- Email: [email protected]
Settlement trough parameters for tunnels in Irish glacial tills
Tunnelling and Underground Space Technology 27 (2012) 1–12
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Settlement trough parameters for tunnels in Irish glacial tills B.A. McCabe a,⇑, T.L.L. Orr b, C.C. Reilly b, B.G. Curran a a b
Department of Civil Engineering, National University of Ireland, Galway, Ireland Department of Civil, Structural and Environmental Engineering, Trinity College Dublin, Ireland
a r t i c l e
i n f o
Article history: Received 1 December 2010 Received in revised form 30 May 2011 Accepted 9 June 2011 Available online 16 July 2011 Keywords: Transverse settlement profiles Glacial till Trough width Tunnelling Volume loss Boulders
a b s t r a c t Glacial till (or boulder clay) is the most widespread sediment on the island of Ireland. The behaviour of these tills, especially Dublin Boulder Clay, is now better understood as a result of ground investigations and associated testing for major construction projects, particularly in the east of the country over the past decade. Despite an increase in tunnelling activity in the country over the same time period, there is very little documented evidence on the settlements induced by tunnelling operations in glacial till. In this paper, transverse surface settlement data from two glacial till sites are presented. Four profiles are presented for the ‘soft ground’ TBM-bored section of the Dublin Port Tunnel. Nine profiles are reported for pipe-jacked microtunnels constructed for a sewerage scheme in Mullingar in the Midlands; one of which was measured at the top of a railway embankment under which the pipeline passes. The measured settlements have been interpreted using a standard Gaussian error function, and trough width parameters show dependence on the fraction of the till, fine or coarse, that governs its behaviour. In addition, conservative design estimates of maximum trough settlement and volume loss have been provided, and the impact of boulders is discussed. This paper provides the first empirical guidance for predicting surface settlements above tunnels in Irish soils. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Irish tunnelling experience has developed considerably since the Orr and Farrell (1996) report on tunnelling activity in the country 15 years ago, at which time tunnels had yet to be used to carry roads or motorways. The Jack Lynch Tunnel under Cork’s River Lee was opened in May 1999; this submerged twin-bore tunnel is 610 m long and is estimated to have cost €133 m. The €752 m Dublin Port Tunnel (DPT), when opened in December 2006, became the longest urban tunnel in Europe (4.7 km twin bore) and the nation’s largest ever civil engineering project. An important coupled outcome of the DPT project was the improved grasp of both the stratigraphy and engineering properties of Dublin Boulder Clay that arose from handling the extensive geotechnical data generated (e.g. Long and Menkiti, 2007). Another highly significant project, the €660 m and 675 m long twin-bore submerged Limerick tunnel under the River Shannon, was opened in September 2010. In parallel with the development of large scale tunnels, the application of pipe-jacking to the construction of pipelines for utilities such as water, sewage and gas in a variety of soil and rock types has escalated. Curran (2010) interprets jacking force data for slurry-shield pipe-jacked microtunnels at eleven Irish sites encompassing a variety of soil and rock types. ⇑ Corresponding author. Tel.: +353 91 492021; fax: +353 91 494507. E-mail address: [email protected] (B.A. McCabe). 0886-7798/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tust.2011.06.002
Lochaden et al. (2008) report on successful efforts to model settlements above the DPT using finite element software, with the predictions relatively insensitive to the constitutive soil model used. In practice, however, predictions of surface settlement troughs over tunnels in Irish soils have to date relied heavily on experience from the UK and beyond, such as the recommended Gaussian trough width parameters published by Mair and Taylor (1997) for tunnels in sands/gravels and clays with generally a small range of particle sizes, and the achievable volume losses suggested by Mair (1996) and others. Some Irish soils have particular characteristics that significantly affect their behaviour; for example, Irish boulder clays are often well-graded and composed of a wide range of particle sizes from clay-size to gravel-size, so that some of these soils are characterised as fine soils while others are characterised as coarse soils (Farrell, 2010). For this reason, it was deemed important to analyse tunnel-induced settlements in these soils, and also to consider the impact of boulders that are sometimes encountered. In this paper, settlement data is provided for (i) large diameter open-face shielded tunnels in glacial till for the DPT project, and (ii) small diameter closed-face pipe-jacked tunnels in a glacial till at Mullingar, Co. Westmeath; including data for a section of pipeline passing beneath a high railway embankment. The inferred Gaussian parameters are an important record of Irish tunnelling-induced settlement experience and also, in the case of the Dublin Port Tunnel, supplement the limited data published internationally for tunnels with low cover to diameter ratios. All of the aforemen-
2
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Nomenclature AVN cu D DBC DPT i k MH N NTc
remote control tunnelling machine with slurry material removal undrained shear strength tunnel diameter (including overcut) Dublin Boulder Clay Dublin Port Tunnel value of y corresponding to point of inflection of Gaussian trough Gaussian trough width parameter manhole stability number stability number at collapse
tioned site locations are illustrated in Fig. 1; in addition to a Belfast site with older data which is referred to later in the paper. 2. Tunnelling projects at Dublin and Mullingar 2.1. Dublin Port Tunnel The Dublin Port Tunnel is a 4.7 km long urban road transport twin-tunnel constructed to ease congestion in Dublin city centre by removing heavy vehicular traffic to and from Dublin Port and rerouting it from the centre of Ireland’s capital to the outer orbital M50 motorway. Construction of the DPT began in 2001 and the tunnel was opened in 2006. Construction methods used included cut and cover, pipe jacking, hard rock TBM tunnelling and ‘soft ground’ TBM tunnelling. The section of the project considered in this paper is the 325 m section of twin tunnel bored in boulder clay at Whitehall, towards the northern end of the scheme (Fig. 2). An 11.77 m diameter Herrenknecht open-face shielded back-actor machine was used for this section. Summary information on the surface settlement profiles is provided in Table 1.
Belfast Mullingar
Dublin Limerick
Cork
Fig. 1. Map showing locations of Irish tunnelling projects referred to in this paper.
N60 sv smax
rt rv TBM Vl Vs Vt y z0
standard penetration test value vertical settlement at any point in trough vertical settlement over centreline of tunnel (at y = 0) tunnel internal support pressure vertical total stress at the depth of the tunnel axis Tunnel Boring Machine volume loss volume of Gaussian settlement trough per metre length volume of tunnel per metre length transverse horizontal distance from tunnel centreline depth to tunnel axis
The data published here, for four different chainages, relate to the southbound bore which was the first to be constructed. Interpretation of the settlement data measured during construction of the northbound bore will be the subject of a future publication. Goto et al. (2004) make general observations from settlement profiles in the Whitehall area as part of an overview paper on the DPT scheme. Additional data and a fuller interpretation is provided in this paper. 2.2. Mullingar Approximately 750 m length of pipe-jacked tunnels were constructed in Mullingar between July and October 2009 as part of the Mullingar Sewerage Improvement Scheme. Herrenknecht AVN (remote-controlled slurry-shield tunnelling) machines deploying a ‘mixed’ head (incorporating both scraping teeth and cutting discs) of the type shown in Fig. 3 were used. Two different sized Tunnel Boring Machines (TBMs) were used, AVN 1200 and AVN 1800, with external diameters of 1505 mm and 2150 mm respectively. The relevant particulars of each are shown in Table 2a. The forces and penetration rates recorded during pipe-jacking in this contract have been interpreted elsewhere (Curran, 2010). The settlement data reported here pertain to distinct areas of the scheme, hereafter referred to as Areas A–C (see Table 2b and Fig. 4): Area A. Area A includes the Supervalu car park and access road to it from Austin Friars Street in the town centre. In this area, transverse profiles 1–7 and 13–19 were situated between manholes MH22 and MH23, and transverse profiles 26–32, 36–43 and 47–53 were situated between manholes MH23 and MH24. The pavement in the area was typically 200 mm thick and the tunnel axis is at a depth of between 3.6 m and 5.4 m. Area B. Area B is immediately west of the River Brosna, near the Mullingar Greyhound Stadium. In this area, transverse profiles 117–123, 127–133 and 159–165 all lie to the southern end of the section of pipeline between manholes MH9 and MH11 and the tunnel axis is at a depth of between 3.5 m and 4.1 m. Profiles 127–133 and 159–165 were measured on a concrete slab. Area C. Area C is immediately north of the Mullingar Greyhound Stadium. In this area, the pipeline between manholes MH11 and MH13 passes approximately 4 m below an 8 m high embankment supporting the Dublin–Sligo railway line. Settlements were recorded on points marked on both rails of each of three railway tracks. Note that the direction of the pipeline is approximately 13° off perpendicular to the embankment centreline at that point.
3
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Fig. 2. Map of Dublin Port tunnel showing soft ground TBM section at Whitehall.
