With the redevelopment of inner-city sites it is increasingly necessary and often beneficial to reuse existing substructures. More often than not there is no technical reason why good use should not be made of what already exists. It often saves the client a lot of money, not to mention saving the planet.
For successful reuse it is necessary to understand how old foundations are likely to behave when their loading condition is changed, especially in comparison to any additional foundations which may be required. An appropriate strategy needs to be devised to distribute new loads between old and new foundations and minimise the risk of significant ground movement, based on an understanding of the behaviour of the mass of the soil. The knowledge is available to do this, but requires some boldness.
Engineers are trained to think of foundations in terms of their load-carrying capacity and to use high factors of safety. But the challenge of reuse forces them to give primary consideration to the actual performance of the foundations and the ground. Will the structural elements, which may have deteriorated with time, take the load changes? Will the ground displacements be compatible with the new loading conditions of the superstructure?
In calculating ground strains, it is necessary to understand the behaviour of ground during both first-time loading and during unloading and reloading cycles. It quickly becomes clear that simple linear elastic models are insufficient.
Increasing use of existing foundations must go hand in hand with advances in the understanding of the behaviour of soils at small strains and a greater ability to replicate the effects of time and strain-dependent behaviour using numerical modelling.
Another issue is the increasing evidence that the ultimate capacity of piles increases with time. Although this is generally accepted, there is no agreement about why and no model to predict the magnitude of the effect.
Although there is increasing awareness of the longterm behaviour of materials, it is not easy to be certain about the quality of the product in the first place and detailed assessment is very difficult. It is possible to test isolated elements, but a pragmatic view often has to be taken.
If a building has performed satisfactorily for many years, the foundations must be good for at least the loads that have already been applied, provided the materials are not deteriorating. But without knowing their size and ground conditions, it is difficult to predict behaviour.
Contemporary records of the design and construction of the foundations can help determine their size. However it is not always easy to find records, even for relatively recent buildings. At Athene Place, Holborn, the redevelopment team were very lucky to find a sealed capsule with as-built drawings in a trial pit (see box).
The likely behaviour of existing foundations can be calculated or measured. Measurement can be the testing of individual elements - for example, a pile - or by observing behaviour during demolition of the building. In many situations, however, neither of these is feasible, and it may be necessary to revert to a combination of theory and a pragmatic approach.
An example of this is the assessment of acceptable loads to be taken on existing walls for refurbished buildings. Often the calculated bearing pressures from the original building are far greater than would normally be acceptable for a new structure.
Small load increases are unlikely to result in large movements because ageing gives the soil an initially higher reloading stiffness. Analysis with small strain models, without the effects of ageing, also demonstrates that for clay soils - particularly soft clays, but also stiff over-consolidated clays - undrained bearing capacity is greater after the ground beneath the foundation has fully consolidated. For a typical shallow foundation in London Clay, it is reasonable to increase bearing pressure by about 10% of the original pressure without significant settlement. If increases are much more than this, it is sensible to spread the load, for example by widening the footing or tying it into some sort of supplementary foundation.
It is not uncommon to find large buildings with basements, built before it became common practice to use bored piles, which are a series of mass concrete pads or even rafts, with heavy reinforcement or grillages of steel beams.
The walls of these basements are also mass concrete with large bases, designed to provide temporary support and to take the loads from the external walls and columns. Unlike modern basements, which are designed to take car parks, in older buildings the basements have loads distributed through relatively closely spaced columns.
This means the building loads are fairly uniformly distributed in the ground. Unless the new building is significantly bigger than the old, it is tempting to use rafts or pads for the new foundations and put loads directly on to the walls.
The most convenient solution, although not often achieved, is to make do with a shallower basement and to distribute new loads through a raft built over the old foundations. Constructing new foundations at a similar depth to the old is cumbersome and expensive.
Economies can be made in raft construction by using isolated piles beneath heavily loaded columns to act as settlement reducers, as can be done in new basements (Burland and Kalra 1982). It is necessary to have a means of modelling what is happening in the ground, particularly if facades are to be retained and there is a risk of significant differential movements during demolition and reconstruction.
Alternatively, provision must be made for differential movements to be tolerated, at least in the short term, both within the facade itself (for example by the construction of movement joints) and in the connection between the existing and the new building. It is often necessary to consider the latter anyway because of the difference in the response of old and new materials to temperature changes.
In buildings where basements were designed to have large open spans, for example to house printing works, the constraints to redevelopment are greater than where the basement was general purpose. Under these circumstances the reuse of the entire substructure is attractive.
One example is the old Daily Mirror building, where the original structure, designed in the 1950s (Foot 1962), has a 15m deep basement surrounded by a mass concrete retaining wall constructed in hand-dug trenched panels before the main top-down excavation. The columns were founded on reinforced concrete bases, installed in shafts before bulk excavation. These bases and the wall foundations were tied into a raft in the final condition. With such a configuration it was attractive to refurbish the old basement rather than to reconstruct it. It was not, however, possible to reuse all the basement floor slabs (except the bottom one) and temporary support was needed to the walls while intermediate floors were removed.
It became necessary to demonstrate that the existing walls were capable of taking the changes in bending and shear. Although the majority of vertical loads could be taken using the existing large steel columns, some had to be taken on new columns on to the raft and it was necessary to demonstrate that this would be acceptable. One of the significant factors in the design of the new basement was the control of water pressures.
