown satisfying the settlement and volume loss criteria for input into the

2D finite element analysis.

, applied to the tunnel crown satisfying the settlement and volume loss criteria for input into the 2D finite element analysis.

Figure 13. Pressures during TBM tunnelling.

3.2. TBM confinement pressure control during TBM tunnelling

3.2.1. TBM confinement pressure control during TBM excavation

TBM S880, TBM S881 and TBM S882 were closed face slurry (mixshield) TBMs and

capable of tunnelling through soft ground and mixed ground conditions within acceptable

volume loss limits. The mixshield technology developed by the TBM manufacturer,

Herrenknecht AG, is an advance on conventional slurry technology with the support pressure

in the excavation chamber precisely managed using an automatically controlled air cushion.

Figure 14 illustrates the general arrangement of the mixshield TBM.

Figure 13. Pressures during TBM tunnelling.


3.2. TBM confinement pressure control during TBM tunnelling

3.2.1. TBM confinement pressure control during TBM excavation

TBM S880, TBM S881 and TBM S882 were closed face slurry (mixshield) TBMs and capable of tunnelling through soft ground and mixed ground conditions within acceptable volume loss limits. The mixshield technology developed by the TBM manufacturer, Herrenknecht AG, is an advance on conventional slurry technology with the support pressure in the excavation chamber precisely managed using an automatically controlled air cushion. Figure 14 illustrates the general arrangement of the mixshield TBM.

The range of minimum and maximum confinement pressure along the tunnel alignment at the crown level was calculated using the methods mentioned in Section 3.1. At the start of each excavation cycle, the designed confinement pressure was manually set using the Samson system in the TBM and displayed on the slurry circuit control panel which was automatically controlled by the Samson air pressure regulation system. This system kept the controlled compressed-air cushion at the exact set point pressure precisely and fluctuations in the bentonite slurry circuit could therefore compensated efficiently. Stone crusher and grid (Figure 15) were installed in front of the intake to keep the slurry suction line free of hard lumps of soil and large rocks. The parameters were checked against the ring

by ring parameter sheets and continuous monitoring of the slurry pressure and throughout the excavation cycle.

Figure 15. Stone crusher and grid.

There was an incident during the TBM break-in operation at the NLF (Figure 16).

When the TBM excavated through the diaphragm wall and break-in plug, the concrete or

cement was pumped to the slurry treatment plant (STP) and led to foaming and overflow of

the slurry tank. This incident reminds that the bentonite particles are flocculated by cement,

resulting in increased rheology and fluid loss. The severity of flocculation depends on the

quantity and the quality that solids present and the solubility of the Ca++ ion. Chemical

treatment such as sodium bicarbonate could be used to remove cement contamination and

control the pH while removing the calcium and excess calcium hydroxide in cement. For the

overflow of slurry tank, the main objective was to reduce the tank volume which could be

achieved by delivering the slurry to the centrifuge or filter press after flocculating the slurry at

STP.

Figure 15. Stone crusher and grid.

There was an incident during the TBM break-in operation at the NLF (Figure 16). When the TBM excavated through the diaphragm wall and break-in plug, the concrete or cement was pumped to the slurry treatment plant (STP) and led to foaming and overflow of the slurry tank. This incident reminds that the bentonite particles are flocculated by cement, resulting in increased rheology and fluid loss. The severity of flocculation depends on the quantity and the quality that solids present and the solubility of the Ca++ ion. Chemical treatment such as sodium bicarbonate

Figure 14. General arrangement of the mixshield TBM.

The range of minimum and maximum confinement pressure along the tunnel

alignment at the crown level was calculated using the methods mentioned in Section 3.1. At

the start of each excavation cycle, the designed confinement pressure was manually set using

the Samson system in the TBM and displayed on the slurry circuit control panel which was

automatically controlled by the Samson air pressure regulation system. This system kept the

controlled compressed-air cushion at the exact set point pressure precisely and fluctuations in

the bentonite slurry circuit could therefore compensated efficiently. Stone crusher and grid

(Figure 15) were installed in front of the intake to keep the slurry suction line free of hard

lumps of soil and large rocks. The parameters were checked against the ring by ring parameter

sheets and continuous monitoring of the slurry pressure and throughout the excavation cycle.

Figure 14. General arrangement of the mixshield TBM.



could be used to remove cement contamination and control the pH while removing the calcium and excess calcium hydroxide in cement. For the overflow of slurry tank, the main objective was to reduce the tank volume which could be achieved by delivering the slurry to the centrifuge or filter press after flocculating the slurry at the STP. 3.2.2. TBM confinement pressure control during TBM cutterhead maintenance

Regular TBM cutterhead inspection is required to monitor the cutter tool wear and replace worn tools. During the cutterhead interventions, compressed air was used to balance the hydrostatic and earth pressures, creating a hyperbaric environment in the TBM excavation chamber.

