Joseph Newhouse, Mott MacDonald (formerly Imperial College London)
Arup Tech Paper logo (NB)
Ground movement due to shaft construction can damage surrounding assets. A comprehensive synthesis of case studies of ground surface movement due to the excavation of circular shafts is presented herein. The data highlight the importance of considering shaft construction method when making empirical predictions.
After subtracting significant settlement due to dewatering, two shafts at Crossrail’s Limmo Peninsula tunnelling site have been added to the existing data set. This has confirmed that when wall installation occurs prior to excavation, shafts typically experience less ground surface settlement than when installation and excavation are concurrent.
New design lines which differentiate between construction method are presented.
High-quality predictions of ground movement due to circular shaft construction are of importance for projects in urban environments. Such movement can result in unacceptable damage to surrounding structures and utilities, and so one must understand which assets are at risk. However, overestimation of movement can lead to the implementation of unnecessary and costly mitigation measures.
Empirical formulae are a key tool for making predictions in line with best practice guidelines such as Crossrail’s Civil Engineering Design Standards (Crossrail, 2010b). However, for circular shafts, historically the data set of case studies on which predictions may be based has been very small, with a single shaft underlying the formula presented by New and Bowers,1994, which for many years was the primary tool for empirical predictions. Subsequent formulae have been developed to take account of shaft geometry (Mott MacDonald, 2013; McNicoll, 2013; Pairaudeau, 2011; GCG, 2007), but the limited number of case studies has prevented a detailed assessment of the influence of construction method.
However, a growing body of case studies, including two shafts at Limmo Peninsula tunnelling site presented herein (and expanded upon in Newhouse, 2017), has offered the opportunity to update and improve current prediction methods.
2.0 Categories of shaft and cause of ground movement
Following Faustin et al, 2017, three categories of shaft may be defined:
- Pre-lined; the wall lining is installed prior to excavation eg secant, sheet pile and diaphragm walls.
- Concurrently-lined; installation of the wall lining and excavation are concurrent eg caisson, underpinning and sprayed concrete lining (SCL).
- Combined; the upper portion of the shaft is pre-lined and the lower portion is concurrently-lined.
There are three general causes of ground movement due to shaft construction that encompass these categories; installation of the wall, excavation in front of the wall, and changes in groundwater regime. When concurrently-lined, movement due to installation and excavation are coincidental.
The focus of the present study is on the maximum vertical ground surface movement due to circular shaft excavation. Details of other facets of ground movement, including settlement trough extent, can be found in Newhouse, 2017.
A sketch of the balance between heave due to unloading and factors causing ground settlement is presented in figure 1. The figure shows two forms of shaft construction used at Limmo Peninsula tunnelling site (section 4); diaphragm wall and SCL.
3.0 Synthesis of case studies
The body of case studies has grown significantly over the last five years, with both Mott MacDonald, 2013, and Schwamb, 2014, presenting details of the increasing catalogue. These, along with subsequent case studies, are summarised in Newhouse, 2017. The 19 case studies are primarily located in London, and have predominantly competent ground conditions that typically comprise stiff clays.
For each of the case studies, a “best-estimate” of the maximum vertical ground surface movement due to excavation, δv.max, has been defined for the available settlement data. δv.max has been normalised by the excavation depth, H, to give α. Categorised by shaft construction method, the values of α are plotted against shaft internal diameter, D, in figure 2, and against D/H, in figure 3. Values of α for the main shaft and auxiliary shaft at Limmo Peninsula tunnelling site (Section 4) at the completion of excavation (H=Hfinal) are included in figure 2 and throughout their excavation in figure 3.
figure 3 rev
The following may be observed:
- In figure 2, the data are presented against the design line proposed by GCG, 2007, for the Crossrail project. The line is close to an upper limit of the data. Thus, in most cases, it would overpredict ground surface settlement. For pre-lined shafts, the overprediction would be considerable.
- Construction method is shown to be of key importance. For common values of D or D/H concurrently-lined shafts (eg underpinning or caisson, with SCL follow-on) show the greatest values of α. Pre-lined shafts (eg secant pile or diaphragm walls) show the smallest values of α. Combined shafts (eg jet grout, secant pile, or sheet pile walls, with SCL follow-on) lie between the two extremes.
- Considering separately pre-lined shafts, and concurrently-lined shafts, the geometry of a shaft clearly has an impact on the observed ground movement; greater values of D and D/H typically exhibiting larger values of α.
