The Philosophy of Underpinning

Clive Richardson


  Shrinkable Clay Victorian heating pipes under the Horniman Museum dried out the clay subsoil, causing shrinkage to a depth of 5.5m. Construction of a new basement eliminated the problem and gave the museum additional floorspace

According to the London Weather Centre, the summer of 1996 had 23 per cent below average rainfall, 21 per cent above average sunshine and 10 per cent above average temperatures. As many buildings in the UK are built on clays that shrink when their moisture content reduces, these warm, dry weather conditions have led to an increase in building movement; more cracked buildings are being reported and insurance claims for subsidence are on the increase.

However, the usual remedy, underpinning, is far from being a universal panacea. Not only is it expensive, but it is now widely recognised that underpinning can also be traumatic for both the buildings and their occupants. Where cracks are caused by subsidence (not all are), underpinning should be specified only with great caution.

This article looks at subsidence in low-rise and medium-rise buildings, which usually have shallow foundations or cellars up to about 3m deep. Subsidence and settlement are taken as interchangeable terms, meaning the sinking of ground on which a structure is founded. Ground can also rise, in which case the movement is known as 'heave'. Depending on the cause of subsidence or heave, horizontal stretching or
squeezing of the ground can accompany the vertical movement.

When movement is, or is likely to become excessive, so that the use or safety of the building is compromised, then underpinning may be a solution. It is usually achieved by digging underneath shallow footings and pouring concrete to extend the reach of the foundations down to a stable stratum. Other forms of underpinning include piles (column-like foundations). In some circumstances, pressure grouting (the use of a fluid cement to consolidate the ground itself) provides an alternative to underpinning.


While some early medieval timber buildings had their main posts dug into the ground, almost all surviving buildings have timber sill beams resting on low masonry plinths. Masonry buildings of the period had walls which sat straight into the ground with no attempt being made to spread the load over a broad foundation. These buildings largely succeeded because their builders followed tried and tested construction techniques and because they could be more selective in choosing suitable ground on which to build.

In later buildings, masonry walls were sometimes supported on one or two wider courses in a series of steps (a form of 'corbelling') to provide a better distribution of the load on the soil. Where the ground was reliable, this practice continued until the First World War, sometimes on a shallow strip of concrete cast into the trench about 500mm below ground. In poor ground, short timber piles were driven before starting the masonry, or cellars were constructed to reach stable ground below the surface.

With the advent of modern mild steel and reinforced concrete at the turn of the century, the design of foundations became more sophisticated. These and the forms seen today fall into three broad categories (and many hybrids):

  • Shallow spread footings: where surface soils have sufficient load bearing capacity and stability, concrete pads, strips, or rafts, are used to spread the forces from the superstructure over the ground, just deep enough to avoid unstable topsoil, frost penetration and seasonal moisture variations
  • Piles: where the surface soils are so weak that even a raft foundation under the entire footprint of the building is insufficient, piles are used to reach deeper and stronger strata
  • Basements/box foundations: as an alternative to piles, a basement box of reinforced concrete can be constructed which, in effect, floats in the ground. The weight of the ground removed to form the basement compensates for the weight of the new building. To a certain extent, old brick cellars, vaults and crypts behave similarly.


During construction, buildings settle as the ground adjusts to the new weight imposed upon it. Where built on rocks, gravel or sands, constructional settlement is substantially complete by the end of construction. For clays, silts and peat however, settlement takes many years. Once constructional settlement is complete it will not recur, unless the status quo alters. Constructional settlement is not usually detrimental provided the structure settles uniformly or is robust enough to accommodate differential settlement.

Variable ground can produce excessive differential settlement. For example, when part of a terrace of houses straddles an old river bed, that part is likely to settle by a different amount from the rest of the terrace.

Constructional settlement may also occur when existing structures are substantially extended or underpinned, as the stress in the ground is increased at a greater depth than before.

Similarly, there is a risk of differential settlement occurring between a building which has been disturbed and neighbouring parts which have not, such as adjoining buildings which may have finished their constructional settlement years ago. The settlement can be difficult to control due to the constraints of the existing fabric. If structural damage does occur, then it should be monitored and repaired at the end of the settling-in period. Provision should always be made within the project costs to pay for the monitoring and repair of any distress which may occur.

Constructional settlement does not always stop. Old buildings with overloaded footings on soft clay, including some Georgian and Victorian houses, have never quite achieved equilibrium and are still sinking slightly today, due to the nature of clay.


Anything which substantially disturbs the balance between the ground and the structure can promote new settlement. Tunnelling, mining and deep excavations, or altering loads - by building on old foundations, for example - can promote new movement in all types of ground. Ground-specific causes include frostheave, ground vibrations, changing water-tables, leaking drains, droughts and trees (see table 'Causes of Ground Movement', below).


A building's response to ground movement depends upon the continuity, ductility, and stiffness of its structure.