The basement had been designed with a drainage layer beneath the raft. Water was found to be leaking through and under the walls, resulting in a steady flow into sumps and the need for frequent pumping.
Although it is more difficult to check the integrity of piled foundations, reuse of timber, steel or concrete piles is often desirable because of the difficulty of working around them. They can be removed by overcoring but this is both expensive and disruptive to the ground. In the case of a piled raft, where the piles have been installed in a uniform grid beneath a thick raft, and the columns of the original building are not directly over the piles, the acceptability of a new column layout may be only a matter of ensuring that the raft is sufficient.
Given that the foundations of the original building would have been regarded as acceptable, it does not seem logical to assume that they may not be acceptable for a new building with the same or a lesser mass. Where the new building has locally higher loads than the original, an additional pile or piles can be installed through the slab, preferably directly under the column in question. Such a pile will work if sufficient load can be mobilised at a displacement which is compatible with the performance of the existing raft under load.
If all loads in the old piles are less than the previous loads, the deflection at working load will be small for most pile types (this needs to be checked for piles that are largely end bearing). Demonstration of the acceptability of a 'hybrid' raft needs to be done with care using a structural raft analysis making a range of assumptions about the relative stiffnesses of the old and the new piles.
An example of the reuse of a piled raft is Thames Court on Upper Thames Street at the former site of Bull Wharf. The 1960s building was founded on a thick concrete raft on straight shafted bored piles through alluvial deposits and Thames Gravel into the London Clay. The top of the raft was just below mean water level and the raft had been designed without piles within a box formed on two sides by the river walls. The use of piles had only been implemented at a very late stage, when it was found that the alluvium was soft. The piles were installed through the raft.
When the site was redeveloped in the 1980s, the new building plot included a substantial area outside the raft away from the river. The original scheme was to construct a large deep basement within new bored pile retaining walls and to remove the piles. A revised scheme with a shallower basement was eventually adopted which used the existing raft and constructed a new basement next to it linking to the raft. Additional piles, longer than the original ones, were added to the raft at isolated points beneath heavily loaded columns.
The raft was sufficiently stiff for it to behave more or less rigidly, and therefore the new piles had to mobilise a large enough load at low deflections to provide sufficient support to ensure that the steel in the raft was sufficient. The detailing of the link between the raft and the new pile was a critical issue.
Redevelopment of buildings founded on underreamed bored piles is an increasingly common problem as the 1960s'generation of buildings which saw their introduction comes to the end of its functional life. Underreamed piles present particular difficulties because of the area occupied by their bases, and because they are shorter than the equivalent straight-shafted piles. The result of the latter is that their response to unloading and reloading is much less predictable. If used in combination with new piles the matching of pile response thus becomes more difficult. In the case of most underreamed piles, because shaft friction is mobilised before the base resistance, the effect of unloading is to reverse the direction of the shaft friction and to lock in the base pressure. The stiffness on reloading is then likely to be high, but may decrease very rapidly once the old load is exceeded.
Juxon House, in the precinct of St Paul's Cathedral, was built in about 1963 on some of the first underreamed piled foundations in London. Problems arose because it was discovered that, in spite of the old buildings being some of the few in the vicinity to have survived the Blitz, they were built over a deep scour hollow in the surface of the London Clay. As a result some of the underreams could not be formed and it was necessary to replace them with straight-shafted piles.
Contemporary records and details of a large number of boreholes carried out both before and during piling gave full details of both pile types and the ground conditions encountered. The building was satisfactorily completed and recent surveys have shown no sign of any differential settlements. The site is now to be redeveloped on a slightly larger footprint with larger floor spans. New piles will be installed around the old ones, and both new and old are to be incorporated in new pile caps which will replace the old ones.
The response of the piles will be observed during demolition in order to confirm that they do have a high stiffness. The new ones will be deep CFA piles extending through the disturbed soils within the scour hollow and into the in-situ clay beneath. Additional site investigation has been carried out to confirm that the original records were representative and to fill in some of the missing information.
Hugh St John is director of Geotechnical Consulting Group Last year, while investigating the foundations to a car park in Shoe Lane, central London for developer Jones Laing LaSalle, consultant Whitby Bird and Partners uncovered a time capsule planted by the previous construction team 40 years earlier.
The 10cm diameter steel canister was buried by FG Mitchell on 21 September 1961 and included a scroll signed by the construction team, a note describing the development, a copy of The Times and the Financial Times, stamps, two model cars and a drawing showing the foundation layout.
It would be extremely useful if all redevelopments could unearth foundation drawings, loadings and as-built construction records as easily. This could allow rapid reassessment of foundation capacities, long-term performance and durability.
CDM regulations on storing asbuilt records are a major advance.
However, information can be lost and destroyed and may be incompatible with future software. The best location for the information is in the building itself, be it buried in a time capsule, stored in a safe in the basement or even electronically in a simple future-proof format on a silicon chip embedded in the basement wall.
Standardisation of drawing and data formats and systematic archiving procedures would be of great value for future developers, enabling them to redevelop quickly, safely and economically, recycling and reusing as many components of the building as possible. Only through this process can we ensure future sustainable use of ground and regeneration of urban areas.
Fiona Chow, Geotechnical Consulting Group and Alister Harwood, Interior plc