Table 1 Settlement profiles at DPT. Drive ID
Settlement profile chainage (m)
Location
Depth to tunnel axis, z0 (m)
Tunnel diameter, D (m)
z0/d
Average penetration rates (m/day)
Predominant overburden material
194 194 194 194
2045 2060 2080 2120
Carpark Carpark Carpark Carpark
15.12 15.60 16.12 17.41
11.77 11.77 11.77 11.77
1.28 1.32 1.37 1.48
4 3.5 2 2
Boulder Boulder Boulder Boulder
SB SB SB SB
Precise electronic levelling, with an accuracy of 0.1 mm, was used to record tunnelling-induced settlements at both the DPT and Mullingar sites. Since the DPT site was mainly a rough carpark, the levels were not reinstated immediately after tunnelling. Instead the whole area was reinstated as part of the ancillary road construction works associated with the DPT. Cement grout was injected into the overcut to prevent further settlement at Areas A–C in Mullingar. 3. Ground conditions 3.1. Dublin port tunnel Ground conditions encountered at Whitehall were Dublin Boulder Clay (DBC), a stiff lodgement till. The engineering properties of
clay clay clay clay
DBC are covered extensively elsewhere (Farrell and Wall, 1990; Skipper et al., 2005; Long and Menkiti, 2007) while Orr and Farrell (1996) give a comprehensive description of its properties which have greatest relevance to tunnelling. DBC is a well-graded, stiff glacial outwash till containing about 35% fines (particles smaller than 0.06 mm), large stone boulders and lenses of sand or gravel. A representative DBC grading curve (once cobbles and boulders have been removed) approximates to a straight line on a semilog particle size distribution plot as shown by the mean particle size distribution plots in Fig. 5 that were obtained for four units within the Dublin Boulder Clay. A fines content in excess of 35% results in the soil being classified as fine and is sufficient to impart the behavioural characteristics of a fine grained soil, but the fines content in a glacial till can vary from site to site and with depth,
4
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
3.2. Mullingar Mullingar is a regional town with a population of under 20,000. Finch (1977) classified Co. Westmeath’s ground conditions, primarily for agricultural purposes, and referred to the sediment around Mullingar as the Rathowen Phase, a till derived predominantly from limestone and shale. Besides this work, there is little published on the geotechnical characteristics of this till. The till material present in the top 3–5 m, which governed the settlement over the jacked pipes, was found to comprise clay and gravel, but other constituents such as made ground and peat were also discovered, and the fines content varied from location to location. In general, it was found that standard penetration test N60 values were highly variable in the Mullingar till which is also a feature of Dublin Boulder Clay, as noted by Lehane and Simpson (2000). In subsequent interpretations of the Mullingar data, where the overburden to the pipes consists of more than one soil type, it has generally been the case that one soil type has a dominant influence (usually owing to greater thickness) than the other(s), unless stated otherwise. In general, the site investigation points for this sewerage scheme were chosen to coincide with the proposed manhole positions, but the jacking force and penetration rate records for the drives (Curran, 2010) were helpful in some cases for establishing the ground conditions between the manholes. A summary of the ground conditions in Areas A–C are given in the following paragraphs and these should be read in conjunction with the tunnel axis depths (z0) and diameters (D) shown in Table 2b:
Fig. 3. Mixed TBM head used at Mullingar.
Table 2a Mullingar site drive parameters. Machine
AVN 1200
AVN 1800
Pipe int. dia. (mm) Pipe ext. dia. (mm) TBM ext. dia. (mm) Torque (kNm) Revolutions per min Force per cylinder (kN) Max drive length (m) Shield length (mm) Shield weight (kN) Pipe weight (kN)
1200 1490 1505 195 0–6 753 200 3200 103 31.9
1800 2120 2150 445 0–3 1272 300 4200 245 58.0
as is evident in Fig. 5. The significant granular content causes difficulties during sampling operations and makes predictions of its overall behaviour difficult to interpret from laboratory tests on small samples. The high stiffness of DBC has meant that settlements over tunnels have not been much of an issue to date and in some cases are close to the limits of measurability. For the particular section of the ‘soft ground’ TBM drive considered in this paper, the ground conditions encountered during tunnelling showed some variation across the 11.77 m face. Tunnelling records describe the glacial till encountered as stiff brown sandy clay with fine to coarse subangular to rounded gravels occurring towards the crown, tending towards dense sandy gravels with some cobbles at the invert.
Area A. The MH23–MH22 and MH23–MH24 jacking records (Curran, 2010) indicate that the tunnel face was always in clayey and/or sandy gravel. Site investigation data shows that the natural overburden is predominantly gravel in the vicinity of profiles 1–7, 26–33 and 47–53 with grading curves indicating the following approximate proportions of the different soil types: gravel 65–70%, sand 15–20%, fines 10–20%. Near profile 13–19, the overburden appears to have a lower gravel content than at other locations, with 1.2 m thickness described as very silty, very sandy gravel (gravel 47%, sand 27%, fines 26%) and 0.3 m thickness described as clayey peat. A trial pit in the vicinity of profile 36–43 indicates an overburden comprising 1.6 m of slightly sandy slightly gravelly clay (with a representative grading indicating gravel 38%, sand 23% and fines 39%). Area B. For profiles 117–123 and 127–133, site investigation evidence suggests that the overburden is predominantly gravel (gravel 67%, sand 24%, fines 9%), although the soil in the tunnel face may actually straddle the division between fine and coarse soils. However, at profile 159–165, there is a significant thickness (1.5 m) of clayey peat and organic silt in the overburden material. There is also approximately 1.0 m of made ground at all three profile locations.
Table 2b Settlement profiles at Mullingar.
*
Area
Drive ID
Settlement profile ID
Location
Depth to tunnel axis, z0 (m)
Tunnel diameter, D (m)
z0/d
Average penetration rates for drive (mm/min)
Best estimate of predominant overburden material
A A A A A B B B C
MH23–MH22 MH23–MH22 MH23–MH24 MH23–MH24 MH23–MH24 MH9–MH11 MH9–MH11 MH9–MH11 MH11–MH13
1–7 13–19 26–32 36–43 47–53 117–123 127–133 159–165 –
Road Road Carpark Carpark Carpark Road Yard Yard Rail lines
5.39 4.28 5.05 3.60 5.05 4.09 3.95 3.5 12.05
1.505 1.505 1.505 1.505 1.505 2.150 2.150 2.150 2.150
3.58 2.84 3.36 2.39 3.36 1.90 1.84 1.63 5.60
12.8 12.8 32.3* 32.3* 32.3* 28.6 28.6 33.7 33.7
Gravel Gravel + peat Gravel Clay Gravel Gravel Gravel Clay/silt + peat Embankment clay + gravel
Penetration rate records only available for first 20 m of this drive.
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
5
Fig. 4. Location of settlement pins at MH23–MH24 and MH23–MH22 (Area A); MH9–MH11 (Area B) and MH11–MH13 (Area C, railway embankment).
Area C. From boreholes either side of the embankment and one carried out an angle beneath the embankment, the tunnelled face is interpreted to have passed through a combination of silty sandy gravel and gravelly silt, with gravel the prevalent overburden to the pre-embankment ground level. The embankment itself, constructed in 1847/48 from material imported to the site from cuttings elsewhere on the Dublin–Sligo railway around Mullingar, comprises sandy gravelly clay with gravel and
cobbles; i.e. a fine material. This embankment material was expected to have had the most influence on settlement of the rails above. 4. Gaussian settlement profiles Several authors (i.e. Peck, 1969; O’Reilly and New, 1982) have shown that the immediate surface settlement profile or trough
6
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
100
Percentage < d
80
60
40
Brown BC Upper Black BC Sandy Brown BC Lower Black BC
20 (Fine soil)
0 0.001
(Coarse soil)
0.01
0.1
1
10
100
Particle size d (mm) Fig. 5. Particle size distribution for Dublin Boulder clay.
Horizontal distance from tunnel centreline, y (m) -6
-4
-2
0 -5
2
4
6
i =kz0 =1.67
Settlement, s v (mm)
0
5
10 point of inflection 15
20 hatched volume (per metre run) is volume loss Vl
Settlement profile 13-19 in Area A, after 22 days
25
Eqn [1], with smax=25.1mm, z0=4.28m, k=0.39
smax 30
Fig. 6. Gaussian settlement trough with typical measured data.
above tunnels on a greenfield site can be represented adequately by a simple Gaussian or error function of the form:
sv 2 ¼ expðy2 =2i Þ smax
ð1Þ
where sv is the vertical settlement at a horizontal distance at y from the tunnel centre line, smax is the settlement at the centreline (y = 0) and i is the value of y corresponding to the point of inflection of the function in Eq. (1). The function is plotted in Fig. 6 and annotated with the relevant Gaussian parameters. O’Reilly and New (1982) also proposed a linear relationship between i and z0, the depth to the tunnel axis.
i ¼ kzo
ð2Þ
Recent research (Jones 2010) has shown that the relationship between i and z0 is not linear for z0 values exceeding 35 m, which is greater than the tunnels depths considered in this paper. The constant of proportionality, k, is known as the trough width parameter and is believed to be largely independent of the construction method (Mair and Taylor, 1997).
Mair and Taylor (1997) summarised a wide range of field data and concluded that 0.4 < k < 0.6 for clays with k showing no dependence on the clay’s stiffness. Corresponding data presented for tunnels in sands and gravels exhibit greater scatter, with the majority of the data bounded by 0.25 < k < 0.45. The volume of the settlement trough Vs per unit length can be determined by integrating Eq. (1). Volume loss (Vl), often used as a measure of quality control of the tunnelling process, is defined as the ratio of Vs to the volume of the tunnel per metre length, Vt:
Vl ¼
Vs ¼ Vt
pffiffiffiffiffiffiffi z s 2pismax max 2 3:192k 0 pD D D
ð3Þ
4
Based on a review of settlement data for bored tunnels, Mair (1996) concluded that: (i) For open-face tunnelling in stiff clays, such as London Clay, Vl values of between 1–2% are achievable. (ii) For closed-face tunnelling, using earth pressure balance or slurry shields, Vl values of 0.5% in sands and 1–2% in soft clays (excluding consolidation settlements) are achievable.