Since the tunnel depth was up to 55 m below the sea level with the deepest seabed level at -21 mPD, the intervention pressure could be up to six bars and exemption was granted from the Labour Department, the HKSAR Government to carry out hyperbaric works above 3.45 bars. The following scenarios were considered for the TBM cutterhead intervention:

Figure 16. Foaming and overflow of the slurry tank.

3.2.2. TBM confinement pressure control during TBM cutterhead maintenance

Regular TBM cutterhead inspection is required to monitor the cutter tool wear and

replace worn tools. During the cutterhead interventions, compressed air was used to balance

the hydrostatic and earth pressures, creating a hyperbaric environment in the TBM excavation

chamber.

Since the tunnel depth was up to 55 m below the sea level with the deepest seabed level at

-21 mPD, the intervention pressure could be up to six bars and exemption was granted from

the Labour Department, the HKSAR Government to carry out hyperbaric works above 3.45

bars. The following scenarios were considered for the TBM cutterhead intervention:

Figure 16. Foaming and overflow of the slurry tank.

• Avoidance of the intervention in adverse ground conditions such as mixed face of CDG/ rock, porous layer of alluvial sand or marine deposit under consolidation process.

• Monitoring and review of the ground conditions by the material discharged from the STP, TBM

penetration rates, quality, pressure of annular grout and relevant geology data.

• Review of the slurry quality, air pressure for the intervention, slurry level in excavation chamber and anticipated intervention duration.

Prior to the TBM cutterhead intervention, fresh bentonite and/ or polymers was added to improve the slurry properties to develop a good bentonite filter-cake on the excavation face and around the TBM shield to ensure the face stability.

Intervention with compressed air under high pressure will pose occupational safety and health (OSH) risk to the hyperbaric workers due to the narcosis effects and high dosage of oxygen. The contractor installed a few special features to minimise the manned intervention and monitor the TBM cutterhead and excavation face conditions. Mobydic was installed on some of the cutters as a predictive monitoring system to monitor the geological conditions, cutterhead operation and wear. This system provided real-time visualisation of the excavation face and wearing condition of cutters that allowed the TBM pilot to adjust the TBM parameters in a timely manner to minimise the damages to the cutting tools.

Snake was a remotely operated, endoscopic inspection device which allowed inspection of the tunnel face and the cutterhead without the need for manned intervention. Telemach was a robotic arm for the remote replacement of worn TBM cutters without the need for manned intervention. These devices were developed as prototypes for TM-CLKL tunnel and further research and improvement would be required for future use.

4. Conclusion

The TM-CLKL project demonstrated the success of large TBMs tunnelling through newly reclaimed land and sub-sea. Understanding the tunnel behaviour at different ground conditions is the key to the success of the TM-CLKL project. The geotechnical risks in terms of soil condition, ground cover, consolidation behaviour, face stability, sensitive structures, ovalisation of segment linings, volume loss and settlement have all been identified early in the project and proactive mitigation measures implemented before the tunnels arrived. Detailed ground investigation and highly intensive numerical tunnel modelling have also been carried out to identify the geotechnical risks. Significant ground improvement methods such as break-in plug, break-out plug, parking plug, deep band drains and barrettes have been constructed together with the close monitoring implemented during the excavation, the tunnelling part of the project has now been successfully completed within the anticipated programme.


Acknowledgements

The authors are grateful for the support of the Highways Department, the HKSAR Government. Their support contributes in expediting the approval process. However, the contents of this paper do not necessarily reflect the views and policies of this supporting organisation, nor does the mention of trade names and commercial products constitute the endorsement or recommendation for use.

Notes on Contributors

Mr Lucas Kwan Lok Chan is a Project Engineer at the AECOM Asia Company Limited, working on TM-CLKL project tha t involves geotechnical-related works, including geotechnical parameters interpretation, instrumentation and monitoring works, design and site supervision of TBM

tunnelling and caterpillar C&C tunnel. Mr Chan was responsible for the site supervision of C&C tunnel of Central – Wanchai Bypass and Island Eastern Corridor Link project that involved diaphragm wall construction, excavation and lateral support system (ELS) and tunnel construction using top-down slabs.

Mr Chan was graduated from The Hong Kong Polytechnic University and is a member of the Institution of Civil Engineers. He worked in contractor firms before joining the AECOM.

