Pile Capacity: Shaft Friction, End Bearing, and the Sign That Reverses

Two mechanisms, mobilised at different displacements

A pile carries load two ways: friction along its SHAFT, and bearing at its TOE. The ultimate capacity is their sum, less the pile's own weight, and that much is uncontroversial.

What the sum hides is that the two do not arrive together. Shaft friction is fully mobilised after a very small movement — a few millimetres, essentially independent of the pile's size. End bearing needs far more, commonly around a tenth of the pile's diameter, which on a large-diameter pile is a settlement nobody would accept in service.

So at working load a pile is usually carrying most of its load on friction, with the toe barely engaged. Designing a large-diameter pile on its ultimate end bearing and checking nothing else produces a pile that is safe against collapse and unacceptable in settlement — which is the same serviceability-governs-strength pattern that runs through shallow foundations.

Diameter helps the toe twice as fast as the shaft

The shaft term contains the pile's PERIMETER, which grows linearly with diameter. The toe term contains its AREA, which grows with the square. So making a pile fatter buys end bearing much faster than it buys friction, and making it longer buys friction and nothing else.

That is the whole of the choice between a long slender friction pile and a short fat end-bearing one, and the ground decides it. Where a competent stratum exists at a reachable depth, an end-bearing pile taken down to it is efficient. Where there is no such stratum — deep soft deposits — the pile has to earn its capacity along its length, and that means length and number rather than diameter.

It also explains why boring records matter more than pile schedules. The unit friction differs in every layer the pile passes through, so a capacity is a sum over a specific sequence of soils, and a pile installed where the sequence turns out different has a different capacity even though it is the same pile.

The sign reverses: negative skin friction

Shaft friction resists the pile moving down through the soil. If instead the SOIL moves down past the pile — because a fill was placed over a compressible layer, or the water table was lowered, or the ground is consolidating for any reason — the friction acts in the other direction and drags the pile down with it.

That is negative skin friction, or downdrag, and it is a LOAD rather than a resistance. It has to be added to the structural load the pile carries, and it is applied over whatever depth the soil is settling more than the pile. The capacity available to resist it comes only from below that depth.

The consequence is severe and routinely missed: a pile through a settling layer can be carrying a substantial permanent load that appears on no structural drawing. The mitigations are physical rather than arithmetic — a bitumen slip coat on the shaft, a sleeve, or preloading the ground so the settlement happens before the piles go in.

An adfreeze uplift is the same mechanism turned upside down. Frozen ground grips a pier's shaft and frost heave lifts it, so the friction becomes an uplift load and the resistance is whatever weight and anchorage exist below the frozen zone.

Groups: ten piles do not carry ten times one

Piles in a group interact. Their zones of influence overlap, so the soil between them is being asked to work for several piles at once, and the group's capacity is less than the sum of its members' individual capacities. The ratio is the group EFFICIENCY, and for friction piles in clay at typical spacings it is meaningfully below one.

There is also a second failure mode that a single-pile calculation cannot see: BLOCK FAILURE, where the group and the soil between it move as one mass, shearing on the outside of the block. The governing capacity is the lesser of the sum-with-efficiency and the block, and the block governs for tightly spaced groups.

The block's capacity has two terms. Round its sides the failure runs through soil rather than along the piles, so the full undrained strength acts over the perimeter and the pile length with no adhesion factor; under its base the clay bears at Skempton's Nc, which grows with depth to two and a half widths and is higher under a square than a strip — 9 under a deep square block, 7.5 under a long narrow one.

Spacing is therefore a design variable rather than a detailing one. Wider spacing raises efficiency and enlarges the cap; tighter spacing shrinks the cap and loses capacity. Around three diameters is the conventional compromise, and it is a compromise rather than a rule.

Where it fails, and what replaces the calculation

Every unit friction and end-bearing value on these pages is an estimate from a correlation — with SPT blow counts, cone resistance, or an undrained strength — and the correlations carry real scatter. Installation changes the soil too: a driven pile displaces and densifies the ground around it, a bored pile relaxes it, and a pile bored under bentonite may leave a softened interface. The same soil gives different friction depending on how the pile got there.

Secant and contiguous pile walls add a geometric constraint on top of the capacity one: overlapping bores have to actually overlap, and verticality tolerance over depth is what decides whether they still do at the toe. A wall that interlocks at the surface and opens at depth is a leak, and the check is a geometry problem driven by construction tolerance rather than by soil strength.

The alternative to all of it is a LOAD TEST. Static tests load a pile and measure what it does, and instrumented tests separate shaft from toe so the two terms above can be confirmed independently. Dynamic testing and integrity testing are the faster, cheaper proxies used on the rest of the population. On any significant piling job the design values are preliminary until the first test pile has been loaded — which is the honest position these pages take.