7
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12 Table 3a Settlement parameters at DPT. Settlement profile chainage (m)
z0/D
Time since TBM passed (days)
i
i/D
k
smax (mm)
Vl (%, from Eq. (3))
Predominant overburden material
2045 2060 2080 2120
1.28 1.32 1.37 1.48
183 188 200 240
7.65 7.77 7.53 9.53
0.65 0.66 0.64 0.81
0.51 0.50 0.47 0.55
11.8 19.6 34.3 75.2
0.21 0.35 0.64 1.66
Boulder Boulder Boulder Boulder
clay clay clay clay
Table 3b Settlement parameters at Mullingar. Settlement profile ID
z0/D
Time since TBM passed (days)
i
i/D
k
smax (mm)
Vl (%, from Eq. (3))
Best estimate of predominant overburden material
1–7 13–19
3.58 2.84 2.84 2.84 2.84 2.84 2.84 3.36 3.36 2.39 3.36 1.90 1.90 1.90 1.84 1.84 1.84 1.63 5.60
14 4 14 22 53 63 81 19 33 29 35 4 6 8 2 6 16 14 1–3
1.51 1.71 1.80 1.67 1.88 1.80 1.84 1.52 1.57 1.51 1.41 1.1 1.02 0.9 0.91 0.95 0.87 1.47 5.6–7.2
1.00 1.14 1.19 1.11 1.25 1.19 1.22 1.01 1.04 1.00 0.94 0.51 0.48 0.42 0.42 0.44 0.40 0.68 2.6–3.4
0.28 0.40 0.42 0.39 0.44 0.42 0.43 0.30 0.31 0.42 0.28 0.27 0.25 0.22 0.23 0.24 0.22 0.42 0.5–0.6
18.4 13.8 21.7 25.1 31.2 30.2 31.7 10.8 12.5 7.1 3.8 19.9 23.9 28.7 21.3 23.0 26.2 8.6 5.0–7.0
3.91 3.32 5.5 5.9 8.3 7.65 8.22 2.31 2.76 1.51 0.76 1.52 1.69 1.79 1.34 1.57 1.57 0.89 2.3–3.5
Gravel Gravel + peat
26–32 36–43 47–53 117–123
127–133
159–165 Railway
In overconsolidated clays, Vl is often estimated using the method of Macklin (1999):
N V l ¼ 0:23 exp 4:4 NTc
ð4Þ
In Eq. (4), Vl depends on the ratio of present stability number N to the stability number at collapse NTc. The stability number N is defined as:
N¼
rv rt cu
ð5Þ
where rv is the vertical total stress at tunnel axis level, rt is the internal support pressure and cu is the undrained shear strength of the clay. Macklin (1999) and Devriendt (2010) provide empirical guidance on the choice of NTc drawn from various experimental studies which, when combined, span a wide range of possible cover to diameter and unsupported advance length to diameter ratios.
5. Interpretation of settlement data 5.1. Interpretation methods Only settlement profiles showing a high level of symmetry about the vertical axis and with minimal measurement errors at the extremities (where settlements were small) are included in Table 3a (DPT) and Table 3b (Mullingar) and the discussion which follows. A typical measured settlement profile, 13–19 in Area A, is shown in Fig. 6. The following steps were taken in the interpretation:
Gravel Clay Gravel Gravel
Gravel
Clay/silt + peat Emb. clay + gravel
(i) The maximum value of settlement recorded (i.e. at y = 0) was set as smax in Eq. (1), whereupon the value of k was varied to provide the best match between the measured data and the profile predicted by Eq. (1). In general, the subjectivity of the match for k is believed to be ±0.02. (ii) The settlement profile data were plotted on Fig. 7 in Peck’s (1969) normalised format, as i/D against z0/D. The plot is overlain with lines representing constant values of k. (iii) The Vl values quoted in Table 3a and 3b were computed using Eq. (3) adopting the best fit values of k. 5.2. Discussion of k values In Fig. 7, the plotted data are classified according to the soil type in the overburden. The following observations can be drawn: (i) The settlement profiles for the tunnels in the Mullingar glacial gravels tend to be represented by k values that lie consistently in the range 0.2 < k < 0.3, which is at the lower end of the range indicated by Mair and Taylor (1997) for tunnels in sands and gravels. The exception is the k = 0.42 value for profile 13–19, which is probably due to the higher fines content in the gravel and/or the presence of peat in the overburden at that location. (ii) The settlement profiles for the tunnels in boulder clays tend to be modelled by k values in the range 0.4 < k < 0.6, which is also consistent with Mair and Taylor (1997). The DPT data (enclosed by a dashed loop to identify them) show marginally higher k values than the Mullingar data (excluding that at z0/D = 5.6), and are largely consistent with k values inferred from hand-excavated shield tunnelling in Belfast
8
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Normalised horizontal distance to point of inflection (i/D) 0
0.5
1
1.5
2
2.5
3
3.5
4
0
DPT data
All data is from Mullingar unless specified
1
fine soils
Belfast data (Glossop et al. 1979)
coarse soils
Normalised depth z0/D
2
3
range for Area C: railway embankment
some peat in overburden 4
5
k = 0.7 6
k = 0.6 k = 0.2
k = 0.3
k = 0.4
k = 0.5
7 Fig. 7. Plot of i/D against z0/D for Mullingar, DPT and Belfast sites.
Normalised maximum settlement smax /D 0.000 0
0.005
0.010
0.015
0.020
0.025 fine soils coarse soils
1
183
200
Normalised depth z0 /D
240 14
188
2
2
6
4
16 6
boulders encountered 8 63 & 81
29
3
19
33 4
14
22
53
35 14
4
5 1-3
6
points annotated with time (days) since TBM passed; see Table 3, arrows track variation over time
Fig. 8a. Normalised maximum surface settlement versus normalised tunnel depth for Mullingar and DPT sites.
soft estuarine clay (sleech) published by Glossop et al. (1979). In general, there is no clear relationship between the strength of these clays and the k value which models its settlement trough. (iii) At Area C, the deviation of the tunnel alignment from perpendicular to the embankment longitudinal axis was approximately allowed for by correcting the horizontal distance measured along the rails by cos(13°). However, due to the small smax values recorded, it was not possible to assess the k value relevant to the settlement profiles recorded at the top of the railway embankment (z0/D = 5.6) to the same accuracy as the k values assessed from the other data. Interestingly, the inferred range 0.5 < k < 0.6 confirms
that the fine material in the embankment is more influential in determining the shape of the settlement trough than the underlying gravel overburden. (iv) Some of the Mullingar data in Fig. 7 were recorded at a single point in time, whereas in other cases ‘range’ bars are shown for profile data where reliable profiles were recorded over a period of time (see Table 3a and 3b for full detail). In general, the ranges of k values shown were found to be small, within the range of natural variation and are not believed to be due to systematic time effects. (v) In clay soils incorporating significant consolidation settlements, there is evidence that Eq. (1) does not model the shape of a settlement trough well (e.g. Glossop et al.,
9
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
5.3. Discussion of maximum settlements
Normalised maximum settlement smax/D 0
0.01
0.02
0.03
0.04
0.05
Values of smax, which are required along with k in Eq. (3) to predict Vl show no obvious dependency on depth z0 when both are plotted in normalised form with respect to the tunnel diameter in Fig. 8a. Interestingly, the smax/D ranges for fine and coarse soil types are clearly distinct for these Irish data. In Fig. 8b, the authors have recast the smax data for settlement profiles of tunnels in clays and sands/gravels tabulated by Mair and Taylor (1997) into the same format as Fig. 8a, which shows that such a clear separation is not typical. Limiting values of smax/D are not theoretically justifiable based on Eq. (3) but are affected by the tunnelling method, the soil type and the geometry of the design situation, being linearly dependent on Vl and inversely dependent on k and z0/D. Higher smax/D values (between 0.007 and 0.022) are associated with the soils containing boulders for the Mullingar tunnelling, highlighted in Fig. 8a. The times elapsing between the passing of the TBM and each set of settlement profile measurements are included on Fig. 8a. The time dependence of smax in the coarse soils is somewhat unexpected. The authors believe that the most likely explanation relates to the bentonite lubricant used to support the excavation and prevent the ground from completely collapsing onto the pipeline. During drives in gravel, and especially where boulders were encountered, a high concentration of bentonite was mixed with water, forming a bentonite fluid with Marsh Funnel times in excess of 100 s. This highly viscous fluid would take a long time to dissipate and might explain the continued settlement over days and weeks. The data in Fig. 8a is reproduced in Fig. 9a (fine) and Fig. 9b (coarse) in an alternative format with the normalised surface settlement shape function k(smax/D) replacing smax/D on the horizontal axis, allowing contours of volume loss (based on Eq. (3)) to be superimposed on these figures. The relevant values of volume loss pertinent to each datapoint can be estimated visually from the Vl contours in Figs. 9a and 9b. These datapoints show distinct ranges for the volume losses when tunnelling in fine and coarse soils with Vl generally less 1% in fine soils and generally greater than 1% in coarse soils. The data of Mair and Taylor (1997) is also provided in Fig. 9c for comparison.
0
1
2
Normalised depth z0/D
3
4
5
6
7
8 fine soils
9
coarse soils
10 Fig. 8b. Normalised maximum surface settlement versus normalised tunnel depth for Mair and Taylor (1997) data.
1979; Mair and Taylor, 1997). While there is insufficient data to draw any inferences on consolidation settlements in the fine soils at Mullingar, these are believed to be small and the Gaussian functions capture the settlement troughs very well. Likewise, consolidation settlements in stiff DBC are not considered significant to this analysis.
Normalised surface settlement shape function k(s max /D) 0.0001 0
0.001
Normalised depth z0 /D
1
183
188
0.01
200
0.1
240
14
2 29
3
0.1% 4
V l values 5 1-3
6
0.5%
1%
2%
points annotated with time (days) since TBM passed; see Table 3
5%
10%
Fig. 9a. Normalised surface settlement shape function versus normalised tunnel depth for fine soils at Mullingar and DPT sites.
10
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Normalised surface settlement shape function k(smax /D) 0.0001 0
0.0 01
0. 01
0. 1
1
Normalised depth z 0 /D
6 & 16 2
2
6&8
4
53, 63 & 81 4
3
14 & 22
19 33
0.1%
boulders encountered
35 14
4
V l values 5 points annotated with time (days) since TBM passed; data linked by brackets pertain to same settlement profile 6
0.5%
1%
2%
5%
10%
Fig. 9b. Normalised surface settlement shape function versus normalised tunnel depth for coarse soils at Mullingar and DPT sites.
Normalised surface settlement shape function k(smax/D) 0.0001 0
0.0 01
0.0 1
0.1
1
Normalised depth z 0/D
2 3
0.1%
4 5 6
V l values
7 8 fine soils coarse soils
9 10
0.5%
1%
2%
5%
10%
20%
50%
Fig. 9c. Normalised surface settlement shape function versus normalised tunnel depth for Mair and Taylor (1997) data.