Ir Dr Alan Kwok Lun Kwong is a Senior Resident Engineer at the AECOM Asia Company Limited, working on TM-CLKL project that involves significant geotechnical and tunnel design and construction controls. Ir Dr Kwong was responsible for the remediation of the Harbour Area Treatment Scheme (HATS)

Stage I tunnel excavation that involved the appraisal of geotechnical and geological conditions of the ground during tunnelling as well as assessing different options of ground support, stabilisation measures and methods of pre-grouting to reduce inflow.

From 2002 to 2009, Ir Dr Kwong worked as a Senior Teaching Consultant at The University of Hong Kong, teaching M.Sc. courses in Foundation Engineering, Underground Excavation, Tunnelling, Ground Investigation, Soil Testing and Rock Mechanics.

Ir Dr Kwong was the Principal Geotechnical Engineer of Montgomery Watson Harza (1999-2002) and the Associate Director of Golder Associates Hong Kong Ltd (1995-1999).

References

[1] Almog E, Mangione M and Cachia G (2015). Ground relaxation in segmental lining design using the convergence-confinement method. In: Proceeding of the Underground Design and Construction Conference 2015. Hong Kong: The Institute of Materials, Minerals and Mining (IOM3). Hong Kong Branch, pp. 335-345.

[2] Ananostou G and Kovari K (1996). Face stability in slurry and EPB shield tunnelling. In: Proceedings of the Symposium on Geotechnical Aspects of Underground Construction in Soft Ground. Rotterdam: Balkema, pp. 453-458.

[3] Aristaghes P and Autuori P (2003). Confinement efficiency concept in soft ground bored tunnels. In: Proceedings of ITA World tunnelling Congress. (Re)Claiming the underground space. Amsterdam: Balkema, pp. 909-913.

[4] Geotechnical Engineering Office (2009). Ground control for slurry TBM tunnelling. Hong Kong: Civil Engineering and Development Department (CEDD), the HKSAR Government, pp. 57.

[5] Horn N (1961). Horizontaler erddruck auf senkrechte abschlussflachen von tunnelrohren. In: Landeskonferenz der Ungarischen Tiefbauindustrie. Deutsche Uberarbeitung durch STUVA: Dusseldorf, pp. 7-16.

[6] Kimura T and Mair RJ (1981). Centrifugal testing of model tunnels in soft clay. In: Proceedings of the 10th International Conference on Soil Mechanics and Foundation Engineer. Rotterdam: Balkema, pp. 319-322.

[7] Panet M (2001). The Convergence - Confinement method. The French Tunnelling and Underground Space Association (AFTES) Recommendations, 2002(174), pp. 1-11.


Mr Lucas Kwan Lok Chan is a Project Engineer at the AECOM Asia Company Limited, working on TM-CLKL project that involves geotechnical-related works, including geotechnical

parameters interpretation, instrumentation and monitoring works, design and site supervision of TBM tunnelling and caterpillar C&C tunnel. Mr Chan was responsible for the site supervision of C&C tunnel of Central – Wanchai Bypass and Island Eastern Corridor Link project that involved diaphragm wall construction, excavation and lateral support system (ELS) and tunnel construction using top-down slabs.


Ir Dr Alan Kowk Lun Kwong is a Senior Resident Engineer at the AECOM Asia Company Limited, working on TM-CLKL project that involves significant geotechnical and tunnel design and construction controls. Ir Dr

Kwong was responsible for the remediation of the Harbour Area Treatment Scheme (HATS) Stage I tunnel excavation that involved the appraisal of geotechnical and geological conditions of the ground during tunnelling as well as assessing different options of ground support, stabilisation measures and methods of pre-grouting to reduce inflow.




Mr Lucas Kwan Lok Chan is a Project Engineer at the AECOM Asia Company Limited, working on TM-CLKL project that involves geotechnical-related works, including geotechnical

parameters interpretation, instrumentation and monitoring works, design and site supervision of TBM tunnelling and caterpillar C&C tunnel. Mr Chan was responsible for the site supervision of C&C tunnel of Central – Wanchai Bypass and Island Eastern Corridor Link project that involved diaphragm wall construction, excavation and lateral support system (ELS) and tunnel construction using top-down slabs.


Ir Dr Alan Kowk Lun Kwong is a Senior Resident Engineer at the AECOM Asia Company Limited, working on TM-CLKL project that involves significant geotechnical and tunnel design and construction controls. Ir Dr

Kwong was responsible for the remediation of the Harbour Area Treatment Scheme (HATS) Stage I tunnel excavation that involved the appraisal of geotechnical and geological conditions of the ground during tunnelling as well as assessing different options of ground support, stabilisation measures and methods of pre-grouting to reduce inflow.



Geotechnical risks, mitigation measures and performance of ...

Documents

Transcript of Geotechnical risks, mitigation measures and performance of ...