The basic concept of effects of tunnelling on buildings describes the transverse settlement trough caused. The trough conforms to a classical Gaussian equation that can be derived from a fundamental parameter in settlement assessments, the volume loss. A volume loss of around one per cent has been established in many cases in London Clay.

Volume losses within the 1.5km EPBM-bored London Clay section of HS1 contract 220 were nearly always less than one per cent. A number of factors control volume loss, but controlling the face pressure is an important element in controlling ground movement. Mixed ground conditions can give rise to fluctuations in face pressure but London Clay is more uniform and it is easier to keep face pressures broadly constant.

For London Clay there is some influence of TBM chamber/face pressure on the volume loss in EPB tunnelling. If for seven quite well instrumented cases, volume loss is plotted against the average face pressure, divided by the total overburden pressure at tunnel axis (sv0), then there appears to be a relationship. As the ratio of face pressure / sv0 rises, volume loss lowers (Mair, 2008).

The depth of the tunnel will have an influence, but what can be seen is that in London Clay we should be getting reliably less than one per cent volume loss with well controlled EPB tunnelling.

Structural stiffness
If a building is considered in relation to the settlement trough, the deflection ratio and horizontal strains are important factors to consider. The stiffness of the building modifies the settlement shape and it may not experience much deflection ratio. A modification factor (M – the deflection ratio of the building/Greenfield deflection ratio) can be established. For a fully flexible building M=1 and for a fully rigid building M=0.

Centrifuge modelling tests have examined the modification factor in some detail. Structures of different stiffness and tunnels at different depths and positions in relation to the structure were idealised. If extracting fluid from the annulus of the modelled tunnel accurately controls volume losses, Greenfield settlement troughs can be produced which, with the addition of a flexible building, are similar. Increasing the stiffness of the building gives an increasingly stiff response to the settlement trough. The relative stiffness of a building (?*) has been established (Potts and Addenbrooke, 1997).

This can be simplified for design purposes as ?hog and ?sag (Mair, 2011). If the centrifuge results are plotted on a graph of modification factors against the log of the relative stiffness of the building ?hog or ?sag , a useful ‘band’ for design can be seen, which is consistent with finite element analyses undertaken by Potts and Addenbrooke (1997) and Franzius et al (2006). A fully rigid response (M=0) is seen if ?hog or ?sag > 1 or a fully flexible response (M=1) is seen if ?hog or ?sag < 10-4.

The horizontal strains measured in a building are often negligible compared to the greenfield theoretical value of horizontal strain.

If a building is assumed to follow the greenfield settlement this can significantly overestimate the damage category. Assuming that the horizontal strain induced in a building is small reduces the damage category, and for a rigid building the category of damage reduces much further. The need for protective measures such as compensation grouting may often then be overestimated.

Drainage
With tunnelling there is an immediate undrained settlement, based on volume loss. If the tunnel acts as a drain there will be a long-term slow development of flow through the clay into the tunnel causing long-term settlement (the tunnel will also squat).

For old London Underground (LU) tunnels pore pressure can be zero at the lining rising rapidly with distance from the tunnel, with exceptions. The tunnel lining system’s permeability is critical, particularly the grout. Older grouts often have a much higher permeability than modern. This can be due to long leaching and carbonation.

If a tunnel acts impermeably there will be very little long-term settlement. A long term dimensionless settlement (DS) can be defined (Wongsoroj, 2005; Mair, 2008), this would vary from zero for an impermeable tunnel to one for a fully permeable tunnel. If DS is plotted against the log of relative permeability (RP) tunnels with an RP<0.1 may be classed as impermeable and there will be no long term settlement. Those with an RP>100 may be classed as fully permeable and there will be significant long-term settlement.

The initial ground pore pressures are a key factor to consider for long-term settlement. The degree of anisotropy of soil permeability is also important; the horizontal permeability of soil is usually greater than the vertical. London Clay permeability is highly variable.

Identification of London Clay divisions resulted from studies on Jubilee Line works passing beneath St James’s Park. Large volume losses were recorded. It was suspected that tunnelling had an influence but also that geology played an important role. The London Clay beneath Westminster can be divided into Unit B, Unit A3 (which is split up into two Units; A3ii and A3i) and Unit A2, which sits above the Lambeth Group. Each of these Units has a different structure and has a range of permeabilities.

For long term monitoring extending over a period of many years, the instrumentation and systems need to be robust and durable. Long-term settlements can occur up to 100m away from the tunnel works, which makes the positioning of a datum problematic. There are not only vertical movements to be measured but also horizontal displacements and pore pressures.

The magnitude of movements depend on: the type of ground (the more compressible the greater the likelihood of movement); the groundwater conditions; the presence of existing tunnels; the type of lining; the relative ground to lining permeability; time; relative consolidation and creep processes.