5.4. Discussion of volume losses Tunnelling operations in DBC were successful in keeping volume losses comfortably within or below the range of 1–2% for stiff clays suggested by Mair (1996). Eq. (4) yields Vl 0.45–0.5% based upon N 0.8 (assuming a constant cu = 450 kPa for DBC quoted by Farrell et al. (1995) and NTc 4.8 estimated from Devriendt (2010). This predicted range falls in the middle of a wider range of Vl values in Table 3a; three of the four profiles have Vl values in the range 0.21–0.64%, while at chainage 2120, where the highest value of smax = 75 mm was recorded, a Vl value of 1.66% was noted. Volume losses at Mullingar were more variable and reflect the variety of ground conditions encountered there. In general, values
of Vl were computed to lie between 0.7% and 1.8% approximately, except: Beneath the railway embankment at Area C, where volume losses in the range 2.3–3.5% apply. These can be explained by the higher z0/D (=5.6) value for the pipeline at this location, compared to nearby profile 159–165 in Area B for instance (Vl = 0.89%, z0/D = 1.63). These relative volume losses are consistent with the proportional relationship between Vl and z0/D implied in Eq. (3). In sections where boulders were encountered (profiles 1–7, 13–19, 26–32 in Area A) Vl values range from 2.3–8.2%. It was possible to deduce when boulders were encountered from the
11
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
30
3000 Face Resistance
25
2000
20
1500
15
1000
10
500
5
0
Pressure generating Cutter Head Torque (MPa)
Face Resistance (kN)
Max Torque
2500
0 0
10
20
30
40
50
60
70
Jacked Distance (m) Fig. 10a. Interpreted face resistance and cutter head torque pressure for the drive from MH23 to MH22.
60
Penetration Rate (mm/min)
50
40
30 moving average (every 5 datapoints averaged)
20
10
0 0
10
20
30
40
50
60
70
Jacked Distance (m) Fig. 10b. Moving average of penetration rate for the drive from MH23 to MH22.
interpreted face resistance, cutter head torque and penetration rate records, such as those for drive MH23–MH22 adapted from Curran (2010). Substantial increases in face resistance are seen to occur for jacked distances in excess of 42 m (Fig. 10a) due to the presence of boulders, with a corresponding drop in penetration rate (Fig. 10b). The higher penetration rate beyond 50 m, reflecting a change in driving style, requires an increase in hydraulic pressure providing the torque at the cutter head beyond a standard target of approximately 10–12 MPa (Fig. 10a).
5.5. Simple design guidelines While Figs. 9a and 9b provide a very convenient framework for summarising the new data presented in this paper, both k and smax must be estimated if these figures are to be used as design charts to estimate volume loss. Simplified expressions have been developed
hereunder to allow an estimate of volume loss based solely on the tunnel depth and soil type (assumed k value) only. On the basis of Fig. 8a, conservative preliminary estimates of smax for Irish tills, using current tunnelling methods and in situations where boulders do not adversely affect the tunnelling process, are proposed in the following equations:
smax ¼ 0:005D in fine soils ði:e: boulder claysÞ
ð6Þ
smax ¼ 0:0125D in coarse soils ði:e: gravelsÞ
ð7Þ
If boulders are encountered (as evidenced by measurements of jacking forces and penetration rates during driving), larger values of smax may be appropriate, but are difficult to predict. Assuming, from Fig. 7, that k = 0.275 and k = 0.5 are representative trough width parameters for coarse and fine soils respectively, then conservative estimates of Vl when using current tunnelling methods may be made by substituting the relevant k value and either Eqs. (6) and (7) as appropriate into Eq. (3) to arrive at:
12
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
z 0 V l ¼ 0:008 in fine soils D
While Eqs. (8) and (9), like any other empirical expressions, should be used with caution, they can provide a useful first estimate of likely volume loss due to tunnelling in Irish ground conditions using current tunnelling methods. Once again, these values may be expected to increase if boulders are encountered.
information presented in this paper relating to the Dublin Port Tunnel and Mullingar Sewerage Improvement Scheme, respectively. Supplementary information provided by Mr. Conor O’Donnell of AGL Consulting is appreciated. The authors would also like to acknowledge J.B. Barry and Partners, Consulting Engineers, who employed Ciarán Reilly during the construction of the Mullingar Scheme, and Ward and Burke Construction, the microtunnelling contractors for the Mullingar Sewerage Improvement Scheme, who employed Brien Curran while carrying out a research Masters at NUI Galway.
6. Conclusions
References
This paper has aimed to collate Irish experience of settlements above tunnels in glacial tills, with a view to providing preliminary pre-contract estimates of the amount and extent of surface settlement that is likely to arise in a tunnelling contract. The following conclusions may be drawn from this research:
Curran, B.G., 2010. Analysis of Jacking Loads and Ground Movements for Microtunnelling in Irish Ground Conditions. MEngSc. Thesis, Dept. of Civil Engineering, NUI Galway. Devriendt, M., 2010. Risk analysis for tunnelling ground movement assessments. In: Proceedings of ICE Geotechnical Engineering, vol. 163, no. GE3, pp. 109–118. Farrell, E.R., 2010. Lessons learned – problems to solve. In: Proceedings of the Symposium on Bridge and Infrastructure Research in Ireland, pp. 181–198 (keynote). Farrell, E., Lehane, B., O’Brien, S. and Orr, T. 1995. Stiffness of Dublin black boulder clay. Proc. XIth.ECSMFE, Copenhagen, 1, 1-109 – 1-117. Farrell, E., Wall, D., 1990. Soils of Dublin. Transactions of the Institution of Engineers of Ireland, pp. 78–97. Finch, T.F., 1977. Soils of Co. Westmeath, Soil Survey Bulletin No. 33, An Foras Talúntais. Glossop, N.H., Saville, D.R., Moore, J.S., Benson, A.P., Farmer, I.W., 1979. Geotechnical aspects of shallow tunnel construction in Belfast estuarine deposits. In: Proc. Tunnelling ’79, London, pp. 45–56. Goto, J., Thompson, S., Squires, S., Mizuno, S., 2004. Dublin Port tunnel – excavation of a 11.8 m diameter urban motorway tunnel. In: Chinese Taipei Tunnelling Association Conference, pp. D091–D098. Jones, B.D., 2010. Low-volume-loss tunnelling for London ring main extension. In: Proceedings of ICE Geotechnical Engineering, vol. 163, no. GE3, pp. 167–185. Lehane, B.M., Simpson, B., 2000. Modelling glacial till under triaxial conditions using a BRICK soil model. Canadian Geotechnical Journal 37 (5), 1078–1088. Lochaden, A.L.E., Farrell, E.R., Orr, T.L.L., 2008. Comparing numerically modelled tunnel induced ground movements with field measurements. In: Proceedings of BCRI ’08 Symposium, NUI Galway. Long, M., Menkiti, C.O., 2007. Geotechnical properties of Dublin Boulder Clay. Geotechnique 57 (7), 595–611. Macklin, S.R., 1999. The prediction of volume loss due to tunnelling in overconsolidated clay based on heading geometry and stability number. Ground Engineering 32 (4), 30–33. Mair, R.J., 1996. Settlement effects of bored tunnels. In: Mair, R.J., Taylor, R. N., (Eds.), Session Report, Proc. Int. Symposium on Geotechnical Aspects of Underground Construction in Soft Ground. London, Balkema, pp. 43-53. Mair, R.J., Taylor, R.N., 1997. Bored tunnelling in the urban environment, state-ofthe-art report and theme lecture. In: Proceedings of the 14th International Conference on Soil Mechanics and Foundation Engineering, vol. 4. Hamburg, Balkema, pp. 2353–2385. O’Reilly, M.P., New, B.M., 1982. Settlements above tunnels in the United Kingdom: their magnitude and prediction. In: Proceedings of Tunnelling ’82 Symposium, London, pp. 173–181. Orr, T.L.L., Farrell, E.R., 1996. Geotechnical aspects of tunnelling in soft ground in Ireland. In: Mair, Taylor (Ed.), Geotechnical Aspects of Underground Construction in Soft Ground, pp. 301–305. Peck, R.B., 1969. Deep excavations and tunnelling in soft ground. In: Proceedings of the 7th International Conference on Soil Mechanics and Foundation Engineering, vol. 3. Mexico City, pp. 311–376. Skipper, J., Follett, B., Menkiti, C., Long, M., Clarke-Hughes, J., 2005. The engineering geology and characterisation of Dublin Boulder Clay. In: Quarterly Journal of Engineering Geology and Hydrogeology, pp. 171–187.
V l ¼ 0:011
z 0 in coarse soils D
ð8Þ ð9Þ
1. For glacial till soils, it is important to establish whether the predominant overburden material governing the settlement trough has the characteristics of a fine or coarse material since the k values for each have been shown to be quite different. The correct choice is important to obtain a reliable prediction of volume loss. 2. The values of the trough width parameter, k, for coarse soils lie in the range 0.2–0.3, which is at the low end of the range given in the Mair and Taylor (1997) database for sands and gravels. 3. The values of k for fine soils lie in the range 0.4–0.6, which is consistent with the Mair and Taylor (1997) database for clays. There is no obvious distinction between the k values obtained for tunnels in soft and stiff clays. 4. An increase in smax with time in the Mullingar coarse soils is believed to be related to dissipation of the bentonite fluid used to support the overcut to the pipelines. There is no systematic variation in k with time in these soils. None of the boulder clays referred to were soft with the result that consolidation settlement was not a significant issue. 5. The data presented has enabled reasonably confident upper limits to be placed on the smax values likely to be obtained for tunnels in fine and coarse soils using current tunnelling methods, which is useful in obtaining a preliminary estimate of the volume loss. However, the presence of boulders, as well as reducing the advance rate, can result in high values of smax and high volume losses which are difficult to predict with accuracy.