In figure 2 and figure 3, new design lines for pre-lined and concurrently-lined shafts are tentatively presented. The design lines represent moderately conservative upper and lower extents of the limited data sets, as well as an average between these two thresholds.
4.0 Overview of Limmo Peninsula tunneling site
Two shafts were constructed at Crossrail’s Limmo Peninsula tunnelling site; the main shaft and auxiliary shaft. Structural and simplified geological details for the shafts are presented in table 1 and are shown along with details of the site hydrogeology, and dewatering and piezometric installations in figure 4.
Table 1: Details of the shafts at Limmo Peninsula Tunnelling Site
|Shaft||Construction method||Wall depth (m)||Primary lining wall thickness (m)||Final excavation depth, Hfinal (m)||Internal Diameter, D (m)|
16.8m sheet pile, 22m SCL
SCL0.6 to 0.8
*Depth reached prior to pause in excavation; final depth considered in the analysis presented herein
Generalised geological stratigraphy for both shafts: 13.5m made ground, alluvium and River Tarrace Deposits; 30m London Clay; 1m Harwich Formation; 17.5m Lambeth Group; 26.5m Thanet Sand; Chalk
Source: DSJV, 2012b, Crossrail, 2011, and Crossrail, 2010a
The site lends itself to the analysis of ground movement, as movement is anticipated to have been relatively unaffected by historic or existing structures. However, extensive dewatering during construction resulted in significant ground settlement, obscuring movement due to excavation. Piezometric level data collected during the shaft construction are presented in figure 5.
Key points of note are listed below. The data these points refer to are indicated in figure 5 adjacent to a corresponding arrow:
- A pumping test was conducted using the external Chalk wells, prior to the excavation of the shafts.
- Drawdown in the lower aquifer was primarily a response to dewatering from the seven external Chalk wells (see figure 5 – 2a) and drawdown in the middle aquifer was primarily a response to dewatering from the six external Thanet wells (2b).
- Piezometric levels gradually fell in the low permeability London Clay, in response to the drawdown in the middle aquifer.
- There was no dewatering and consequently no drawdown in the upper aquifer.
- Varying instruments exhibit gaps in readings (5a), termination of changeable readings (5b), erratic behaviour (5c) and sudden spikes (5d). These were all accounted for in the analysis described below.
5.0 Analysis of ground movement at Limmo Peninsula tunneling site
To allow for comparison of the ground movement due to excavation of the main shaft and auxiliary shaft with the case studies presented in section 3, settlement due to dewatering has been isolated from the total vertical ground surface movement observed. To calculate the settlement due to dewatering, a three-dimensional numerical model of the site was developed. The model was developed assuming:
- The aquifers (shown in figure 4) to be confined, and flow in the aquifers to be horizontal.
- The drawdown from each well to be axisymmetric, and the principle of superposition to apply when calculating the total drawdown.
- A transient drawdown response in the aquifers in response to the measured dewatering flow rate (Preene, Roberts and Powrie, 2016).
- Vertical flow in the London Clay aquiclude, with the drawdown calculated using the finite difference method (Knappett and Craig, 2012), whereby drawdown in the middle aquifer was used to define a variable boundary condition.
- Settlement due to dewatering to be exclusively due to one-dimensional consolidation of the London Clay in response to the change in effective stress resulting from drawdown.
The drawdown was calibrated against the observed piezometric readings presented in section 4.
In figure 6, the predicted dewatering settlement is presented against the actual total settlement for five of the 57 levelling points analysed. The difference between the actual total settlement and predicted dewatering settlement is considered to be due to excavation. Generally, there is good agreement between the two, indicating that the majority of movement was due to dewatering.
For both the main shaft and auxiliary shaft, “best‑estimate” values of α for the movement attributed to excavation are presented for increasing excavation depth, H in figure 3. In agreement with the existing case studies:
- For the main shaft (constructed with a pre-lining of diaphragm walls), in accordance with the other pre-lined shafts, the values of settlement were small, and there is general trend of a decrease in α with decreasing D/H.
- For the auxiliary shaft (a combined shaft constructed with a pre-lining of sheet piles over its upper half, and concurrently-lined with SCL thereafter), the values of settlement were small while the excavation remained within the sheet piles, but increased significantly as excavation transitioned to SCL, up to values that may have been projected in line with the existing concurrently-lined shafts.