Cases
St James’s Park is underlain by Units B and A3 and the tunnel was constructed with an open face shield with expanded concrete segments. The westbound tunnel was in Unit A3, with A3ii lying just above the alignment. The eastbound tunnel was in Units A3ii and B. The lining was permeable – a walk through the tunnel showed the tunnel lining to be wet in one Unit and then dry in the adjacent Unit. At St James’s Park there is almost a hydrostatic profile. Settlement profiles over time show a progressive deepening and widening of the trough with around five times the short-term movements recorded. Consolidation settlement continues for 16 years following construction.

The larger 8m-diameter CTRL tunnels at Dagenham were constructed in the Unit A2 by EPBM and grouted segmental linings. Consolidation movements took place within quite a short period of time with 12mm in the first month reducing quite drastically. The long-term settlements were of similar magnitude to the short term and took place very rapidly. The tunnel was located within the Unit A2, which has the highest permeability and allowed pore pressure to essentially equilibrate. As for St James’s Park, this was a greenfield site with hydrostatic pore pressures and no other tunnels nearby. The smaller settlements recorded can perhaps be attributed to a grouted lining.

Monitoring for Elizabeth House on the JLE showed that cumulative settlements from running tunnels and cross over passage, constructed in sprayed concrete, are 4.5 to five times smaller than those at St James’s Park and are levelling out. The reason is that this is not a greenfield site; the area is crossed by number of existing tunnels. The pore water distribution at this site shows a certain amount of under drainage compared to the hydrostatic profile found at St James’s Park. The Elizabeth House tunnels are also constructed in Unit A3i, the less permeable Unit. The lining is very different with St James’s Park being permeable and the ones at Elizabeth House almost impermeable.

Long-term settlements measured at the Treasury Building for four years before it was renovated show a settlement of about 25mm, compared to the 30mm measured at St James’s Park over a similar time period. Settlement took place regardless of the compensation grouting.

Long-term monitoring
Field monitoring in the long term is vital for understanding the effects of consolidation and creep and that more thought about who does the monitoring is needed. The long term response is dependent on a number of factors: the type of ground; the London Clay Units; the ground water conditions – whether hydrostatic or underdrained; the presence of any existing tunnels; the type of lining and the relative ground to lining permeability; and also whether there is just consolidation or creep as well. Creep plays a major factor and this should be considered for analysing long-term movements accurately. Long-term displacements can be several times greater than the short term, and the trough widths widen considerably as at St James’s Park.

Long term settlements at Dagenham for the CTRL and at Elizabeth House were much smaller than at St James’s Park. Long-term settlements are also evident for a structure that has been compensation grouted. It would also be useful to gain information on long-term horizontal movements so we can ascertain the long-term displacement field in conjunction with pore pressure measurements.

LU behaviour observations
Many of LU’s tunnels are over 100 years old, in over-consolidated clays, with many elements that would not satisfy modern design requirements. These structures are usually robust, long lived and have low maintenance, with occasional failures, an opportunity to study long-term behaviour.

Changes in LU’s tunnel structures include: changes in lining loadings due to tunnel excavation, other works or soil changes; deformation of tunnel linings by loading changes; groundwater change with recent concerns relating to perched aquifers; thermal changes such as heat generated by trains; and soil changes such as dessication or acid generation.

Data from an instrumented ring at Regent’s Park shows that 35 years after its construction, lining loads are currently at nearly 60 per cent of overburden and are still increasing slowly. An instrumented ring at St James’s Park shows that the lining load increases significantly during the first decade asymmetrically between 70 per cent and 90 per cent of overburden.

Circularity surveys by Tubelines found that despite uncertainty over build shapes, a large data set confirms that most London Clay tunnels squat by up to one per cent.

A consequence could be a slow change leading to equilibrium, possibly with ‘locked-in’ stresses, an example of which could be tunnel squat. Alternatively there could be slow changes, possibly at a reducing rate, which could lead to the onset of a progressive failure: an example is damage to an expanded pre-cast concrete lining first noted 30 years after construction, caused by movements due to soil desiccation increasing lining loads.

Such processes develop over decades, which means that determining when intervention is needed is difficult. Slow changes may eventually lead to a sudden failure. In London’s clay tunnels these events may occur after many decades of satisfactory performance.

Change management
Despite the minimal maintenance, these structures don’t necessarily exist in perfect equilibrium. There is change going on in much of the network that can give rise to changes later in the service life. A couple of those problems arose at about 30-40 years into the life of the structure, while another arose 100 years into the life of the structure. This would not necessarily be the time to expect this to happen.

LU needs to understand what the structures could be capable of and ensure they have appropriate systems in place to track any changes that develop.