Acknowledgements The authors would like to thank Nishimatsu Construction Co. Ltd. and Westmeath County Council for the permission to use
Contents lists available at ScienceDirect
Tunnelling and Underground Space Technology journal homepage: www.elsevier.com/locate/tust
Settlement trough parameters for tunnels in Irish glacial tills B.A. McCabe a,⇑, T.L.L. Orr b, C.C. Reilly b, B.G. Curran a a b
Department of Civil Engineering, National University of Ireland, Galway, Ireland Department of Civil, Structural and Environmental Engineering, Trinity College Dublin, Ireland
a r t i c l e
i n f o
Article history: Received 1 December 2010 Received in revised form 30 May 2011 Accepted 9 June 2011 Available online 16 July 2011 Keywords: Transverse settlement profiles Glacial till Trough width Tunnelling Volume loss Boulders
a b s t r a c t Glacial till (or boulder clay) is the most widespread sediment on the island of Ireland. The behaviour of these tills, especially Dublin Boulder Clay, is now better understood as a result of ground investigations and associated testing for major construction projects, particularly in the east of the country over the past decade. Despite an increase in tunnelling activity in the country over the same time period, there is very little documented evidence on the settlements induced by tunnelling operations in glacial till. In this paper, transverse surface settlement data from two glacial till sites are presented. Four profiles are presented for the ‘soft ground’ TBM-bored section of the Dublin Port Tunnel. Nine profiles are reported for pipe-jacked microtunnels constructed for a sewerage scheme in Mullingar in the Midlands; one of which was measured at the top of a railway embankment under which the pipeline passes. The measured settlements have been interpreted using a standard Gaussian error function, and trough width parameters show dependence on the fraction of the till, fine or coarse, that governs its behaviour. In addition, conservative design estimates of maximum trough settlement and volume loss have been provided, and the impact of boulders is discussed. This paper provides the first empirical guidance for predicting surface settlements above tunnels in Irish soils. Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction Irish tunnelling experience has developed considerably since the Orr and Farrell (1996) report on tunnelling activity in the country 15 years ago, at which time tunnels had yet to be used to carry roads or motorways. The Jack Lynch Tunnel under Cork’s River Lee was opened in May 1999; this submerged twin-bore tunnel is 610 m long and is estimated to have cost €133 m. The €752 m Dublin Port Tunnel (DPT), when opened in December 2006, became the longest urban tunnel in Europe (4.7 km twin bore) and the nation’s largest ever civil engineering project. An important coupled outcome of the DPT project was the improved grasp of both the stratigraphy and engineering properties of Dublin Boulder Clay that arose from handling the extensive geotechnical data generated (e.g. Long and Menkiti, 2007). Another highly significant project, the €660 m and 675 m long twin-bore submerged Limerick tunnel under the River Shannon, was opened in September 2010. In parallel with the development of large scale tunnels, the application of pipe-jacking to the construction of pipelines for utilities such as water, sewage and gas in a variety of soil and rock types has escalated. Curran (2010) interprets jacking force data for slurry-shield pipe-jacked microtunnels at eleven Irish sites encompassing a variety of soil and rock types. ⇑ Corresponding author. Tel.: +353 91 492021; fax: +353 91 494507. E-mail address: [email protected] (B.A. McCabe). 0886-7798/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tust.2011.06.002
Lochaden et al. (2008) report on successful efforts to model settlements above the DPT using finite element software, with the predictions relatively insensitive to the constitutive soil model used. In practice, however, predictions of surface settlement troughs over tunnels in Irish soils have to date relied heavily on experience from the UK and beyond, such as the recommended Gaussian trough width parameters published by Mair and Taylor (1997) for tunnels in sands/gravels and clays with generally a small range of particle sizes, and the achievable volume losses suggested by Mair (1996) and others. Some Irish soils have particular characteristics that significantly affect their behaviour; for example, Irish boulder clays are often well-graded and composed of a wide range of particle sizes from clay-size to gravel-size, so that some of these soils are characterised as fine soils while others are characterised as coarse soils (Farrell, 2010). For this reason, it was deemed important to analyse tunnel-induced settlements in these soils, and also to consider the impact of boulders that are sometimes encountered. In this paper, settlement data is provided for (i) large diameter open-face shielded tunnels in glacial till for the DPT project, and (ii) small diameter closed-face pipe-jacked tunnels in a glacial till at Mullingar, Co. Westmeath; including data for a section of pipeline passing beneath a high railway embankment. The inferred Gaussian parameters are an important record of Irish tunnelling-induced settlement experience and also, in the case of the Dublin Port Tunnel, supplement the limited data published internationally for tunnels with low cover to diameter ratios. All of the aforemen-
2
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Nomenclature AVN cu D DBC DPT i k MH N NTc
remote control tunnelling machine with slurry material removal undrained shear strength tunnel diameter (including overcut) Dublin Boulder Clay Dublin Port Tunnel value of y corresponding to point of inflection of Gaussian trough Gaussian trough width parameter manhole stability number stability number at collapse
tioned site locations are illustrated in Fig. 1; in addition to a Belfast site with older data which is referred to later in the paper. 2. Tunnelling projects at Dublin and Mullingar 2.1. Dublin Port Tunnel The Dublin Port Tunnel is a 4.7 km long urban road transport twin-tunnel constructed to ease congestion in Dublin city centre by removing heavy vehicular traffic to and from Dublin Port and rerouting it from the centre of Ireland’s capital to the outer orbital M50 motorway. Construction of the DPT began in 2001 and the tunnel was opened in 2006. Construction methods used included cut and cover, pipe jacking, hard rock TBM tunnelling and ‘soft ground’ TBM tunnelling. The section of the project considered in this paper is the 325 m section of twin tunnel bored in boulder clay at Whitehall, towards the northern end of the scheme (Fig. 2). An 11.77 m diameter Herrenknecht open-face shielded back-actor machine was used for this section. Summary information on the surface settlement profiles is provided in Table 1.
Belfast Mullingar
Dublin Limerick
Cork
Fig. 1. Map showing locations of Irish tunnelling projects referred to in this paper.
N60 sv smax
rt rv TBM Vl Vs Vt y z0
standard penetration test value vertical settlement at any point in trough vertical settlement over centreline of tunnel (at y = 0) tunnel internal support pressure vertical total stress at the depth of the tunnel axis Tunnel Boring Machine volume loss volume of Gaussian settlement trough per metre length volume of tunnel per metre length transverse horizontal distance from tunnel centreline depth to tunnel axis
The data published here, for four different chainages, relate to the southbound bore which was the first to be constructed. Interpretation of the settlement data measured during construction of the northbound bore will be the subject of a future publication. Goto et al. (2004) make general observations from settlement profiles in the Whitehall area as part of an overview paper on the DPT scheme. Additional data and a fuller interpretation is provided in this paper. 2.2. Mullingar Approximately 750 m length of pipe-jacked tunnels were constructed in Mullingar between July and October 2009 as part of the Mullingar Sewerage Improvement Scheme. Herrenknecht AVN (remote-controlled slurry-shield tunnelling) machines deploying a ‘mixed’ head (incorporating both scraping teeth and cutting discs) of the type shown in Fig. 3 were used. Two different sized Tunnel Boring Machines (TBMs) were used, AVN 1200 and AVN 1800, with external diameters of 1505 mm and 2150 mm respectively. The relevant particulars of each are shown in Table 2a. The forces and penetration rates recorded during pipe-jacking in this contract have been interpreted elsewhere (Curran, 2010). The settlement data reported here pertain to distinct areas of the scheme, hereafter referred to as Areas A–C (see Table 2b and Fig. 4): Area A. Area A includes the Supervalu car park and access road to it from Austin Friars Street in the town centre. In this area, transverse profiles 1–7 and 13–19 were situated between manholes MH22 and MH23, and transverse profiles 26–32, 36–43 and 47–53 were situated between manholes MH23 and MH24. The pavement in the area was typically 200 mm thick and the tunnel axis is at a depth of between 3.6 m and 5.4 m. Area B. Area B is immediately west of the River Brosna, near the Mullingar Greyhound Stadium. In this area, transverse profiles 117–123, 127–133 and 159–165 all lie to the southern end of the section of pipeline between manholes MH9 and MH11 and the tunnel axis is at a depth of between 3.5 m and 4.1 m. Profiles 127–133 and 159–165 were measured on a concrete slab. Area C. Area C is immediately north of the Mullingar Greyhound Stadium. In this area, the pipeline between manholes MH11 and MH13 passes approximately 4 m below an 8 m high embankment supporting the Dublin–Sligo railway line. Settlements were recorded on points marked on both rails of each of three railway tracks. Note that the direction of the pipeline is approximately 13° off perpendicular to the embankment centreline at that point.
3
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Fig. 2. Map of Dublin Port tunnel showing soft ground TBM section at Whitehall.