Field data showing ground movement due to the excavation of two shafts at Limmo Peninsula tunnelling site have been added to a synthesis of existing case studies. Together these highlight the importance of accounting for construction method when making empirical predictions. For given values of shaft internal diameter, D, and diameter normalised by excavation depth, D/H, shafts where wall installation and excavation are concurrent exhibit significantly greater normalised settlement than shafts where wall installation occurs prior to excavation. New design lines have been presented which differentiate between these construction methods. For many shafts, the design lines will facilitate a reduction in predicted settlement, offering the potential to save on costly mitigation measures.
This work is testament to collaboration across the civil engineering industry; from construction to research. I would like to thank Jamie Standing of Imperial College London for his guidance in producing both this paper and my underlying MSc dissertation (Newhouse, 2017). I would also like to thank Michael Williamson and Rob Talby of Mott MacDonald and Mike Black of Crossrail.
Support from the British Geotechnical Association Fund and the Mott MacDonald Research and Further Education Fund was also gratefully received.
Crossrail. (2011) C121 - SCL Tunnels, Limmo Peninsula Shaft, Geotechnical Design Report - Part 2, Geotechnical Design Summary. 2.0 Report number: C121-MMD-C2-RGN-CR144-SH011-00002.
Crossrail. (2010a) C123 - Intermediate Shafts, LIMMO, Geotechnical Design Report Part 2: Geotechnical Design Summary Report. 2.0 Report number: C123-JUL-C2-RGN-CR144_SH011_Z-00004.
Crossrail. (2010b) Civil Engineering Design Standard, Part 7, Ground Movement Prediction. 6.0 London, Report number: CR-STD-303-7.
DSJV. (2013) C305: Eastern Running Tunnels, I+M Weekly Summary Report, Year 3, Period 7, Week 4. 90.0 Report number: C305-DSJ-C2-RGN-CRG03-50002.
DSJV. (2012a) C305: Eastern Running Tunnels, Dewatering at Limmo Peninsula. 2.0 Report number: C305-DSJ-C-GMS-CR144_WS155-50008.
DSJV. (2012b) Limmo Shaft, Shift Review Group Reports from 14/02/12 to 18/10/12. Report number: C305-DSJ-C2-MRC-CR144_WS155-50002.
Faustin, N, Mair, R, Elshafie, M, Menkiti, C and Black, M, (2017) Field measurements of ground movements associated with circular shaft construction. Proceedings of the 9th International Symposium on Geotechnical Aspects of Underground Construction in Soft Ground, TC204 ISSMGE. 2017, Sao Paulo.
GCG. (2007) Settlement Estimation Procedure: Box Excavations & Shafts. Rev B London, Crossrail. Report number: 1D0101-G0G00-01004.
Knappett, J and Craig, R, (2012) Craig’s Soil Mechanics. 8th edition. Abingdon, Spon Press.
McNicoll, S, (2013) Study into ground induced settlement caused by shaft construction on the London Power Tunnels. The Harding Prize. British Tunnelling Society.
Mott MacDonald. (2015) Ground Movement Impact Assessment - Stage 3 Refinements, Confidential Project. 2.0 Report number: MMD-N001-2360000-GEO-RPT-00024.
Mott MacDonald. (2013) Impacts of Tunnels in the UK. London, HS2.
Muramatsu, M and Abe, Y, (1996) Considerations in shaft excavation and peripheral ground deformation. In: Mair, R and Taylor, R, (eds.) Geotechnical Aspects of Underground Construction in Soft Ground. Balkema, Rotterdam. pp.173-178.
New, B and Bowers, K, (1994) Ground movement model validation at the Heathrow Express trial tunnel. Tunnelling ’94. 301-329.
Newhouse, J, (2017) Ground movement due to shaft construction. MSc. Imperial College London.
Pairaudeau, H, (2011) Construction and Performance of Shafts. MEng. University of Cambridge.
Preene, M, Roberts, T and Powrie, W, (2016) Groundwater control: design and practice. 2nd edition. London, CIRIA. Report number: C750.
Schwamb, T, (2014) Performance Monitoring and Numerical Modelling of a Deep Circular Excavation. PhD. University of Cambridge.
Wong, R and Kaiser, P, (1988) Behaviour of vertical shafts: Reevaluation of model test results and evaluation of field measurements. Canadian Geotechnical Journal. 25 (2), 338-352.