Good structural continuity (or 'tensile connection') can be provided by timber, steel and reinforced concrete frames which enable buildings to flex without coming apart at the seams. However, the lack of structural continuity or 'togetherness' of most pre-1970 unframed masonry structures permits joints to open and cracks and instability to occur more readily. After 1970, the Building Regulations and British Standards were amended to provide continuity in the wake of the progressive collapse of Ronan Point in 1968.

Ductile structural materials such as steel and properly detailed reinforced concrete can accommodate large deformations without breaking. In contrast, a brittle material such as unreinforced masonry set in cement mortar can only deform within its elastic limit. Historic unframed masonry structures can accommodate large distortions without cracking due to the 'creep' of the lime mortar, if movement is not too fast. (Creep is the continuing deformation or 'strain' of a material under constant stress). Modern cement mortars do not creep.

If a structure is sufficiently stiff it may be able to ride out the ground movement, moving or tilting as a whole, and heavily braced frames and compact crosswall structures with only small openings may have sufficient rigidity to disperse localised ground movement.

If ground movement is anticipated, say from tunnelling, then structural damage may be mitigated by installing temporary tie-bars and bracing door and window openings.


Not all distortions and cracks in buildings are necessarily due to ground movement. Symptoms of distress can also be caused by inadequate strength of materials, inadequate structural togetherness, material decay, dimensional instability (caused by thermal and moisture movement), overall instability, alterations, misuse and accidental loads (Editor's note: see this author's article, 'Structural Movement: Is it Really a Problem?' in The Building Conservation Directory 1996).

There are no foolproof rules for distinguishing between the causes of movement in buildings, and correct assessment can only be made with experience and by following good surveying practice. It is essential to be thorough: examine every part of the structure and every possible cause of failure; consult geological maps; record all individual symptoms; and keep an open mind. The most probable causes may be determined by a process of elimination. If symptoms are consistent with ground movement as well as other causes, further investigations must be made to distinguish between them, including trial-pits, boreholes, drain testing, and movement monitoring.


Underpinning is a messy, noisy and traumatic operation for buildings and their occupants alike. Unless sophisticated and expensive jacking systems are incorporated, the underpinning will almost inevitably promote some additional subsidence as the works settle in. If a structure is partially underpinned, for example one house in a terrace, then future damage may recur as the rest of the non-underpinned structure continues settling. For these reasons, underpinning should be avoided if at all possible.

Underpinning is not necessary from a purely engineering viewpoint in the following situations:

  • where the cause of the ground movement has ceased and is unlikely to recur, repairing the damage should be sufficient
  • where the rate and total magnitude of anticipated ground movement is unlikely to significantly threaten the structural strength, stability or integrity of a building during its required lifespan, periodic repairs and redecoration should suffice. Doors and windows may have to be eased from time to time or changed for other types which are more tolerant of frame distortion.

When ground movement is expected to do structural damage, it may still be possible to reduce movement sufficiently to avoid underpinning, for example by:

  • pollarding and root-pruning trees
  • repairing leaking drains
  • modifying the superstructure
  • pressure-grouting the ground.

In certain cases, such as when a building is to be sold, an owner may be compelled to underpin in order to attract a purchaser even though it may be unnecessary in engineering terms.

Table 1 Typical causes and effects


    -Structural alterations
    -Roof extensions
    -Adjacent new structures
    -Adjacent demolition works


    -Excavations and cuttings
    -Yielding retaining walls
    -Tunnelling and mining
    -Settlement of soft ground in swallow-holes and fissures


    -Heavy vibrating machinery


    -Pumping ground water
    -Land drainage
    -Tree growth

    -Tree removal
    Leaking drains
    Frost heave (caused by the expansion of moisture as it freezes)
    Decline in aquifer pumping



Most types of underpinning involve digging holes under buildings in confined spaces. The existing structure is expected to defy gravity and temporarily arch over the excavation. Collapses can occur. The risks must be identified and managed in accordance with CDM legislation.

  • Investigate services before digging
  • Check that underpinning pits cannot flood or be gassed
  • Strengthen superstructure before digging
  • Check that walls above are strong enough to support themselves over pits
  • Support sides of excavations
  • Ensure that workers can escape from pits easily
  • Use threaded couplers instead of dowel bars to connect reinforcement rods between sections of shallow mass concrete underpinning
  • Ensure safe access and ventilation to pits
  • Use a Banksman to oversee safety.



Recommended Reading

  • C Richardson, 'The AJ Guide to Structural Surveys', The Architect's Journal, 1985
  • Guide to Subsidence of Low Rise Buildings, The Institution of Structural Engineers, 1994
  • 'Desiccation in Clay Soils', BRE Digest 412, Feb 1996


This article is reproduced from The Building Conservation Directory, 1997


CLIVE RICHARDSON BSc, CEng, FICE, FIStructE, ACIArb is a Chartered Engineer and Associate Director of JAMES Consulting Engineers. He is also Engineer to the Dean and Chapter of Westminster Abbey, and Visiting Lecturer in building conservation at The Architectural Association School of Architecture, London.

Further information





Structural Engineers

Site Map