Table 1 Settlement profiles at DPT. Drive ID
Settlement profile chainage (m)
Location
Depth to tunnel axis, z0 (m)
Tunnel diameter, D (m)
z0/d
Average penetration rates (m/day)
Predominant overburden material
194 194 194 194
2045 2060 2080 2120
Carpark Carpark Carpark Carpark
15.12 15.60 16.12 17.41
11.77 11.77 11.77 11.77
1.28 1.32 1.37 1.48
4 3.5 2 2
Boulder Boulder Boulder Boulder
SB SB SB SB
Precise electronic levelling, with an accuracy of 0.1 mm, was used to record tunnelling-induced settlements at both the DPT and Mullingar sites. Since the DPT site was mainly a rough carpark, the levels were not reinstated immediately after tunnelling. Instead the whole area was reinstated as part of the ancillary road construction works associated with the DPT. Cement grout was injected into the overcut to prevent further settlement at Areas A–C in Mullingar. 3. Ground conditions 3.1. Dublin port tunnel Ground conditions encountered at Whitehall were Dublin Boulder Clay (DBC), a stiff lodgement till. The engineering properties of
clay clay clay clay
DBC are covered extensively elsewhere (Farrell and Wall, 1990; Skipper et al., 2005; Long and Menkiti, 2007) while Orr and Farrell (1996) give a comprehensive description of its properties which have greatest relevance to tunnelling. DBC is a well-graded, stiff glacial outwash till containing about 35% fines (particles smaller than 0.06 mm), large stone boulders and lenses of sand or gravel. A representative DBC grading curve (once cobbles and boulders have been removed) approximates to a straight line on a semilog particle size distribution plot as shown by the mean particle size distribution plots in Fig. 5 that were obtained for four units within the Dublin Boulder Clay. A fines content in excess of 35% results in the soil being classified as fine and is sufficient to impart the behavioural characteristics of a fine grained soil, but the fines content in a glacial till can vary from site to site and with depth,
4
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
3.2. Mullingar Mullingar is a regional town with a population of under 20,000. Finch (1977) classified Co. Westmeath’s ground conditions, primarily for agricultural purposes, and referred to the sediment around Mullingar as the Rathowen Phase, a till derived predominantly from limestone and shale. Besides this work, there is little published on the geotechnical characteristics of this till. The till material present in the top 3–5 m, which governed the settlement over the jacked pipes, was found to comprise clay and gravel, but other constituents such as made ground and peat were also discovered, and the fines content varied from location to location. In general, it was found that standard penetration test N60 values were highly variable in the Mullingar till which is also a feature of Dublin Boulder Clay, as noted by Lehane and Simpson (2000). In subsequent interpretations of the Mullingar data, where the overburden to the pipes consists of more than one soil type, it has generally been the case that one soil type has a dominant influence (usually owing to greater thickness) than the other(s), unless stated otherwise. In general, the site investigation points for this sewerage scheme were chosen to coincide with the proposed manhole positions, but the jacking force and penetration rate records for the drives (Curran, 2010) were helpful in some cases for establishing the ground conditions between the manholes. A summary of the ground conditions in Areas A–C are given in the following paragraphs and these should be read in conjunction with the tunnel axis depths (z0) and diameters (D) shown in Table 2b:
Fig. 3. Mixed TBM head used at Mullingar.
Table 2a Mullingar site drive parameters. Machine
AVN 1200
AVN 1800
Pipe int. dia. (mm) Pipe ext. dia. (mm) TBM ext. dia. (mm) Torque (kNm) Revolutions per min Force per cylinder (kN) Max drive length (m) Shield length (mm) Shield weight (kN) Pipe weight (kN)
1200 1490 1505 195 0–6 753 200 3200 103 31.9
1800 2120 2150 445 0–3 1272 300 4200 245 58.0
as is evident in Fig. 5. The significant granular content causes difficulties during sampling operations and makes predictions of its overall behaviour difficult to interpret from laboratory tests on small samples. The high stiffness of DBC has meant that settlements over tunnels have not been much of an issue to date and in some cases are close to the limits of measurability. For the particular section of the ‘soft ground’ TBM drive considered in this paper, the ground conditions encountered during tunnelling showed some variation across the 11.77 m face. Tunnelling records describe the glacial till encountered as stiff brown sandy clay with fine to coarse subangular to rounded gravels occurring towards the crown, tending towards dense sandy gravels with some cobbles at the invert.
Area A. The MH23–MH22 and MH23–MH24 jacking records (Curran, 2010) indicate that the tunnel face was always in clayey and/or sandy gravel. Site investigation data shows that the natural overburden is predominantly gravel in the vicinity of profiles 1–7, 26–33 and 47–53 with grading curves indicating the following approximate proportions of the different soil types: gravel 65–70%, sand 15–20%, fines 10–20%. Near profile 13–19, the overburden appears to have a lower gravel content than at other locations, with 1.2 m thickness described as very silty, very sandy gravel (gravel 47%, sand 27%, fines 26%) and 0.3 m thickness described as clayey peat. A trial pit in the vicinity of profile 36–43 indicates an overburden comprising 1.6 m of slightly sandy slightly gravelly clay (with a representative grading indicating gravel 38%, sand 23% and fines 39%). Area B. For profiles 117–123 and 127–133, site investigation evidence suggests that the overburden is predominantly gravel (gravel 67%, sand 24%, fines 9%), although the soil in the tunnel face may actually straddle the division between fine and coarse soils. However, at profile 159–165, there is a significant thickness (1.5 m) of clayey peat and organic silt in the overburden material. There is also approximately 1.0 m of made ground at all three profile locations.
Table 2b Settlement profiles at Mullingar.
*
Area
Drive ID
Settlement profile ID
Location
Depth to tunnel axis, z0 (m)
Tunnel diameter, D (m)
z0/d
Average penetration rates for drive (mm/min)
Best estimate of predominant overburden material
A A A A A B B B C
MH23–MH22 MH23–MH22 MH23–MH24 MH23–MH24 MH23–MH24 MH9–MH11 MH9–MH11 MH9–MH11 MH11–MH13
1–7 13–19 26–32 36–43 47–53 117–123 127–133 159–165 –
Road Road Carpark Carpark Carpark Road Yard Yard Rail lines
5.39 4.28 5.05 3.60 5.05 4.09 3.95 3.5 12.05
1.505 1.505 1.505 1.505 1.505 2.150 2.150 2.150 2.150
3.58 2.84 3.36 2.39 3.36 1.90 1.84 1.63 5.60
12.8 12.8 32.3* 32.3* 32.3* 28.6 28.6 33.7 33.7
Gravel Gravel + peat Gravel Clay Gravel Gravel Gravel Clay/silt + peat Embankment clay + gravel
Penetration rate records only available for first 20 m of this drive.
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
5
Fig. 4. Location of settlement pins at MH23–MH24 and MH23–MH22 (Area A); MH9–MH11 (Area B) and MH11–MH13 (Area C, railway embankment).
Area C. From boreholes either side of the embankment and one carried out an angle beneath the embankment, the tunnelled face is interpreted to have passed through a combination of silty sandy gravel and gravelly silt, with gravel the prevalent overburden to the pre-embankment ground level. The embankment itself, constructed in 1847/48 from material imported to the site from cuttings elsewhere on the Dublin–Sligo railway around Mullingar, comprises sandy gravelly clay with gravel and
cobbles; i.e. a fine material. This embankment material was expected to have had the most influence on settlement of the rails above. 4. Gaussian settlement profiles Several authors (i.e. Peck, 1969; O’Reilly and New, 1982) have shown that the immediate surface settlement profile or trough
6
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
100
Percentage < d
80
60
40
Brown BC Upper Black BC Sandy Brown BC Lower Black BC
20 (Fine soil)
0 0.001
(Coarse soil)
0.01
0.1
1
10
100
Particle size d (mm) Fig. 5. Particle size distribution for Dublin Boulder clay.
Horizontal distance from tunnel centreline, y (m) -6
-4
-2
0 -5
2
4
6
i =kz0 =1.67
Settlement, s v (mm)
0
5
10 point of inflection 15
20 hatched volume (per metre run) is volume loss Vl
Settlement profile 13-19 in Area A, after 22 days
25
Eqn [1], with smax=25.1mm, z0=4.28m, k=0.39
smax 30
Fig. 6. Gaussian settlement trough with typical measured data.
above tunnels on a greenfield site can be represented adequately by a simple Gaussian or error function of the form:
sv 2 ¼ expðy2 =2i Þ smax
ð1Þ
where sv is the vertical settlement at a horizontal distance at y from the tunnel centre line, smax is the settlement at the centreline (y = 0) and i is the value of y corresponding to the point of inflection of the function in Eq. (1). The function is plotted in Fig. 6 and annotated with the relevant Gaussian parameters. O’Reilly and New (1982) also proposed a linear relationship between i and z0, the depth to the tunnel axis.
i ¼ kzo
ð2Þ
Recent research (Jones 2010) has shown that the relationship between i and z0 is not linear for z0 values exceeding 35 m, which is greater than the tunnels depths considered in this paper. The constant of proportionality, k, is known as the trough width parameter and is believed to be largely independent of the construction method (Mair and Taylor, 1997).
Mair and Taylor (1997) summarised a wide range of field data and concluded that 0.4 < k < 0.6 for clays with k showing no dependence on the clay’s stiffness. Corresponding data presented for tunnels in sands and gravels exhibit greater scatter, with the majority of the data bounded by 0.25 < k < 0.45. The volume of the settlement trough Vs per unit length can be determined by integrating Eq. (1). Volume loss (Vl), often used as a measure of quality control of the tunnelling process, is defined as the ratio of Vs to the volume of the tunnel per metre length, Vt:
Vl ¼
Vs ¼ Vt
pffiffiffiffiffiffiffi z s 2pismax max 2 3:192k 0 pD D D
ð3Þ
4
Based on a review of settlement data for bored tunnels, Mair (1996) concluded that: (i) For open-face tunnelling in stiff clays, such as London Clay, Vl values of between 1–2% are achievable. (ii) For closed-face tunnelling, using earth pressure balance or slurry shields, Vl values of 0.5% in sands and 1–2% in soft clays (excluding consolidation settlements) are achievable.
7
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12 Table 3a Settlement parameters at DPT. Settlement profile chainage (m)
z0/D
Time since TBM passed (days)
i
i/D
k
smax (mm)
Vl (%, from Eq. (3))
Predominant overburden material
2045 2060 2080 2120
1.28 1.32 1.37 1.48
183 188 200 240
7.65 7.77 7.53 9.53
0.65 0.66 0.64 0.81
0.51 0.50 0.47 0.55
11.8 19.6 34.3 75.2
0.21 0.35 0.64 1.66
Boulder Boulder Boulder Boulder
clay clay clay clay
Table 3b Settlement parameters at Mullingar. Settlement profile ID
z0/D
Time since TBM passed (days)
i
i/D
k
smax (mm)
Vl (%, from Eq. (3))
Best estimate of predominant overburden material
1–7 13–19
3.58 2.84 2.84 2.84 2.84 2.84 2.84 3.36 3.36 2.39 3.36 1.90 1.90 1.90 1.84 1.84 1.84 1.63 5.60
14 4 14 22 53 63 81 19 33 29 35 4 6 8 2 6 16 14 1–3
1.51 1.71 1.80 1.67 1.88 1.80 1.84 1.52 1.57 1.51 1.41 1.1 1.02 0.9 0.91 0.95 0.87 1.47 5.6–7.2
1.00 1.14 1.19 1.11 1.25 1.19 1.22 1.01 1.04 1.00 0.94 0.51 0.48 0.42 0.42 0.44 0.40 0.68 2.6–3.4
0.28 0.40 0.42 0.39 0.44 0.42 0.43 0.30 0.31 0.42 0.28 0.27 0.25 0.22 0.23 0.24 0.22 0.42 0.5–0.6
18.4 13.8 21.7 25.1 31.2 30.2 31.7 10.8 12.5 7.1 3.8 19.9 23.9 28.7 21.3 23.0 26.2 8.6 5.0–7.0
3.91 3.32 5.5 5.9 8.3 7.65 8.22 2.31 2.76 1.51 0.76 1.52 1.69 1.79 1.34 1.57 1.57 0.89 2.3–3.5
Gravel Gravel + peat
26–32 36–43 47–53 117–123
127–133
159–165 Railway
In overconsolidated clays, Vl is often estimated using the method of Macklin (1999):
N V l ¼ 0:23 exp 4:4 NTc
ð4Þ
In Eq. (4), Vl depends on the ratio of present stability number N to the stability number at collapse NTc. The stability number N is defined as:
N¼
rv rt cu
ð5Þ
where rv is the vertical total stress at tunnel axis level, rt is the internal support pressure and cu is the undrained shear strength of the clay. Macklin (1999) and Devriendt (2010) provide empirical guidance on the choice of NTc drawn from various experimental studies which, when combined, span a wide range of possible cover to diameter and unsupported advance length to diameter ratios.
5. Interpretation of settlement data 5.1. Interpretation methods Only settlement profiles showing a high level of symmetry about the vertical axis and with minimal measurement errors at the extremities (where settlements were small) are included in Table 3a (DPT) and Table 3b (Mullingar) and the discussion which follows. A typical measured settlement profile, 13–19 in Area A, is shown in Fig. 6. The following steps were taken in the interpretation:
Gravel Clay Gravel Gravel
Gravel
Clay/silt + peat Emb. clay + gravel
(i) The maximum value of settlement recorded (i.e. at y = 0) was set as smax in Eq. (1), whereupon the value of k was varied to provide the best match between the measured data and the profile predicted by Eq. (1). In general, the subjectivity of the match for k is believed to be ±0.02. (ii) The settlement profile data were plotted on Fig. 7 in Peck’s (1969) normalised format, as i/D against z0/D. The plot is overlain with lines representing constant values of k. (iii) The Vl values quoted in Table 3a and 3b were computed using Eq. (3) adopting the best fit values of k. 5.2. Discussion of k values In Fig. 7, the plotted data are classified according to the soil type in the overburden. The following observations can be drawn: (i) The settlement profiles for the tunnels in the Mullingar glacial gravels tend to be represented by k values that lie consistently in the range 0.2 < k < 0.3, which is at the lower end of the range indicated by Mair and Taylor (1997) for tunnels in sands and gravels. The exception is the k = 0.42 value for profile 13–19, which is probably due to the higher fines content in the gravel and/or the presence of peat in the overburden at that location. (ii) The settlement profiles for the tunnels in boulder clays tend to be modelled by k values in the range 0.4 < k < 0.6, which is also consistent with Mair and Taylor (1997). The DPT data (enclosed by a dashed loop to identify them) show marginally higher k values than the Mullingar data (excluding that at z0/D = 5.6), and are largely consistent with k values inferred from hand-excavated shield tunnelling in Belfast
8
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Normalised horizontal distance to point of inflection (i/D) 0
0.5
1
1.5
2
2.5
3
3.5
4
0
DPT data
All data is from Mullingar unless specified
1
fine soils
Belfast data (Glossop et al. 1979)
coarse soils
Normalised depth z0/D
2
3
range for Area C: railway embankment
some peat in overburden 4
5
k = 0.7 6
k = 0.6 k = 0.2
k = 0.3
k = 0.4
k = 0.5
7 Fig. 7. Plot of i/D against z0/D for Mullingar, DPT and Belfast sites.
Normalised maximum settlement smax /D 0.000 0
0.005
0.010
0.015
0.020
0.025 fine soils coarse soils
1
183
200
Normalised depth z0 /D
240 14
188
2
2
6
4
16 6
boulders encountered 8 63 & 81
29
3
19
33 4
14
22
53
35 14
4
5 1-3
6
points annotated with time (days) since TBM passed; see Table 3, arrows track variation over time
Fig. 8a. Normalised maximum surface settlement versus normalised tunnel depth for Mullingar and DPT sites.
soft estuarine clay (sleech) published by Glossop et al. (1979). In general, there is no clear relationship between the strength of these clays and the k value which models its settlement trough. (iii) At Area C, the deviation of the tunnel alignment from perpendicular to the embankment longitudinal axis was approximately allowed for by correcting the horizontal distance measured along the rails by cos(13°). However, due to the small smax values recorded, it was not possible to assess the k value relevant to the settlement profiles recorded at the top of the railway embankment (z0/D = 5.6) to the same accuracy as the k values assessed from the other data. Interestingly, the inferred range 0.5 < k < 0.6 confirms
that the fine material in the embankment is more influential in determining the shape of the settlement trough than the underlying gravel overburden. (iv) Some of the Mullingar data in Fig. 7 were recorded at a single point in time, whereas in other cases ‘range’ bars are shown for profile data where reliable profiles were recorded over a period of time (see Table 3a and 3b for full detail). In general, the ranges of k values shown were found to be small, within the range of natural variation and are not believed to be due to systematic time effects. (v) In clay soils incorporating significant consolidation settlements, there is evidence that Eq. (1) does not model the shape of a settlement trough well (e.g. Glossop et al.,
9
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
5.3. Discussion of maximum settlements
Normalised maximum settlement smax/D 0
0.01
0.02
0.03
0.04
0.05
Values of smax, which are required along with k in Eq. (3) to predict Vl show no obvious dependency on depth z0 when both are plotted in normalised form with respect to the tunnel diameter in Fig. 8a. Interestingly, the smax/D ranges for fine and coarse soil types are clearly distinct for these Irish data. In Fig. 8b, the authors have recast the smax data for settlement profiles of tunnels in clays and sands/gravels tabulated by Mair and Taylor (1997) into the same format as Fig. 8a, which shows that such a clear separation is not typical. Limiting values of smax/D are not theoretically justifiable based on Eq. (3) but are affected by the tunnelling method, the soil type and the geometry of the design situation, being linearly dependent on Vl and inversely dependent on k and z0/D. Higher smax/D values (between 0.007 and 0.022) are associated with the soils containing boulders for the Mullingar tunnelling, highlighted in Fig. 8a. The times elapsing between the passing of the TBM and each set of settlement profile measurements are included on Fig. 8a. The time dependence of smax in the coarse soils is somewhat unexpected. The authors believe that the most likely explanation relates to the bentonite lubricant used to support the excavation and prevent the ground from completely collapsing onto the pipeline. During drives in gravel, and especially where boulders were encountered, a high concentration of bentonite was mixed with water, forming a bentonite fluid with Marsh Funnel times in excess of 100 s. This highly viscous fluid would take a long time to dissipate and might explain the continued settlement over days and weeks. The data in Fig. 8a is reproduced in Fig. 9a (fine) and Fig. 9b (coarse) in an alternative format with the normalised surface settlement shape function k(smax/D) replacing smax/D on the horizontal axis, allowing contours of volume loss (based on Eq. (3)) to be superimposed on these figures. The relevant values of volume loss pertinent to each datapoint can be estimated visually from the Vl contours in Figs. 9a and 9b. These datapoints show distinct ranges for the volume losses when tunnelling in fine and coarse soils with Vl generally less 1% in fine soils and generally greater than 1% in coarse soils. The data of Mair and Taylor (1997) is also provided in Fig. 9c for comparison.
0
1
2
Normalised depth z0/D
3
4
5
6
7
8 fine soils
9
coarse soils
10 Fig. 8b. Normalised maximum surface settlement versus normalised tunnel depth for Mair and Taylor (1997) data.
1979; Mair and Taylor, 1997). While there is insufficient data to draw any inferences on consolidation settlements in the fine soils at Mullingar, these are believed to be small and the Gaussian functions capture the settlement troughs very well. Likewise, consolidation settlements in stiff DBC are not considered significant to this analysis.
Normalised surface settlement shape function k(s max /D) 0.0001 0
0.001
Normalised depth z0 /D
1
183
188
0.01
200
0.1
240
14
2 29
3
0.1% 4
V l values 5 1-3
6
0.5%
1%
2%
points annotated with time (days) since TBM passed; see Table 3
5%
10%
Fig. 9a. Normalised surface settlement shape function versus normalised tunnel depth for fine soils at Mullingar and DPT sites.
10
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
Normalised surface settlement shape function k(smax /D) 0.0001 0
0.0 01
0. 01
0. 1
1
Normalised depth z 0 /D
6 & 16 2
2
6&8
4
53, 63 & 81 4
3
14 & 22
19 33
0.1%
boulders encountered
35 14
4
V l values 5 points annotated with time (days) since TBM passed; data linked by brackets pertain to same settlement profile 6
0.5%
1%
2%
5%
10%
Fig. 9b. Normalised surface settlement shape function versus normalised tunnel depth for coarse soils at Mullingar and DPT sites.
Normalised surface settlement shape function k(smax/D) 0.0001 0
0.0 01
0.0 1
0.1
1
Normalised depth z 0/D
2 3
0.1%
4 5 6
V l values
7 8 fine soils coarse soils
9 10
0.5%
1%
2%
5%
10%
20%
50%
Fig. 9c. Normalised surface settlement shape function versus normalised tunnel depth for Mair and Taylor (1997) data.
5.4. Discussion of volume losses Tunnelling operations in DBC were successful in keeping volume losses comfortably within or below the range of 1–2% for stiff clays suggested by Mair (1996). Eq. (4) yields Vl 0.45–0.5% based upon N 0.8 (assuming a constant cu = 450 kPa for DBC quoted by Farrell et al. (1995) and NTc 4.8 estimated from Devriendt (2010). This predicted range falls in the middle of a wider range of Vl values in Table 3a; three of the four profiles have Vl values in the range 0.21–0.64%, while at chainage 2120, where the highest value of smax = 75 mm was recorded, a Vl value of 1.66% was noted. Volume losses at Mullingar were more variable and reflect the variety of ground conditions encountered there. In general, values
of Vl were computed to lie between 0.7% and 1.8% approximately, except: Beneath the railway embankment at Area C, where volume losses in the range 2.3–3.5% apply. These can be explained by the higher z0/D (=5.6) value for the pipeline at this location, compared to nearby profile 159–165 in Area B for instance (Vl = 0.89%, z0/D = 1.63). These relative volume losses are consistent with the proportional relationship between Vl and z0/D implied in Eq. (3). In sections where boulders were encountered (profiles 1–7, 13–19, 26–32 in Area A) Vl values range from 2.3–8.2%. It was possible to deduce when boulders were encountered from the
11
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
30
3000 Face Resistance
25
2000
20
1500
15
1000
10
500
5
0
Pressure generating Cutter Head Torque (MPa)
Face Resistance (kN)
Max Torque
2500
0 0
10
20
30
40
50
60
70
Jacked Distance (m) Fig. 10a. Interpreted face resistance and cutter head torque pressure for the drive from MH23 to MH22.
60
Penetration Rate (mm/min)
50
40
30 moving average (every 5 datapoints averaged)
20
10
0 0
10
20
30
40
50
60
70
Jacked Distance (m) Fig. 10b. Moving average of penetration rate for the drive from MH23 to MH22.
interpreted face resistance, cutter head torque and penetration rate records, such as those for drive MH23–MH22 adapted from Curran (2010). Substantial increases in face resistance are seen to occur for jacked distances in excess of 42 m (Fig. 10a) due to the presence of boulders, with a corresponding drop in penetration rate (Fig. 10b). The higher penetration rate beyond 50 m, reflecting a change in driving style, requires an increase in hydraulic pressure providing the torque at the cutter head beyond a standard target of approximately 10–12 MPa (Fig. 10a).
5.5. Simple design guidelines While Figs. 9a and 9b provide a very convenient framework for summarising the new data presented in this paper, both k and smax must be estimated if these figures are to be used as design charts to estimate volume loss. Simplified expressions have been developed
hereunder to allow an estimate of volume loss based solely on the tunnel depth and soil type (assumed k value) only. On the basis of Fig. 8a, conservative preliminary estimates of smax for Irish tills, using current tunnelling methods and in situations where boulders do not adversely affect the tunnelling process, are proposed in the following equations:
smax ¼ 0:005D in fine soils ði:e: boulder claysÞ
ð6Þ
smax ¼ 0:0125D in coarse soils ði:e: gravelsÞ
ð7Þ
If boulders are encountered (as evidenced by measurements of jacking forces and penetration rates during driving), larger values of smax may be appropriate, but are difficult to predict. Assuming, from Fig. 7, that k = 0.275 and k = 0.5 are representative trough width parameters for coarse and fine soils respectively, then conservative estimates of Vl when using current tunnelling methods may be made by substituting the relevant k value and either Eqs. (6) and (7) as appropriate into Eq. (3) to arrive at:
12
B.A. McCabe et al. / Tunnelling and Underground Space Technology 27 (2012) 1–12
z 0 V l ¼ 0:008 in fine soils D
While Eqs. (8) and (9), like any other empirical expressions, should be used with caution, they can provide a useful first estimate of likely volume loss due to tunnelling in Irish ground conditions using current tunnelling methods. Once again, these values may be expected to increase if boulders are encountered.
information presented in this paper relating to the Dublin Port Tunnel and Mullingar Sewerage Improvement Scheme, respectively. Supplementary information provided by Mr. Conor O’Donnell of AGL Consulting is appreciated. The authors would also like to acknowledge J.B. Barry and Partners, Consulting Engineers, who employed Ciarán Reilly during the construction of the Mullingar Scheme, and Ward and Burke Construction, the microtunnelling contractors for the Mullingar Sewerage Improvement Scheme, who employed Brien Curran while carrying out a research Masters at NUI Galway.
6. Conclusions
References
This paper has aimed to collate Irish experience of settlements above tunnels in glacial tills, with a view to providing preliminary pre-contract estimates of the amount and extent of surface settlement that is likely to arise in a tunnelling contract. The following conclusions may be drawn from this research:
Curran, B.G., 2010. Analysis of Jacking Loads and Ground Movements for Microtunnelling in Irish Ground Conditions. MEngSc. Thesis, Dept. of Civil Engineering, NUI Galway. Devriendt, M., 2010. Risk analysis for tunnelling ground movement assessments. In: Proceedings of ICE Geotechnical Engineering, vol. 163, no. GE3, pp. 109–118. Farrell, E.R., 2010. Lessons learned – problems to solve. In: Proceedings of the Symposium on Bridge and Infrastructure Research in Ireland, pp. 181–198 (keynote). Farrell, E., Lehane, B., O’Brien, S. and Orr, T. 1995. Stiffness of Dublin black boulder clay. Proc. XIth.ECSMFE, Copenhagen, 1, 1-109 – 1-117. Farrell, E., Wall, D., 1990. Soils of Dublin. Transactions of the Institution of Engineers of Ireland, pp. 78–97. Finch, T.F., 1977. Soils of Co. Westmeath, Soil Survey Bulletin No. 33, An Foras Talúntais. Glossop, N.H., Saville, D.R., Moore, J.S., Benson, A.P., Farmer, I.W., 1979. Geotechnical aspects of shallow tunnel construction in Belfast estuarine deposits. In: Proc. Tunnelling ’79, London, pp. 45–56. Goto, J., Thompson, S., Squires, S., Mizuno, S., 2004. Dublin Port tunnel – excavation of a 11.8 m diameter urban motorway tunnel. In: Chinese Taipei Tunnelling Association Conference, pp. D091–D098. Jones, B.D., 2010. Low-volume-loss tunnelling for London ring main extension. In: Proceedings of ICE Geotechnical Engineering, vol. 163, no. GE3, pp. 167–185. Lehane, B.M., Simpson, B., 2000. Modelling glacial till under triaxial conditions using a BRICK soil model. Canadian Geotechnical Journal 37 (5), 1078–1088. Lochaden, A.L.E., Farrell, E.R., Orr, T.L.L., 2008. Comparing numerically modelled tunnel induced ground movements with field measurements. In: Proceedings of BCRI ’08 Symposium, NUI Galway. Long, M., Menkiti, C.O., 2007. Geotechnical properties of Dublin Boulder Clay. Geotechnique 57 (7), 595–611. Macklin, S.R., 1999. The prediction of volume loss due to tunnelling in overconsolidated clay based on heading geometry and stability number. Ground Engineering 32 (4), 30–33. Mair, R.J., 1996. Settlement effects of bored tunnels. In: Mair, R.J., Taylor, R. N., (Eds.), Session Report, Proc. Int. Symposium on Geotechnical Aspects of Underground Construction in Soft Ground. London, Balkema, pp. 43-53. Mair, R.J., Taylor, R.N., 1997. Bored tunnelling in the urban environment, state-ofthe-art report and theme lecture. In: Proceedings of the 14th International Conference on Soil Mechanics and Foundation Engineering, vol. 4. Hamburg, Balkema, pp. 2353–2385. O’Reilly, M.P., New, B.M., 1982. Settlements above tunnels in the United Kingdom: their magnitude and prediction. In: Proceedings of Tunnelling ’82 Symposium, London, pp. 173–181. Orr, T.L.L., Farrell, E.R., 1996. Geotechnical aspects of tunnelling in soft ground in Ireland. In: Mair, Taylor (Ed.), Geotechnical Aspects of Underground Construction in Soft Ground, pp. 301–305. Peck, R.B., 1969. Deep excavations and tunnelling in soft ground. In: Proceedings of the 7th International Conference on Soil Mechanics and Foundation Engineering, vol. 3. Mexico City, pp. 311–376. Skipper, J., Follett, B., Menkiti, C., Long, M., Clarke-Hughes, J., 2005. The engineering geology and characterisation of Dublin Boulder Clay. In: Quarterly Journal of Engineering Geology and Hydrogeology, pp. 171–187.
V l ¼ 0:011
z 0 in coarse soils D
ð8Þ ð9Þ
1. For glacial till soils, it is important to establish whether the predominant overburden material governing the settlement trough has the characteristics of a fine or coarse material since the k values for each have been shown to be quite different. The correct choice is important to obtain a reliable prediction of volume loss. 2. The values of the trough width parameter, k, for coarse soils lie in the range 0.2–0.3, which is at the low end of the range given in the Mair and Taylor (1997) database for sands and gravels. 3. The values of k for fine soils lie in the range 0.4–0.6, which is consistent with the Mair and Taylor (1997) database for clays. There is no obvious distinction between the k values obtained for tunnels in soft and stiff clays. 4. An increase in smax with time in the Mullingar coarse soils is believed to be related to dissipation of the bentonite fluid used to support the overcut to the pipelines. There is no systematic variation in k with time in these soils. None of the boulder clays referred to were soft with the result that consolidation settlement was not a significant issue. 5. The data presented has enabled reasonably confident upper limits to be placed on the smax values likely to be obtained for tunnels in fine and coarse soils using current tunnelling methods, which is useful in obtaining a preliminary estimate of the volume loss. However, the presence of boulders, as well as reducing the advance rate, can result in high values of smax and high volume losses which are difficult to predict with accuracy.
Acknowledgements The authors would like to thank Nishimatsu Construction Co. Ltd. and Westmeath County Council for the permission to use